Article https://doi.org/10.1038/s41467-023-35936-z Pollen tube emergence is mediated by ovary-expressed ALCATRAZ in cucumber Received: 28 October 2021 Zhihua Cheng1,4, Xiaofeng Liu1,4, Shuangshuang Yan1, Bin Liu2, Yanting Zhong1, Weiyuan Song1, Jiacai Chen1, Zhongyi Wang1, Gen Che1, Liu Liu1, Ao Ying1, Accepted: 9 January 2023 Hanli Lv1, Lijie Han1, Min Li1, Jianyu Zhao1, Junqiang Xu3, Zhengan Yang3, Zhaoyang Zhou 1 & Xiaolan Zhang 1 Check for updates Pollen tube guidance within female tissues of flowering plants can be divided into preovular guidance, ovular guidance and a connecting stage called pollen 1234567890():,; 1234567890():,; tube emergence. As yet, no female factor has been identified to positively regulate this transition process. In this study, we show that an ovary-expressed bHLH transcription factor Cucumis sativus ALCATRAZ (CsALC) functions in pollen tube emergence in cucumber. CsALC knockout mutants showed diminished pollen tube emergence, extremely reduced entry into ovules, and a 95% reduction in female fertility. Further examination showed two rapid alkalinization factors CsRALF4 and CsRALF19 were less expressed in Csalc ovaries compared to WT. Besides the loss of male fertility derived from pre- cocious pollen tube rupture as in Arabidopsis, Csralf4 Csralf19 double mutants exhibited a 60% decrease in female fertility due to reduced pollen tube dis- tribution and decreased ovule targeting efficiency. In brief, CsALC regulates female fertility and promotes CsRALF4/19 expression in the ovary during pol- len tube guidance in cucumber. Seed production is critically important for sexual propagation in guidance (guidance from funiculus to micropyle and embryo sac) agricultural crops. Successful delivery of sperm cells (contained within (Fig. 1c)5,7,8. Between preovular and ovular guidance, there is a crucial the pollen tube) to female gametes enclosed inside the ovary is a transition phase called pollen tube emergence, in which pollen tubes prerequisite for seed genesis1–3. After germinating on the stigma, pol- exit from the transmitting tract and move onto the septum sur- len tubes extend and navigate along the female reproductive tract face (Fig. 1c). (transmitting tract: TT) towards ovules1,4 (Fig. 1a). During this process, Accurate pollen–pistil interactions occur during the long journey the pollen tube perceives signal substances from both female spor- from stigma to ovules, functional disruption of either male gameto- ophytic and gametophytic tissues of the pistil, which guide it towards phytes or female reproductive tissues will bring obstacles to oriented the embryo sac, where the pollen tube penetrates into the synergid pollen tube growth, and thus lead to failure of double fertilization9. and releases two sperm cells, with one fusing with the egg cell and the Upon landing on the stigma, Arabidopsis compatible pollen grains other with the central cell5,6. Pollen tube guidance can be divided into germinate via precise interactions between the pollen and pistil preovular guidance (before entering into the ovary) and ovular gui- involving POLLEN COAT PROTEIN B-class peptides and ANJEA- dance (Fig. 1b); the latter step is further divided into funicular guidance FERONIA receptor kinase complex10. In tomato (Lycopersicon escu- (guidance from the septum surface to the funiculus) and micropylar lentum), LeSTIG1, a cysteine-rich protein secreted from the pistil, can 1 State Key Laboratories of Agrobiotechnology, Joint International Research Laboratory of Crop Molecular Breeding, Beijing Key Laboratory of Growth and Developmental Regulation for Protected Vegetable Crops, Department of Vegetable Sciences, China Agricultural University, 100193 Beijing, China. 2Centre for Research in Agricultural Genomics (CRAG), CSIC-IRTA-UAB-UB, Campus Universitat Autònoma de Barcelona, Bellaterra 08193, Spain. 3College of Horticulture and Landscape, Yunnan Agricultural University, 650201 Kunming, Yunnan, China. 4These authors contributed equally: Zhihua Cheng, Xiaofeng Liu. e-mail: zyzhou@cau.eu.cn; zhxiaolan@cau.edu.cn Nature Communications | (2023)14:258 1 Article https://doi.org/10.1038/s41467-023-35936-z bind to the pollen-specific receptor kinases LePRK1 and LePRK2 to receptors glycosylphosphatidylinositol-anchored proteins (GPI-APs), stimulate pollen tube growth11. Arabidopsis pollen tube-specific LORELEI-like-GPI-anchored protein 2 (LLG2) and LLG316, are demon- cysteine-rich peptides (CRPs), Rapid Alkalinization Factor 4 (RALF4), strated to control pollen tube integrity and growth. Pollen tubes in and RALF1912,13, together with their pollen tube located Catharanthus double mutant ralf4 ralf1912,13, anx1 anx214,15, bups1 bups212, llg2 llg316 all roseus Receptor Like Kinase 1 Like (CrRLK1L) receptors BUDDA’S burst precociously at stigma or not beyond the style, resulting in PAPER SEAL1 (BUPS1)/ BUPS212, and ANXUR1 (ANX1)/ ANX214,15, and co- extension cutoff at preovular stage. Nature Communications | (2023)14:258 2 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 1 | CsALC expression pattern in reproductive organs and defective female placenta, sep septum, ov ovule. Scale bars = 200 μm. The experiment was repeated fertility of Csalc mutants by CRISPR-Cas9. a–c Schematic diagram of pollen tube three times with similar results. k Mutation forms of three homozygous T1 Csalc guidance during the process of fertilization in flowering plants. a Schematic dia- mutant lines by CRISPR-Cas9 system. The target sequence and PAM are marked by gram of a flower showing the pollen tube penetrates through the stigma and style, a black line and a red line, respectively. Deletions/ insertions are indicated as red enters the ovary, and targets an ovule. b Two phases of pollen tube guidance: dashes/ nucleotides, and corresponding nucleotide sizes are marked after the preovular guidance and ovular guidance. c One pollen tube emerges from the sequence as a minus or plus, respectively. The right dotted box shows the corre- transmitting tract (pollen tube emergence process) by ovule attraction, then sponding protein premature termination. The green/gray boxes indicate the extends towards the embryo sac under funicular guidance and micropylar gui- protein-coding region and missense sequences resulting from frameshift muta- dance. d, e qPCR analysis of CsALC expression in different cucumber organs. R tions, respectively. The orange box in WT represents the bHLH conserved domain roots, S stems, T tendrils, YL young leaves, MB male flower buds, FB female flower in CsALC. aa amino acid. l Seed number of wild type (WT) and Csalc mutant lines by buds, MF male flowers, FF female flowers, O ovaries at anthesis, TT ovary trans- self-pollination. Pictures represent the mature fruit, its longisection, and all plump mitting tract at anthesis, Ovu ovules at anthesis, Po pollens at anthesis, PT pollen seeds harvested from it, with the corresponding number at the bottom. Scale tubes at 3 h after germination. The vertical axis refers to the relative CsUBI bars = 5 cm. m Reciprocal cross between Csalc #1 mutant and WT. Scale bars = 5 cm. (CsaV3_5G031430) expression. n = 3 biologically independent samples. Error bars The experiments were repeated three times with similar results (l, m). represent mean ± SD. f–j In situ hybridization of CsALC in cucumber flower buds. f A n Quantitative analysis of seed numbers from (l, m). Upmost/lowest lines indicate male flower bud at stage 9. g, h Longitudinal sections of female flower buds at stage the maximum/minimum values, box limits indicate the upper and lower quartiles, 8-1 (g), and 8-2 (h). i Ovary transection of a female flower bud at stage 9. j Ovary and lines in boxes indicate the median values of these data. From left to right, n = 5, transection of a female flower bud at stage 8 hybridized with sense probe of CsALC 11, 5, 10, 9, 5, 7, 6, 5, 8 mature fruits. ***p < 0.001, compared to WT (two-sided as the negative control. se sepal, pe petal, sta stamen, sti stigma, ca carpel, pla Student’s t-test). “Ovular guidance” attracts one pollen tube to its target ovule, ALC, functioning in fruit opening, was searched in Cucurbit Genomics forming a specific one-to-one relationship5. Besides Arabidopsis MAGA- Database (CuGenDB: http://cucurbitgenomics.org/), and named as TAMA1 (MAA1: an unknown type protein) and MAA3 (a kind of CsALC (Csa2G356640.1). ALC was reported to be the closest relative to helicase)8,17, several attractant peptides produced from female gameto- SPATULA (SPT)29. Phylogenetic analysis showed there were SPT lineage phytes have been identified to govern ovular guidance, including the and ALC lineage within core eudicots (Supplementary Fig. 1a). CsALC Gramineae-specific small peptide EGG APPARATUS1 (EA1) in maize18,19, gene contains six exons and five introns, and its encoded protein is defensin-like CRPs LUREs and XIUQIUs in Torenia