Jasmonate regulates male sterility in cucumber

Abstract

Male sterility affects various agronomic traits. Although jasmonate (JA) has been shown to influence fertility in several plant species, most research on JA-mediated regulation of fertility has focused on bisexual flowers, leaving the role of this phytohormone in unisexual flowering plants largely unexplored. Cucumber (Cucumis sativus) has unisexual flowers and is an ideal model plant for studying male and female gametophyte development. In this study, we determined that JA regulates male and female flower development as well as male fertility in cucumber. In the JA biosynthesis-deficient mutant Csopr3, both female and male flower petals failed to open properly, anthers showed defective dehiscence, and pollen in male flowers ruptured and was nonviable. MeJA treatment restored normal petal opening and anther dehiscence in Csopr3, and the pollen regained the ability to germinate. Further analysis revealed that CsXTH33, a gene involved in cell-wall structure, negatively regulates stamen fertility in cucumber. Transcriptional regulation assays indicated that CsMYC2 binds to the promoters of CsWRKY45 and CsWRKY57 and activates their expression. In turn, CsWRKY45 and CsWRKY57 bind to the promoter of CsXTH33, inhibiting its expression and promoting normal stamen development. This study thus reveals the role of JA in cucumber unisexual flower development and deepens our understanding of the mechanisms by which JA regulates plant fertility.

1. Introduction

Male sterility is widely regarded as a foundational technique for hybrid breeding. Advantages of male sterility include eliminating the need for artificial emasculation, reducing costs for seed production, and ensuring the purity of F1 hybrids [1]. Male-sterile mutants serve as critical models for analyzing complex processes involved in male reproductive development. Studies of these mutants have revealed core regulatory networks governing pollen development, including programmed cell death [2], phytohormone signaling pathways [3,4], and dynamic cell-wall remodeling [5], providing valuable insights into reproductive biology. Plant hormones, including jasmonates (JAs) [4], auxin [3], and gibberellin [6], function in a coordinated manner to regulate male sterility. Auxin plays a crucial role in anther morphogenesis during the early stages of reproductive development, establishing the foundation for reproductive structure formation [7]. By contrast, gibberellins play dominant roles in the biosynthesis of the inner pollen wall during the middle stages of pollen development, a process critical for the structural integrity of pollen [6]. JAs play the most important role, as they regulate anther dehiscence and pollen release at the mature stage, which directly influence reproductive success [4]. Defects in the JA signaling pathway result in anther indehiscence and pollen retention, the most critical defects affecting fertility [4,8]. JAs also play central roles in the molecular mechanisms governing male sterility in crops by interacting with auxin and gibberellin via synergistic and antagonistic relationships in a tightly regulated network that controls anther development [3,9].

JA is biosynthesized from α-linolenic acid, which is converted into 12-oxo-phytodienoic acid (OPDA) through the sequential actions of lipoxygenase (LOX), allene oxide synthase (AOS), and allene oxide cyclase (AOC) [10]. OPDA is then converted to JA by OPDA reductase 3 (OPR3) and three rounds of oxidation [[10], [11], [12]]. Bioactive JA-Ile is then synthesized from JA through catalysis by the jasmonate synthetases JASMONATE RESISTANT1 (JAR1)/GH3.11 and GH3.10 [13,14]. In the JA signaling pathway of Arabidopsis (Arabidopsis thaliana), the receptor CORONATINE-INSENSITIVE1 (COI1) interacts with SKP-LIKE1/2 (ASK1/2) and associates with Cullin1 and Rbx1 to form a ubiquitin ligase complex [[15], [16], [17]]. This complex recognizes JA-Ile and recruits JASMONATE-ZIM-DOMAIN (JAZ) proteins for ubiquitination and degradation, thus releasing their repression of MYC2, encoding a central regulator of JA signaling [[18], [19], [20]].

The role of JA in regulating male sterility has been extensively studied in plants with bisexual flowers, including Arabidopsis [14,21,22], tomato (Solanum lycopersicum) [23], and rice (Oryza sativa) [24,25]. These studies often revealed abnormalities in stamen development in JA-deficient mutants, including shorter filaments, defective anther dehiscence, abnormal spikelets, and reduced pollen viability [9,[26], [27], [28], [29], [30], [31], [32]]. In Arabidopsis, the R2R3-type MYB transcription factor (TF) genes MYB21 and MYB24 are expressed in flowers and function downstream of JA [8]. These TFs form complexes with subgroup IIIe basic helix-loop-helix (bHLH) TFs, including MYC2/3/4/5, which activate MYB108 to regulate stamen development [4]. JAZ proteins interact with these bHLH-MYB complexes to regulate stamen development [4]. A recent study revealed that the H3K27me3 effector EMF1 interacts with MYB26, negatively regulating the transcription of NAC SECONDARY WALL THICKENING PROMOTING FACTOR1 (NST1) and NST2, which mediate secondary cell-wall thickening in anthers during JA-regulated stamen maturation [8].

