Liu et al. Horticulture Research (2021)8:2 https://doi.org/10.1038/s41438-020-00430-w Horticulture Research www.nature.com/hortres ARTICLE Open Access Plant buffering against the high-light stress- induced accumulation of CsGA2ox8 transcripts via alternative splicing to finely tune gibberellin levels and maintain hypocotyl elongation Bin Liu1,2, Shuo Zhao1, Pengli Li1, Yilu Yin1, Qingliang Niu1, Jinqiang Yan3,4 and Danfeng Huang1 Abstract In plants, alternative splicing (AS) is markedly induced in response to environmental stresses, but it is unclear why plants generate multiple transcripts under stress conditions. In this study, RNA-seq was performed to identify AS events in cucumber seedlings grown under different light intensities. We identified a novel transcript of the gibberellin (GA)-deactivating enzyme Gibberellin 2-beta-dioxygenase 8 (CsGA2ox8). Compared with canonical CsGA2ox8.1, the CsGA2ox8.2 isoform presented intron retention between the second and third exons. Functional analysis proved that the transcript of CsGA2ox8.1 but not CsGA2ox8.2 played a role in the deactivation of bioactive GAs. Moreover, expression analysis demonstrated that both transcripts were upregulated by increased light intensity, but the expression level of CsGA2ox8.1 increased slowly when the light intensity was >400 µmol·m−2·s−1 PPFD (photosynthetic photon flux density), while the CsGA2ox8.2 transcript levels increased rapidly when the light intensity was 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; >200 µmol·m−2·s−1 PPFD. Our findings provide evidence that plants might finely tune their GA levels by buffering against the normal transcripts of CsGA2ox8 through AS. Introduction DS-3, which encodes a DELLA protein, results in semi- Gibberellins (GAs) are known to be the dominant dwarf oilseed rape plants5. hormone governing hypocotyl elongation1,2. GAs regulate Light is among the most essential factors for plant cell expansion, thus causing changes in hypocotyl length growth and development6. Seedlings grown under dark or plant height via the GA signaling pathway. GA-GID1- conditions are characterized by relatively long hypocotyls DELLA is a widely accepted signaling module through compared with those grown under light conditions7,8. which bioactive GAs regulate plant growth and develop- One of the reasons for this is that under light conditions, ment2. For example, GIBBERELLIN INSENSITIVE the GA level decreases, while under dark or shaded con- DWARF 1 (GID1)-inactivation mutants show reduced ditions, the GA level increases9–11. Most related studies plant height3,4, and substitution of proline to leucine in have focused mainly on low-light (shading)-induced GA biosynthesis for signaling12, and only a few studies have shown high-light-induced excessive GA deactivation. However, it remains unknown how plants respond to Correspondence: Jinqiang Yan (yan.jinqiang@163.com) or high-light stress and finely tune their GA levels to Danfeng Huang (hdf@sjtu.edu.cn) 1 School of Agriculture and Biology, Shanghai Jiao Tong University, Key maintain growth. Laboratory of Urban Agriculture (South), Ministry of Agriculture, Dongchuan Alternative splicing (AS) is a widespread mechanism in Road, Shanghai 200240, China 2 eukaryotes in which different protein variants are gener- Department of Plant Genomics, Centre for Research in Agricultural Genomics (CRAG), CSIC-IRTA-UAB-UB, Bellaterra 08193, Spain ated from a single gene via different combinations of Full list of author information is available at the end of the article © The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Liu et al. Horticulture Research (2021)8:2 Page 2 of 10 splicing sites. AS is a prevalent approach in which plants by quantitative reverse-transcription PCR (qRT-PCR) respond to environmental stress, but it is counterintuitive (Fig. 1d). These results suggested that the expression of why plants would invest in protein synthesis under con- CsGA2ox8 might be affected by light intensity. To test this ditions of declining energy supplies13,14. Chaudhary et al. hypothesis, we treated young cucumber seedlings with proposed that, under stress conditions, plants buffer seven different light intensities (40, 80, 120, 160, 200, 240, against normal protein synthesis levels via AS to decrease and 280 µmol·m−2·s−1 PPFD) and quantified the expres- the translation of a significant portion of the tran- sion of CsGA2ox8 in seedlings grown under