To over-express OsHOX24 in rice, PCR amplified complete open reading frame (ORF) [using gene specific primers (Supplementary Table S1)], was cloned in modified pCAMBIA1301 vector in BamHI/KpnI restriction sites. The confirmed clone was transformed in Agrobacterium strain LBA4404. Rice seeds [Oryza sativa cultivar Pusa Basmati 1 (PB1)] were used as background for generation of transgenic plants. The transformation of embryogenic calli derived from the scutellum of dehusked mature rice seeds was done as described previously (Sharma et al., 2013). Rice transformants were confirmed by Southern blotting. Seeds obtained from T1 generation plants were screened on Murashige–Skoog (MS) media supplemented with hygromycin (40 mg/L). Further, segregation ratio of each confirmed transgenic line was estimated and transgenics were grown till homozygous stage to obtain seeds for future analyses.

To study the effect of transgene in over-expression rice transgenics, phenotypic characterization of wild-type (WT) and rice transgenics over-expressing OsHOX24 was carried out. Three-week-old rice seedlings grown in the culture room, were transferred to pots and grown till maturity with optimum supply of water. Growth of plants was monitored and several phenotypic parameters, like shoot length, flag-leaf area, number of panicles and tillers per plant were documented in three independent biological replicates, consisting of at least 13–15 plants per line.

To assess the performance of rice transgenics under abiotic stress conditions, seed germination assays were carried out. Seeds from transgenic and WT plants were plated on MS medium without or with ABA (5 and 10 μM), 200 mM NaCl, 200 mM mannitol and -0.4 MPa PEG 6000. Seed germination was recorded after 3 days of transferring the plated seeds to culture room except for PEG 6000 treatment (germination was recorded after 7 days). The number of germinated seeds (considering radicle emergence as seed germination parameter) was expressed as percentage of total number of seeds plated, as described earlier (Dansana et al., 2014). Each experiment was repeated at least three times.

For stomatal opening/closure assays, 7-day-old rice seedlings were kept in Yoshida medium (experimental control) and subjected to exogenous ABA treatment (100 μM) for 3 h. After incubation, leaf sections were visualized under scanning electron microscope (Zeiss EVO LS10, Germany) and stomatal status (opened/partially opened/closed) were analyzed for each sample. The experiment was repeated at least three times.

For evaluating the effect of salinity and desiccation stresses (DSs) on seedlings, WT and OsHOX24 rice transgenics (H1, H49 and H74) were grown on MS medium supplemented without or with NaCl (200 mM) and PEG 6000 (20%) in the culture room for 12 days. The root and shoot lengths of seedlings grown under control and stress conditions were measured. The relative average root and shoot lengths under salinity and DS conditions were expressed as percentage of root and shoot lengths of the seedlings under control condition. The experiments were performed in at least three independent biological replicates.

To assess the effect of DS, rice transgenics and WT plants, grown in greenhouse for 2 months till mature vegetative stage (6 weeks) and reproductive stage (15 weeks) were subjected to DS by withholding water for 5 weeks, followed by 2 weeks of recovery. WT and transgenic plants of same age served as experimental controls. Plant growth was monitored at all stages and phenotype of plants under control and DS followed by recovery phase was documented. To determine the effect of DS, chlorophyll content in transgenic and WT leaves under desiccation and control conditions at mature and reproductive stages after recovery phase, was estimated using a Chlorophyll Meter (SPAD-502, Japan). To assess the effect of DS, number of turgid leaves was counted and survival percentage of plants was calculated after recovery. In addition, rate of water loss in transgenic and WT leaves at mature and reproductive stages were estimated as described (Dansana et al., 2014). The experiments were performed in at least two independent biological replicates.

To determine the effect of DS on rice transgenics and WT, leaf disk assays were carried out. Healthy and fully expanded leaves from 2-month-old WT and rice transgenics were detached and 4–5 leaf disks were floated in sterile water (experimental control) or 20% PEG 6000 solution and samples were incubated under culture room conditions for 3 days. The effect of DS was assessed by monitoring phenotypic changes and measuring chlorophyll content of leaves relative to control samples. Chlorophyll was extracted from leaf samples and the amount of chlorophyll a, chlorophyll b and total chlorophyll was calculated as described previously (Sharma et al., 2013).

All the experiments were performed in at least two or three independent biological replicates and standard error (SE) was computed. For estimation of statistical significance, Student’s t-test was performed. Statistically significant differences between WT and transgenics or between control and stress conditions (^∗P ≤ 0.05 and ^∗∗P ≤ 0.01) were denoted by asterisks.

Total RNA isolated from 3-week-old transgenic and WT seedlings subjected to 3 h of control and DS treatment were used for microarray analysis and quality control of samples was carried out as described (Sharma et al., 2013). Microarray analysis for rice seedlings was performed using Affymetrix WT PLUS Reagent kit (Affymetrix, Santa Clara, CA, USA) in three independent biological replicates, according to manufacturer’s instructions. Normalization of microarray data was done by Robust Multi-array Average (RMA) algorithm implemented in Genespring software version 12.6. The microarray data has been submitted in the Gene Expression Omnibus database at NCBI under series accession number GSE79212. Differential gene expression analysis was performed as described earlier (Sharma et al., 2013).

