The cotton cultivar CCRI10 (premature senescence) was used for gene cloning and subjected to different treatments. Healthy and uniform seeds were selected and cultivated in a culture room at 25°C with a 16-h light/8-h dark cycle. Arabidopsis thaliana ecotype Col-0 was used as the parent strain for Arabidopsis transformation. Surface-sterilized seeds were sown in 1/2 MS medium and incubated at 4°C for 3 days in dark conditions to break dormancy, and the plants were then cultivated at 22°C with a 16-h light/8-h dark cycle and a relative humidity of approximately 80%.

Cotton was subjected to drought and salinity treatments in three ways. First, detached true leaves were submerged in water for 1 day and then transferred into a 20% PEG6000 or 200 mM NaCl solution, and controls were submerged in water. Samples were collected at 0, 2, 4, 6, 8, 12, and 24 h. Second, roots were submerged in a 20% PEG6000 or 200 mM NaCl solution, and the same amount of water was used for the control. The leaves were collected at 0 Day post anthesis (DPA), 5, 10, 15, and 20 Days. Third, healthy seeds were germinated in sealed glass jars containing either 200 mM mannitol or 200 mM NaCl, and the leaves and roots were collected separately once the seedlings showed an obvious phenotype. The control consisted of only the medium.

Two plant hormone treatments were carried out. First, seedlings at the three-leaf stage were sprayed with 50 μM ABA, 200 μM meJA or 200 μM ethephon (ethephon releases ethylene when dissolved in water), and controls were sprayed with water. The leaf samples were collected at 0, 4, 12, 24, and 48 h after treatment. Second, 25 μM ABA, 100 μM meJA and 100 μM ethephon were individually applied to the sealed glass jars, and then the leaf and root samples were collected separately. The control consisted of only MS medium.

The cotyledons, true leaves, stems, roots and flowers of cotton seedlings were collected for gene expression analyses. Fibers at 0, 5, 10, 15, 20, and 25 DPA were obtained as described previously (Singh et al., 2009). To evaluate the function of GhNAC79 during cotyledon development, two short-season cotton cultivars, CCRI10 (premature senescence) and Liao4086 (no premature senescence), were selected and cultivated under normal water and nitrogen management. Cotyledon samples were collected every 7 days in each of eight developmental stages from the non-senescent stage to the completely senescent stage. There were three biological replicates for each sample.

The enzymatic solution was prepared as follows: 0.5 g samples were quickly ground in a cold mortar with 8 μl of cold extraction medium (0.05 mol/L, pH = 7.8, Na₂HPO₄-NaH₂PO₄), and all the liquid was then transferred into a 25 ml centrifuge tube and centrifuged at 20,000 rpm for 30 min at 2°C. The supernatants were collected for further use as the enzymatic solution.

The MDA content was determined according to the method described by Jingqing Zhao et al. (2012) with slight modification. A mixture of 1.5 ml of buffer (0.05 mol/L, pH = 7.8, Na₂HPO₄-NaH₂PO₄) and 1.5 ml of enzymatic solution was placed in a 10 ml centrifuge tube (the control consisted of 1.5 ml of buffer and 1.5 ml of H₂O), and 2.5 ml of 0.5% thiobarbituric acid (TBA) was then added to each tube. Finally, the reaction mixture was mixed well, incubated at 100°C for 20 min, and then quickly put on ice. After centrifugation at 1800 g for 10 min, the absorbances of the supernatant at 450, 532, and 663 nm were determined with a spectrophotometer. The soluble protein was determined according to the method described by Bradford (1976).

Total RNA was extracted with an RNAprep Pure kit (Tiangen, China), which was suitable for plants rich in polyphenols and amylase. The quality and concentration of the RNA were confirmed by 1% agarose gel electrophoresis and a spectrophotometer, and DNaseI was used to remove the genomic DNA. cDNA synthesis was performed by strictly following the manufacturer’s protocol for ReverTra Ace qPCR RT Master Mix (TOYOBO, Japan). Total genomic DNA was extracted from the cotton leaf tissues by the CTAB method (Song et al., 1998).

The primers used for gene cloning, qRT-PCR, VIGS and promoter cloning were designed with OLIGO7, as shown in Table 1. The conserved domain of GhNAC79 was searched for in NCBI¹. Using the G. hirsutum genome database (Zhang et al., 2015), GhNAC79 was pitched at scaffold42.1, and the intron and exon structure was analyzed by comparing the genomic and coding sequences. At the same time, an unrooted phylogenetic tree was constructed with 126 AtNACs by MEGA 5.1 (Tamura et al., 2011).