these con- scriptome and produce the protein isoforms necessary for ditions. As shown in Fig. 1e, the transcript level of adaptation to stresses; however, the mechanisms of this CsGA2ox8 increased with increasing light intensity. Pio- process are not yet known14. neering studies have shown that light modulates In this study, we performed RNA-seq to identify dif- photomorphogenesis-related hypocotyl elongation mainly ferentially expressed genes and AS events in cucumber by interacting with GAs, and many genes are involved in (Cucumis sativus L.) seedlings grown under different light this light-GA system10. Our results suggested that the intensities. We focused on GA-related genes and found a expression of the GA deactivation factor CsGA2ox8 was novel transcript of Gibberellin 2-beta-dioxygenase upregulated by increased light levels, which might be (CsGA2ox8), CsGA2ox8.2, with intron retention (IR) involved in light-regulated hypocotyl elongation. between the second and third exons. Gibberellin 2-beta- dioxygenases compose a family whose members can cat- Differential AS (DAS) analysis for identifying novel alyze the deactivation of bioactive GAs or their pre- transcripts of CsGA2ox8 cursors15. Most studies have shown that increasing AS is a prevalent approach used by plants in response to transcripts of GA2oxs can reduce plant height by reg- stress and involves the generation of multiple transcripts ulating endogenous bioactive GA levels15–20. We found from the same gene21. In this study, five types of AS that under high-light stress, cucumber seedlings tended to events, namely, RIs (retained introns), A5SSs (alternative generate more nonfunctional CsGA2ox8.2 but not func- 5′ splice sites), A3SS (alternative 3′ splice sites), MXEs tional canonical CsGA2ox8.1. In other words, the plants (mutually exclusive exons), and SEs (skipped exons), were buffered against normal transcript levels of CsGA2ox8 via identified by RNA-seq analysis (Fig. 2a, Supplementary AS to limit the inactivation process of GAs under high- Table S2). Among them, SEs and RIs accounted for more light stress, and this strategy may allow seedlings to than 25% of all AS events (Fig. 2b), which is similar to that maintain a certain GA level and sustain hypocotyl reported in Arabidopsis22,23. elongation. Among all RI events, we found a novel transcript of CsGA2ox8 with an IR between the second and third Results and discussion exons. The original transcript of CsGA2ox8 was defined as CsGA2ox8 expression is upregulated in response to CsGA2ox8.1, and the isoform with IR was defined as increasing light intensity CsGA2ox8.2 (Fig. 2c). To validate the existence of We recorded the changes in cucumber hypocotyl CsGA2ox8.2, the coding regions of CsGA2ox8 of plants lengths after the cotyledons emerged from the soil under under treatment X2 were amplified and then sequenced. treatments involving two light intensities: X1 Two distinct bands were observed after RT-PCR-based (180 µmol·m−2·s−1 PPFD) and X2 (40 µmol·m−2·s−1 amplification (Fig. 2c). After sequencing, the two ampli- PPFD). The hypocotyl lengths of seedlings grown under cons were found to be 1056 and 1655 bp in length (Fig. treatment X1 were shorter than those grown under 2c). Sequence comparison showed the same result as that treatment X2 (Fig. 1a), and a significant difference was after RNA-seq, and a schematic diagram of this event is observed after 40 h of light treatment (Fig. 1b). To illustrated in Fig. 2c. RIs promoted the disruption of the investigate the molecular differences of seedlings grown open reading frame of CsGA2ox8, resulting in an mRNA under the two light conditions, a transcriptome analysis with a premature stop codon (TGA), and leading to a was performed. In total, 279 upregulated and 321 down- truncated protein (Fig. 2c). Compared with other GA2ox regulated genes were identified between the two proteins in Arabidopsis, the AtGA2ox7 and AtGA2ox8 treatments (Fig. 1c, Supplementary Table S1). In this proteins have a unique region15. Alignment of the amino study, we focused on GA-related genes because bioactive acid sequences of AtGA2ox8 and CsGA2ox8.1 revealed GAs have a large effect on hypocotyl elongation in that the CsGA2ox8.1 protein had a conserved motif (Fig. cucumber1. We found that CsGA2ox8, a gene whose 2d, denoted by a red box), which may define the specificity product catalyzes the deactivation of bioactive GAs, was of the reactions