fournieri20 and located in the nucleus, with 379 amino acids and a typical bHLH Arabidopsis21,22. In Arabidopsis, the leucine-rich repeat receptor-like domain (Supplementary Fig. 1b–d). qPCR analysis showed that CsALC kinase (LRR-RLK) PRK6, localized at the pollen tip, was identified as the transcripts accumulate predominately in reproductive organs, receptor for LURE1s23. Unlike that knock-down of EA1 gene resulted in including male flowers and female flowers (Fig. 1d). During fertiliza- ovule sterility in maize, neither Arabidopsis allure1 septuple mutants nor tion, CsALC was expressed in transmitting tract (TT), ovules, pollens, prk6 mutant exhibited defects in fertility. Simultaneous knockout of and pollen tubes (Fig. 1e). In situ hybridization further localized CsALC both 7 LURE1s and 4 XIUQIUs led to only a 20% decrease in female transcripts in petals, stamens, and degenerated carpel primordia in fertility, indicating the redundant roles between LURE1s and XIUQIUs, male flower buds (Fig. 1f). In female flower buds, there was substantial and additional attractants in Arabidopsis micropylar guidance22,24. enrichment of CsALC signals in developing carpels, especially placenta Pollen tube emergence as the transition phase between preovular and stigmas, as well as in petals and degraded stamen primordia and ovular guidance refers to the first re-orientation of pollen tubes (Fig. 1g, h). Ovary transverse sections showed that compared to the extending from the transmitting tract to the septum surface before negative control, prominent CsALC signals were localized in ovules, proceeding onto funiculus, which was speculated to be affected by placenta, and septum (two parts where transmitting tracts are gener- ovules to some extent5. In Arabidopsis, pollen tubes that lost the ated for fertilization in cucumber) at female flower bud stage (Fig. 1i, j). function of two endoplasmic reticulum-localized cation/proton exchangers CHX21 and CHX23 exhibited normal growth but were Knockout of CsALC resulted in reduced female fertility in unable to emerge from the transmitting tract, showing a defect in cucumber pollen tube guidance25. Recently, pollen tube-expressed RALF6, 7, 16, To gain insight into CsALC function in cucumber, the first exon of 36, and 37 were shown to interact with female tissue-located receptor- CsALC was targeted by CRISPR-Cas9 gene-editing system (Supple- like kinases FERONIA, ANJEA, and HERCULES RECEPTOR KINASE 1 mentary Fig. 1b). Three homozygous frameshift mutant lines were (HERK1) to control the emergence of a single pollen tube at the septum obtained for further characterization (Fig. 1k). Compared to the wild for each ovule26. The multiple mutants of RALFs or RLK genes showed type (WT), Csalc mutant lines exhibited similar plant growth and fruit polytubey phenotype (arrival of multiple pollen tubes to one ovule)26. length (Supplementary Fig. 2a–c). However, the seed number per fruit Although it was proposed there exists long-range attraction from in Csalc mutant lines was dramatically reduced after self-pollination ovules, no female sporophytic genes have been identified to positively (Fig. 1l), and the few seeds were randomly distributed in the fruit regulate pollen tube emergence in planta so far5. chamber of Csalc mutants. Considering the higher expression of CsALC ALCATRAZ (ALC) encodes a bHLH transcription factor that med- in male gametophytes (Fig. 1e), we first investigated the pollen devel- iates fruit opening by driving the formation of a separation layer at the opment and viability by Alexander staining and in vitro germination valve/replum border that constitutes the dehiscence zone with the assays. As shown in Supplementary Fig. 2d, Csalc pollen grains were lignified layer in Arabidopsis27. Cucumber (Cucumis sativus L.) is an morphologically normal and able to germinate and grow on a medium. important vegetable crop bearing pepo fruit cultivated worldwide for Quantitative analysis showed the slightly reduced viability of Csalc over 3000 years28. In this study, we found cucumber ALC homolog pollens (90.4 ± 1.6%, 90.6 ± 1.7%, 90.6 ± 1.1%, p < 0.001 versus WT (CsALC) functioned specifically in pollen tube emergence, regulated 97.9 ± 1.4% by Alexander staining assay; 94.9 ± 1.3% p < 0.01, female fertility, and promoted TT-expressed CsRALF4/19 during pollen 94.9 ± 0.7% p < 0.001, 95.2 ± 1.8% p < 0.05 versus WT 99.2 ± 0.5% by tube growth in cucumber. in vitro germination assay) (Supplementary Fig. 2e, f). This mild decrease in pollen viability has a negligible impact on the cucumber Results seed set since an excess of pollens is usually available during open CsALC is highly expressed in reproductive organs of cucumber pollination. To identify putative fruit developmental regulators in the fleshy pepo Next, reciprocal crosses between Csalc mutants and WT were fruit, the cucumber homolog of Arabidopsis bHLH transcription factor performed. As shown in Fig. 1m, n, when Csalc pollens were pollinated Nature Communications | (2023)14:258 3 Article https://doi.org/10.1038/s41467-023-35936-z to WT, high seed set was acquired (244.6 ± 35.3, 269.6 ± 39.5, Csalc mutants is from the defect in pollen tube emergence from 277.0 ± 49.6 seeds per fruit), similar to that in WT self-crossing lateral TT to terminal TT. No difference was observed in tissue (292.5 ± 70.9). In contrast, when Csalc mutants were fertilized with WT structure and cell morphology between Csalc mutant and WT pollens, a drastically lower seed set (17.2 ± 22.1, 15.3 ± 19.2, 12.5 ± 13.3, transmitting tract, while the extracellular matrix (ECM) content p < 0.001 versus WT self-crossing) was obtained, equivalent to those in in ovary TT detected by Alcian blue staining, an indicator of Csalc mutant self-pollination (18.3 ± 13.9, 13.7 ± 15.9, 11.0 ± 9.6, acidic polysaccharides, displayed a reduction in Csalc mutants p < 0.001 versus WT self-crossing), resulting in an average of 95% (Fig. 2x–y1”). ECM, containing various glycoproteins, has been pro- (respective 94.1%, 94.8%, 95.7%) decrease in seed number per fruit posed to mediate the attraction, nutrition, and stimulation of pollen compared to WT. The above data suggested Csalc mutants have a tube growth in Arabidopsis4. remarkable defect in female fertility. When a heterozygous+/Csalc mutant self-crossing, the +/+, +/Csalc, Csalc/Csalc in the 213 progeny CsRALF4/19 transcripts declined at the transmitting tracts in seedlings showed segregation of 53:117:43, approximating the ratio of Csalc mutants 1:2:1, suggesting that the defect in Csalc mutant is localized to female To explore putative downstream targets of CsALC during fertiliza- sporophytic tissues. tion, transcriptomic analyses were performed on WT and Csalc ovaries at 32 h after pollination (Supplementary Fig. 4a). Differen- Pollen tubes displayed difficulty in targeting ovules in Csalc tially expressed genes (DEGs) were identified using the false dis- mutants covery rate <0.05 and fold change larger than 2 as thresholds. A total Successful fertilization in flowering plants requires accurate pollen of 8 and 20 genes were up- and down-regulated in Csalc ovaries, tube guidance in pistils, which can be divided into preovular guidance, respectively (Supplementary Fig. 4b, c). Among them, a gene ovular guidance, and a transition phase called pollen tube emergence5. encoding rapid alkalinization factor (Csa6G484570, CsRALF19) was As shown in Fig. 2a, we illustrated these processes in cucumber. Unlike remarkably decreased (log2 fold change = 3.46, P = 3.49E−04) in the dehiscent silique with two cavities in Arabidopsis, cucumber pro- Csalc mutants (Supplementary Fig. 4d). qPCR analysis using unpol- duces fleshy fruit, which is generally developed from three dorsally linated ovary at anthesis showed that CsRALF19 and its paralogous fused carpels, and the ovules encased by flesh are born at both edges gene CsRALF4 (Csa6G199790) were expressed significantly lower in of each carpel. The distinct carpel fusion form in cucumber gives rise Csalc mutants than WT (Fig. 3a, b). Phylogenetic analysis indicated to the unique morphology of the transmitting tract (TT), which is that the two peptides clustered with Arabidopsis RALF4/19, which divided into medial TT, lateral TT, and terminal TT in ovary30 (Fig. 2a). function in pollen tube integrity (Supplementary Fig. 5a)12,13. Unlike After entering into ovaries, pollen tube bundles firstly extend along the pollen-specific expression of Arabidopsis RALF4/19, CsRALF4/19 medial TT to the nearby lateral TT, then after a short extension, pollen also have a certain degree of expression in the ovary, indicating their tubes exit from lateral TT to terminal TT, resembling Arabidopsis putative new function in female tissues in cucumber. CsRALF4 and pollen tube emergence process in which pollen tubes start to depart CsRALF19 both contain a single exon and encode a 