Physiological, biochemical, and transcriptomic alterations in Castor (Ricinus communis L.) under polyethylene glycol-induced oxidative stress

Experimental sample collection

The castor CSR•181 seeds were provided by the Tongliao Agricultural Science Research Institute. Castor seeds were sown in 7 × 7 cm square nursery pots with vermiculite as the germination substrate and watered every three days with Hoagland’s nutrient solution. After the castor seedlings emerged, they were transferred to a hydroponic system containing Hoagland’s nutrient solution and cultured for three weeks. The Hoagland’s nutrient solution was supplemented with 15% PEG-6000 to stress the castor seedlings for 3 and 7 d, followed by rehydration in hydroponic Hoagland’s nutrient solution for 4 d (for a total of 11 d). The control group seedlings were cultured in hydroponic Hoagland’s nutrient solution throughout. The nutrient solution was refreshed twice weekly. All castor seedlings were cultured in a cultivation room with a dark/light cycle of 8/16 hours, a temperature of 25 °C, and 60% humidity. Root samples from both the treatment and control groups were collected at each time point, with six biological replicates per time point. The samples were frozen in liquid nitrogen and stored at -80 °C.

Determination of physiological and biochemical indicators of castor root systems

In this study, six plants were randomly selected from each treatment group for analysis. The castor growth parameters examined included plant height, stem diameter, and leaf area. Both plant height and stem diameter were measured using a caliper. The leaf area was measured using a leaf area meter (Li-3100 C Area Meter). The Li-3100 C Area Meter was turned on, after which zero-point calibration and full-scale calibration were performed, and the leaf area of the fourth true leaf of the plant was measured.

The respiration rate was measured via an oxygen electrode (Hansatech). A 5 cm segment of castor root was finely chopped and placed into the reaction chamber, followed by the addition of 2 ml of buffer solution (10 mM HEPES, 10 mM MES, 2 mM CaCl₂; pH 7.5). Then, 50 µmol of salicylhydroxamic acid (SHAM) was added after 1 min, and 100 µmol of NaN₃ solution was added after 2 min. The measurement was completed at 3 min.

Photosynthetic parameters were measured via a Li-6400 portable photosynthesis system. On the day of the experiment, at 10:00 AM, the youngest fourth true leaf was selected, and the photosynthetic rate was measured at the middle part of the leaf using the Li-6400 instrument, where the distance between chambers was < 10 cm. The Li-6400 in-chamber conditions were as follows: LED light intensity of 1,200 µmol m^− 2 s^− 1, reference CO₂ concentration of 500 ppm, flow rate of 400 µmol s^− 1, and VPD controlled at 2.1 ± 0.2 kPa.

Hydrogen peroxide (H₂O₂) was detected spectrophotometrically as described by Willekens et al. [15]. The 100 µmol/L H₂O₂ acetone reagent was prepared by diluting a 10 mmol/L H₂O₂ solution to 100 µmol/L with acetone. The standard curve for H₂O₂ concentration determination was performed as follows: the reaction buffer was prepared by mixing precooled acetone (mL), 5% titanium sulfate (mL), and concentrated ammonia (mL) in a 10:1:2 ratio. A total of 1.3 mL of the reaction buffer was added to each of seven centrifuge tubes. Next, 0 mL, 0.1 mL, 0.2 mL, 0.4 mL, 0.6 mL, 0.8 mL, or 1.0 mL of the prepared 100 µmol/L H₂O₂ acetone reagent was added to the tubes. The mixture was allowed to stand for 10 min and then centrifuged at 12,000 × g for 10 min. The supernatant was discarded, and the precipitate was retained. Five milliliters of 2 mol/L sulfuric acid was added to the precipitate. After the precipitate dissolved, the volume was increased to 10 mL with ddH₂O. The absorbance at 415 nm was measured via a Bio-Rad Microplate Reader 550, and the standard curve was plotted. To prepare the sample extraction liquid, 3 g of castor roots was mixed with 3 mL of precooled acetone and a small amount of quartz sand, the mixture was ground to a homogeneous slurry, and the slurry was transferred to a centrifuge tube. The mixture was centrifuged at 12,000 × g for 10 min, and the supernatant was collected. Then, 0.1 mL of 5% titanium sulfate and 0.2 mL of concentrated ammonia were added to 1 mL of the sample extraction liquid. The mixture was allowed to stand for 10 min and then centrifuged at 12,000 × g for 10 min. The supernatant was discarded, and the precipitate was retained. Five milliliters of 2 mol/L sulfuric acid was added to the precipitate. Once dissolved, the absorbance at 415 nm was measured via a microplate reader, and the H₂O₂ content in the roots was calculated on the basis of the standard curve.

