The candidate gene “GhGLK1” was identified in the BC₂F₂ population among the wild male donor parent, i.e., Gossypium Tomentosum, which is noted for its drought tolerance, and the recurrent parent, i.e., Gossypium hirsutum, a high yielding but drought susceptible species via genotype by sequencing (Magwanga et al., 2018). A. thaliana (Col-0) was screened from T0 to T3. The sterilized seeds were grown on a 1/2 murashige and skoog (MS) medium and then allowed to grow in suitable 22°C (16 h light/8 h dark) conditions. Then, the plants were transplanted into a nutrient mixture of nutrient soil and vermiculite with a 1:1 ratio. For cold stress treatment, after 3–4 weeks of transplantation, we selected the consistent WT and overexpression T3 generation plants of the Arabidopsis and placed them in a refrigerator at 4°C. After 7 days of cold treatment, the survival rates of plants were recorded. To simulate drought, 16% PEG6000 solution was applied, and germination rate and root length were determined using the 1/2 MS medium containing mannitol. Mannitol concentrations were set to 100, 200, and 300 mM in three concentration gradients, with the optimum concentration of 200 mM mannitol selected after pretesting. Each plastic canopy contains five Arabidopsis seedlings, each set of 12 boxes (Lu et al., 2019).

Marie Galante-85, a wild species of terrestrial cotton, was used in the experiment, and the carrier was TRV2. Seeds were soaked, the filter paper method was used for sowing, and 4 days after sprouting were transplanted into a hydroponic setup filled with Hoagland nutrients. Greenhouse conditions were 28°C during the day, 25°C at night with 16 h/8 h light/dark cycle, and relative humidity of 60–70%. The plants were moved to a 4°C growth chamber for the cold stress treatment, and then samples were taken at 0, 3, 6, 12, and 24 h treatment time. Immediately after sampling, the leaf samples are collected in liquid nitrogen and stored at a −80°C refrigerator until RNA was extracted. The RNAprep plant kit, which is supplied by Tiangen Biotech¹, was used during RNA extraction by following the kit procedures.

To clone the GhGLK gene, BamHI, and SacI, restriction enzymes with forward primer “GAGAACACGGGGGACTCTAGAATG CTAGCTGTGTCACCTTTGAGG” and reverse primer “AC GGGGGACTCTAGAGGATCCCCTCAAAGGTGACACAGCT AGCAT” were used. The vector used in the construction of overexpression recombinant was PBI121, and the specific method of constructing recombinant carrier using the frozen-melt method to transform recombinant plasmids into Agrobacterium GV3101 receptor, coated on lysogeny broth (LB) media containing Kanamycin and Rifamycin. After shaking culture for 48 h, a single bacterial colony was picked for PCR and gel electrophoresis confirmations. Floral dipping is used to apply the suspension solution in WT Arabidopsis (Lu et al., 2019).

Using cDNA as a template, the target gene is amplified according to specific primers (forward: GTGAGTAAGGTTA CCGAATTCAGTGAAGGTGGATTGGACGC and reverse: CGTGAGC GGTACCGGATAGCCCCCCCCGCATATGATTGC TCTG). The VIGS vector construction was made using the double enzyme cutting method of EcoRI and BamHI enzymes (Yang et al., 2019). The required cultures are TRV2:GhGLK1, empty TRV2:00, and TRV2: PDS. The Agrobacterium solution OD600 concentration is approximately 1.5. After injecting cotton leaves, the plants were placed in the dark for 24 h, and later, plants were exposed to normal growth conditions (Mehari et al., 2021). After gene silencing, plants were grown to three leaves for treatment, using a nutrient solution containing 17% PEG6000 to simulate drought treatment, and for cold treatment, 4°C growth chamber temperature at 16 h/8 h light/dark conditions.