116- and a 112- from the main growth axis to target ovules (Fig. 2a). Therefore, this amino-acid-long peptide, respectively (Supplementary Fig. 5b). first re-orientation in cucumber is designated as “pollen tube emer- qPCR analysis showed that CsRALF4/19 were highly expressed in gence”. The second re-orientation from terminal TT to funiculus is mature stamens (Fig. 3c, d). However, unlike the differential expres- depicted as funicular guidance (Fig. 2a). After extremely short navi- sion of CsRALF4/19 in female tissues of Csalc mutants and WT, the gation along the funiculus, a third re-orientation is likely triggered by expression of these two genes in stamens was unchanged between micropylar guidance cues at a much closer range from micro- Csalc mutants and WT (Supplementary Fig. 6a, b). Tissue subdivision pyle (Fig. 2a). showed that CsRALF4/19 transcripts were enriched in pollen grains and To dissect the defective step of fertilization in Csalc mutants, pollen tubes. Notably, the expression of CsRALF4/19 in the transmitting Aniline blue staining was used to visualize pollen tube elongation tract was higher than that in ovules (Fig. 3e, f). To compare the and penetration in pistils. Compared to the WT ovary without pol- expression pattern in WT and Csalc mutants, in situ hybridization of lination (Fig. 2b), obvious fluorescence signals showed the move- CsRALF4/19 and CsALC was performed on the transections of ovaries ment trail of pollen tubes from the stigma to style, ovary chamber, before pollination and after fertilization. In WT unpollinated ovaries, and finally the ovules in WT self-crossing or hybridized with Csalc CsRALF4/19 and CsALC transcripts were enriched at transmitting tract pollens (Fig. 2c, d). In Csalc ovaries, pollen tubes spread normally and ovules (Fig. 3g–i’), and CsRALF19 exhibited much higher expres- downward along the medial transmitting tract and maintained sion than CsRALF4, consistent with qPCR results (Fig. 3e, f). In Csalc similar fluorescence intensity with those in WT (Fig. 2e, f). By local unpollinated ovaries, CsRALF4/19 transcript signals seemed weaker at longitudinal sections, abundant pollen tubes could be observed TTs compared to WT (Fig. 3k, l’). The TT identity and morphology emerging from lateral TT and entering ovules in WT (Fig. 2g–g’), appeared unaltered, and the TT marker gene SPATULA (CsSPT) was while rare in Csalc mutants (Fig. 2h–j’). Quantitative analysis showed expressed normally in Csalc ovaries compared to WT30 (Supplemen- that 83.5 ± 7.7% pollen tubes successfully targeted ovules in WT, tary Fig. 6c). Considering the almost abolished pollen tube extension while only 7.1 ± 1.4%, 10.8 ± 2.4%, 10.1 ± 0.7% in Csalc #1, #2, #3 to terminal TT in Csalc ovaries (Fig. 2g–w), we speculated CsALC may mutants, respectively (Fig. 2k). Transverse sections of stigmas and regulate pollen tube emergence at least partially through CsRALF4/19 styles showed comparable fluorescence intensity of pollen tubes at function in cucumber. While after fertilization, transcripts of all three the stigma and style transmitting tract in Csalc mutants to that in WT genes showed reduced accumulation (Fig. 3m–o’), indicating the (Fig. 2l–s). While, the fluorescence signals in Csalc mutants were potential specific roles of these genes in ovary TTs during the fertili- almost abolished at the terminal TT where funiculus attached zation process. (Fig. 2t–w), and pollen tubes were barely observed approaching and penetrating into ovules as compared to WT (Fig. 2t’–w'). By a time- Positive regulation of CsALC on CsRALF4/19 expression course observation (2, 5, 10, 20 h), we found WT pollen tubes grew in To examine any direct regulation of CsALC on CsRALF4/19 expression, Csalc transmitting tract at a similar rate to those in WT (Supple- the cis-elements in CsRALF4/19 genes were analyzed. The bHLH tran- mentary Fig. 3a). However, pollen tubes at lateral TT extended scription factor was shown to bind to the G-box of target genes31. almost downward, and rarely spread to terminal TT or turned G-box elements in the promoter of 28 DEGs between Csalc mutants growth direction to where ovules located in Csalc mutants (Sup- and WT were searched on the PlantCARE website (http:// plementary Fig. 3b, c). Therefore, the decreased female fertility in bioinformatics.psb.ugent.be/webtools/plantcare/html/) Nature Communications | (2023)14:258 4 Article https://doi.org/10.1038/s41467-023-35936-z (Supplementary Data 4). CsRALF4 has one G-boxes in the promoter sequence (1811 bp for CsRALF4 and 3326 bp for CsRALF19), both region, and one in the genomic region, and CsRALF19 has two G-box in pCsRALF4:LUC and pCsRALF19:LUC could be activated upon co- the promoter region, and one in the genomic region (Supplementary infiltration of CsALC protein in Nicotiana benthamiana leaves, Fig. 7a). Yeast one hybrid (Y1H) assay showed that CsALC was unable to increasing to 2.03 and 1.66 fold, respectively (Supplementary Fig. 7c). bind to any G-box containing fragments in CsRALF4 and CsRALF19 These results indicated the positive regulation of CsALC on CsRALF4/19 (Supplementary Fig. 7b). However, using the complete promoter expression in cucumber, consistent with the downregulation of Nature Communications | (2023)14:258 5 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 2 | Pollen tubes are less likely to reach ovules in the Csalc ovary. a Schematic k Quantitative analysis of pollen tubes entering ovules in WT and Csalc mutants. diagram of pollen tube paths along the female reproductive tract in cucumber, in n = 3 biologically independent samples. Error bars represent mean ± SD. which the ovary transmitting tract (TT) consists of medial TT [pale green (long- ***p < 0.001, compared to WT (two-sided Student’s t test). l–w’ Transverse sections itudinal section) and orange (transverse section)], lateral TT [gray (longitudinal displaying the diminished distribution of pollen tubes at terminal transmitting tract section) and light brown (transverse section)] and terminal TT (dark brown). in Csalc ovaries, and pollen tubes unable to target ovules. l–o stigma, p–s style, b–f Longitudinal sections showing pollen tubes extend along the transmitting tract t–w the top quarter of the ovary. t’–w’ Enlarged view of red dotted boxes in (t–w). from top to bottom in WT and Csalc ovaries. Composite images were integrated The ovules are delineated by yellow dots. Orange arrows refer to pollen tubes and spliced from local images taken separately. Scale bars = 500 μm. g–j’ Typical entering ovules. Scale bars = 500 μm. x–y1” Cell morphology and structural com- images showing WT pollen tubes enter ovules in WT and Csalc mutants by long- position of the transmitting tract in styles (x, y) and ovaries (x1, y1) of WT (top row) itudinal sections. Red arrow heads indicate pollen tubes targeting ovules. g’–j’ and Csalc mutant (bottom row) by Alcian blue staining. (x1’, y1’) and (x1”, y1”) are Enlarged view of the orange boxes in (g–j). Targeted ovules are delineated by enlargement of red and orange boxes in (x1) and (y1), respectively. Scale yellow dots, and pollen tubes are marked by red lines. Scale bars = 500 μm. bars = 200 μm. CsRALF4/19 transcripts in Csalc mutants as detected by qRT-PCR and at length of 172.0 ± 71.6 μm and 183.6 ± 67.1 μm (p < 0.001, Fig. 6a, c). in situ hybridization (Fig. 3). Similarly, Csralf19 pollen tube length was 115.2 ± 67.2 μm (#1) and 122.0 ± 77.3 μm (#2) versus WT 233.0 ± 80.2 μm (p < 0.001, Fig. 6a, d). Knockout of both CsRALF4 and CsRALF19 led to reduced female In addition, the weakened fluorescence intensity by Aniline blue fertility and a complete loss of male fertility staining indicated poor extension of Csralf4 and Csralf19 pollen To characterize the function of CsRALF4/19 in cucumber, single tubes in WT ovaries, especially at lateral- and terminal-TT (Fig. 6e). In mutants Csralf4 (obtained by targeting the CsRALF4 gene) and Csralf19 Csralf4 Csralf19 double mutants, pollen tubes burst immediately (separated from Csralf4 Csralf19 double mutants), as well as double upon germination in vitro and on the stigmatic surface, which mutant Csralf4 Csralf19 (generated by simultaneously targeting resembles the Arabidopsis ralf4 ralf19 mutant phenotypes (Fig. 6a, CsRALF4 and CsRALF19), were obtained by CRISPR-Cas9 system. Two e), suggesting the conserved function of CsRALF4/19 in pollen tube homozygous lines for each mutant type were chosen for further integrity32. characterization. Sequencing data showed that all mutant lines pos- sessed frameshift mutations that resulted in protein premature ter- CsALC was selected during cucumber domestication mination (Figs. 4a, d, and 5a). Both Csralf4 and Csralf19 single mutants Our findings showed the important function of CsALC in cucumber showed a reduced seed set when self-crossing (seed number per fruit: seed production and sustainable