However, the mechanisms by which JA regulates fertility in unisexual flowers remain relatively unexplored, with most studies focusing on maize (Zea mays) [33]. A JA-deficient maize mutant exhibits non-dehiscent anthers and nonviable pollen grains [33], resembling findings in Arabidopsis [4,8]. In addition to defects in male fertility, abnormal sex determination has also been observed in such mutants in maize [34], and sorghum (Sorghum bicolor) [35]. By contrast, the Cplox3a and Cpjar1b mutants in Cucurbita pepo exhibit petal closure but normal fertility, suggesting that the role of JA in unisexual flower development may be species-specific [36,37]. The presence of multiple LOX and JAR1 genes in C. pepo may contribute to functional redundancy, potentially obscuring mutant phenotypes [36,37]. By contrast, the role of JA in flower development in other cucurbit plants remains to be elucidated, as JA-related mutants have not yet been identified in these species.

CsGL2-LIKE regulates male fertility in cucumber (Cucumis sativus L.) and participates in the JA pathway by interacting with CsJAZ1 [38], suggesting that JA signaling may play a role in male flower development in cucumber [38]. Cucumber, both a monoecious and dioecious crop in the Cucurbitaceae family, is a well-established model for studying unisexual flower development [39]. However, compared with model plants like Arabidopsis and rice, the mechanisms regulating fertility in cucumber remain poorly understood. Further studies are needed to clarify the role of JA in reproductive development and its associated regulatory mechanisms in cucumber.

In this study, we generated mutants deficient in JA biosynthesis and signaling. Analysis of these mutants revealed that JA deficiency impairs petal opening in both male and female flowers and causes anthers to remain indehiscent, producing nonviable pollen grains, resulting in male sterility. Methyl jasmonate (MeJA) treatment restored normal petal opening and anther dehiscence while rescuing pollen grain viability in the JA-deficient mutant. We also identified CsXTH33, a xyloglucan endotransglucosylase/hydrolase gene, which negatively regulates male fertility in cucumber. CsMYC2 binds to the promoters of CsWRKY45 and CsWRKY57, inducing their expression. In turn, these TFs repress CsXTH33 expression by binding to its promoter. These findings highlight a molecular link between JA and enzymes involved in cell-wall modification in regulating flower development and male sterility in cucumber.

2. Results

2.1. JA deficiency affects petal opening and male sterility in cucumber

Arabidopsis produces bisexual flowers containing both pistils and stamens (Fig. 1A), whereas cucumber develops unisexual flowers, with each flower bearing either pistils or stamens (Fig. 1A). Both male and female cucumber flowers have petals (Fig. 1A). To investigate the roles of JA in male sterility in cucumber, we generated two homozygous Csopr3 mutants using CRISPR/Cas9-mediated gene editing in the inbred line Cu2 background. The Csopr3-1 and -2 mutants carry 2 bp and 4 bp deletions, respectively, both leading to premature termination of CsOPR3 translation (Fig. 1B).

Fig. 1.

JA deficiency affects petal opening and male sterility in cucumber. A Bisexual flower structure of Arabidopsis. Pe, petal; St, stamen; Sy, style. Female flower structure of cucumber. Scale bars: 1 cm (left, male flower), 5 mm (right). B Mutation sequences of homozygous Csopr3 mutants generated via CRISPR/Cas9: Target sequences (red) and protospacer adjacent motif (PAM) sequences (bold). Black dashes indicate deletions, denoted by a minus sign (−) followed by the number of deleted nucleotides above the sequence. The dashed box highlights the premature termination codon in the corresponding protein. Green boxes mark the Csopr3 protein-coding region, and gray boxes indicate missense mutations caused by frameshift mutations. C JA-Ile content of cucumber leaves after 1 h wounding treatment. FW, fresh weight. Significance analysis was performed with the two-tailed Student's t-test (∗∗P < 0.01). Values are means ± SD (n = 3). D Male and female flower phenotypes of Cu2 and Csopr3 mutants at different developmental stages. Scale bar: 1 cm. E Seeds from female flowers following self-pollination and hybridization. Self-pollination uses male flowers from the same plant, and hybridization uses male flowers from Cu2. Scale bar: 1 cm. F The seeds germination phenotype of Cu2 and Csopr3 mutants. Scale bar: 1 cm.

We measured JA-Ile levels of cucumber leaves 1 h after wounding treatment. Wild-type Cu2 plants showed an average JA content of 78.9 ng/g fresh weight (FW), while both Csopr3 mutants exhibited nearly undetectable levels of JA (Fig. 1C), confirming the successful disruption of JA biosynthesis.

Phenotypic analysis revealed that in Cu2 plants, petals of male and female flowers transitioned from the pre-flowering (before flowering; BF) to the flowering (F) stage within a single day, fully opening and entering the after-flowering (AF) stage by one day later (Fig. 1D). Petals then began to partially close before undergoing apoptosis (Fig. 1D). By contrast, Csopr3 petals remained closed throughout the AF stage (Fig. 1D).

Self-pollination of Cu2 produced plump seeds, while self-pollination of Csopr3 resulted in shriveled seeds (Fig. 1E). However, cross-pollination of Csopr3 with Cu2 pollen yielded plump seeds, similar to those from Cu2 (Fig. 1E). Germination tests revealed that only plump seeds from Cu2 and the cross-pollination group germinated normally, whereas seeds from self-pollinated Csopr3 failed to germinate and lacked embryos (Fig. 1F). These results demonstrate that JA deficiency disrupts petal development and compromises pollen fertility in cucumber, while stigma fertility remains unaffected.