catalyzed by these enzymes15. The significantly upregulated in treatment X1 compared to absence of amino acids in CsGA2ox8.2 is indicated using treatment X2. This expression difference was confirmed a purple line in Fig. 2d. Liu et al. Horticulture Research (2021)8:2 Page 3 of 10 * * * f e d c b a a Fig. 1 Expression of CsGA2ox8 is positively regulated by light intensity. a Phenotypes of cucumber seedlings before (a1) and after (a2) treatment with fluorescent lights at two intensities (X1, 180 µmol·m−2·s−1 PPFD; X2, 40 µmol·m−2·s−1 PPFD) for 56 h. b Hypocotyl lengths of cucumber seedlings between the two treatments. The error bars indicate the standard deviations (SDs; n = 5). The asterisks indicate significant differences in hypocotyl lengths between X1 and X2 (P < 0.05). c Volcano plot of the transcriptome between X1- and X2-treated cucumber seedlings. d Quantitative reverse-transcription PCR results (three replications) validating the different expressions of CsGA2ox8 between the two treatments. The asterisks indicate significant differences in gene expression between X1 and X2 (P < 0.05). e Relative expression of CsGA2ox8 under different light intensities. The lowercase letters indicate that CsGA2ox8 transcription was modified by light intensity (P < 0.05, three replications) Liu et al. Horticulture Research (2021)8:2 Page 4 of 10 a b Percentage (%) c 40 µmol·m-2·s-1 ppfd 180 µmol·m-2·s-1 ppfd d AtGA2ox8 CsGA2ox8.1 Consensus AtGA2ox8 CsGA2ox8.1 Consensus AtGA2ox8 CsGA2ox8.1 Consensus AtGA2ox8 CsGA2ox8.1 Consensus AtGA2ox8 CsGA2ox8.1 Consensus Fig. 2 (See legend on next page.) Liu et al. Horticulture Research (2021)8:2 Page 5 of 10 (see figure on previous page) Fig. 2 Analysis of differential alternative splicing (DAS) for the identification of novel transcripts of Gibberellin 2-beta-dioxygenase 8 (CsGA2ox8). a Identification of five AS events in the RNA-seq datasets of cucumber seedlings grown under two light conditions. RI, retained intron; A5SS, alternative 5′ splice site; A3SS, alternative 3′ splice site; MXE, mutually exclusive exon; and SE, skipped exon. Black boxes, exons flanking AS events; white boxes, differentially included exons; dotted lines, sequences spliced together. b Stacked bars showing five AS events and their frequency in each replication under the two light conditions. c Validation of the intron retention (IR) of CsGA2ox8 via reverse-transcription PCR (RT- PCR). A termination codon, TGA (red line), was detected in the retained intron (gray box). M, Trans2K Plus DNA Marker. d Alignment of the amino acid sequences of AtGA2ox8 and CsGA2ox8.1. The unique conserved domain of GA2ox8.1 is denoted by the red box above. Purple line, the lost amino acid region of CsGA2ox8.2 Overexpression of CsGA2ox8.1 but not CsGA2ox8.2 with the results of Schomburg et al. (2003), who reported decreases GA levels and the length of the hypocotyl in that GA9, GA20, and GA4 levels increased in AtGA2ox8 Arabidopsis thaliana overexpression lines15. GA12 is the precursor of GA9 and GA2ox genes function in reducing GAs15–20,24. Given GA4, and GA53 is the precursor of GA20 and GA125. that CsGA2ox8.2 lost the third and fourth exons from Therefore, it is possible that decreasing GA12 and GA53 CsGA2ox8.1 (Fig. 2c, d), we questioned whether levels would affect the synthesis of GA9, GA4, GA20, and CsGA2ox8.2 still maintains the same function as that of GA1. Taken together, these results suggest that CsGA2ox8.1. To verify this, CsGA2ox8.1 and CsGA2ox8.2 CsGA2ox8.1, rather than its CsGA2ox8.2 isoform, regu- were overexpressed in Arabidopsis Col-0. CsGA2ox8.1 lated hypocotyl elongation, and this regulation was overexpression lines (OE8.1) had shorter hypocotyl achieved mostly via decreasing endogenous GA levels. lengths than Col-0 under both light and dark conditions (Fig. 3a–c). The RT-PCR results suggested that GA2ox8.1 Plants tend to generate nonfunctional CsGA2ox8.2 but not was overexpressed in OE8.1 but not in Col-0 (Fig. 3d), functional CsGA2ox8.1 under high-light stress suggesting that GA2ox8.1 was functional for hypocotyl As one type of AS, IR mostly results in the production elongation. However, the hypocotyl length of CsGA2ox8.2 of nonfunctional transcripts carrying premature termi- overexpression lines (OE8.2) was not significantly differ- nation codons targeted for degradation via the nonsense- ent from that of Col-0 (Fig. 3e–g), even though the mediated mRNA decay pathway27,28. However, it