The MDA content was estimated via the method developed by Wang et al. [16]. Four milliliters of PBS (136.9 mmol NaCl, 2.7 mmol KCl, 10.1 mmol Na₂HPO₄, 1.8 mmol KH₂PO₄; pH 7.4) and a small amount of quartz sand were added to 0.5 g of castor root, the mixture was ground to a homogeneous mixture, and the mixture was transferred to a centrifuge tube. The mixture was centrifuged at 12,000 × g for 10 min, after which the supernatant was collected. One milliliter of the supernatant was mixed with 2 mL of 0.5% thiobarbituric acid, heated at 100 °C for 15 min, and immediately cooled on ice to stop the reaction. The mixture was centrifuged again at 12,000 × g for 10 min, and the supernatant was collected. The MDA content was quantified by measuring the absorbance at 450, 532, and 600 nm. The MDA content is expressed as µmol g^− 1 FW.

SOD (EC1.15.1.1) activity was measured according to the methods of Beauchamp and Fridovich [17]. Four milliliters of PBS (136.9 mmol NaCl, 2.7 mmol KCl, 10.1 mmol Na₂HPO₄, 1.8 mmol KH₂PO₄; pH 7.4) and a small amount of quartz sand were added to 0.5 g of castor root, the mixture was ground to a homogeneous mixture, and the mixture was transferred to a centrifuge tube. The mixture was centrifuged at 12,000 × g for 10 min, after which the supernatant was collected. The reaction mixture included 200 µl of enzyme extract, 400 µl of KH₂PO₄ (100 mM, pH 7.0), 50 µl of methionine (260 mM), 80 µl of NBT (4.2 mM), 10 µl of EDTA (10 mM), and 170 µl of riboflavin (130 µM). The mixture was illuminated for 15 min with a cool white fluorescent lamp providing light at 60 µmol m^− 2 s^− 1. The absorbance of each sample was recorded at a wavelength of 560 nm via a microplate reader. SOD activity is expressed as units of SOD activity per mg of protein per minute (U mg^− 1 protein min^− 1).

POD (EC 1.11.1.7) activity was measured via the method developed by Ashraf et al. [18]. Preparation of 0.01 mol of phosphate buffer was performed by mixing 61.0 mL of Na₂HPO₄ (0.1 mol) and 39.0 mL of KH₂PO₄ (0.1 mol). Then, ddH₂O was added, and the volume was increased to 1000 mL. A total of 0.5 g of castor root was added to 4 mL of PBS (0.01 mol) along with a small amount of quartz sand, and the mixture was ground to a homogeneous paste. The mixture was transferred to a centrifuge tube and centrifuged at 12,000 × g for 10 min, after which the supernatant was collected. The reaction mixture included 100 µl of enzyme extract, 50 µl of guaiacol (0.02 mol), 2.9 mL of phosphate buffer (0.01 mol), and 10 µl of H₂O₂ (0.04 mol). The absorbance of each sample was measured at a wavelength of 470 nm via a microplate reader. POD activity is expressed in units of POD activity per mg of protein per min (U mg^− 1 protein min^− 1).

CAT (EC 1.11.1.6) activity was assayed according to the methods of Havir and Mchale with some modifications [19]. First, 0.5 g of castor root was added to 4 mL of PBS (0.2 mol, pH 7.8) along with a small amount of quartz sand. The mixture was ground to a homogeneous paste and then transferred to a centrifuge tube. The mixture was then centrifuged at 12,000 × g for 10 min, and the supernatant was collected. The enzyme extract (2.5 mL) was added to 2.5 mL of H₂O₂ (0.1 mol) and incubated in a 30 °C water bath for 10 min. Then, 2.5 mL of H₂SO₄ (10%) was added, and the mixture was titrated with 0.1 mol of KMnO₄ until a pink color appeared. The absorbance of each sample was recorded at a wavelength of 240 nm via a spectrophotometer. CAT activity is expressed as the amount of H₂O₂ decomposed per mg of protein per min (µmol min⁻¹ pro mg⁻¹).