According to the domain features of GhNAC79, the entire open reading frame (ORF) was divided into four parts, which were inserted into the pGBKT7 vector with EcoRI and BamHI. After sequencing with the T7 universal primer, all the constructed plasmids were transferred into a yeast strain (Y₂H) by strictly following the Clontech method². The positive yeast strain was selected using SD/-Trp/X-α-gal/25 mM 3-AT medium, and PCR was used to verify the results. The primers were designed based on the sequences of the four fragments and are shown in Table 2.

For promoter cloning, approximately 1.0 kb sequences of the 5′ UTR were isolated based on the information from the genome database. Using XbaI and ScaI cutting sites, the promoter was inserted into the pBI121::GUS expression vector and transferred into Arabidopsis through Agrobacterium tumefaciens strain LBA4404 with GUS as the reporter. At the same time, PlantCARE³ was used to predict the important elements in the promoter.

Histochemical staining for GUS activity was conducted by incubating fresh Arabidopsis tissues in the GUS staining solution (Jefferson et al., 1987) with the wild type as the control. After one night of incubation at 37°C, the stained tissues were bleached with ethanol ranging from a high concentration to a low concentration until the wild type became white. Tissues were then photographed directly under a stereomicroscope.

The primer was designed with OLIGO7 according to the C-terminus of the sequence, and the specificity of the primer was ensured by the melting peaks and dissociation curves. The reactions were performed with Go Taq qPCR Master Mix (Promega, United States) on an ABI7500 instrument (Applied Biosystems, United States) with GhHis3 as the endo-reference gene (Tu et al., 2007). A total reaction volume of 20 μl was used for qRT-PCR as follows: 2 μl of diluted cDNA, 1.0 μl of both the forward and reserve primers, 6 μl of sterile H₂O, and 10 μl of Go Taq qPCR Master Mix. The running procedure was strictly established as described in the manual, and each collected sample contained three technical replicates. The entire operation was performed under low-light conditions, and the results were calculated using the 2^-ΔΔC_T method (Livak and Schmittgen, 2001), where ΔC_t1 = Ct_(GhNAC79)-Ct_(GhHis3) and ΔC_t2 = Ct_1(GhNAC79)-Ct_{1(GhNAC79of control)}. Statistical significance was calculated by analysis of variance (ANOVA) or t-tests using SAS software (Knapp et al., 1985).

The 35S::GhNAC79 plasmid was constructed by inserting the coding region into a binary vector, pBI121, with XbaI and SacI cutting sites. After sequencing, the 35S::GhNAC79 plasmid was transformed into A. tumefaciens strain LBA4404 by the heat shock method (Bhuiyan et al., 2011), and Arabidopsis was transfected with an improved inflorescence-dip method (Clough and Bent, 1998). The T₀ transgenic plants were germinated on 1/2 MS medium with kanamycin. After 14 days, the green plants were transferred to nutritive soil in a culture room, and PCR was used to analyze the positive plants.

The transformation of cotton was completed at the Institute of Cotton Research of the Chinese Academy of Agricultural Sciences (CAAS). Seeds of CCRI24 (an upland cotton cultivar) were sterilized and germinated on 1/2 MS medium at 28°C in a culture room, and hypocotyls cut from sterile seedlings served as the transformation receptors. The construction of 35S::GhNAC79 was transferred into cotton via Agrobacterium (Zhang, 2008). The positive cotton lines were verified with PCR and kanamycin, and the T₁ generations of the GhANC79-overexpression lines were used for further analysis. The primer used for PCR was designed based on the 35S promoter and the GhNAC79 sequence, as shown in Supplementary Table S1.

To improve the robustness of the virus-induced cotton GhNAC79 silencing results, we used two silencing systems: pCLCrVA-pCLCrVB and pYL156-pYL192. Approximately 200 bp of the GhNAC79 sequence was inserted into the pCLCrVA vector with SpeI and AscI cutting sites and pCLCrVB as the helper vector (Gu et al., 2014). pCLCrVA::GhNAC79 and pCLCrVB were transferred into cotton cotyledons through Agrobacterium as previously described (Kumagai et al., 1995) with pLCrVA::PDS as the indicator and pLCrVA-infected plants as the negative control. At the same time, 200 bp of the GhNAC79 sequence was inserted into the pYL156 vector with EcoRI and BamHI with pYL192 as the helper vector, pYL156::PDS as the indicator, and pYL156-infected plants as the negative control (Unver and Budak, 2009). The cotton used for infection was cultivated at 22°C until the cotyledons flattened, and the infected cotton seedlings were grown at 22°C under low humidity.