is transcript level of GA2ox8.2 was as high as that of unclear why plants perform AS under stress conditions. GA2ox8.1 (Figs. 3d, h). Next, the GA levels were quanti- Given that CsGA2ox8.2 was detected under relatively fied in the Col-0, OE8.1, and OE8.2 lines. Based on the high-light conditions and that CsGA2ox8.2 seemed to be a GA synthetic pathway, GA2ox genes encoding enzymes nonfunctional transcript, we hypothesized that that convert GA12, GA9, GA53, GA20, GA1, and GA4 to CsGA2ox8.2 was generated for buffering against normal GA110, GA51, GA97, GA29, GA8, and GA34, respec- CsGA2ox8.1 transcripts under high-light stress to finely tively25,26 (Fig. 3i). In this study, we found that GA12, GA9, tune GA levels and maintain plant growth. Chaudhary GA53, GA20, GA1, and GA4 significantly decreased in et al. posited the same idea and proposed that, under OE8.1 compared with Col-0, and no significant differ- stress conditions, plants buffer against normal protein ences in GA levels were detected between OE8.2 and Col- synthesis levels via AS to decrease translation of a sig- 0 (Fig. 3i), suggesting that CsGA2ox8.1 is a functional nificant portion of the transcriptome, but the mechanisms transcript but that CsGA2ox8.2 is not. of this process have not yet been elucidated14. AS is a prevalent approach in which plants respond to To test our hypothesis, the relative abundance of the environmental stress. One hypothesis is that plants buffer two mRNA forms of CsGA2ox8 in plants under 50, 100, against normal protein synthesis levels via AS to decrease 150, 200, 400, 800, and 1200 µmol·m−2·s−1 PPFD was the translation of a significant portion of the tran- analyzed by RT-PCR. The results showed that both scriptome under stress conditions14. Therefore, for future transcripts were positively regulated by light, and studies, it is important to know how CsGA2ox8.1 and CsGA2ox8.1 was the dominant transcript under all light CsGA2ox8.2 work at the protein level. conditions (Fig. 4a). However, the expression level of The homolog of CsGA2ox8 in Arabidopsis is AtGA2ox8, CsGA2ox8.1 increased slowly when the light intensity was whose product functions in hydroxylating the C20-GA >400 µmol·m−2·s−1 PPFD, while that of CsGA2ox8.2 precursors GA12 and GA5315. In this study, consistent increased rapidly when the light intensity was with the results from overexpression of AtGA2ox815, >200 µmol·m−2·s−1 PPFD (Fig. 4a). Furthermore, the GA12 and GA53 levels dramatically decreased in OE8.1, inclusion/exclusion (In/Ex) ratio (CsGA2ox8.2/ suggesting that CsGA2ox8 has a function similar to that of CsGA2ox8.1) significantly increased when the light AtGA2ox8. However, we found that the content of GA9, intensity was >400 µmol·m−2·s−1 PPFD (Fig. 4b). In other GA20, GA1, and GA4 also decreased, which is inconsistent words, the CsGA2ox8.2 expression level was negligible Liu et al. Horticulture Research (2021)8:2 Page 6 of 10 a Dark b Light e Dark f Light Col-0 OE8.1 Col-0 OE8.1 Col-0 OE8.2 Col-0 OE8.2 c CsGA2ox8.1 g CsGA2ox8.2 Hypocotyl length (cm) Hypocotyl length (cm) * * d h i Fig. 3 (See legend on next page.) Liu et al. Horticulture Research (2021)8:2 Page 7 of 10 (see figure on previous page) Fig. 3 CsGA2ox8.1 but not CsGA2ox8.2 had a role in the deactivation of bioactive gibberellins (GAs). a, b CsGA2ox8.1 overexpression (OE8.1) lines had shortened hypocotyls under both dark and light conditions. Scale bars, 5 mm. c Hypocotyl lengths of the indicated genotypes in (a) and (b). The error bars indicate the standard deviations (SDs, n > 20). The asterisks indicate significant differences between the hypocotyl lengths of OE8.1 lines and Col-0 (P < 0.05). d GA2ox8.1 is overexpressed in OE8.1 but is not expressed in Col-0. e, f CsGA2ox8.2 overexpression (OE8.2) lines had hypocotyl lengths similar to those of Col-0 under both dark and light conditions. g Hypocotyl lengths of the indicated genotypes in (e) and (f). The error bars indicate SDs (n > 20). h GA2ox8.2 is overexpressed in OE8.2 but is not expressed in Col-0. i Profile of endogenous GAs in Col-0, OE8.1, and OE8.2 lines a b * * Relative expression * * * * * Light intensity (µmol/m2·s ppfd ) Light intensity (µmol/m2·s ppfd ) c Fig. 4 Plants tend to generate more CsGA2ox8.2 but not CsGA2ox8.1 under high-light stress. a Relative expression levels of CsGA2ox8.1 and CsGA2ox8.2 under different light conditions as measured via quantitative reverse-transcription