Glutathione reductase (GR, EC 1.6.4.2) activity was determined by NADPH oxidation [20]. A total of 0.5 g of castor root was added to 4 mL of PBS (50 mM, pH 7.8) along with a small amount of quartz sand. The mixture was ground to a homogeneous paste and then transferred to a centrifuge tube. The mixture was then centrifuged at 12,000 × g for 10 min, and the supernatant was collected. The reaction mixture included 200 µL of enzyme extract, 1.7 mL of HEPES, 0.1 mL of NADPH (2.4 mM), and 0.1 mL of GSSG (10 mM). The absorbance of each sample was measured at 340 nm via a microplate reader. GR activity is expressed as the amount of NADPH decomposed per mg of protein per min (µmol min⁻¹ pro mg⁻¹).

An ascorbate peroxidase (APX, EC1.11.1.11) activity assay was performed according to the methods of Nakano and Asada with some modifications [21]. Four milliliters of PBS (50 mM) and a small amount of quartz sand were added to 0.5 g of castor root. The mixture was ground to a homogenate and transferred to a centrifuge tube. The mixture was then centrifuged at 12,000 × g for 10 min, after which the supernatant was collected. The reaction mixture included 100 µL of enzyme extract, 1.7 mL of EDTA-Na₂ (0.1 mM), 100 µL of ASA (5 mM), and 100 µL of H₂O₂ (20 mM). The absorbance of each sample was recorded via a microplate reader at a wavelength of 290 nm. APX activity is expressed as the amount of ASA decomposed per mg of protein per min (µmol min⁻¹ pro mg⁻¹).

The physiological and biochemical data for castor roots were subjected to one-way analysis of variance (ANOVA), followed by Duncan’s multiple range test to evaluate the mean differences at a significance level of 0.05. The significance levels (p) were used to evaluate the significance of the parameters estimated.

RNA extraction, Illumina sequencing, and data quality control

Total RNA from castor roots was extracted via the TRIzol method [22]. A Dynabeads™ mRNA Purification Kit (Thermo Fisher) was used to isolate and purify mRNA from the total RNA of castor roots. The enriched mRNA was fragmented in fragmentation buffer and reverse transcribed to cDNA with random primers. Second-strand cDNA was synthesized by DNA polymerase I, RNase H, dNTPs, and buffer. After the cDNA fragments were purified with a QIAquick PCR Purification Kit, the ends were repaired, poly(A) tails were added, and the fragments were ligated to Illumina sequencing adapters. The ligation products were size-selected via agarose gel electrophoresis, amplified by PCR, and sequenced with an Illumina HiSeq™ 2500 system by Gene Denovo Biotechnology Co. (Guangzhou, China).

Gene annotation and functional characterization

The sequencing reads were subjected to quality control analysis for the removal of short and low-quality reads (those containing adapters, more than 10% unknown nucleotides, or more than 50% low-quality (Q value ≤ 20) bases). All raw data were deposited in the NCBI Gene Expression Omnibus (GEO) with the accession number GSE252706. To identify gene transcripts, all of the reconstructed transcripts were aligned to the reference genome (NCBI: ASM1957865v1) and divided into twelve categories via Cuffcompare. The genes with the class code “u” were unknown transcripts or transcripts from intergenic regions. The edge R package (http://www.rproject.org/) was used to identify differentially expressed genes (DEGs) across samples or groups. Significant DEGs were defined as genes with a |log2FC|>1 and an FDR < 0.05. GO enrichment analysis was performed on the Gene Ontology Consortium website (http://geneontology.org/). KEGG pathway enrichment analyses of DEGs were performed using the KEGG pathway database (https://www.kegg.jp/kegg/pathway.html).

qRT‒PCR (real-time PCR) analysis

The qRT‒PCR primers used were designed via NCBI (https://www.ncbi.nlm.nih.gov/) Primer-BLAST (Table S1). The settings used for primer design in Primer-BLAST were as follows: product length: 100–200 bp, spanning exons and introns; organism: castor bean (taxid: 3988); Tm: 58–62 °C. qRT‒PCR was performed with SYBR Premix TB Green™ Premix Ex Taq™ II (TaKaRa) via an ABI PRISM 7500 Sequence Detection System (Applied Biosystems). The qRT‒PCR mixture contained the following: 100 ng of cDNA, 0.25 µM forward and reverse primers, 10 µL of SYBR, and ddH₂O to a total volume of 20 µL. The reaction program was as follows: 95 °C for 5 min; 40 amplification cycles of 95 °C for 10 s, 60 °C for 30 s, and 72 °C for 15 s. The relative expression levels of the DEGs were measured via the ΔΔCT method [23], and the 18 S rRNA gene was used as a reference for data normalization.