2.2. JA deficiency results in indehiscent anthers and nonviable pollen

To examine pollen viability, we cultured pollen grains in vitro and performed I₂-KI staining. Most pollen grains from Csopr3 stained normally, similar to those from Cu2 (Fig. 2A). However, Csopr3 pollen grains showed significantly larger diameters than Cu2 pollen grains (Fig. 2A and B). However, in vitro germination tests showed that none of the pollen grains from Csopr3 germinated, indicating a complete loss of viability (Fig. 2C and D). Furthermore, most Csopr3 pollen grains exhibited an abnormal ruptured phenotype during in vitro germination (Fig. 2C–E), suggesting potential defects in pollen cell-wall homeostasis. MeJA treatment for 7 days before flowering rescued the pollen viability of the mutants, allowing pollen germination without any pollen rupture (Fig. 2C–E). These results confirm that JA deficiency results in nonviable pollen grains and suggest that JA may regulate cell-wall homeostasis in pollen grains during germination.

Fig. 2.

JA deficiency results in nonviable pollen grains. A Pollen grains I₂-KI staining phenotype. Scale bar: 200 μm. B Statistic analysis of pollen grains diameter. Significance analysis was performed with the two-tailed Student's t-test (∗∗∗P < 0.001). Values are means ± SD (n ≥ 93). CIn vitro pollen grains' germination phenotype at the flowering stage. The black arrows represent normal pollen grains, the red arrows represent raptured pollen grains. The scale bar is 200 μm. D Statistic results of pollen grains germination in vitro of that in Fig. C. Significance analysis was performed with the two-tailed Student's t-test (∗P < 0.05, ∗∗∗P < 0.001). Values are means ± SD (n = 3). E Statistic analysis of pollen grains ruptured rate during germination in vitro of that in Fig. C. Significance analysis was performed with the two-tailed Student's t-test (∗∗∗P < 0.001). Values are means ± SD (n = 3). nd, no detected. F Anther phenotype of Cu2 and Csopr3 mutants at different development stage. Left: entire plants; right: transverse section of anther structure. Scale bars: 500 μm (left), 250 μm (right).

To further investigate the effects of JA deficiency on anther development, we conducted detailed phenotypic observations of whole anthers and transverse anther sections. In Cu2, anthers exhibited paler coloration and smoother surfaces during the BF stage, with anther dehiscence beginning in the F stage and becoming more pronounced in the AF stage (Fig. 2F). Analysis of transverse sections of Cu2 anthers in the BF stage revealed that pollen grains were enclosed within intact anther locules, but locular dehiscence at the F stage facilitated pollen release (Fig. 2F). By contrast, Csopr3 anthers retained structural integrity and smooth surfaces from the BF to AF stages, failing to dehisce (Fig. 2F). The anther locules in Csopr3 remained intact throughout both the BF and F stages, preventing pollen release (Fig. 2F, Fig. S1). These observations suggest that JA deficiency influences cell-wall development in anthers, thus disrupting anther dehiscence and pollen dispersal in cucumber.

To confirm the function of JA, we performed a rescue experiment using MeJA treatment. MeJA treatment restored both petal opening (Fig. S2A–B) and anther dehiscence (Fig. S2C) in Csopr3, effectively reversing the JA deficiency-induced defects.

2.3. CsXTH33 acts downstream of the JA pathway and negatively regulates pollen viability

To identify the downstream genes regulated by JA, we performed RNA-seq of Cu2 and Csopr3 anthers from male flowers at the BF and F stages. Principal component analysis (PCA) clearly distinguished the Cu2 and Csopr3 groups, revealing distinct gene expression profiles between the two genotypes (Fig. S3A). In Cu2, the anthers at the BF stage were white and compact (Fig. 2F). After entering the F stage, the anthers of Cu2 turned yellow and cracked to release pollen grains. By contrast, the Csopr3 samples from the BF and F stages clustered together, suggesting minimal changes in gene expression between these two stages (Fig. S3A), which was consistent with the similar anther phenotypes observed in Csopr3 at both stages (Fig. 2F).

Analysis of differentially expressed genes (DEGs) between Cu2 and Csopr3 revealed 1890 down-regulated and 2043 up-regulated genes in Cu2 during the F stage, while the Csopr3 mutant displayed only 73 down-regulated and 81 up-regulated genes during this stage (Fig. S3B). Gene Ontology (GO) analysis of the up- and down-regulated genes in Cu2 revealed that genes associated with the cell wall were predominantly identified in the down-regulated set (Fig. S3C), whereas the up-regulated set was primarily linked to photosynthesis (Fig. S3D). Given the abnormal apoptosis in cell walls in the anther chamber of Csopr3 (Fig. 2F), we focused on genes related to cell wall development, i.e., down-regulated genes. Venn diagram analysis revealed 1862 down-regulated DEGs potentially responsive to JA between the Cu2 and Csopr3 groups (Fig. 3A). GO analysis of these DEGs indicated that the biological process, cellular component, and molecular function categories included GO terms related to the cell wall (Fig. 3B), implying that these genes function in cell wall development, which is consistent with our previous hypothesis.

Fig. 3.