PCR (three replications). The asterisks indicate significant differences in gene expression between CsGA2ox8.1 and CsGA2ox8.2 (P < 0.05). b High light promotes the expression of CsGA2ox8.2. The RNA products are quantified via the CsGA2ox8.2/CsGA2ox8.1 [inclusion/exclusion (In/Ex)] ratio. c Model depicting that plants buffer against high-light stress-induced functional CsGA2ox8.1 via alternative splicing of a nonfunctional CsGA2ox8.2 isoform to finely tune gibberellin (GA) levels and maintain hypocotyl elongation Liu et al. Horticulture Research (2021)8:2 Page 8 of 10 under low-light conditions, but under increasing light The germinated seeds were then sown in autoclaved soil intensity, especially that corresponding to high-light at 24 °C. Single cucumber seedlings with cotyledons stress, plants tended to generate more nonfunctional emerging from the soil were chosen for light treatment. CsGA2ox8.2 transcripts and limit the increase in func- A. thaliana Col-0 was used in this study. The seeds were tional CsGA2ox8.1 levels (Fig. 4c). In this way, plants disinfected and germinated on Murashige–Skoog (MS) could produce less functional GA2ox and could maintain media consisting of 1% sucrose and 0.2% Phytagar at 4 °C a certain level of GA for hypocotyl elongation under high- for 3 days before being moved into a growth chamber at light conditions. 22 °C31. AS is known to promote plant stress tolerance13. For example, stress imposed by high salinity promotes the use Light treatment and hypocotyl measurements of noncanonical splice sites29. In addition, AS events finely Light intensities were measured using a Lighting Pas- tune gene expression by regulating the ratio of active and sportTM ALP-01 Handheld Spectrometer (HESON, nonactive isoforms30. Similarly, we showed that high-light Shanghai, China). Cucumber seedlings were grown in a stress increased the CsGA2ox8.2/CsGA2ox8.1 ratio in this greenhouse with “DE-HPS 1000 W” lights. To maintain study, suggesting a new mechanism for the GA- the light conditions within an appropriate range, the deactivating-enzyme response to high-light stress. How- plants were placed at different heights below the fixtures. ever, it is not clear how high light modulates AS events. To For example, to generate a high-light condition, the plants our knowledge, AS events are mediated by core spliceoso- were moved closer to the lamps; to generate low-light mal components (small nuclear ribonucleoproteins) as well conditions, the plants were moved farther from the lamps as RNA-binding proteins (serine/arginine-rich proteins and or were covered with film. For Arabidopsis, the plates with heterogeneous nuclear ribonucleoproteins)30. One hypoth- disinfected seeds were treated at 4 °C for 72 h and then esis is that high-light stress affects the expression or protein placed under white light for 10 h to induce uniform ger- structure of spliceosome genes, thereby modulating AS mination. Afterward, the plates were transferred to dark events at the whole-genome level. or light conditions and then incubated at 22 °C for Our results suggest that IR may have been employed to 3–6 days for hypocotyl measurements. The hypocotyl finely tune gene expression by altering the ratio between lengths of the seedlings were measured using a tapeline. functional and nonfunctional variants. Following a BLAST search of The Arabidopsis Information Resource GA level analysis (TAIR, https://www.arabidopsis.org/) and the MSU Rice The GA level was analyzed as previously reported, with Genome Annotation Project Database (http://rice. moderate changes32. Briefly, for product purification, plantbiology.msu.edu/index.shtml), AT4G21200 and ~0.5 g of the sample was quickly immersed in liquid LOC_Os02g41954 were identified has sharing high nitrogen, homogenized, and extracted with 5 mL of pre- sequence homology (57.42% and 49.03%, respectively) cooled 80% methanol for 20 h at 4 °C. Afterward, the with that of CsGA2ox8. As shown in TAIR, AT4G21200 samples were shaken at 4 °C for 30 min and then cen- has one splice variant with an IR between the second and trifuged at 18,000 × g for 20 min at 4 °C. The supernatant third exons, which is the same as CsGA2ox8.2, with the RI was collected and immersed in liquid nitrogen. Then, the promoting the disruption of the open reading frame of samples were dissolved in 100 µL of chromatography- AT4G21200, resulting in the generation of an mRNA with grade methanol