Gene ontology (GO) enrichment was performed using BiNGO. The stress response pathway analysis was carried out using MapMan (version 3.5.1)¹ with P-value cut-off of ≤0.05. Venn diagrams and heatmaps were generated using online tools, VENNY² and MeV (version 4.9), respectively.

For gene expression profiling, total RNA isolation, first-strand cDNA synthesis and qRT-PCR analysis were performed as described previously (Sharma et al., 2013). Experiments were conducted in at least two biological replicates for each sample and three technical replicates were analyzed for each biological replicate. ΔΔC_T calculation method was used to calculate relative expression level of each gene. For normalizing the relative mRNA level of individual gene in various RNA samples, Ubiquitin 5 (UBQ5) was used as internal control gene (Jain et al., 2006). Results of microarray experiments were validated by qRT-PCR analysis of selected differentially expressed genes. The list of primer sequences used for qRT-PCR analysis has been provided in Supplementary Table S1.

The complete coding region of OsHOX24 was over-expressed under the control of ubiquitin (UBQ) promoter (UBQ::OsHOX24) in rice (Figure 1A). A total of 18 hygromycin-resistant T0 transgenic plants were obtained. Three of these lines, exhibiting 3:1 segregation ratio in T1 generation and confirmed by Southern blotting, were used for further analyses. The enhanced expression level of transgene in all the selected transgenic lines as compared to WT was detected by real-time PCR analysis (Figure 1B). The relative expression level of OsHOX24 was found to be highest in H74 line followed by H49 and H1 lines.

We did not observe any detectable phenotypic differences in the OsHOX24 rice transgenics as compared to WT at the seedling stage under control conditions (Supplementary Figure S1). However, over-expression of OsHOX24 resulted in significant alteration in the phenotype of rice transgenics (as compared to WT) at the reproductive stage (Supplementary Figure S2A). Transgenic lines showed significantly reduced shoot length (87–89%) (Supplementary Figure S2B) and smaller flag-leaf area (61–81%) as compared to WT (Supplementary Figure S2C). Moreover, the number of panicles (Supplementary Figure S2D) and tillers (Supplementary Figure S2E) were also found to be lesser in the transgenic lines as compared to WT. In transgenic lines, panicle and tiller numbers ranged from 65–73% and 69–73%, respectively, as compared to WT (Supplementary Figures S2D,E). These observations indicated the role of OsHOX24 in modulating plant phenotype during reproductive development in rice.

To evaluate the effect of stress hormone, ABA and abiotic stress conditions (osmotic, salinity and desiccation) on the transgenic lines and WT, seed germination assays were carried out. The percentage germination of transgenic lines was much lesser as compared to WT on MS medium supplemented with different concentrations of ABA (5 and 10 μM), NaCl (200 mM), mannitol (200 mM) and PEG 6000 (-0.4 MPa) [equivalent to 20% PEG 6000] (Figure 2). Exogenous ABA treatment resulted in greater susceptibility of transgenics as compared to WT. For instance, transgenics showed only about 53% germination after 3 days on 5 μM ABA as compared to 66% for WT. At 10 μM ABA, WT showed 61% germination in comparison to 28–45% germination of transgenics (Figure 2A). Under salinity stress (200 mM), WT exhibited 62% germination, whereas transgenic lines exhibited significantly reduced germination (28–42%) (Figure 2B). Similarly, the effect of osmotic stress (200 mM mannitol) on germination was found to be more prominent on transgenics (Figure 2B), which showed only 22–24% germination in comparison to WT (63%). Under DS (20% PEG 6000 treatment), transgenic lines showed 63% germination as compared to 81% in WT (Figure 2B). Overall, these results indicated that OsHOX24 over-expression altered seed germination in the transgenic plants under abiotic stress conditions. Among the transgenic lines, H49 exhibited greater susceptibility to different abiotic stresses.

Further, we investigated the effect of ABA on stomatal closure in leaves of transgenics and WT. Under control conditions, transgenics and WT showed similar number of opened and closed stomata. However, the fraction of completely closed stomata in transgenics was considerably lesser (53–74%) as compared to WT (82%) under exogenous ABA treatment (Figure 2C). This observation indicated that transgenics possess significantly reduced ability of stomatal closure under stress condition.

The effect of desiccation and salinity stresses on growth of transgenic and WT seedlings were evaluated. A significant reduction in root and shoot growth was observed in transgenics as compared to WT seedlings under salinity (200 mM NaCl) and desiccation (20% PEG 6000) stresses, whereas all the seedlings appeared healthy under control conditions (Figure 3A). The average root growth of transgenics was found to be 4–8% as compared to 16% for WT under salinity stress relative to control condition. Similarly, average root growth of transgenics was found to be 13–14% as compared to 30% in WT under DS (Figure 3B). The average shoot growth of transgenics was 4–5% as compared to 12% in WT under salinity stress, and 14–15% as compared to 36% in WT under DS (Figure 3C).