Screening and Functional verification of CsXTH33.A The venn diagram of down-regulated gene-set of Cu2 and Csopr3. The red number means genes unique of Cu2. B The top 10 GO results from biological process (BP), cellular component (CC), and molecular function (MF) of 1862 DEGs from (A). The GO terms marked with red were used for further analysis. C Expression profiles of two genes from the GO terms marked with red from (B). D The log₂FC indicate expression changes from BF to F stage in female flowers. E Relative expression of CsXTH33 in cu2 and Csopr3 treated with or without MeJA. F Relative expression of CsXTH33 from Cu2 and the overexpression lines at anthers from flowering stage. OE1, OE-XTH33-1, OE3, OE-XTH33-3. G Anthers phenotype under microscope of Cu2 and the overexpression lines of CsXTH33. Scale bar: 500 μm. H I₂-KI staining phenotype of pollen grains from Cu2 and OE-XTH33 lines. Scale bar: 200 μm. IIn vitro pollen grains germination phenotype of Cu2 and the overexpression lines of CsXTH33. Scale bar: 200 μm. J Statistical analysis of pollen grains abortion rate from (H). K Statistical analysis of pollen grains germination rate from (I). L Statistical result of pollen grains ruptured rate during germination process in vitro from (H). Significance analysis was performed with the two-tailed Student's t-test (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001). Values are means ± SD (n = 3). nd, no detected.

We further analyzed the enriched pathways related to the cell wall. Among these, in the plant-type cell-wall organization or biogenesis pathway, only two genes were highly expressed in male flowers, namely CsXTH33 (xyloglucan/xyloglucosyl transferase 33) and CSLB03 (cellulose synthase-like B3) (Fig. 3C, Fig. S4). CSLB03 is required for pollen wall development and is a direct downstream target gene of bHLH010 [40], verifying the validity of our screening approach. We then focused on CsXTH33 (Fig. 3D). RT-qPCR analysis confirmed that CsXTH33 was significantly down-regulated in Cu2 at the F stage, which is consistent with the results of RNA-seq (Fig. 3D, Fig. S5B–C). By contrast, CsXTH33 was significantly up-regulated in Csopr3 (Fig. 3E). MeJA treatment reduced CsXTH33 expression in Csopr3 (Fig. 3E), suggesting that CsXTH33 is a downstream target of the JA pathway in regulating fertility. To investigate the function of CsXTH33, we generated transgenic lines overexpressing (OE) this gene (Fig. 3F). The OE lines showed swollen but dehiscent anthers compared to Cu2 (Fig. 3G and H). However, pollen from these OE lines exhibited abnormal germination and nonviability (Fig. 3I-L), similar to Csopr3 pollen (Fig. 2C and D). These results demonstrate that CsXTH33 negatively regulates pollen germination and is a key downstream gene in the JA pathway.

2.4. CsMYC2 participates in JA signaling and regulates pollen viability in cucumber

MYC2, a central regulator of JA signaling, mediates JA-dependent responses in plants [41,42]. To investigate how JA regulates CsXTH33, we identified the functional ortholog of AtMYC2 in cucumber and generated a CsMYC2 mutant in the Cu2 background using CRISPR/Cas9. We obtained one homozygous line carrying a 101 bp deletion in CsMYC2 causing premature termination of translation (Fig. 4A). The Csmyc2 mutant exhibited the same petal closure phenotype as Csopr3 but retained normal anther dehiscence compared to Cu2 from stage BF to F (Fig. 4B). In addition, I₂-KI staining showed that most pollen grains in Csmyc2 stained normally and were significantly larger than those in Cu2 (Fig. S9). However, in vitro germination tests revealed a complete loss of pollen viability in Csmyc2, with no germination occurring (Fig. 4C and D). Moreover, most Csmyc2 pollen grains ruptured during in vitro germination (Fig. 4C–E), similar to the Csopr3 mutants (Fig. 2C–E). These results suggest that CsMYC2 is essential for pollen germination in cucumber.

Fig. 4.

Functional verification of CsMYC2 in cucumber. A CRISPR/Cas9-mediated mutation in a homozygous Csmyc2 mutant. Target sequences marked with red and PAM sequences marked with bold. Black dashes indicate deletions, with the number of deleted nucleotides marked as a minus sign (−) above the sequence. Dashed box highlights the premature termination codon in the corresponding protein. Green boxes represent CsMYC2 protein-coding regions; gray boxes indicate missense mutation caused by frameshift mutations. B Flower and anther phenotypes of Cu2 and Csmyc2 mutant at different developmental stages. Scale bars: 1 cm (male flower), 500 μm (anther). CIn vitro pollen grains germination phenotypes of Cu2 and Csmyc2 mutant. Scale bar: 200 μm. D Statistical analysis of pollen grains germination rate from (C). E Statistical analysis of pollen grains ruptured rate during germination process in vitro from (C). Significance analysis was performed with the two-tailed Student's t-test (∗∗∗P < 0.001). Values are means ± SD(n = 3). nd, no detected.

2.5. CsMYC2 inhibits the expression of CsXTH33 by promoting the expression of CsWRKY45 and CsWRKY57

To explore whether CsMYC2 regulates CsXTH33 expression, we conducted a dual-luciferase reporter assay in Nicotiana benthamiana leaves. CsMYC2 did not directly repress CsXTH33 expression (Fig. 5A), suggesting that other TFs might be involved in this regulation. To identify the responsible TFs, we re-analyzed the RNA-seq data from Csopr3 and Cu2 at the F stage. Among the 2464 DEGs, 111 TF genes were identified (Fig. 5B). Among these TFs, the NAC (14%), ERF (12%), and WRKY (9%) families were the most abundant (Fig. 5C). Analysis of the cis-elements in the 2 kb XTH33 promoter sequence revealed 15, 11, and 7 binding sites for WRKY, NAC, and ERF TFs, respectively (Fig. 5D). Given the predominance of WRKY binding sites, we focused our analysis on the WRKY TF family.