propagation. During cucumber Csralf4 mutants 219.6 ± 49.6, 203.0 ± 42.0, p < 0.05 versus WT domestication, the fruit is elongated while seed number per unit fruit 338.2 ± 101.6; Csralf19 mutants 217.0 ± 37.7, 227.7 ± 43.9, p < 0.05 ver- length is generally decreased33. To explore any selection footprint on sus WT 295.4 ± 61.7). When Csralf4 or Csralf19 mutant ovaries polli- CsALC, the genomic region of CsALC plus its 3 kb promoter sequences nated by WT pollens, seed numbers (Csralf4 mutants as maternal (Chr02: 16738402–16744826) was analyzed in the published 115 parents: 337.9 ± 78.4 and 296.9 ± 49.0; Csralf19 mutants as maternal cucumber accessions consisting of five populations: Indian wild gen- parents: 317.9 ± 81.6 and 300.0 ± 62.3) were comparable to WT self- otypes (INW), Indian domestic (IND), Xishuangbanna (XSBN), East crossing, suggesting the female fertility in both single mutants is Asian (EA) and Eurasian (EU)28. The results showed that CsALC exhibited normal. However, when Csralf4/ Csralf19 mutants as pollen donors to sharply reduced nucleotide diversity (π) in modern varieties (EA and WT ovaries, the resultant seed set was significantly reduced EU), as compared to wild (INW) or semi-wild cucumber (XSBN) (Sup- (194.5 ± 62.8, 183.2 ± 38.5, p < 0.05 for Csralf4 mutants, and plementary Fig. 8a). We also calculated the pairwise difference in allele 214.4 ± 30.1, 226.5 ± 31.1, p < 0.05 for Csralf19 mutants) versus WT self- frequency (FST) between domesticated groups and wild population crossing, suggesting the defective male fertility in Csralf4 and Csralf19 INW, the results showed that CsALC was selected in Xishuangbanna, single mutants (Fig. 4b, c, e, f). Further, Csralf4 Csralf19 double mutants East Asian and Eurasian groups (Supplementary Fig. 8b), implying the completely lost male fertility and none seed was produced when as importance of CsALC in cucumber domestication and breeding. pollen donors (Fig. 5b). Intriguingly, when Csralf4 Csralf19 double mutants hybridized with WT pollens, the seed number per fruit Discussion (125 ± 51.3, 122.3 ± 63.8, p < 0.001) was significantly fewer than WT self- The ovary encloses and nourishes ovules, and ample evidence sug- pollination (295.4 ± 61.7), accounting for ~60% decrease in female fer- gested that the ovary must provide an environment that supports tility (Fig. 5c, d). The distribution of full seeds in the ovary chamber of pollen tube growth and guidance to the ovule34. Some female Csralf4 Csralf19 double mutants was scattered, similar to that in Csalc molecules, e.g. Ca2+ ions35, γ aminobutyric acid (GABA)36,37, mutants. The average seed number per unit fruit length (1 cm) in brassinosteroids38, carbohydrates4, and various peptides39, have Csralf4 Csralf19 double mutants was 12.7 ± 4.5 (p < 0.001) and 15.5 ± 7.1 been characterized to stimulate pollen germination and pollen tube (p < 0.001), significantly reduced as compared to WT (29.6 ± 7.0) growth in preovular guidance process. These early stages are gen- (Fig. 5c, e). Moreover, only 38.5 ± 5.9% (p < 0.001) and 37.7 ± 5.4% erally regulated by sporophytic genes such as those in the trans- (p < 0.001) ovules were observed to be targeted by pollen tubes in two mitting tract, e.g., Transmitting-Tissue-Specific Arabinogalactan gene Csralf4 Csralf19 double mutant lines, while the targeting ratio was (TTS) in tobacco40, SPT29, HECATEs (HEC1/2/3)41, HALF FILLED (HAF)42 76.7 ± 8.1% in WT (Fig. 5f, g). The pollen tube density at the lateral- and in Arabidopsis and CsSPT in cucumber30. In Csspt mutants, the dra- terminal-transmitting tract in Csralf4 Csralf19 ovaries was significantly matically decreased polysaccharide at the style transmitting tract led reduced versus WT (Fig. 5h–j). These data suggested a new role of to much less pollen tubes penetrating into ovaries and thus reduced CsRALF4/19 in female fertility in cucumber. seed set and female fertility30. Unlike that stimulant molecules pro- Besides, defects in male gametophytes were detected in Csralf4, mote the movement of pollen tubes at stigma or styles at preovular Csralf19 single mutants, and Csralf4 Csralf19 double mutants. Pollen guidance stage, the attractant cues secreted from ovules generally viability was significantly reduced in all mutants (Csralf4: 92.3 ± 2.2%, trigger the re-orientation of pollen tube during the ovular guidance 92.9 ± 1.1%; Csralf19: 92.7 ± 1.9%, 93.0 ± 1.4%; Csralf4 Csralf19: process (late stage of pollen tube guidance), e.g. the roles of CRP 83.2 ± 4.3%, 81.8 ± 4.7%; all p < 0.001) as compared to WT (96.6 ± 0.8%) peptide LUREs and synergid cells during short range attraction6,20,43. (Fig. 6a, b). WT pollen tubes ruptured after reaching 242.9 ± 89.3 μm Meanwhile, male factors such as receptor PRK6 and ion transporters length, while pollens tubes in two Csralf4 lines ruptured precociously CHX21/23 mediate pollen tube responsive behavior and competence Nature Communications | (2023)14:258 6 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 3 | Reduced transcript accumulation of CsRALF4 and CsRALF19 in Csalc normalized with the CsUBI (CsaV3_5G031430) expression (a-f). Error bars represent ovaries. a, b Expression of CsRALF4 (a) and CsRALF19 (b) in unpollinated ovaries at mean ± SD (a–f). g–o’ Transcript accumulation of CsALC (g, g’, m, m’), CsRALF4 anthesis of WT and Csalc mutants by qPCR. *p < 0.05, **p < 0.01 (two-sided Stu- (h, h’, k, k’, n, n’), and CsRALF19 (i, i’, l, l’, o, o’) at transmitting tract of unpollinated dent’s t-test). c, d CsRALF4 (c) and CsRALF19 (d) expression in different cucumber WT and Csalc ovaries (g–l’) and fertilized WT ovaries one day after anthesis (32 h organs. S stems, MB male flower buds at stage 1050; FB female flower buds at stage after pollination) (m–o’) by in situ hybridization. j–j’ SP6, Sense CsALC probe was 1050; O ovaries at anthesis; Sta stamens at anthesis. e, f CsRALF4 (e) and CsRALF19 (f) hybridized as the negative control. g’–o’ are enlargements of red boxes in (g–o). expression in female sporophytes and male gametophytes. TT ovary transmitting Red arrows indicate transcript signals at lateral- and terminal-TTs in WT before tract at anthesis, Ovu ovules at anthesis, Po pollens at anthesis, PT pollen tubes at pollination, and blue arrows refer to the weakened transcript signals at lateral- and 3 h after germination. The left/right vertical axis refers to the data of the blue/ terminal-TTs in Csalc mutants, or in WT after fertilization. Scale bars = 500 μm. The orange column, respectively, due to the magnitude difference of expression in experiment was repeated three times with similar results. different tissues. n = 3 biologically independent samples (a–f). The values were for ovular attractions23,25. However, none female sporophytic gene In this study, we identified a transmitting tract-expressed gene has been reported yet to positively regulate pollen tube CsALC that participates in pollen tube emergence in cucumber. re-orientation at the transition stage between preovular guidance Loss-of-function Csalc mutants displayed almost absence of pollen and ovular guidance, namely the pollen tube emergence process. tube distribution at terminal TT and thus nearly 95% reduction in Nature Communications | (2023)14:258 7 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 4 | Decreased seed set in Csralf4 and Csralf19 single mutants. a, d Mutation from frameshift mutations. aa amino acid. b, e Mature fruits, corresponding forms of two homozygous T1 Csralf4 lines (a) and two homozygous T1 Csralf19 longisection, and harvested seeds from reciprocal crosses between Csralf4 lines (d). Target sequences and PAMs are indicated below the black and red lines, mutants (b)/Csralf19 mutants (e) and WT. Scale bars = 5 cm. c, f Quantitative ana- respectively. Deletions/insertions are indicated as red dashes/nucleotides, and lysis of seed number from (b, e). Upmost/lowest lines indicate the maximum/ deletion/insertion sizes (nucleotides) are marked after the sequence as a minus or minimum values, box limits indicate the upper and lower quartiles, and lines in plus, respectively. The right dotted box shows the corresponding protein pre- boxes indicate the median values of these data. From left to right, n = 6, 5, 4, 7, 5, 5, mature termination. The blue/orange boxes indicate CsRALF4/CsRALF19 protein 7 mature fruits in (c) and 7, 5, 5, 7, 6, 6, 6 mature fruits in (f), respectively. *p < 0.05, coding region, respectively, and the gray indicates missense sequences resulting **p < 0.01 (two-sided Student’s t-test). female fertility (Fig. 1). The movement of pollen tubes from top to that the transmitting tract circumstance in Csalc mutants is altered, bottom within Csalc pistils appeared normal while extending pollen in which main component polysaccharides and two TT-expressed tubes within Csalc ovaries failed to turn towards ovules, showing a peptides