Over-expression of OsHOX24 affected the rate of water loss in transgenic rice. The transgenic leaves exhibited comparatively greater water loss at mature and reproductive stages of development than WT till 210 min of air-drying (Supplementary Figure S3). Detached leaves of WT from the mature and reproductive stage of development retained 69 and 60% of fresh weight as compared to 46 and 47% of fresh weight retained by the transgenic line (H49) following 90 min of air drying, respectively (Supplementary Figure S3). After 210 min of air drying, detached WT leaves taken at the mature and reproductive stage of development retained 34–40% of fresh weight, in contrast to transgenic line (H49) which retained only 26–30% of fresh weight (Supplementary Figure S3).

To assess the effect of DS at mature vegetative stage, leaf disk assays were performed for transgenics and WT plants using 20% PEG 6000. The transgenic leaves exhibited greater chlorophyll loss as compared to WT (Supplementary Figure S4). The chlorophyll content of transgenic lines, H49 (60%) and H74 (61%), was lesser in comparison to WT (86%) under DS relative to control condition (Supplementary Figure S4). Further, 2-month-old UBQ::OsHOX24 transgenic lines and WT were subjected to DS by withholding water for 5 weeks followed by recovery for 15 days. The transgenic lines showed greater wilting of leaves and poor recovery as compared to WT (Figure 4A). The extent of chlorosis was more prominent in the leaves of transgenic lines as compared to WT (Figure 4B). The chlorophyll content in transgenic lines was significantly lesser than WT under DS (Figure 4B). The transgenic lines possessed lesser number of turgid leaves than WT under DS (Figure 4C). The above observations indicated the higher susceptibility of OsHOX24 over-expressing transgenics under DS as compared to WT at the mature stage of development as well.

To evaluate the effect of DS at the reproductive stage, 15-week-old WT and UBQ::OsHOX24 transgenic line (H49) were subjected to DS by withholding water for 30 days followed by recovery for 15 days. Transgenics (H49) showed more compromised growth, greater wilting of leaves and poor recovery than WT (Figure 4D). The actual chlorophyll content in the transgenic line was evidenced to be much lesser than WT under DS and control condition (Figure 4E). The chlorophyll content of transgenic line was only 35% as compared to WT under DS (Figure 4F). In addition, under control condition too, the chlorophyll content of transgenic line was lesser (only 69%) in comparison to WT (Figure 4E). In addition, the transgenic line exhibited lower survival percentage as compared to WT (Figure 4F). Only 25% of the transgenic plants survived as compared to WT (50%) under DS (Figure 4F). The above observations confirmed the susceptible nature of UBQ::OsHOX24 transgenics under DS as compared to WT at the reproductive stage of development too.

To study the effect of OsHOX24 over-expression in rice, microarray analysis of transgenic line (H49) and WT plants, under control condition and subjected to DS treatment, was performed. The differential gene expression analysis, revealed a total of 3108 significantly (≥twofold, corrected P-value ≤ 0.05) differentially regulated genes in the transgenic line as compared to WT under control and/or DS. A total of 1247 and 1896 genes were found to be up-regulated and down-regulated, respectively, in at least one of the conditions analyzed (Figure 5A and Supplementary Table S2). These genes were found to be involved in diverse biological processes. GO enrichment of down-regulated genes in the transgenic line under DS revealed that several cellular metabolic and biosynthetic processes (primary metabolic processes and nucleic acid metabolic processes) were significantly enriched (Figure 5B). About 7% of the differentially expressed genes belonged to 51 different TF families (Supplementary Figure S5). At least 39 and 23 TF-encoding genes were uniquely differentially expressed in the transgenic line (H49) and WT, respectively, whereas 158 genes were found to be commonly differentially regulated in the transgenic line and WT under DS (Figure 6A). GO enrichment analysis of these TF-encoding genes revealed their involvement in crucial biological processes, including gene expression, developmental processes, metabolic processes, response to abiotic stress and hormone stimulus (Figures 6B–F). Further, biotic and abiotic stress response pathway overview indicated that genes involved in signaling pathways, hormone signaling (auxin, ABA, brassinosteroid and ethylene), stress-responsive transcription regulatory components (including TFs like ERF, bZIP, WRKY and MYB), redox metabolism and respiratory burst, defense responses, secondary metabolism, proteolysis and cell wall, were enriched in the transgenic line under DS (Supplementary Figure S6). This indicated the role of OsHOX24 in biotic stress responses, besides mediating abiotic stress responses in transgenic rice.

Overall, the differentially expressed genes were found to be involved in diverse metabolic and developmental processes. The differential expression patterns of selected stress-responsive rice genes, known to be involved in abiotic stress responses, including LOC_Os07g14610 encoding for IAA-amino acid hydrolase, LOC_Os01g07120 encoding for OsDREB2A, LOC_Os07g37400 encoding for OsFBX257, LOC_Os05g37060 encoding for MYB TF and LOC_Os12g03050 encoding for NAM TF were validated via real-time PCR analysis (Supplementary Figure S7). Overall, the transcriptome analysis of transgenic line and WT under control and DS revealed alteration of several developmentally important and stress-related genes, which explain the susceptible phenotype of rice transgenics as compared to WT.