Fig. 5.

CsMYC2-WRKY-XTH33 cascade regulation module in cucumber. A Statistic analysis of dual-luciferase reporter assay between CsMYC2 and CsXTH33 promoter. ns, no significance. B The DEG numbers in Csopr3 vs. Cu2 at flowering stage. C Top 10 TF family numbers of the 111 DEGs from (B). D Total numbers of binding domain in the top 3 TF families in Fig. C. E Heatmap of 10 WRKY TFs TPM value in Cu2 and Csopr3 mutant at BF and F stage. F Statistical analysis of dual-luciferase reporter assay between CsXTH33 and CsWRKY45/57 promoter, respectively. G, H Statistical analysis of dual-luciferase reporter assay results between CsMYC2 and CsWRKY45(G) and CsWRKY57 promoter (H). Significance analysis was conducted with the two-tailed Student's t-test (∗P < 0.05, ∗∗∗P < 0.001). Values are means ± SD (n ≥ 3). I, J Yeast one hybrid assay results between CsMYC2 and promoter of CsWRKY45 (I) or CsWRKY57 (J). K Yeast one hybrid assay results between CsWRKY45 or CsWRKY57 and CsXTH33 promoter. The number above represents the dilution multiple. The numbers below indicate the concentration of AbA, with the unit of nanogram.

We identified 10 WRKY TF genes whose expression gradually increased in Cu2 during flowering (Fig. 5E). However, in Csopr3, the expression levels of these genes were relatively low and did not significantly change during flowering (Fig. 5E). In dual-luciferase reporter assays, among these 10 WRKYs, only CsWRKY45 and CsWRKY57 bound to the CsXTH33 promoter and repressed its expression (Fig. 5F, Fig. S7). Furthermore, CsMYC2 bound to the CsWRKY45 and CsWRKY57 promoters and enhanced their transcription (Fig. 5G), which is consistent with their expression patterns revealed by RNA-seq (Fig. 5E). In a yeast one-hybrid (Y1H) assay (Fig. 5I–K), CsMYC2 bound to the CsWRKY45 and CsWRKY57 promoters, and both CsWRKY45 and CsWRKY57 bound to the CsXTH33 promoter. Consistent with these findings, in an electrophoretic mobility shift assay (EMSA), both CsWRKY45 and CsWRKY57 bound to the CsXTH33 promoter (Fig. S7B). These results suggest that JA regulates male sterility in cucumber through the MYC2-WRKY45/57-XTH33 regulatory module (Fig. 5).

3. Discussion

Bisexual flowering plants such as rice [32,43,44] and Arabidopsis [8,9,28,29,31] produce flowers containing both pistils and stamens (Fig. 1A), whereas unisexual flowering plants such as Cucurbitaceae species and maize bear either staminate or pistillate flowers (Fig. 6A) [33,36]. Cucumber, a representative member of the Cucurbitaceae, serves as an excellent model for investigating the development of unisexual flowers [38].

Fig. 6.

A model diagram illustrating the regulation of flower development and stamen fertility by JA in Arabidopsis and cucumber. A Phenotype of JA mutants. In Arabidopsis, JA deficiency results in shorter filaments, delayed dehiscence anther, and pollen grains nonviable. In tomato, JA deficiency results in anther dehiscent prematurely, pollen grains viability reduce and female sterility. In rice, JA deficiency results in abnormal spikelet development, reduced stamens, duplicated glumes, multiple carpels, and indehiscent anthers. In tobacco, JA deficiency leads to shorter filaments, delayed anther dehiscence and closed petal. In maize, JA deficiency results in anther indehiscent. In cucumber, JA deficiency results in closed petals both in the male and female flowers and indehiscent anthers. B The MYC2-WRKYs-XTH33 module regulates male sterility in cucumber. In the presence of JA, JA promotes MYC2 release, which activates CsWRKY45/57 expression and subsequently suppresses CsXTH33 transcription, resulting in low CsXTH33 abundance (low CsXTH33) and enabling normal flower development and pollen grains germination. In the absence of JA, MYC2 is inhibited by JAZ proteins, thereby ceasing MYC2-mediated activation of WRKYs and relieving repression of CsXTH33 (high XTH33). Elevated CsXTH33 levels lead to complete loss of pollen viability. , activation; , inhibition. Arrow thickness represents the strength of promoting or inhibitory effects.

JA regulates multiple reproductive processes, including floral organ development, anther dehiscence, pollination, and seed formation [25,30,45]. In this study, we obtained compelling evidence that JA is indispensable for proper flower opening, anther dehiscence, and pollen viability in cucumber (Fig. 1, Fig. 2), resembling observations in other plant species (Fig. 6A) [4,33,46]. In cucumber, JA-deficient mutants exhibited non-dehiscent anthers and nonviable pollen grains (Fig. 2), resembling findings in Arabidopsis, rice, tobacco, and maize (Fig. 6A), highlighting the conserved requirement for JA in regulating pollen viability across angiosperms [4,8,9,23,24,[26], [27], [28], [29], [30], [31], [32],43,[47], [48], [49], [50], [51], [52],45].