Phenotypic changes in castor seedlings under PEG-6000 stress

Under PEG stress, the growth of castor plants was inhibited, and the plant height, leaf area, and stem diameter were lower than those in the control group (Fig. S1). Additionally, as the duration of PEG-6000 stress increased, the respiration rate of castor roots continuously increased (Fig. 1H). To examine the effects of PEG stress on photosynthesis (Pn, Gs, Ci, and Tr) in castor plants, we measured the dynamic parameters Pn, Gs, Ci, and Tr (Table 1). Photosynthesis (Pn, Gs, Ci, and Tr) under PEG stress on day 7 was significantly lower than that on day 3, and recovery of photosynthesis (Pn, Gs, Ci, and Tr) was observed after PEG was removed (11 d).

The contents of H₂O₂ and MDA in the roots of castor plants significantly increased under PEG-6000 stress for 3 and 7 d compared with those in the control and decreased after 4 d of rehydration (11 d) (Fig. 1A and B). The contents of POD, GR, and SOD in the roots of castor plants under PEG-6000 stress for 3 and 7 d gradually increased (Fig. 1C and D, and Fig. 1G). However, under PEG-6000 stress, the enzymatic activities of APX and CAT in the castor roots of castor initially increased but then decreased. Although the difference between 3 d and 7 d was not significant, the changes during these two periods were very significant compared with those at 0 d (Fig. 1E and F).

Transcriptomic analysis of castor seedlings under PEG-600 stress

To evaluate the changes in gene expression in castor roots under PEG-6000 stress, we conducted RNA-seq analysis on roots subjected to PEG-6000 stress for 3 and 7 d and rehydrated for 4 d (for a total of 11 d). We used TRIzol™ to extract total RNA from each sample and sequenced the RNA on an Illumina HiSeq™ 2500 platform. Initially, transcriptome sequencing generated ~ 64.29 million raw reads and ~ 49.34 million clean reads for all the samples (Table S2). To identify genes related to drought resistance, we used edgeR (FDR < 0.05 and |log₂FC|>1) to perform differential gene expression analysis between the treatment and control groups (Table S3). By comparing Ck3d with T3d, we identified 2,926 DEGs in castor roots, of which 620 were upregulated and 2,306 were downregulated. By comparing CK7d and T7d, we identified 1,507 DEGs, of which 431 were upregulated and 1076 were downregulated. A total of 111 DEGs were identified between Ck11d and T11d, of which 67 were upregulated and 44 were downregulated (Fig. 2).

GO enrichment and KEGG pathway analysis of DEGs

We performed GO functional enrichment analyses of the DEGs in the castor root systems (Fig. 2 and Table S4-1). GO analyses of the 3, 7, and 11 d DEGs revealed 43, 44, and 28 enriched GO terms, respectively. The GO terms significantly enriched (P < 0.05) in the roots under PEG-6000 stress for 3 and 7 d included response to stimulus (GO:0050896), signaling (GO:0023052), biological regulation (GO:0065007), antioxidant activity (GO:0016209), and nucleic acid binding transcription factor activity (GO:0001071). The significant increase in antioxidant activity indicates that antioxidants play an important role in the response of castor plants to drought stress.

The metabolic pathways related to drought stress were identified via the KEGG database (Fig. 2 and Table S4-2). In castor roots subjected to PEG-6000 stress for 3 or 7 d, the metabolic pathways significantly enriched (P < 0.05) with DEGs included glutathione metabolism (ko:00480), fatty acid metabolism (ko:01212), and plant hormone signal transduction (ko:04075). The main metabolic pathway-enriched DEGs in the root system of castor after 4 d (11 d) of rehydration included valine, leucine, and isoleucine degradation (ko:00280); pentose and glucuronate conversion (ko:0040); phenoline metabolism (ko:00360); and phenylpropanoid biosynthesis (ko:00940).