GhNAC79 (KU963586, unpublished) was cloned from a senescent cotyledon of G. hirsutum. According to the genome information, GhNAC79 was located on scaffold42.1 and contained 3 exons and 2 introns. It further contained a conserved NAM domain and encoded a protein consisting of 231 amino acids with a molecular weight of 26.972 kDa.

The GhNAC79 sequence was divided into four fragments. The first fragment was from 23 to 485 bp and contained the conserved NAM domain. The second fragment was slightly longer than the first fragment, from 23 to 613 bp, which was selected to verify the definite transcriptional activation domain. The third fragment was from 479 to 767 bp, which avoided the conserved domain. Finally, the fourth fragment contained the complete ORF. The constructed plasmids pGBKT7-fra1, pGBKT7-fra2, pGBKT7-fra3, and pGBKT7-fra4 were transformed into Y₂H, and the results are shown in Figure 1. The positive control stains (pGBKT7-p53 + pGADT7-largeT) appeared blue in color on X-α-Gal substrate, whereas the negative control (pGBKT7-laminC + pGADT7-largeT) exhibited no change in color. The GhNAC79 fragments, except pGBKT7-fra1, appeared blue in color, which indicated that GhNAC79 functioned as a transcriptional activator and that its transactivation domain was located at the C-terminus. To ensure the accuracy of the results, all of the stains were affirmed by PCR.

To detect the special expression patterns of GhNAC79 in different tissues, the roots, cotyledons, stems, true leaves, flowers and fiber at different stages were collected. As shown in Figure 2A, GhNAC79 was mainly expressed in the cotyledon and later stages of fiber development, such as 20 and 25 DPA. In contrast, the expression level was relatively low in the stem and the initial fiber developmental stage.

The GhNAC79 promoter was cloned, and the sequenced promoter contained parts of the ORF sequences, which demonstrated consistency from the ORF to 5′ UTR. Based on the PlantCARE analysis, the 1.0 kb sequences contained all basal elements, such as the TATA box, CAAT box and 5 UTR Py-rich stretch. Additionally, the promoter contained some stress- and regulation-related cis-elements, such as those related to light, circadian rhythm, drought and ethylene, as shown in Table 3. The reconstructed vector _promoterGhNAC79::GUS was transferred into Arabidopsis, and histochemical staining was conducted to detect the GUS activity. As shown in Figures 2B–F, there was no GUS activity in the wild type, while GUS activity was preferentially detected in the cotyledons and cotyledon bases in transgenic Arabidopsis, which is presented in Figures 2G–K.

The promoter of GhNAC79-induced GUS was especially expressed in the cotyledon, which was consistent with the expression patterns of GhNAC79 in cotton tissues. The high expression level of GhNAC79 in the cotyledon and fiber indicated its special function in these tissues.

During development, the color of the cotyledon became yellow in both cultivars (Figure 3A), and the MDA content sharply increased in CCRI10 but was relatively steady in Liao4086, as shown in Figure 3B. As a product of membrane lipid peroxidation, MDA is an important indicator of cell damage (Kramer et al., 1991), and the high level of MDA in CCRI10 indicated premature senility. In terms of soluble protein, the two cultivars did not differ greatly, and the soluble protein contents of both decreased during leaf senescence, as shown in Figure 3C.

Figure 3D shows that GhNAC79 was very highly expressed in CCRI10 and Liao4086 in later cotyledon development stages, and the highest changes in expression level were approximately 170-fold in CCRI10 and 40-fold in Liao4086. The expression of GhNAC79 was significantly higher in CCRI10 than in Liao4086. As CCRI10 was more sensitive to senescence than Liao4086, the higher expression of GhNAC79 in CCRI10 indicated that it might function in cotyledon senescence.

To explore the role of GhNAC79 in response to different stresses, cotton seeds were sown in sealed glass jars containing MS medium subjected to different treatments. As shown in Figures 4A,B, GhNAC79 was induced in leaves by the meJA and drought treatments, but its expression in roots was repressed in all treatments, especially the meJA, ethylene and ABA treatments, by approximately 5∼10 fold. These results indicated that GhNAC79 might be involved in responses to abiotic stresses.