Physiological index determination was taken for dehydration rate [excised leaf water loss (ELWL)], relative moisture content [relative leaf water content (RLWL)], ion permeability [cell membrane stability (CMS)], and chlorophyll content [soil plant analysis development (SPAD)] at 0 and 24 h time points. Biochemical index evaluation, mainly SOD, PRO, hydrogen peroxide (H₂O₂), and MDA leaf samples, was performed at 0 and 24 h after treatment, and the Solarbio (Solarbio, Beijing, China) test kit was used to determine the enzyme assay according to specific procedures shown in the instructions.

Trypan blue and 3,3′-diaminobenzidine (DAB) experiment was performed to see the cell damage of control and treated plants. Taking the leaves of WT and overexpressed Arabidopsis plants, before and after drought and cold stress, soaking them in trypan blue dye solution (10 ml of lactic acid, 10 ml of glycerin, 10 g of phenol, 1 0 mg of trypan blue, and 10 ml of distilled water), placing in boiling water bath for 2 min, and later cooling in hydrated chloral (2.5 g/ml), daily replacement of decolorization solution was performed. At least three biological replications were made for each group of three leaf blades. The DAB dyeing experiment used a DAB color rendering kit from the Nanjing Institute of Bioengineering. DAB color rendering work liquid preparation includes reagent A and reagent B in the 1:19 ratio according to the use of mixing and ready to use. Color rendering time is 8–12 h for WT and overexpression Arabidopsis leaves in the prepared DAB color rendering liquid. During this time, observe staining to prevent excessive staining. Use alcohol for gradient discoloration. Wash off the blade color 2–3 times with ethanol for 30 min at a time. Use another 70% ethanol to continue to wash off the leaf color 2–3 times for 30 min at a time. Each group of three blades is repeated three times (Li et al., 2019).

Using vector pCAMBIA2300-eGFP-Flag, the sequence of GhGLK1 was amplified by forward primer (GAGAACA CGGGGGACTCTAGAATGCTAGCTGTGTCACCTTTGAGG) and reverse primer (ACCCATGTTAATTAAGGATCCACCCAT TGTGGGTGGAAGCC) restriction enzymes containing Xba1 and BamH1 cutting sites. Then, the agroinfiltration approach was used to transform the generated recombinant gene into tobacco leaves. Finally, to determine the location of the candidate gene, the transformed tobacco leaves were incubated at 25°C for 24 h in the dark and viewed using fluorescence microscopy (Mehari et al., 2021).

Leaf samples were taken at 0 and 24 h for the extraction of RNA. RNA was extracted by using the Tiangen Biotech (see text footnote 1) kit. The GhActin gene forward sequence ATCCTCTCTTGACCTTG and the reverse sequence TGTCCGCAGGCAACTACTCAT were used as a reference standard for qRT-PCR analysis. For RT-qPCR primer design, the NCBI² website was used. RT-qPCR analysis using AceQ Universal SYBR Green qPCR Master Mix (Vazyme, Nanjing, China) is performed, as detailed in the instruction manual. Specific primers were designed, and RT-qPCR analysis was also performed in four stress-responsive genes, namely, COR15A, RD29A, KIN1, and COR47, to determine their expression in plants after stress treatment (Supplementary Table 1).

The statistical analysis was performed via the SPSS package at the 5% probability level, and when the significant variation was observed, the mean comparison was performed using the Duncan multiple range test (DMRT). The error bars in each graph represent the mean values and SE of three replicates. GraphPad Prism was used to do the graphs of this study. In this study, all the experiments were carried out in three replications.

Based on the RT-qPCR results of the T2 generation, we selected 3 out of the 16 overexpressing lines with higher levels of GhGLK1 expression. In this study, we selected OE-5, OE-14, and OE-16 for further study (Figure 1). Further screening of T3 generations of expression of pure seeds for phenotype identification and functional verification of the target gene was performed under drought and cold stress.