Weighted gene coexpression network analysis

We utilized the R package to perform weighted gene coexpression network analysis (WGCNA) on the DEGs to identify different coexpression modules in castor roots under PEG stress. All the DEGs were assigned to fifteen modules, of which the navajowhite1 module contained the greatest number of genes (1521) (Fig. 3A and B). The navajowhite1 module genes were positively correlated with drought stress and are expected to be upregulated during drought stress to achieve drought tolerance (Fig. 3C and D).

We plotted the log10(FPKM) (fragments per kilobase of exon model per million mapped fragments) values of the genes belonging to the navajowhite1 coexpressed module along with the eigengene expression values to further understand the specificity of the navajowhite1 coexpressed module with respect to the corresponding expression patterns in the different datasets (Fig. 3D). We observed that, in the navajowhite1 module, the gene expression level under drought stress was greater than that in the control, and there was no difference in the gene expression level between the rehydrated group and the control; these results confirmed the positive correlation observed earlier. The genes in the navajowhite1 module were more active early in the response to drought.

GO enrichment and pathway analysis of navajowhite1 module genes

GO term and KEGG pathway enrichment analyses were also conducted to determine the biological functions of the coexpressed DEGs in the navajowhite1 module. As shown in Table S5, the DEGs were significantly enriched in the terms “antioxidants”, “signal transduction”, and “molecular function modulators”. These findings clearly show that the redox process in castor is highly active under PEG-6000 stress. We conducted KEGG pathway enrichment analysis on the coexpressed DEGs in the navajowhite1 module and identified several significantly enriched metabolic pathways, including alanine, aspartate, and glutamate metabolism; amino acid biosynthesis; unsaturated fatty acid biosynthesis; glutathione metabolism; and the spliceosome (Fig. 4). As shown in Fig. 4, DEGs associated with these pathways were upregulated under stress induced by PEG-6000.

We constructed a navajowhite1 modular network and identified two core genes: GA2ox4 and PP2C39 (Fig. 5). To experimentally confirm our results, we selected eight genes from the navajowhite1 module and performed qRT‒PCR to measure their expression levels after 15% PEG-6000 stress for 3, 7, and 11 d (Fig. 6). The results revealed that the gene expression trend was consistent with the transcriptomic results.

Physiological and biochemical responses of castor under PEG-6000 stress

The yield and productivity of castor plants are severely impaired after drought stress. Drought causes lipid peroxidation of the cell membrane and the production of many ROS that damage membranes and increase membrane permeability [24]. In this study, the levels of H₂O₂ and MDA in the root system of castor plants stressed with PEG-6000 for 3 and 7 d were significantly greater than those in the control, indicating that the balance between the production and scavenging of ROS within the cells was disrupted. Studies on crops such as cotton [25], barley [26], corn [27], and rice [28] have shown that drought resistance is related to the ability to clear ROS, but this ability is regulated by enzyme systems such as SOD, POD, CAT, APX, GR, and DHAR [29]. Wang et al. reported that the activities of the SOD, POD, CAT, and GR enzymes in Triarrhena sacchariflora tended to increase under mild drought, whereas under moderate and severe drought, they tended to increase, followed by a downward trend, whereas APX activity tended to increase [30]. However, Yin reported that the SOD, GR, and APX activities in Cerasus humilis leaves tended to decrease after drought stress [31]. In the present study, the APX and CAT activities in castor roots first increased but then decreased after drought stress, and the POD, GR, and SOD activities continued to increase. The results of this study revealed that the response of plant antioxidant enzymes to drought stress varies by species. Drought can damage chloroplasts and reduce photosynthesis rates [32]. Stomata are important channels via which plants exchange gas and water with the external environment and play an important role in helping plants cope with drought. The stomatal conductance (Cs) and intercellular CO₂ concentration (Ci) exhibited the same trend, which suggested that the decrease in the photosynthetic rate was caused by stomatal factors [33]. This study revealed that as the duration of PEG stress in castor plants increased, the Pn, Ci, and Cs decreased. Therefore, it is speculated that the decrease in the Pn of castor under PEG stress is caused by stomatal factors.