supplemented with 0.1 M glacial acetic a premature stop codon (TGA), leading to a truncated acid and filtered through a 0.22 µm filter. High- protein23. From the rice database, LOC_Os02g41954 also performance liquid chromatography-based analysis was has one splice variant, which has a premature stop codon performed using an Agilent (CA, USA) Zorbax SB-C18 (TGA) between the second and third exons30. These column (5 µm, 4.6 × 250 mm). The mobile phase A sol- findings suggest that AS of GA2ox8 commonly occurs in vent consisted of 100% methanol, and mobile phase B was plants, and this strategy might also be used by other plant 0.1 M acetic acid. The gradient elution was performed as species to buffer against the stress-responsive tran- follows: 0–40 min, 3% A and 97% B; after 40 min, 67.6% A scriptome and promote efficient adaptation to stress and 32.2% B. The flow rate was 1 mL·min−1, and the conditions. column temperature was set at 30 °C. Chromatograms were quantified using the Agilent chromatography Materials and methods workstation. Each sample involved three replicates from Plant materials and growth conditions independent experiments. Cucumber inbred line 9930 was used in this study. Seeds were surface sterilized and soaked in water for 3 h, RNA-seq analysis and validation after which they were then transferred to Petri dishes Cucumber seedlings were subjected to two light lined with wet filter papers and germinated at 28 °C. conditions (X1, 180 µmol·m−2·s−1 PPFD; X2, Liu et al. Horticulture Research (2021)8:2 Page 9 of 10 40 µmol·m−2·s−1 PPFD) for 56 h, after which the hypo- controls39. The primers used are listed in Supplementary cotyls were harvested for RNA extraction (SV Total RNA Table S3. Isolation System, Promega). Each treatment was repli- cated three times. The RNA quality was measured using Transformation of A. thaliana an Agilent 2100 Bioanalyzer system (Agilent Technolo- Transformation assays were performed as previously gies, CA, USA), with a RIN/RQ (RNA integrity number/ described31. To generate CsGA2ox8.1 and CsGA2ox8.2 RNA quality index) ≥8. Sequencing libraries were pre- overexpression constructs, full-length CsGA2ox8.1 and pared using a NEBNext® UltraTM RNA Library Prep Kit CsGA2ox8.2 cDNAs were cloned and inserted into a pBI121 for Illumina® (NEB, USA). The libraries were sequenced vector between the XbaI and SmaI restriction sites40. The on a HiSeq 4000 platform (Illumina, San Diego, CA). constructs were then introduced into Agrobacterium Thereafter, unknown or low-quality reads, as well as reads tumefaciens strain LBA4404 by electroporation, which were with adapters, were passed through a filter (SOAPnuke then transformed into wild-type A. thaliana (Col-0), as v1.5.2, https://github.com/BGI-flexlab/SOAPnuke) to previously described41. The transgenic plants were screened generate clean reads33. The remaining high-quality clean on MS media supplemented with 50 mg/L kanamycin and reads were mapped to the Cucumber Reference Genome then confirmed via PCR. The primers used are listed in V3 version (ftp://cucurbitgenomics.org/pub/cucurbit/ Supplementary Table S3. genome/cucumber/Chinese_long/v3/)34 using Bowtie2 v2.2.535. Differential expression analyses were subse- Statistical analysis quently carried out using RSEM v1.2.1236 software. Dif- All the treatments mentioned in this study involved at ferentially expressed genes were defined as those having least three independent biological and technical repli- an |log2 fold change (FC) | ≥ 1 and a q-value ≤0.05. DAS cates. The results were analyzed using analysis of variance. analyses were performed using rMATS v3.2.537. Tran- All the analyses were carried out using Statistics Analysis scripts were deemed to be the result of DAS when the System 15.1 for Windows. false discovery rate was ≤0.05. Acknowledgements To validate the GA2ox8 AS isoforms, seedlings were This study was supported by The National Key Research and Development treated with the same light conditions as those applied for Program of China (2019YFD1000300), the International Postdoctoral Exchange RNA-seq (X1, 180 µmol·m−2·s−1 PPFD; X2, Fellowship Program from the China Postdoctoral Council (20170053), the −2 −1 Technology System Construction of Modern Agricultural Industry of Shanghai 40 µmol·m ·s PPFD) for 56 h, after which the hypo- (19Z113040008), and the Presidential Foundation of Guangdong Academy of cotyls were harvested for RNA extraction, cDNA synth- Agricultural Sciences (BZ201901). esis, and PCR-based analysis. The primers used are listed in Supplementary Table 3. Author details 1 School of Agriculture and Biology, Shanghai Jiao Tong University, Key Laboratory of Urban Agriculture (South), Ministry of Agriculture, Dongchuan Sequence alignment Road, Shanghai 200240, China. 