CsRALF4 and CsRALF19 were reduced (Figs. 2 and 3). The defect in pollen tube emergence from lateral-TT to terminal-TT changed TT circumstance may affect the diffusion or distribution of (Figs. 2 and 7). Unlike the hollow pod in Arabidopsis, cucumber some unknown ovular attractant molecules at lateral and terminal bears the fleshy fruit, in which the cell–cell communication of TT in cucumber, causing unresponsive pollen tube behavior. pollen–pistil interactions mainly occurs at the transmitting tracts. Another possibility is that the absence of female-assisted molecules From the physiological and molecular evidence, we can speculate at TTs, like phytohormones, peptides, and glycoproteins, may blunt Nature Communications | (2023)14:258 8 Article https://doi.org/10.1038/s41467-023-35936-z the pollen tube perception for ovular signals or affect the direc- may modify pollen tube status to lose the directional growth tional growth of pollen tubes. The successful guidance for pollen capacity. It may be operated through pollen-specific unknown tubes not only requires the secreted attractants but also the com- receptors, or those genes involved in localized vesicle traffic, actin petency control from female tissues for pollen tube behavior34. It dynamics, or ion balance (e. g. Arabidopsis CHX21 and CHX23 in K+ was reported that an ovary-abundant chemical cue, methyl- homeostasis) that specify pollen tube tip growth and re- glucuronosyl arabinogalactan (AMOR), can promote pollen tube orientation44. competence to respond to ovular attractant LURE peptides in Tor- However, despite a similarly reduced extracellular matrix in enia fournieri34. The third possibility is that the mutant TT in Csalc Csspt ovaries, no pollen tube emergence defect was observed Nature Communications | (2023)14:258 9 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 5 | Reduced female fertility in Csralf4 Csralf19 double mutants. a Mutation right, n = 7, 6, 7 mature fruits in (d) and 8, 5, 6 mature fruits in (e), respectively. forms of two homozygous T1 transgenic Csralf4 Csralf19 lines. The target sequence f Aniline blue staining showing the ovule targeting in WT and Csralf4 Csralf19 mutant and PAM are indicated below a black and a red line, respectively. Deletions/inser- ovaries at 32 h after pollination with WT pollens by transverse (left) and longitudinal tions are indicated as red dashes/nucleotides, and deletion/insertion sizes sections (right). Yellow stars indicate ovules targeted by pollen tubes. Scale bars = (nucleotides) are marked after sequencing as a minus or plus, respectively. The 500 μm. g Quantitative analysis of pollen tubes entering ovules in (f). n = 5, 5, 4 right dotted box shows the corresponding protein premature termination. The biologically independent samples. h Pollen tubes distribution at lateral- and blue/orange boxes indicate CsRALF4/CsRALF19 protein coding region, respec- terminal-TT (pointed by green boxes and blue boxes, respectively) in ovary trans- tively, and the gray indicates missense sequences resulting from frameshift muta- ections from (f). Scale bars = 500 μm. I, j Quantitative analysis of pollen tube fluor- tions. aa, amino acid. b Mature fruits and corresponding longisection from escence intensity at lateral- (i) and terminal-TT (j) in (h). The fluorescence threshold reciprocal crosses between Csralf4 Csralf19 mutants and WT. The corresponding is 20–255. n = 39, 32, 36 lateral TTs in (i), and 235, 141, 120 terminal TTs in (j), seed numbers were shown at the bottom. Scale bars = 5 cm. c All plump seeds and respectively. For these Box-plots, upmost/lowest lines indicate the maximum/mini- fruit transections from WT self-crossing and Csralf4 Csralf19 hybridizing with WT mum values, box limits indicate the upper and lower quartiles, and lines in boxes pollens. Scale bars = 2 cm. d, e Quantitative analysis of seed number. From left to indicate the median values of these data. ***p < 0.001 (two-sided Student’s t-test). (Supplementary Fig. 9)30, suggesting the polysaccharides in TT may Interestingly, CsALC surrounding region is under-selected during not be the critical factor required for proper pollen tube re- the domestication process in cucumber (Supplementary Fig. 8). In wild orientation. CsSPT mainly functions in the development of style TT cucumbers, seedy fruits are advantageous for species reproduction that facilitates pollen tube extension from the stigma to the ovary. and survival. However, in cultivated cucumber, fruits with fewer seeds Mutation of CsSPT resulted in only upper portion ovules fertilized and are preferred for easy consumption. Thus, the CsALC gene may be thus a 60% reduction in seed set30. Considering Csspt Csalc double selected unconsciously during the long history of cucumber cultiva- mutants displayed a complete loss of female fertility derived from the tion. Notably, a segment in the first exon of CsALC (not the bHLH absence of TT, more severe than either Csspt or Csalc single domain) displayed no polymorphism among all the 115 cucumber mutants30, therefore, we conclude that CsALC and CsSPT have germplasms examined, which encoded a serine that is highly con- divergent functions in pollen tube guidance, while redundant roles in served in the Cucurbitaceae family. The functions of this specific ser- TT development in cucumber. ine, as well as the key domestication regions of CsALC await further Besides the extremely high enrichment in pollens, CsRALF4/19 studies. were also expressed in female reproductive tissues in cucumber (Fig. 3). Knockout of both CsRALF4 and CsRALF19 resulted in pollen Methods tube rupture prematurely and reduced female fertility from dimin- Plant materials and growth conditions ished pollen tube distribution at lateral and terminal-TT, and Cucumber (Cucumis sativus L.) inbred line XTMC (North China type) decreased ovule targeting efficiency (Fig. 5). These data suggested that was used in this study. Seeds were soaked and germinated in the dark in addition to the conserved function in pollen tube integrity by pollen- at 28 °C, then grown in the soil in a growth chamber under a 16 h/8 h located CsRALF4/19 peptides12,13, the transmitting tract-expressed and 28 °C/22 °C day/night regime until the roots fully developed and CsRALF4/19 peptides mediate pollen tube growth and extension in true leaves expanded. Seedlings were then transplanted to a standard the cucumber ovary, probably by interaction with pollen tube- greenhouse in the experimental field of China Agricultural University, expressed RALF receptors similar to those in pollen tube integrity Beijing. Water and fertilizer management, and pest control were per- control12 (Fig. 7). Expression analysis indicated the CRPs CsRALF4/19 formed according to standard protocols. transcripts were decreased in Csalc ovaries (Fig. 3). Although no direct interaction was found by yeast one hybrid, in vivo assay of N. ben- Gene cloning and structural analysis thamiana showed the positive regulation of CsALC on CsRALF4/19 An 1140-bp fragment containing the complete CsALC coding sequence expression, perhaps with the assistance of other cofactors (Supple- (CDS) was amplified from male flower buds and open male flowers, and mentary Fig. 7). Interestingly, the expression of CsRALF4/19 was the 351-bp length CsRALF4 CDS and 339-bp length CsRALF19 CDS were unchanged in the stamen of Csalc mutant (Supplementary Fig. 6), obtained in the same way with gene-specific primers (Supplementary implying the positive regulations of CsRALF4/19 by CsALC is female Data 1), based on sequence information in CuGenDB (http:// specific. Considering the crossing data of male sterility and reduced cucurbitgenomics.org/). The gene structure was analyzed using the female fertility in Csralf4 Csralf19 double mutants (Figs. 5 and 6), online software GSDS 2.0 (http://gsds.gao-lab.org/). CsRALF4/19 may have distinct functions in male and female cells in cucumber. Notably, the phenotype of seed set decrease in Csralf4 Phylogenetic trees construction Csralf19 double mutants was not as severe as that in Csalc mutants, To retrieve protein and CDS sequences for phylogenetic analysis, indicating there are additional downstream genes of CsALC besides Arabidopsis SPT and ALC were used as queries to search against the CsRALF4/19 during pollen tube guidance (Fig. 7). genome database Phytozome (https://phytozome.jgi.doe.gov/pz/ As the paralog of SPT, ALC originated from a gene duplication portal.html) and National Center for Biotechnology Information event prior to the emergence of the Brassicaceae29. In the dehiscent (https://blast.ncbi.nlm.nih.gov/Blast.cgi), using BLAST with a para- silique of Arabidopsis, ALC controls fruit opening by promoting cell meter of 1 × 10−10. Full-length protein sequences of sampled species differentiation at separation layers27. In the berry fruit of Solanaceae, were aligned using