We selected 3–4-week-old overexpression lines and WTs with consistent growth in nutritional pots for drought and cold tolerance treatment. In the control group, the phenotype of WT and the three overexpressing lines (i.e., OE-5, OE-14, and OE-16) was the same, but after 20% PEG-6000 treatment for 48 h, the overexpressing lines showed a more resistant phenotype with less wilting symptoms as compared with WT. Whereas, under cold stress, after 7 days of culture at −15°C, the three overexpressing lines somewhat reverted to green, with new leaves, whereas the WT of Arabidopsis almost completely withered and turned yellow (Figure 2A). After drought treatment, the survival rates of the three overexpressing lines (i.e., OE-5, OE-14, and OE-16) were 81.3, 49, and 71%, respectively, compared with 8.3% for WT Arabidopsis. Under cold stress, the survival rates of the three overexpressing lines (i.e., OE-5, OE-14, and OE-16) reached 64.3, 63.7, and 37.7%, respectively, and for WT 11.7%, the survival rate was recorded. RT-qPCR was used to find the expression level of the GhGLK1 gene, and it was found to be meaningfully higher in three overexpressing lines relative to WT, with the highest expression in the OE-5 line (Figures 2B,C).

Under normal conditions, the germination rate of WT, OE-5, OE-14, and OE-16 overexpression seeds was similar, while the germination rate of the WT after drought stress treatment was only 44.4%, as compared with 75.7, 49.7, and 74% for the three overexpressed lines, and there was a significant difference among OE-5, OE-16, and WT (Figure 3A). The germination percentage of the three overexpressing lines was significantly higher than that of WT germination, indicating that the resistance of seeds to drought and cold increased after the expression of the gene GhGLK1 in the A. thaliana (Figure 3B). Root length determination experiments in WT and OE-5, OE-14, and OE-16 lines before treatment were not significantly different, but after drought and cold treatment, the root length of the three overexpressing lines grows meaningfully faster than the WT (Figure 3C). Under normal conditions, the root length of WT and overexpressing lines varies between 25.2 and 26.2 mm, whereas, after drought and cold stress treatments, the root length in WT was recorded with a maximum average of 6.3 mm, and the overexpressing lines scored between 11.2 and 11.6 mm and 9.5 and 12.2 mm in both drought and cold stress treatments, respectively (Supplementary Table 2). The superior performance of overexpressing lines in germination and root growth showed that the GhGLK1 gene has the potential to increase tolerance to both stresses in the overexpressing lines of A. thaliana.

3,3′-Diaminobenzidine imaging exhibited that under the normal conditions, the H₂O₂ production was very low, and the appearance of brown color was almost invisible. The brown region in the overexpressed Arabidopsis leaves was substantially smaller than the WT under drought and cold stress, and the dyeing depth was lighter, and the phenotype was more apparent during drought stress. It is shown that under the same stress, the overexpressing lines suffer less damage (Figure 4A). The results of the trypan blue staining experiment also showed that the area dyed blue in case of overexpressing lines and WT leaves is small under control environments. However, after drought and cold stress treatment, the leaves of overexpressing lines (i.e., OE-5, OE-14, and OE-16) have significantly lesser blue areas than WT leaves. Light blue color indicates that OE-5, OE-14, and OE-16 overexpressing lines suffered less damage than the WT (Figure 4B).

The ion permeability of WT and OE-5, OE-14, and OE-16 overexpression lines were recorded lower under control situations. In stress conditions of drought and cold, plants experienced injury, and as a result, leaves get damaged, and in this study, we found out that overexpressing lines were more resistant to the said stresses as compared with the WT. The ion exchange in the WT leaves was recorded more relative to the overexpressed lines. With the extension of cold and drought stress for 3 h, the ion permeability of WT kept on increasing but in the case of overexpressed lines, ion permeability was recorded stable. So, we assumed that the cell membrane stability of the overexpressing lines was more as compared with the WT, thus, making overexpressing lines more resistant to drought and cold stress. This suggests that overexpression of the GhGLK1 gene boosted plant resistance to cold and drought stress in the overexpressing lines (Figures 5E,F).