DEGs involved in cellular antioxidant activity

Drought stress significantly affects the growth and metabolism of plants, leading to damage to their normal homeostasis. In particular, under water deficit conditions, plants experience oxidative stress and the accumulation of various ROS, such as H₂O₂ and hydroxyl radicals [34]. The regulation of ROS is a key process for plants to adapt to drought, wherein APX and GPX are key enzymes responsible for clearing H₂O₂. Through GO enrichment analysis of the DEGs, we found significant enrichment in “antioxidant activity”. The genes involved in the antioxidant process included APX1 (ID: 8261021), APX2 (ID: 8283829), GPX1 (ID: 8274172), and GPX2 (ID: 8264294). In Arabidopsis, there are eight APXs, three of which are cytAPXs (AtAPX1, AtAPX2, AtAPX6) [34]. APX1 responds to various abiotic stresses, such as selenium, lead, heat, oxidation, and water stress [35–37]. For example, overexpression of APX1 in a cpr5 mutant could clear excess ROS and restore plant growth [36]. Zandalinas et al.. reported that the response of ABA to APX1 is indispensable during water stress [37]. However, research on APX2 has focused mainly on antioxidant activity and light stress, and the response of APX2 to water stress has not been fully elucidated. In this study, APX1 was significantly upregulated in the roots after 3 d of PEG stress, whereas there was no significant difference after 7 d of PEG stress. Additionally, APX2 showed no significant difference in expression in the roots after 3 d of PEG stress but was significantly upregulated after 7 d. Our qRT‒PCR analysis of the APX1 and APX2 genes revealed that, in castor roots under PEG stress, the expression trends of these genes in the treatment group compared with those in the control group aligned with the transcriptomic data (Fig. 6). The changes in the expression levels of APX1 and APX2 in castor roots under PEG stress may be more conducive to maintaining the homeostasis of ROS within the cells. However, the molecular mechanisms of APX1 and APX2 under drought stress still need further investigation. The GPX gene family plays an important role in different stress responses in plants. For example, the upregulation of bZIP68 oxidation mediated by rice GPX1 plays a role in ABA-independent osmotic stress signaling [38]. Arabidopsis plants transformed with the TaGPX1-D gene exhibit increased tolerance to salt and drought stress [39]. Two wheat GPX genes (W69 and W106) endow Arabidopsis plants with tolerance to salt, H₂O₂, and ABA [40]. In this study, transcriptomic analysis revealed that GPX2 was significantly upregulated in roots after 3 d of PEG stress, whereas GPX1 was significantly upregulated in roots after 7 d of PEG stress. We conducted qRT‒PCR analysis of GPX1 and GPX2 and found that the changes in the expression levels of these two genes in the treatment group compared with those in the control group were consistent with the transcriptomic results (Fig. 6). These findings suggest that GPX2 is more sensitive to early PEG stress. These findings help to elucidate how plants respond to drought stress by regulating APX and GPX genes, providing new directions for further research.

DEGs involved in glutathione metabolism

Growing evidence suggests that glutathione metabolism is closely related to the drought stress response [41, 42]. Vuković reported that the level of glutathione disulfide (GSSG) significantly increased and that glutathione (GSH) was strongly depleted but recovered to a lesser extent under drought stress in winter wheat [43]. Previous studies have shown that GSH helps eliminate excess ROS in plants under oxidative stress [44, 45]. Therefore, the GSH content in plants is an important indicator of plant antioxidant activity. In this study, 6 DEGs were involved in glutathione metabolism (Fig. 4). Isocitrate dehydrogenase (IDHC) and GPX2 are involved in the generation of GSSG from GSH. RIR1 and RIR2 are involved in the generation of trypanothione disulfide from GSH. Glutathione S-transferase X6 (GSTX6) and puromycin-sensitive aminopeptidase (PSA) are involved in the production of glycine from GSH. The higher the GSH/GSSG ratio is, the stronger the ability of a plant to clear ROS in the body its [46, 47]. In the present study, GPX2 was upregulated in the root system of castor plants under PEG-6000 stress for 3 d but downregulated at 7 d, indicating that the generation of GSSG by GSH was inhibited under severe drought in these plants, which may have increased the GSH/GSSG ratio, allowing the plants to cope with drought stress. The specific GSH/GSSG ratio may also depend on the consumption of GSH through the other two pathways. Therefore, further research is needed on the ability of GPX2 to clear ROS by regulating the GSH/GSSG ratio.