2Department of Plant Genomics, Centre for Sequence alignments were performed based on the Research in Agricultural Genomics (CRAG), CSIC-IRTA-UAB-UB, Bellaterra 08193, Spain. 3Vegetable Research Institute, Guangdong Academy of Agricultural amino acid sequences of CsGA2ox8.1, CsGA2ox8.2, and Sciences, Guangzhou 510640, China. 4Guangdong Key Laboratory for New AtGA2ox8. Multiple sequence alignment was performed Technology Research of Vegetables, Guangzhou 510640, China using DNAMAN for Windows (Lynnon Corporation, Author contributions Quebec). D.H. and B.L. proposed the project; B.L. and J.Y. designed the research; B.L., J.Y., S.Z., Y.Y., and P.L. performed the experiments; and B.L. and J.Y. analyzed the Gene expression analysis data and wrote the paper, with the participation of Q.N. and D.H. Gene expression analysis methods were used as pre- Conflict of interest viously reported31. Total RNA was extracted from the The authors declare that they have no conflict of interest. hypocotyls using an SV Total RNA Isolation System (Promega), and cDNA was synthesized using a Multi- Supplementary Information accompanies this paper at (https://doi.org/ 10.1038/s41438-020-00430-w). ScribeTM Reverse Transcriptase Kit (Applied Biosystems) according to the manufacturer’s instructions. Quantitative Received: 12 May 2020 Revised: 13 September 2020 Accepted: 17 October RT-PCR was performed using a SYBR® Premix Ex TaqTM 2020 Kit (TaKaRa, China) in conjunction with an Applied Biosystems 7500 Real-time PCR system. For RT-PCR, PCR primers were designed to amplify two isoforms of different sizes, and their RNA products were quantified References 1. Lopez-Juez, E., Kobayashi, M., Sakurai, A., Kamiya, Y. & Kendrick, R. E. Phyto- and reported as an In/Ex ratio38. Cucumber α-TUBULIN chrome, gibberellins, and hypocotyl growth (A study using the cucumber (TUA) and Arabidopsis ACTIN 2 were used as internal (Cucumis sativus L.) long hypocotyl mutant). Plant Physiol. 107, 131–140 (1995). Liu et al. Horticulture Research (2021)8:2 Page 10 of 10 2. Sun T-p. The molecular mechanism and evolution of the GA–GID1–DELLA 22. Filichkin, S. A. et al. Genome-wide mapping of alternative splicing in Arabi- signaling module in plants. Curr. Biol. 21, R338–R345 (2011). dopsis thaliana. Genome Res. 20, 45–58 (2010). 3. Griffiths, J. et al. Genetic characterization and functional analysis of the GID1 23. Marquez, Y., Brown, J. W., Simpson, C., Barta, A. & Kalyna, M. Transcriptome gibberellin receptors in Arabidopsis. Plant Cell 18, 3399–3414 (2006). survey reveals increased complexity of the alternative splicing landscape in 4. Cheng, J. et al. A single nucleotide mutation in GID1c disrupts its interaction Arabidopsis. Genome Res. 22, 1184–1195 (2012). with DELLA1 and causes a GA-insensitive dwarf phenotype in peach. Plant 24. Plackett, A. R. et al. Analysis of the developmental roles of the Arabidopsis Biotechnol. J. 17, 1723–1735 (2019). gibberellin 20-oxidases demonstrates that GA20ox1, -2, and -3 are the 5. Zhao, B. et al. Brassica napus DS-3, encoding a DELLA protein, negatively dominant paralogs. Plant Cell 24, 941–960 (2012). regulates stem elongation through gibberellin signaling pathway. Theor. Appl. 25. Hedden, P. & Proebsting, W. M. Genetic analysis of gibberellin biosynthesis. Genet. 130, 727–741 (2017). Plant Physiol. 119, 365–370 (1999). 6. Kami, C., Lorrain, S., Hornitschek, P., & Fankhauser, C. in Current Topics in 26. Wang, G.-L. et al. Morphological characteristics, anatomical structure, and gene Developmental Biology (Elsevier, 2010). expression: novel insights into gibberellin biosynthesis and perception during 7. García-Martinez, J. L. & Gil, J. Light regulation of gibberellin biosynthesis and carrot growth and development. Horticulture Res. 2, 1–10 (2015). mode of action. J. Plant Growth Regul. 20, 354–368 (2002). 27. Filichkin, S. A. et al. Environmental stresses modulate abundance and timing of 8. Gawronska, H. et al. Effects of low irradiance stress on gibberellin levels in pea alternatively spliced circadian transcripts in Arabidopsis. Mol. Plant 8, 207–227 seedlings. Plant Cell Physiol. 