Four stress-responsive genes were investigated in GhGLK to see the relative expression of these genes under both stresses in the wild and overexpressing lines. The results of RT-qPCR revealed that the expression levels of the four stress response genes (i.e., COR15A, RD29A, KIN1, and COR47) were significantly higher in the three overexpressing lines, i.e., OE-5, OE-14, and OE-16, than the WT (Figures 5A–D).

The GhGLK1 gene was localized in the nucleus based on online predictions. To further verify whether GhGLK1 was expressed in the nucleus or not, we constructed the pCAMBIA2300-eGFP-Flag-GhGLK1 fusion expression vector. After transformation to the leaves, we observed the subcellular localization via fluorescence microscope and found that the empty vector was localized in the nuclear membrane, while the vector carrying the target gene is expressed in the nucleus (Figure 6B).

The albino phenotype appeared approximately 8 days after injecting PDS to the cotton leaves, and after 23 days, the albino phenotype was still evident, indicating that gene silencing was successful and the effect was stable (Figure 7A). The expression of GhGLK1 in WT, TRV2:00, and TRV2:GhGLK1 plants was measured using RT-qPCR. No significant variations were observed in the expression of the GhGLK1 gene in WT and TRV2:00, but the expression of the gene in TRV2:GhGLK1 plants was significantly lower than that of WT and TRV2:00, indicating the success of gene silencing (Figure 7B).

In the WT group, growth was more robust, while plants with TRV2:00, TRV2:GhGLK1 suffered mechanical damage during VIGS injection (Figure 8A). After simulated drought treatment of 24 h, WT and TRV2:00 empty plant leaves showed slight wilting, but TRV2:GhGLK1 seedlings were significantly more vulnerable to drought as compared with WT and empty vector (Figure 8B). Higher water loss and wilting in the silenced plants indicate that the drought resistance of the plant was reduced after GhGLK1 silencing. After cold stress treatment of 24 h, there was no considerable difference among WT and TRV2:00 empty vector, but the plants with silent GhGLK1 gene experienced a severe damage than the WT and the empty vector. The stalks of silenced plants softened, and the plants began to show slight signs of inverting, indicating that the cooling resistance of the plant decreased after the gene was silent (Figure 8C).

Evaluation of the water loss percentage in cotton leaves showed that there was no significant variation among WT, TRV2:00, and TRV2:GhGLK1 plants under normal conditions, but after drought and cold stress treatment, the water loss rate in TRV2:GhGLK1 plants was significantly higher than WT and empty vector, indicating a decrease in the water holding capacity of plants after the silencing of the candidate gene. When plants are under stress, damage to the cell membrane causes ions to seep out of the cell membrane. The ion permeability increased significantly in TRV2:GhGLK1 plants than WT and empty vectors, indicating that cell membrane damage was more severe in TRV2:GhGLK1 plants after drought and cold stress. SPAD meter can instantly measure the relative content of chlorophyll in cotton leaves, and the leaf chlorophyll content can indicate the growth of cotton to a certain extent. We observed that under drought stress, the chlorophyll contents of the silenced plants were reduced more as compared with the WT and empty vector plants. After drought and cold stress, the relative water content of silenced plants decreased significantly, compared with WT and plants injected with empty vector (Figure 9). Current results support a decrease in drought and cold resistance of plants after silencing of GhGLK1.

The SOD contents of plants with the silenced gene were significantly lower than WT and empty vectors under both drought and cold stress conditions. After applying cold and drought stress, the proline content of WT, TRV2:00, and TRV2:GhGLK1 plants increased, but it was significantly lower in TRV2:GhGLK1 gene silent plants than in WT and empty vector plants. After treatment, the H₂O₂ content of WT, TRV2:00, and TRV2:GhGLK1 plants increased as compared with control, but it can be seen that the content of H₂O₂ in plants with silent gene was significantly higher than WT and empty vector. After stress, MDA levels were significantly higher in plants with TRV2:GhGLK1 silent gene than WT and empty vector, indicating that lipid oxidation was more severe (Figure 10). These results further indicate that plants after silencing the GhGLK1 gene suffer more severe damage and are less resistant to drought and cold stress.