DEGs involved in fatty acid metabolism

Fatty acid metabolism plays an important role in the response of plants to drought stress [13, 48, 49]. Previous studies have shown that Arabidopsis, Salvia officinalis, Vitis vinifera L., and young Cocos nucifera L. plants present increased levels of unsaturated fatty acids under drought stress [50–53]. An increase in unsaturated fatty acid levels helps maintain the fluidity and integrity of cell membranes [54, 55]. In addition, the double bonds in unsaturated fatty acids can react with free radicals, thereby terminating the free radical chain reaction and preventing cellular oxidative damage [56]. In this study, four of the DEGs were involved in fatty acid metabolism: SDR1 (short-chain dihydrogenase), THIK2 (peroxisomal 3-ketoacyl CoA thiolase), PSA (puromycin-sensitive aminopeptidase), and FAD6C (omega-6 fatty acid deficiency) (Fig. 4). SDR1, STAD, and FAD6C are involved in linoleoyl-CoA formation. Linoleoyl-CoA is a long-chain unsaturated fatty acid that contains two double bonds and can be further converted into other types of unsaturated fatty acids through specific enzymatic reactions, such as fatty acid desaturase and fatty acid elongation enzymes [57]. SDR1 and THIK2 are involved in docosapentaenoyl-CoA formation. Although there is currently no direct evidence to suggest that docosapentaenoyl-CoA is a substrate for the formation of unsaturated fatty acids, it is an intermediate in the formation of unsaturated fatty acids [58]. In this study, the expression of SDR1, STAD, FAD6C, and THIK2 was upregulated in castor roots stressed by PEG-6000, which may promote the synthesis of unsaturated fatty acids, thereby playing a crucial role in maintaining cell membrane stability and preventing oxidative damage to cells.

DEGs involved in plant hormone signaling

Studies have shown that certain components of phytohormone signaling pathways are involved in drought stress. For example, water deficit stress in plants triggers the synthesis and accumulation of ABA in roots and the transport of ABA to aboveground parts and regulates stomatal closure. There exists a dynamic balance between the synthesis and degradation of endogenous ABA in plants. The degradation of ABA is achieved through ABA 8´‑hydroxylase (CYP707A) [59]. Previous studies have shown that inhibition of the CYP707A gene can increase plant drought and salinity tolerance [60–63]. Inhibition of the CYP707A gene in apple seedlings and grapes enhances plant drought resistance [60–62]. In this study, three CYP707A genes, namely, CYP707A4 (ID: 8278867), CYP707A4 (ID: 8262889), and CYP707A1 (ID: 8275341), were differentially expressed (Fig. 7). CYP707A1 is downregulated in castor roots at both 3 and 7 d of PEG stress. However, the two CYP707A4 genes were upregulated in the roots at 3 d of PEG stress and downregulated at 7 d. The changes in the expression of CYP707A1 and CYP707A4 may help maintain the dynamic balance of ABA. CYP707A1 and CYP707A4 were both downregulated in the castor root system under PEG stress for 7 d, promoting the accumulation of ABA and aiding in the clearance of cellular ROS [64, 65]. Type 2 C protein phosphatase (PP2C) is one of the three main components of the ABA signaling pathway and is responsible for regulating various physiological processes, including stomatal movement and the drought response [66]. In the absence of stress, PP2C inhibits the PYR/PYL/RCAR components in the ABA signaling pathway, placing plants in a nonstressed state. However, under stress conditions, the accumulation of ABA inhibits the activity of PP2C, allowing the transmission of ABA signals and activating stress responses [67–69]. In this study, 8 PPC2 genes were differentially expressed, among which 4 (PP2C25, PP2C30, PP2C39, and PP2C06) were downregulated in the root system of castor plants after 7 d of stress. Nevertheless, the expression of PP2C72, PP2C14, PP2C04, and PP2C29 was downregulated in castor root system after 3 d of stress. WGCNA indicated that PP2C39 is one of the core genes of the navajowhite1 module (Fig. 5). The downregulation of PP2C39 expression in the root system of castor plants after 7 d of stress (Fig. 6) may promote the accumulation of ABA and improve the antioxidant capacity of the plant.