36, 1361–1367 (1995). (2015). 9. De Lucas, M. et al. A molecular framework for light and gibberellin control of 28. Kalyna, M. et al. Alternative splicing and nonsense-mediated decay modulate cell elongation. Nature 451, 480–484 (2008). expression of important regulatory genes in Arabidopsis. Nucleic Acids Res. 40, 10. Feng, S. et al. Coordinated regulation of Arabidopsis thaliana development by 2454–2469 (2012). light and gibberellins. Nature 451, 475–479 (2008). 29. Feng, J. et al. SKIP confers osmotic tolerance during salt stress by controlling 11. Li, K. et al. DELLA-mediated PIF degradation contributes to coordination of alternative gene splicing in Arabidopsis. Mol. Plant 8, 1038–1052 (2015). light and gibberellin signalling in Arabidopsis. Nat. Commun. 7, 1–11 (2016). 30. Kawahara, Y. et al. Improvement of the Oryza sativa Nipponbare reference 12. Colebrook, E. H., Thomas, S. G., Phillips, A. L. & Hedden, P. The role of gib- genome using next generation sequence and optical map data. Rice 6, 4 berellin signalling in plant responses to abiotic stress. J. Exp. Biol. 217, 67–75 (2013). (2014). 31. Liu, B. et al. Silencing of the gibberellin receptor homolog, CsGID1a, affects 13. Laloum, T., Martín, G. & Duque, P. Alternative splicing control of abiotic stress locule formation in cucumber (Cucumis sativus) fruit. N. Phytologist 210, responses. Trends Plant Sci. 23, 140–150 (2018). 551–563 (2016). 14. Chaudhary, S., Jabre, I., Reddy, A. S., Staiger, D. & Syed, N. H. Perspective on 32. Yang, R. et al. Hormone profiling and transcription analysis reveal a major role alternative splicing and proteome complexity in plants. Trends Plant Sci. 24, of ABA in tomato salt tolerance. Plant Physiol. Biochem. 77, 23–34 (2014). 496–506 (2019). 33. Li, R. et al. SOAP2: an improved ultrafast tool for short read alignment. 15. Schomburg, F. M., Bizzell, C. M., Lee, D. J., Zeevaart, J. A. & Amasino, R. M. Bioinformatics 25, 1966–1967 (2009). Overexpression of a novel class of gibberellin 2-oxidases decreases gibberellin 34. Li, Q. et al. A chromosome-scale genome assembly of cucumber (Cucumis levels and creates dwarf plants. Plant Cell 15, 151–163 (2003). sativus L.). GigaScience 8, giz072 (2019). 16. Busov, V. B. et al. Activation tagging of a dominant gibberellin catabolism 35. Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. gene (GA 2-oxidase) from poplar that regulates tree stature. Plant Physiol. 132, Methods 9, 357 (2012). 1283–1291 (2003). 36. Li, B. & Dewey, C. N. RSEM: accurate transcript quantification from RNA-Seq 17. Huang, J. et al. Activation of gibberellin 2-oxidase 6 decreases active gibber- data with or without a reference genome. BMC Bioinforma. 12, 323 (2011). ellin levels and creates a dominant semi-dwarf phenotype in rice (Oryza sativa 37. Shen, S. et al. rMATS: robust and flexible detection of differential alternative L.). J. Genet. Genomics 37, 23–36 (2010). splicing from replicate RNA-Seq data. Proc. Natl Acad. Sci. USA 111, 18. Otani, M. et al. Overexpression of the gibberellin 2-oxidase gene from Torenia E5593–E5601 (2014). fournieri induces dwarf phenotypes in the liliaceous monocotyledon Tricyrtis 38. Zhou, X. et al. Transcriptome analysis of alternative splicing events regulated sp. J. Plant Physiol. 170, 1416–1423 (2013). by SRSF10 reveals position-dependent splicing modulation. Nucleic Acids Res. 19. Wuddineh, W. A. et al. Identification and overexpression of gibberellin 2‐ 42, 4019–4030 (2014). oxidase (GA 2ox) in switchgrass (Panicum virgatum L.) for improved plant 39. Zhang, Y. et al. A GAMYB homologue CsGAMYB1 regulates sex expression of architecture and reduced biomass recalcitrance. Plant Biotechnol. J. 13, cucumber via an ethylene-independent pathway. J. Exp. Bot. 65, 3201–3213 636–647 (2015). (2014). 20. Hu, Y.-X., Tao, Y.-B. & Xu, Z.-F. Overexpression of Jatropha Gibberellin 2-oxidase 40. Jefferson, A. R. Assaying chimeric genes in plants: the GUS gene fusion system. 6 (JcGA2ox6) induces dwarfism and smaller leaves, flowers and fruits in Ara- Plant Mol. Biol. Report. 5, 387–405 (1987). bidopsis and Jatropha. Front. Plant Sci. 8, 2103 (2017). 41. Clough, S. J. & Bent, A. F. Floral dip: a simplified method for Agro- 21. Chaudhary, S. et al. Alternative splicing and protein diversity: plants versus bacterium-mediated transformation of Arabidopsis thaliana. Plant J. 16, animals. Front. Plant Sci. 10, 708 (2019). 735–743 (1998).