the program Muscle 3.6 (http://www.drive5.com/ transient knock-down of both ALC and SPT genes resulted in increased muscle/), and the resulting alignment was manually adjusted in Gen- fruit lignification in tomato and pepper45. In the fleshy pepo fruit of eDoc 3.2. Based on the protein alignment, a DNA matrix was produced Cucurbitaceae, our study showed the unique role of CsALC in female by PAL2NAL (http://www.bork.embl.de/pal2nal/). Protein sequences of fertility via mediating pollen tube emergence in cucumber (Fig. 2, RALFs from Arabidopsis and cucumber were obtained using the pro- Fig. 7). Therefore, ALC has undergone functional divergence in dif- tein BLAST search in NCBI, TAIR (https://www.Arabidopsis.org/), and ferent fruit types. From an evolutionary point of view, however, ALC CuGenDB. A phylogenetic tree was generated using the Neighbor- homologs from different species commonly contribute to sexual joining method and the bootstrap analysis with 1000 replicates in the reproduction success, such as participating in fruit opening and thus MEGA-X. Multiple sequence alignment was performed using CLUS- facilitating seed dispersal in Arabidopsis46, or controlling pollen tube TALW in MEGA-X. Accession numbers of all sequences used for phy- guidance and enabling seed genesis in cucumber. logenetic analysis are listed in Supplementary Data 2. Nature Communications | (2023)14:258 10 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 6 | Abolished male fertility in Csralf4 Csralf19 double mutants. a Alexander (two-sided Student’s t test). For these Box-plots, upmost/lowest lines indicate the staining of pollens (top row) and in vitro pollen germination assay (bottom row) of maximum/minimum values, box limits indicate the upper and lower quartiles, and Csralf4, Csralf19 and Csralf4 Csralf19 mutants. Scale bars = 100 μm. b–d Quantitative lines in boxes indicate the median values of these data. e Aniline blue staining analysis of pollen viability (b) and pollen tubes length (c, d) from (a). n = 9, 3, 3, 3, 3, showing the extension of pollen tubes by longitudinal (up) and transverse sections 3, 3 biologically independent samples in (b). Error bars represent mean ± SD. n = 115, (bottom) in Csralf4, Csralf19, and Csralf4 Csralf19 mutants. Composite images were 71, 48 pollen tubes in (c) and 240, 97, 83 pollen tubes in (d), respectively. ***p < 0.001 integrated and spliced from local images taken separately. Scale bars = 500 μm. Subcellular localization of CsALC Quantitative real-time PCR The full-length CDS of CsALC without termination codon was fused The young leaves, stems, tendrils, male flower buds at stage 10, female into the pUC-SPYNE vector to generate a CsALC-GFP in-frame fusion flower buds at stage 10, male flowers, female flowers, ovaries and protein. The empty pUC-SPYNE vector was used as a positive control. pollens, pollen tubes at 3 h after germination, transmitting tract, Constructs were bombarded into onion epidermal cells using a PDS- ovules were frozen in the liquid nitrogen and stored at −80 °C until use. 1000/He particle delivery system (Bio-Rad) as described previously47. Unless otherwise indicated, these reproductive tissues and organs After incubation in the dark at 22 °C for 24 h, fluorescence signals were were collected at anthesis (the day for pollination). Transmitting tracts captured using a confocal laser-scanning microscope (Leica SP8, Ger- and ovules were isolated from the middle parts of the ovary. Pollens many) with 488 nm excitation wavelength. from four male flowers were eluted in RNA-extracted lysis buffer. Nature Communications | (2023)14:258 11 Article https://doi.org/10.1038/s41467-023-35936-z Fig. 7 | A working model for CsALC and CsRALF4/19 in pollen tube guidance in 19 peptides, resulting in an approximately 95% reduction in seed number per fruit. cucumber. In WT ovary, pollen tubes (in red) extend successively along the In Csralf4 Csralf19 mutant, the reduced pollen tube extension and decreased ovule medial-, lateral- and terminal-TT and eventually enter ovules, producing hundreds targeting efficiency leads to about a 60% reduction of seed number in fruit. The of seeds per fruit in cucumber. In Csalc mutant, pollen tubes fail to re-orient to much fewer seeds in Csalc mutants than in Csralf4 Csralf19 mutants indicate there ovules and are blocked at the lateral TT, partially due to the decrease of CsRALF4/ must be other downstream target genes. Pollens were cultured in liquid pollen germination medium [15% (w/v) Cucumber transformation sucrose, 0.01% (w/v) boric acid, 1 mM Ca(NO3)2, pH 6.5] for 3 h and The CRISPR-Cas9 knockout vector pKSE402 with GFP fluorescence after centrifugal collection, pollen tubes were re-suspended in RNA- screening marker is provided by Prof. Huang Sanwen. The guide RNA extracted lysis buffer. Total RNA was extracted with Trizol reagent as (gRNA) is designed in the Cas-Designer website (http://www.rgenome. described in the manufacturer’s instructions (Promega, China, http:// net/cas-designer/). Single target primers (for mutation of CsALC, promega.bioon.com.cn/), and cDNA was synthesized using FastQuant CsRALF4) and double target primers (For simultaneous mutation of RT Kit (Tiangen, China, http://www.tiangen.com/). Quantitative real- CsRALF4 and CsRALF19) were designed and synthesized according to time PCR (qPCR) was performed using TB Green™ Premix Ex Taq™ the published methods52. Single target gRNA generated by denaturing (Takara, Japan, http://www.takarabiomed.com.cn/) on the Applied annealed, and double target gRNAs by PCR amplification using vector Biosystems 7500 real-time PCR system and Bio-RAD CFX384 system. pCBC-DT1T2 as template were respectively incorporated into vector Cucumber Ubiquitin extension protein (CsUBI CsaV3_5G031430) was pKSE402. Agrobacterium tumefaciens EHA105 carrying the recombi- used as the internal reference gene48. The relative expression was nant vector was shake-cultured to OD600 0.6–0.8, then adjusted to calculated according to the comparative cycle threshold (CT) OD600 0.2–0.3. The cucumber cotyledon explant was infected by method49. Three biological replicates and three technical replicates Agrobacterium suspensions with a syringe under negative pressure. were performed for each gene. Primer information is listed in Sup- After 3 days of co-culture in darkness, explants were transferred to the plementary Data 1. medium for bud differentiation and bacteriostasis. After 3–4 weeks, the regenerated shoots with GFP fluorescence were screened under a In situ hybridization fluorescence stereomicroscope (Filter Cube: Excitation filters The 25-day-old shoot apex, male and female flower buds, and ovaries at 460–500 nm, Dichroic mirror 505 nm, and Barrier filters 510 nm). The anthesis without pollination and at 32 h after pollination of XTMC and homozygous T1 mutant seedlings without vector (GFP-free) were Csalc mutants were fixed with 3.7% formalin–acetic acid–alcohol (FAA) selected from T0 transgenic lines for further phenotypic analysis and fixative. Sampling and recognition of flower buds at different devel- data statistics. Details about cucumber tissue culture refer to the opmental stages were referred to in the previous study50. Sense and previous study53. antisense probes were amplified with gene-specific primers containing SP6 and T7 RNA polymerase binding sites, respectively. Sample fixa- Alexander staining for pollen viability tion, sectioning, and hybridization were performed as described Stamens from open male flowers, at least 3 for each plant, were washed previously51. Primer information is listed in Supplementary Data 1. to make pollen suspension (each flower stamen with 200 μL H2O). 