Ethylene signaling plays an important role in the response of plants to drought stress. ERFs are members of the AP2/ERF transcription factor superfamily in plants and play a crucial role in the response to drought stress. Under drought stress conditions, ERFs usually positively regulate plant drought resistance, but they also occasionally negatively regulate plant drought resistance [70–72]. ERF109, ERF114, and ERF115 are ROS-responsive factors in Arabidopsis that mediate ROS signaling and control the recognition of the root stem cell niche (SCN) [73]. In quinoa, CqERF24 can increase antioxidant enzyme activity and activate stress-related genes, increasing plant drought resistance [74]. In corn, ZmEREBP60 is a positive regulatory factor in the drought response that regulates H₂O₂ decomposition [75]. A total of nine DEGs were identified in this study; seven genes (ERF25, ERF17, ERF91, ERF113, ERF80, ERF98, and ERF24) were upregulated and two genes (ERF62 and ERF34) were downregulated under PEG-6000 stress (Fig. 7). These differentially expressed ERF genes may respond to drought stress by regulating ROS, but the specific regulatory mechanism requires further in-depth research.

Gibberellin signal transduction plays an important regulatory role in plants, especially in the response to environmental stress. The genes involved in gibberellin signal transduction include gibberellin synthesis genes (GA20ox), gibberellin degradation genes (GA2ox), and gibberellin signaling DELLA protein family genes (GAI, RGA). GA20ox (gibberellin 20 oxidase) is one of the key enzymes in the gibberellin biosynthetic pathway. However, GA2ox (gibberellin 2 oxidase) is a key enzyme in the gibberellin degradation pathway. A functional deficiency in maize ZmGA20ox3 can increase the drought resistance of maize seedlings [76]. Eucalyptus plants overexpressing the EguGA20ox1 and EguGA20ox2 genes exhibit increased sensitivity to abiotic stress, whereas those overexpressing the EguGA2ox1 gene exhibit increased stress resistance [77]. At present, the mechanisms by which GA2ox and GA20ox respond to drought stress are unclear, but there is evidence suggesting that gibberellin signaling responds to abiotic stress by regulating antioxidant defense ability and osmoregulatory substances [78]. GA20ox1 and GA20ox2 were downregulated and GA2ox4 was upregulated in PEG-6000-stressed castor roots in the present study (Fig. 7). The upregulation of GA2ox4 expression (Figs. 6 and 7) contributes to the degradation of endogenous GA in cells, which may increase the antioxidant defense ability of plants through GA signal transduction.

Navajowhite1 module genes involved in the response to drought

We performed WGCNA on the DEGs, and the navajowhite1 module was positively correlated with drought stress; moreover, we constructed a gene network map of navajowhite1. In addition to PP2C39 and GA2ox4, NCA14 is one of the core genes (Fig. 7). Many studies have shown that NAC transcription factors positively regulate the response of plants to drought stress [79–82]. The TwNAC01 gene enhances the antioxidant capacity of Arabidopsis plants under drought stress [83]. In addition, overexpression of ThNAC4, VvNAC17, OsNAC006, and AfNAC1 increased the antioxidant capacity of the plants [79, 84–86]. In the present study, the upregulation of NCA14 (ncbi_8270777) expression in the castor root system under PEG-6000 stress may have contributed to the antioxidant capacity of the plant. The MYB transcription factor family is a class of gene regulatory proteins that are widely found in plants; these transcription factors play important roles in plant growth and development and in the response to environmental stresses. MYB44, a member of the MYB family, has different functions in the response to drought stress. For example, GhMYB44 responds to drought stress by regulating plant antioxidant capacity and stomatal aperture [87]. However, BcMYB44 responds to drought stress by inhibiting the accumulation of anthocyanins and regulating stomatal movement and the ROS balance [88]. In the navajowhite1 module in this study, the MYB family members MYB44 (ncbi_8273467), MB3R1 (ncbi_8265448), and MYB05 (ncbi_8267898) were included. MYB44 is upregulated under PEG stress, which helps improve the antioxidant capacity of plants.

Oxidative stress model of the response of castor plants to drought stress

In this study, an antioxidant stress model of the response of the castor root system to drought stress was constructed on the basis of GO functional enrichment analysis, KEGG pathway analysis, and WGCNA of the DEGs (Fig. 8). First, fatty acid metabolism plays a protective role in maintaining the membrane of castor plants under drought stress and preventing oxidative damage. In addition, antioxidant enzymes, glutathione metabolism, plant hormone signal transduction, NAC transcription factors, and MYB transcription factors play important roles in maintaining the ROS balance in castor root cells under drought stress.

Ethics approval and consent to participate

Not applicable.

Consent for publication

All the authors have agreed to the publication of this study.

Competing interests

The authors declare no competing interests.