Although the JA pathway is broadly conserved, species-specific differences exist [4,32,33,53]. In cucumber, loss of function of CsMYC2 alone led to completely nonviable pollen (Fig. 4), resembling the phenotype of rice Osmyc2 mutants [32,53] and contrasting sharply with the strong MYC gene redundancy observed in Arabidopsis [4]. As anther dehiscence persists in the cucumber myc2 mutant, additional MYC homologs likely function redundantly to regulate this process, a concept that requires further investigation.

Beyond stamen development, JA deficiency in cucumber caused additional phenotypes (Fig. 1). Both JA biosynthesis-deficient and signaling-deficient mutants exhibited closed petals in both male and female flowers at anthesis (Fig. 1, Fig. 4), demonstrating that JA is essential for petal opening in cucumber, resembling JA-deficient flowers in tobacco [47,48] (Fig. 6A). Similarly, in C. pepo, the JA-deficient mutants Cplox3a and Cpjar1B display closed petals but maintain normal pollen viability [36,37], suggesting that closed petals might be a conserved JA-related phenotype within Cucurbitaceae. In addition, JA signaling regulates diurnal flower opening in rice, suggesting that JA plays a conserved role in regulating flowering, but its specific functions might be species-specific [54]. By contrast, CsXTH33-overexpressing cucumber plants exhibited normal petal opening but defective pollen development (Fig. 3), indicating that genes beyond CsXTH33 participate in JA-mediated petal opening. Furthermore, JA-deficient cucumber mutants displayed increased pollen grain diameter (Fig. 2A–B, Fig. S6), a phenotype not previously reported in other species, which requires further investigation.

In addition to reproductive processes, JA also influences defense against insects in cucumber, as well as maize [55], rice, cotton (Gossypium hirsutum), and Arabidopsis [56,45]. The fall armyworm (Spodoptera frugiperda, FAW) is a pernicious pest affecting a wide range of host plants [56,57]. When feeding on Csopr3 leaves, FAW larvae showed significantly increased weight compared with larvae feeding on Cu2 leaves (Fig. S8A and B). These results reveal the functional conservation of JA in plant defense against insects [[55], [56], [57]]. JA also influences the growth rate of cucumbers (Fig. S8C and D). During vegetative growth, the Csopr3 mutants exhibited significantly greater plant height compared to the wild type (Fig. S8C and D). Similarly, overexpressing AtOPR3 decreased plant height in wheat (Triticum aestivum) [58,59]. Overexpressing the LOX gene OsRCI-1 reduced plant height in rice [60], which is consistent with the increase in plant height observed in the cucumber Csopr3 mutant (Fig. S8C and D). Collectively, these findings confirm the notion that JA plays conserved roles in inhibiting plant growth [[58], [59], [60]].

We identified CsXTH33 as a downstream target of the JA pathway that negatively regulates male fertility in cucumber (Fig. 3). CsXTH33 expression decreased sharply at anthesis, and its overexpression resulted in severe defects in pollen viability and malformed male flowers (Fig. 3). CsXTH33 belongs to the xyloglucan endo-transglycosylase/hydrolase (XTH) family, which mediates xyloglucan remodeling and regulates cell-wall flexibility [61,62]. While AtXTH33 is highly expressed in mature Arabidopsis anthers and developing siliques [63], Atxth33 mutants retain normal fertility [61]. Given that fertility in the Atxth33 mutant is normal and cucumbers possess 27 XTH genes with potential functional redundancy (Fig. S5A), we reasoned that knocking out XTH33 in cucumber would lead to a lack of fertility. Therefore, we did not generate an XTH33 knockout mutant in cucumber at this stage.

XTH genes frequently modulate cell-wall structure, affecting organ elongation and various developmental processes [[64], [65], [66]]. For instance, Arabidopsis AtXTH18, AtXTH19, and AtXTH20 affect cell-wall composition and hypocotyl growth [67,68]. However, only AtXTH28 has been implicated in fertility; xth28 mutants exhibit shortened filaments and male sterility [69]. In this study, we identified a novel XTH gene, CsXTH33, as a regulator of male fertility in cucumber. However, the underlying molecular mechanism is still unknown. Based on the known roles of the XTHs in modifying the xyloglucan–cellulose network and the importance of sugar transport for anther and pollen maturation [70,71], we propose that increasing CsXTH33 expression disrupts cell-wall remodeling, interferes with sugar allocation, and thus blocks pollen germination. However, this hypothesis requires further experimental validation. Furthermore, although CsXTH33 has been implicated in male sterility in cucumber, its functional conservation in other crops remains uncharacterized and warrants investigation. Additionally, CsXTH33 overexpression resulted in enlarged petals, sepals, and receptacles (Fig. S6), suggesting broader developmental roles that warrant further study.

Several WRKYs have been linked to pollen development [[72], [73], [74]]. However, the signaling cascade connecting WRKY TFs to the regulation of fertility remains poorly understood. Based on RNA-seq analysis, we identified 10 candidate WRKY TFs in cucumber and demonstrated that CsWRKY45 and CsWRKY57 repress CsXTH33 expression, a process regulated by CsMYC2 (Fig. 5). Both WRKY45 and WRKY57 have been implicated in growth and stress responses [[75], [76], [77], [78], [79], [80]], but their involvement in fertility has not previously been reported. Our findings suggest that CsWRKY45 and CsWRKY57 act downstream of CsMYC2 to suppress CsXTH33 expression, forming a MYC2–WRKY45/57–XTH33 module that mediates JA-dependent regulation of stamen fertility in cucumber. Given the potential functional redundancy among WRKYs, single knockout mutants of WRKY45 or WRKY57 may not exhibit clear defects in fertility. Further studies using double mutants will be essential to determine their precise roles and to uncover whether additional WRKY or non-WRKY TFs participate in the JA-regulated fertility network, possibly producing phenotypes similar to those observed in response to CsXTH33 overexpression.