5 μL Nature Communications | (2023)14:258 12 Article https://doi.org/10.1038/s41467-023-35936-z pollen suspension and 15 μL Alexander dye were mixed on the slide CsRALF19 into the yeast strain EGY48. G-box elements were queried on and stained for 2 min. The pollen grains stained with purplish red CsRALF4/19 promoter and genomic region, as well as other 27 DEGs under the microscope indicate the viable pollens. Each biological between Csalc mutant and WT in the PlantCARE website (http:// replicate counted at least 5 fields. bioinformatics.psb.ugent.be/webtools/plantcare/html/). Yeast trans- formants were then selected on SD/-Trp-Ura agar plates for 4 days at in vitro pollen germination 30 °C. The yeast transformants were further grown on SD/Gal/Raf/- Pollen grains from open male flowers were lightly swept into the pollen Trp-Ura agar plates containing X-Gal (5-bromo-4-chloro-3-indolyl-β-D- germination medium [15% (w/v) sucrose, 0.01% (w/v) boric acid, 1 mM galactopyranoside) for blue color development58. Yeast transforma- Ca(NO3)2, and 0.5% (w/v) agar, pH 6.5] using a soft brush54. After being tion was conducted as described in the Yeast Protocols Handbook cultured in light and humid at 28 °C for 3 h, the germinating pollen (Clontech). tubes were observed and captured under microscope. The pollen tube germination percentage and pollen tube length were counted and LUC/REN assay for protein–DNA interactions in tobacco leaves measured in CellSens standard and ImageJ software55. At least three The CsALC CDS was cloned into pGreenII 62-SK as effectors. The male flowers were counted for each mutant line and five fields were 1811 and 3326 bp sequences upstream of CsRALF4 and CsRALF19 counted for each Petri dish. Pollen tube length longer than pollen grain coding regions were cloned into pGreenII 0800-LUC as reporters. diameter is determined as normal germination. The Agrobacterium GV3101 pSoup strain carrying the above recombinant vector (9: 1 effector: reporter construct cell culture Aniline blue staining of pollen tubes mixture) was introduced into the half of a N. benthamiana leaf. The One female flower was pollinated with two male flowers. Pollinated other half leaf was infiltrated with the strain carrying the empty excised pistils (with superior ovaries) were placed in Petri dishes and effector and CsRALF4/CsRALF19 reporter as a negative control. The wetly cultured in a growth chamber under a 16 h/8 h and 28 °C/22 °C activities of firefly luciferase (LUC) and Renilla reiformis luciferase day/night regime. After 32 h, the pistils were sliced lengthwise or (REN) were measured at 72 h after infiltration on a microplate crosswise (about 0.5–1 mm in thickness) and fixed in 75% ethanol and reader instrument (Molecular Devices SpectraMax i3x, the USA). 25% acetic acid solution for 3–6 h. After softened in 8 M NaOH for 3–6 h, The ratio of LUC to REN was calculated as the measurement result the sample slices were stained with Aniline blue solution [0.1% (w/v) of transcriptional activity of CsALC on CsRALF4/19 with eight bio- Aniline Blue in 0.1 M K3PO4] for at least 2 h. The pistil slice was placed on logical replicates59. a slide and the pollen tube extension in vivo was observed and pho- tographed under a fluorescence microscope with excitation wave- Nucleotide diversity and selective footprint analysis length of 330–385 nm56. The pollen tube density in pistils was indicated Published sequencing data of 115 cucumber germplasms with 37 East by fluorescence intensity and quantitated in ImageJ software55. After Asian (EA), 29 Eurasian (EU), 19 Xishuangbanna (XSBN), 13 Indian wild fluorescence images were split into three channels, the red channel genotypes (INW), and 17 Indian domestic (IND) accessions were used in diagram was chosen and the fluorescence intensity was converted to this analysis28. Nucleotide diversity was calculated as an π value for the the gray value, adjusted, and calculated in ImageJ software. five groups using VCFtools, which was further used to estimate the fixation index (FST) between EA/EU/XSBN/IND and the Indian wild Alcian blue staining of transmitting tracts counterparts (INW). Pistils at anthesis were fixed in 3.7% formaldehyde, 5% acetic acid, and 50% ethanol and subsequently dehydrated through ethanol series Reporting summary (50%, 70%, 85%, 95%, 100%, each 60 min), then hyalinized with xylene Further information on research design is available in the Nature (50% 60 min, 100% 60 min three times) and embedded in paraffin. Portfolio Reporting Summary linked to this article. Samples (styles and ovaries) in paraffin were made for 10 μm thickness sections by microtome. After deparaffinization through xylene (10 min Data availability twice) and ethanol series (100%, 95%, 85%, 70%, 50%, each 5 min), Transcriptomic data in this study can be found under accession sections were hydrated in water twice and stained with 0.5% Alcian number GSE182139 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? blue solution (1 g Alcian blue 8GX in 200 mL H2O) for 30 min. After acc=GSE182139). Cucumber genes reported in this manuscript can be washing, sections were examined under microscopy56. The deeper searched in CuGenDB (http://cucurbitgenomics.org/). All data gener- blue refers to the higher acid polysaccharide content. ated or analyzed during this study are available in this published article and its supplementary information files. Source data are provided with Transcriptomic analysis this paper. Ovaries of WT and Csalc mutant at 32 h after pollination were col- lected for RNA-seq analysis. Three biological replicates were set. References Illumina’s next-generation high-throughput sequencing platform 1. Qu, L. J., Li, L., Lan, Z. J. & Dresselhaus, T. Peptide signalling during and PE150 sequencing strategy were adopted. After quality control, the pollen tube journey and double fertilization. J. Exp. Bot. 66, clean read pairs were obtained and mapped to the cucumber gen- 5139–5150 (2015). ome sequence (http://cucurbitgenomics.org/ Chinese long CDS V2). 2. Russell, S. D. Double fertilization. Int. Rev. Cytol. 140, All genes were quantitatively analyzed to identify the differential 357–387 (1992). genes. The cutoff for DEGs was at least a 2-fold change in expression 3. Raghavan, V. Some reflections on double fertilization, from its dis- and a false discovery rate (FDR) of <0.0557 (listed in Supplementary covery to the present. N. Phytol. 159, 565–583 (2003). Data 3). Sequencing data were deposited with the Gene Expression 4. Crawford, B. C. W. & Yanofsky, M. F. The formation and function of Omnibus (GEO) database at the National Center for Biotechnology the female reproductive tract in flowering plants. Curr. Biol. 18, Information under accession number GSE182139. R972–R978 (2008). 5. Higashiyama, T. & Takeuchi, H. The mechanism and key molecules Yeast one-hybrid assay involved in pollen tube guidance. Annu. Rev. Plant Biol. 66, Vectors pB42AD and pLacZi2u were used for yeast one-hybrid assay. 393–413 (2015). Plasmids for AD-fusions of CsALC were co-transformed with the LacZ 6. Higashiyama, T. et al. Pollen tube attraction by the synergid cell. reporters driven by two fragments from CsRALF4 or 3 fragments from Science 293, 1480–1483 (2001). Nature Communications | (2023)14:258 13 Article https://doi.org/10.1038/s41467-023-35936-z 7. Marton, M. L. & Dresselhaus, T. Female gametophyte-controlled 29. Groszmann, M., Paicu, T., Alvarez, J. P., Swain, S. M. & Smyth, D. R. pollen tube guidance. Biochem. Soc. Trans. 38, 627–630 (2010). SPATULA and ALCATRAZ, are partially redundant, functionally 8. Shimizu, K. K. & Okada, K. Attractive and repulsive interactions diverging bHLH genes required for Arabidopsis gynoecium and between female and male gametophytes in Arabidopsis pollen tube fruit development. Plant J. 68, 816–829 (2011). guidance. Development 127, 4511–4518 (2000). 30. Cheng, Z. et al. SPATULA and ALCATRAZ confer female sterility and 9. Lopes, A. L., Moreira, D., Ferreira, M. J., Pereira, A. M. & Coimbra, S. fruit cavity via mediating pistil development in cucumber. Plant Insights into secrets along the pollen tube pathway in need to be Physiol. 189, 1553–1569 (2022). discovered. J. Exp. Bot. 70, 2979–2992 (2019). 31. Toledo-Ortiz, G., Huq, E. & Quail, P. H. The Arabidopsis basic/helix- 10. Liu, C. et al. Pollen PCP-B peptides unlock a stigma peptide- loop-helix transcription factor family. Plant Cell 15, receptor kinase gating mechanism for pollination. Science 372, 1749–1770 (2003). 171–175 (2021). 32. Kim, M. J., Jeon, B. W., Oh, E., Seo, P. J. & Kim, J. Peptide signaling 11. Tang, W. H., Kelley, D., Ezcurra, I., Cotter, R. & McCormick, S. LeS- during plant reproduction. Trends Plant Sci. 26, 822–835 (2021). TIG1, an extracellular binding partner for the pollen receptor kina- 33. Zhao, J. et al. A functional allele of CsFUL1 regulates fruit length ses LePRK1 and LePRK2, promotes pollen tube growth in vitro. Plant through repressing CsSUP and inhibiting auxin transport in J. 39, 343–353 (2004). cucumber. Plant Cell 31, 1289–1307 (2019). 12. Ge, Z. X. et al. Arabidopsis pollen tube integrity and sperm release 34. Mizukami, A. G. et al. The AMOR arabinogalactan sugar chain are regulated by RALF-mediated signaling. Science 358, induces pollen-tube competency to respond to ovular guidance. 1596–1599 (2017). Curr. Biol. 26, 1091–1097 (2016). 13. Mecchia, M. A. et al. RALF4/19 peptides interact with LRX proteins to 35. Iwano, M. et al. A pollen coat-inducible autoinhibited Ca2+-ATPase control pollen tube growth in Arabidopsis. Science 358, expressed in stigmatic papilla cells is required for compatible pol- 1600–1603 (2017). lination in the Brassicaceae. Plant Cell 26, 636–649 (2014). 14. Boisson-Dernier, A. et al. Disruption of the pollen-expressed FER- 36. Yu, G. H. et al. Exogenous gamma-aminobutyric acid (GABA) affects ONIA homologs ANXUR1 and ANXUR2 triggers pollen tube dis- pollen tube growth via modulating putative Ca2+-permeable charge. Development 136, 3279–3288 (2009). membrane channels and is coupled to negative regulation on 15. Miyazaki, S. et al. ANXUR1 and 2, sister genes to FERONIA/SIRENE, glutamate decarboxylase. J. Exp. Bot. 65, 3235–3248 (2014). are male factors for coordinated fertilization. Curr. Biol. 19, 37. Ling, Y. et al. gamma-Aminobutyric acid (GABA) homeostasis reg- 1327–1331 (2009). ulates pollen germination and polarized growth in Picea wilsonii. 