To verify the response of GhNAC79 to drought and salt, detached cotton leaves were submerged in 20% PEG6000 or 200 mM NaCl. Figure 4C shows that GhNAC79 was sharply induced at 24 h after drought treatment, whereas GhNAC79 remained constant under salt treatment. At the same time, after submerging the roots of cotton seedlings in 20% PEG6000 or 200 mM NaCl, GhNAC79 exhibited an expression pattern indicating sensitivity to drought treatment, as shown in Figure 4D. In terms of the response to different plant hormones, GhNAC79 was very sensitive to ethylene, but its response to ABA and meJA remained relatively steady, as illustrated in Figure 4E.

35S::GhNAC79 was structured (Figure 5A) and transformed into Arabidopsis with selective 1/2 MS medium, and the results were verified by PCR. The positive plants were selected and grown in a culture room, and the T₄-generation transgenic plants were used for phenotype analysis. At the transcriptional level, the expression of GhNAC79 in the four lines was higher than that of the wild type by approximately 5∼80 fold, as shown in Figure 5B. Transgenic Arabidopsis also exhibited an early flowering phenotype, which is shown in Figure 5C, and the flowering time was approximately 5 days earlier than the wild type, as presented in Table 4.

To explore the function of GhNAC79 during drought stress, transgenic and wild type Arabidopsis were treated with 20% PEG6000 after 20 days of sowing. Under normal water management, all plants reached the vegetable developmental stage, and these results are shown in Figure 6A. After drought treatment, transgenic Arabidopsis began bolting, but the wild type was still at the vegetable stage, as shown in Figure 6B. Additionally, transgenic Arabidopsis and wild type were treated with 20% PEG6000 after 14 days of sowing. After 3 days of drought stress, wild type exhibited obvious wilting, while the transgenic Arabidopsis remained fresh. Furthermore, the stomatal aperture of transgenic Arabidopsis was smaller than that of the wild type, as shown in Figures 6C,D.

Transgenic Arabidopsis and wild type were germinated in 1/2 MS medium containing 100 mM mannitol, and after approximately 15 days, the transgenic plants displayed an obvious phenotype. As shown in Figure 6E, the transgenic plants were stronger than the wild type, and their roots were longer. The dry/wet ratio of the transgenic plants was significantly lower than that of the wild type, as visualized in Figure 6F, and the high dry/wet ratio in the wild type implied relatively high water loss. The enhanced drought tolerance in Arabidopsis indicated the positive role of GhNAC79 in drought response.

To study the function of GhNAC79 in cotton during drought stress, two virus-induced gene silencing (VIGS) systems were used (pCLCrVA-pCLCrVB and pYL156-pYL192). After injection, cotton seedlings were covered with black boxes for one night and then grown in a culture room. As the leaves of the indicator plants exhibited an albino phenotype, the expression level of GhNAC79 was detected by qRT-PCR. In the pCLCrVA-pCLCrVB system, the expression level of GhNAC79 was 0.2∼0.6% lower in the infected plants compared with the control (pCLCrVA vector-infected plants), as shown in Figure 7A. After treatment with 20% PEG6000, the infected plants showed severe wilting, and all leaves were soft and drooping. However, the young leaves of the control plants remained fresh, as shown in Figure 7B. For the pYL156-pYL192 system, the expression of GhNAC79 decreased by 0.2∼0.9% relative to the control plants (pYL156 vector-infected plants), and after 20% PEG6000 treatment, the infected plants showed more wilting than the control, as shown in Figures 7C,D.

To analyze the overexpression of GhNAC79 in different cotton lines, the roots of two transgenic cotton lines and a wild type (CCRI24) were submerged in 20% PEG6000, and the results are shown in Figures 7E–G. After 7 days of treatment, the leaves of the wild type appeared wilted or had dropped, while the transgenic cotton showed normal growth. After 14 days, the wild type had dropped three true leaves, but the two transgenic lines had dropped one leaf and one cotyledon and had another unhealthy cotyledon. Furthermore, there was no significant difference in stoma number and the size of the stomatal aperture between the transgenic lines and the wild type under normal water management, but after drought treatment, the stomatal apertures were smaller in the two transgenic lines compared to the wild type.

Two VIGS-induced GhNAC79 silencing systems resulted in the cotton being more sensitive to drought, while overexpression of GhNAC79 increased the drought tolerance of cotton, which demonstrated the positive role of GhNAC79 in drought tolerance in cotton.