4. Materials and methods

4.1. Plant materials and growth conditions

Cucumber inbred line Cu2 was used in this study. Seeds were initially germinated on moist filter paper and subsequently sown in 9 cm × 9 cm plastic pots filled with a 2:1 (v/v) mixture of peat and vermiculite. Plants were cultivated in a greenhouse at the Agricultural Genomics Institute at Shenzhen, Chinese Academy of Agricultural Sciences, in a controlled growth environment under a 16 h day (25 °C)/8 h night (18 °C) photoperiod. All experiments were performed using homozygous lines of transgenic plants. The Nicotiana benthamiana plants used in this study originated from laboratory stock and were cultivated under conditions identical to those used for cucumber plants.

4.2. Plasmid construction and Agrobacterium-mediated transformation

To construct the knockout vector, sgRNA sequences were designed using the online tools CRISPR-P 2.0 (http://cbi.hzau.edu.cn/cgibin/CRISPR2/CRISPR) and CRISPR RGEN Tools (http://www.rgenome.net/cas-designer/). The sgRNAs were inserted into the plasmid pKSE402 as previously described [81]. For the overexpression vector, the full-length coding sequence (without the stop codon) of CsXTH33 was inserted into the linearized plasmid pCAMBIA1305.4 using the ligation enzymes. The resulting plasmid was transformed into Agrobacterium tumefaciens strain EHA105 via heat shock transformation. Agrobacterium-mediated transformation of cucumber was performed as previously described [81].

4.3. RNA isolation and reverse transcription-quantitative PCR (RT-qPCR)

Total RNA was extracted from fresh plant tissue using a TransZol Up Plus RNA Kit (TransGen Biotech, Beijing, China). Briefly, the tissue was flash-frozen in liquid nitrogen, ground to a powder, and mixed with 1 mL of TransZol and 0.2 mL of RNA extraction agent. Subsequent steps followed the manufacturer's protocol. RNA (1 μg) was reverse-transcribed into single-stranded cDNA using NovoScript Plus All-in-one 1st Strand cDNA Synthesis Super Mix with gDNase (Novo protein, Suzhou, China). Quantitative PCR was carried out on a CFX96 Touch system (Bio-Rad) with Hieff UNICON® Universal Blue qPCR SYBR Green Master Mix (YEASEN). Cucumber Ubiquitin mRNA served as the internal control. Each gene was subjected to three technical replicates and three biological replicates. The primers used for RT-qPCR are listed in Table S1.

4.4. Characterization of transgenic plants

GFP fluorescence in the explants was observed under a LUYOR-3415RG light source (Luyor Instrument, Shanghai) during both the tissue culture and rooting stages for transgenic validation. Genomic DNA was extracted from the samples using an SDS-based method. PCR amplification of the target region was performed to confirm the presence of transgenes, and homozygous lines were selected for subsequent experiments. Primer information is shown in Table S1.

4.5. In vitro pollen grain germination assay

Pollen grains were germinated on germination medium consisting of 0.0025% boric acid, 0.25 mM MgSO₄, 1.25 mM CaCl₂, 1.25 mM KCl, 10% sucrose, and 0.015% agar. The pollen was cultured on a shaker at 25 °C for 4 h.

4.6. RNA-seq

RNA-seq was performed by Berry Genomics, with three biological replicates each for Cu2 and Csopr3. Raw reads were sequenced on an Illumina platform. Adaptor sequences and low-quality reads were filtered out using Trimmomatic. Clean reads were mapped to the reference cucumber genome (http://cucurbitgenomics.org/) using HISAT2 (v2.2.1) with default parameters. Read counts were quantified and normalized via StringTie. Differential gene expression analysis, principal component analysis (PCA), and Gene Ontology (GO) enrichment analysis were conducted using R package tools. A P-adjust value < 0.05 and a fold change ≥2 were set as the significance thresholds for defining differential expression.

4.7. Measurement of JA and JA-Ile contents

Leaves were harvested 1 h after wounding treatment and flash-frozen in liquid nitrogen. Each biological replicate consisted of at least three leaves from different plants. For metabolite extraction, 100 ± 2 mg tissue was weighed and transferred to a 2 mL Eppendorf tube. The tissue was mixed with 1.5 mL of isopropyl formate containing 0.5% (v/v) internal standard (JA-d5) and vortexed again for 2 min, followed by centrifugation at 15,871 g for 10 min at 8 °C. A 1.4 mL aliquot of the supernatant was transferred to a new 2 mL Eppendorf tube and spin-dried at 30 °C under a vacuum. The residue was redissolved in 0.8 mL of 85% (v/v) methanol, sonicated in an ice-water bath for 10 min, and centrifuged at 15,871 g for 10 min at 8 °C. The supernatant was loaded onto a pre-activated SPE HLB column (Waters, USA), which was sequentially conditioned with 0.8 mL of anhydrous methanol and 0.8 mL of 85% methanol. The column was eluted with 0.8 mL of 85% methanol, and 1.6 mL of the eluent was collected and spin-dried. The final residue was redissolved in 0.2 mL of 50% (v/v) methanol, centrifuged at 15,871 g for 15 min at 8 °C, and the supernatant transferred to a certified vial for analysis. The sample was redissolved in 0.8 mL of 85% methanol and sonicated in an ice-water bath for 10 min before centrifugation. An SPE HLB column was activated with 0.8 mL of 85% methanol, and the supernatant was passed through the column and eluted with another 0.8 mL of 85% methanol. A total of 1.6 mL of eluent was collected and spin-dried. The residue was redissolved in 0.2 mL of 50% methanol, centrifuged at 15,871 g for 15 min at 8 °C, and the supernatant transferred to a vial for LC analysis. Each replicate included at least three leaves from different plants. JA and JA-Ile contents were measured using gas chromatography-mass spectrometry.