16. Ge, Z. et al. LLG2/3 Are Co-receptors in BUPS/ANX-RALF signaling Planta 238, 831–843 (2013). to regulate Arabidopsis pollen tube integrity. Curr. Biol. 29, 38. Vogler, F., Schmalzl, C., Englhart, M., Bircheneder, M. & Sprunck, S. 3256–3265 (2019). Brassinosteroids promote Arabidopsis pollen germination and 17. Shimizu, K. K., Ito, T., Ishiguro, S. & Okada, K. MAA3 (MAGATAMA3) growth. Plant Reprod. 27, 153–167 (2014). helicase gene is required for female gametophyte development 39. Chae, K. & Lord, E. M. Pollen tube growth and guidance: roles of and pollen tube guidance in Arabidopsis thaliana. Plant Cell Physiol. small, secreted proteins. Ann. Bot.-Lond. 108, 627–636 (2011). 49, 1478–1483 (2008). 40. Cheung, A. Y., Wang, H. & Wu, H. M. A floral transmitting tissue- 18. Marton, M. L., Cordts, S., Broadhvest, J. & Dresselhaus, T. Micropylar specific glycoprotein attracts pollen tubes and stimulates their pollen tube guidance by egg apparatus 1 of maize. Science 307, growth. Cell 82, 383–393 (1995). 573–576 (2005). 41. Gremski, K., Ditta, G. & Yanofsky, M. F. The HECATE genes regulate 19. Marton, M. L., Fastner, A., Uebler, S. & Dresselhaus, T. Overcoming female reproductive tract development in Arabidopsis thaliana. hybridization barriers by the secretion of the maize pollen tube Development 134, 3593–3601 (2007). attractant ZmEA1 from Arabidopsis ovules. Curr. Biol. 22, 42. Crawford, B. C. W. & Yanofsky, M. F. HALF FILLED promotes 1194–1198 (2012). reproductive tract development and fertilization efficiency in Ara- 20. Okuda, S. et al. Defensin-like polypeptide LUREs are pollen tube bidopsis thaliana. Development 138, 2999–3009 (2011). attractants secreted from synergid cells. Nature 458, 43. Takeuchi, H. & Higashiyama, T. Tip-localized receptors control 357–361 (2009). pollen tube growth and LURE sensing in Arabidopsis. Nature 531, 21. Takeuchi, H. & Higashiyama, T. A Species-specific cluster of 245–248 (2016). defensin-like genes encodes diffusible pollen tube attractants in 44. Cheung, A. Y. & Wu, H. M. Structural and signaling networks for the Arabidopsis. PLoS Biol. 10, 12 (2012). polar cell growth machinery in pollen tubes. Annu. Rev. Plant Biol. 22. Zhong, S. et al. Cysteine-rich peptides promote interspecific 59, 547–572 (2008). genetic isolation in Arabidopsis. Science 364, 851 (2019). 45. Ortiz-Ramirez, C. I., Giraldo, M. A., Ferrandiz, C. & Pabon-Mora, N. 23. Takeuchi, H. & Higashiyama, T. Tip-localized receptors control Expression and function of the bHLH genes ALCATRAZ and SPA- pollen tube growth and LURE sensing in Arabidopsis. Nature 531, TULA in selected Solanaceae species. Plant J. 99, 686–702 (2019). 245–248 (2016). 46. Ferrandiz, C. & Fourquin, C. Role of the FUL-SHP network in the 24. Zhong, S. & Qu, L. J. Cysteine-rich peptides: signals for pollen tube evolution of fruit morphology and function. J. Exp. Bot. 65, guidance, species isolation and beyond. Sci. China Life Sci. 62, 4505–4513 (2014). 1243–1245 (2019). 47. Varagona, M. J., Schmidt, R. J. & Raikhel, N. V. Nuclear localization 25. Lu, Y. et al. Pollen tubes lacking a pair of K+ transporters fail to target signal(s) required for nuclear targeting of the maize regulatory ovules in Arabidopsis. Plant Cell 23, 81–93 (2011). protein Opaque-2. Plant Cell 4, 1213–1227 (1992). 26. Zhong, S. et al. RALF peptide signaling controls the polytubey block 48. Wan, H. J. et al. Selection of appropriate reference genes for gene in Arabidopsis. Science 375, 290–296 (2022). expression studies by quantitative real-time polymerase chain 27. Rajani, S. & Sundaresan, V. The Arabidopsis myc/bHLH gene reaction in cucumber. Anal. Biochem. 399, 257–261 (2010). ALCATRAZ enables cell separation in fruit dehiscence. Curr. Biol. 11, 49. Schmittgen, T. D. & Livak, K. J. Analyzing real-time PCR data by the 1914–1922 (2001). comparative C-T method. Nat. Protoc. 3, 1101–1108 (2008). 28. Qi, J. J. et al. A genomic variation map provides insights into the 50. Bai, S. L. et al. Developmental analyses reveal early arrests of the genetic basis of cucumber domestication and diversity. Nat. Genet. spore-bearing parts of reproductive organs in unisexual flowers of 45, 1510–1515 (2013). cucumber (Cucumis sativus L.). Planta 220, 230–240 (2004). Nature Communications | (2023)14:258 14 Article https://doi.org/10.1038/s41467-023-35936-z 51. Zhang, X. L. et al. Transcription repressor HANABA TARANU con- Author contributions trols flower development by integrating the actions of multiple X.Z., Z.C., X.L., and Z.Z. designed the research; Z.C. and X.L. performed hormones, floral organ specification genes, and GATA3 family the experiments.; X.Z., Z.C., and Z.Z. wrote the paper; S.Y., B.L., Y.Z., genes in Arabidopsis. Plant Cell 25, 83–101 (2013). W.S., J.C., Z.W., G.C., L.L., A.Y., H.L., L.H., M.L., J.Z., J.X., and Z.Y. pro- 52. Xing, H. L. et al. A CRISPR/Cas9 toolkit for multiplex genome editing vided experimental assistance; all the authors revised the manuscript. in plants. BMC Plant Biol. 14, 327 (2014). 53. Hu, B. W. et al. Engineering non-transgenic gynoecious cucumber Competing interests using an improved transformation protocol and optimized CRISPR/ The authors declare no competing interests. Cas9 system. Mol. Plant 10, 1575–1578 (2017). 54. Cheng, J. T. et al. Down-regulating CsHT1, a cucumber pollen- Additional information specific hexose transporter, inhibits pollen germination, tube Supplementary information The online version contains growth, and seed development. Plant Physiol. 168, 635–647 supplementary material available at (2015). https://doi.org/10.1038/s41467-023-35936-z. 55. Schindelin, J., Rueden, C. T., Hiner, M. C. & Eliceiri, K. W. The ImageJ ecosystem: an open platform for biomedical image analysis. Mol. Correspondence and requests for materials should be addressed to Reprod. Dev. 82, 518–529 (2015). Zhaoyang Zhou or Xiaolan Zhang. 56. Girin, T. et al. INDEHISCENT and SPATULA interact to specify carpel and valve margin tissue and thus promote seed dispersal in Arabi- Peer review information Nature Communications thanks Said Hafidh dopsis. Plant Cell 23, 3641–3653 (2011). and the other, anonymous, reviewer(s) for their contribution to the peer 57. Jiang, L. et al. Transcriptomic analysis reveals the roles of review of this work. microtubule-related genes and transcription factors in fruit length regulation in cucumber (Cucumis sativus L.). Sci. Rep. 5, Reprints and permissions information is available at 8031 (2015). http://www.nature.com/reprints 58. Wang, R. et al. Cis-regulation of the amino acid transporter genes ZmAAP2 and ZmLHT1 by ZmPHR1 transcription factors in maize ear Publisher’s note Springer Nature remains neutral with regard to jur- under phosphate limitation. J. Exp. Bot. 72, 3846–3863 (2021). isdictional claims in published maps and institutional affiliations. 59. Liu, X. F. et al. CsSPL functions as an adaptor between HD-ZIPIII and CsWUS transcription factors regulating anther and ovule develop- Open Access This article is licensed under a Creative Commons ment in Cucumis sativus (cucumber). Plant J. 94, 535–547 Attribution 4.0 International License, which permits use, sharing, (2018). adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the Acknowledgements source, provide a link to the Creative Commons license, and indicate if We are grateful to Prof. S. Huang for giving CRISPR-Cas9 vectors, to Drs. changes were made. The images or other third party material in this L. Yang and M. Liu for help in cucumber transformation and pollen article are included in the article’s Creative Commons license, unless germination in vitro respectively, to Drs. Michael Scanlon and Kelly indicated otherwise in a credit line to the material. If material is not Dawe for advice in manuscript revision. This work is supported by grants included in the article’s Creative Commons license and your intended from the National Key Research and Development Program of China use is not permitted by statutory regulation or exceeds the permitted [2022YFD1200502], the National Natural Science Foundation of China use, you will need to obtain permission directly from the copyright [32025033, 31930097, and 31872654], the Construction of Beijing Sci- holder. To view a copy of this license, visit http://creativecommons.org/ ence and Technology Innovation and Service Capacity in Top Subjects licenses/by/4.0/. [CEFFPXM2019_014207_000032], Expert Workstation in Yunnan Pro- vince (202205AF150021), and 111 Project [B17043]. © The Author(s) 2023 Nature Communications | (2023)14:258 15