4.8. Phylogenetic analysis

Published Arabidopsis XTH family protein sequences were obtained from TAIR database. Cucumber XTH protein sequences were retrieved from the Cucurbitaceae database website. The cucumber sequences were aligned against the Arabidopsis XTH protein sequences using BLASTP with an E-value threshold of 1e⁻³ to identify homologous sequences. A domain search was performed on these candidates with the Hidden Markov Model (HMMER) using the Pfam domains PF00722 and PF06955 to further refine the selection. The Gene Structure Display Server tool (GSDS) was used to visualize exon-intron structures for each gene family member. Conserved protein motifs were identified using Motif-based-Sequence analysis (MEME), and multiple sequence alignment of the amino acid sequences was conducted with ClustalW or MAFFT. A phylogenetic tree was constructed via the maximum likelihood method in MEGA. The tree was visualized and annotated using interactive Tree of Life (iTOL). Information about the genes used for both sequence alignment and phylogenetic analysis is available in Table S2.

4.9. Dual-LUC assay

The full-length coding sequences of CsMYC2 and CsWRKYs were inserted into the pGreen II 62 SK vector (effector), and the CsWRKY45, CsWRKY57, and CsXTH33 promoters were cloned into the pGreen II 0800-LUC vector (reporter). The vectors were transformed into Agrobacterium tumefaciens strain GV3101 (pSoup-P19) and adjusted to a concentration of OD₆₀₀ = 0.75 using 2-(N-morpholino) ethanesulfonic acid buffer. A mixture of cultures (10:1, effector: reporter) was injected into tobacco leaves. Leaves were harvested three days later, and firefly LUC and REN activities were measured using a Dual-Luciferase Reporter Assay System (Promega) on a microplate reader instrument (Molecular Devices SpectraMax i3x, USA). The primer information is available in Table S1.

4.10. Yeast one-hybrid assay

The CsWRKY45, CsWRKY57, and CsXTH33 promoter sequences were separately inserted into the pAbAi vector. The resulting constructs were transferred into yeast (S. cerevisiae) strain Y1H Gold. The yeast cells were grown on SD/-Ura medium with AbA (Aureobasidin A). The full-length coding sequences of CsMYC2, CsWRKY45, and CsWRKY57 were separately inserted into pGADT7 and transferred into yeast strain Y1H Gold containing pAbAi-CsWRKY45/CsWRKY57/CsXTH33. Interactions of TFs with the promoter fragments were tested by plating yeast on SD/−Ura/−Leu medium containing the tested AbA concentration. Primers used are listed in Table S1.

4.11. EMSA

Conserved motifs in the CsXTH33 promoter were predicted using JASPAR. Oligonucleotide probes including predicted binding sites were synthesized with biotin at the 30-hydroxyl end of the sense strand. The probes used for EMSA are listed in Table S1. To validate the specificity of the shifted band, nuclear proteins were pre-incubated with non-labeled identical or mutated oligonucleotides for 10 min before adding labeled probes. The signals were detected using Chemiluminescent EMSA Kit (Beyotime, Jiangsu, China; GS009). The coding sequences of CsWRKY45/57 were inserted into the pET28a vector to generate the recombinant His-CsWRKY45 and His-CsWRKY57 plasmids. The plasmids were separately transformed into E. coli BL21 (DE3) cells. For protein expression, a single colony was inoculated into 5 mL LB containing 50 μg mL⁻¹ kanamycin, incubated overnight at 37 °C, diluted 1:100 in 500 mL fresh LB, and grown to OD₆₀₀ = 0.6. Expression was induced by adding 0.5 mM IPTG and incubating at 16 °C with shaking for 16 h. Cells were harvested by centrifugation (4 °C, 6000×g, 10 min), and proteins were purified using His-tag nickel columns.

4.12. Statistical analysis

Statistical analysis was performed by two-tailed Student's t-test, with significance defined as P < 0.05. The data represent means ± SD from at least three independent experimental replicates.

CRediT authorship contribution statement

Haifei Liang: Writing – review & editing, Writing – original draft, Validation, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation. Shengchao Liu: Methodology, Data curation. Tongliang Yan: Methodology, Data curation. Huan Wang: Validation, Software. Jiaming Li: Methodology, Investigation. Haidong Chen: Software, Investigation. Xueyong Yang: Resources, Methodology. Hai Zhou: Writing – review & editing, Conceptualization. Suhua Li: Writing – review & editing, Methodology. Jianbin Yan: Writing – review & editing, Supervision, Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

The raw transcriptome data for anthers were deposited in the Sequence Read Archive (SRA) of NCBI under BioProject number PRJNA1366062.