Transcriptome Analyses on Compound Material Regulating Saline Stress and Alkaline Stress on Cotton

Background: Soil salinization and alkalinization are the main factors that affect the agricultural productivity in the world. Evaluating the persistence of the modier applied in eld soils is an important part of the regulation of saline stress and alkaline stress. Result: To determine the molecular mechanism of cotton’s responses to the regulation of saline stress and alkaline stress by the modier, in this study, cotton was planted in the salinized soil (NaCl 8g kg -1 ) and alkalized soil (Na 2 CO 3 8g kg -1 ) after application of the modier, and ion content, physiological characteristics, and transcription and sequencing of new leaves during the owering and boll-forming stage of cotton were analyzed. The results showed that compared with saline stress, alkaline stress increased the content of Na + , K + , SOD, and MDA in leaves, and the application of modier reduced the content of Na + but increased the K + /Na + ratio, the activities of SOD, POD, CAT, and REC. Transcriptome analysis revealed that after the application of the modier, the Na + /H + exchanger gene in cotton leaves was down-regulated, the K + transporter, K + channel and POD genes were up-regulated. Besides, the down-regulation of genes related to lignin synthesis in phenylalanine biosynthesis pathway was consistent with the study results of ion content and physiological characteristics in leaves. The quantitative analysis with PCR proved the reliability of the results of RNA sequencing. Conclusion: These ndings indicate that the modier alleviated saline stress and alkaline stress on cotton by regulating candidate genes in key biological pathways, which improves our understanding of the molecular mechanism of the modier regulating saline stress and alkaline stress.

Cotton (Gossypium spp.) is one of the most important economic crops in the world, among which Gossypium hirsutum L. has been widely planted and its planting area accounts for more than 95% of the global cotton production. Although cotton is of salt tolerance, its growth and development still be affected by saline stress and alkaline stress [14]. According to reports, saline stress and alkaline stress reduced seed germination, seedling growth, root growth, owering, and boll number of cottons, resulting in a great loss of yield [15][16][17]. Facing the increasing of global demand for cotton, the studies on regulating the damages caused by saline stress and alkaline stress to cotton has increased sharply [18]. Some saline-alkaline response genes in cotton has been selected. For example, ion channels and transporters can mitigate Na + toxicity and K + /Na + homeostasis, and overexpression of NHX1 or SOS1 in cotton can improve salt tolerance [19]. GhSOS3 and GhCBL10 are involved in saline stress and alkaline stress, and the GhSOS3/GhCBL10-SOS2 network also plays a central role in Gossypium hirsutum L. responses to saline stress and alkaline stress [20]. Besides, by increasing the detoxi cation capacity of reactive oxygen species, overexpression of the OSCU/Zn-SOD gene can improve the detoxi cation capacity of reactive oxygen species and improve the tolerance of plants under salt stress [21]. However, most of the previous studies were conducted through pot experiments or indoor culture experiments, and few was conducted through eld experiments. Field experiments make the growth conditions of crops very close to their natural growth, which can truly re ect the growth law of crops.
In this study, RNA-seq was used to analyze the transcriptional changes of cotton leaves under saline stress and alkaline stress, and to elucidate the molecular mechanism of the modi er regulating the stress and alkaline stress. We analyzed many genes related to plant antioxidant defense, K + /Na + transport, and lignin biosynthesis, and these genes may be involved in the regulation of saline stress and alkaline stress by the modi er. The main purposes of this experiment are: (1) to determine the differences in the responses of cotton to saline stress and alkaline stress; (2) to determine the differences in the effects of the modi er regulating saline stress and alkaline stress on K + , Na + , and physiological characteristics of cotton leaves; (3) to provide insights on the relevant genes in the process of the regulation of saline stress and alkaline stress by the modi er.

K + , Na + and Physiological Characteristics of Cotton Leaves
To make clear the effect of the application of compound material on saline stress and alkaline stress, the contents of K + and Na + , K + /Na + , superoxide dismutase (SOD) activity, peroxidase (POD) activity, catalase (CAT) activity, malondialdehyde (MDA) content, and relative electric conductivity (REC) of cotton leaves were determined (Fig. 1).
The contents of K + and Na + of leaves for NaCO 3 treatments (CK-J and P-J treatments) were higher than those for NaCl treatments (CK-Y and P-Y treatments) (Fig. 1A). The contents of K + and Na + and K + /Na + for the CK-J treatment were increased by 30.54% (P<0.05), 21.20% (P<0.05), and 2.14%, respectively compared with those for the CK-Y treatment (Fig. 1A). The K + content and K + /Na + for the P-Y and P-J treatments were increased and the Na + content was decreased after the application of compound material compared with the controls (CK-Y and CK-J treatments). For the P-Y treatment, the K + content was increased by 3.46%, the Na + content was decreased by 13.18%, and the K + /Na + ratio was increased by 18.40%, compared with those of the CK-Y treatment (Fig. 1A). For the P-J treatment, the K + content was increased by 6.37%, the Na + content was decreased by 18.26%, and the K + /Na + ratio was increased by 37.11% (P<0.05) compared with those for the CK-J treatment (Fig. 1A). Meanwhile, the contents of K + and Na + and the K + /Na + ratio for the P-J treatment were increased by 35.14% (P<0.05), 14.11%, and 18.27% (P<0.05), respectively compared with those for the P-Y treatment (Fig. 1A).
The SOD activity for the CK-J treatment was increased by 46.29% (P<0.05), while the POD and CAT activities and REC were decreased by 4.09% (P<0.05), 27.60% (P<0.05), and 1.67%, respectively compared with those for the CK-Y treatment (Fig. 1B). The antioxidant enzyme activity was increased, the MDA accumulation was decreased, and the cell membrane permeability was enhanced after the application of compound material (P-Y and P-J treatments) compared with the controls (CK-Y and CK-J treatments) ( Fig. 1B). The SOD, POD and CAT activities and REC for the P-Y treatment were increased by 45.24%, 44.71%, 29.11%, and 24.02%, respectively (P<0.05), while the MDA content was decreased by 0.12% compared with those for the CK-Y treatment. The SOD, POD and CAT activities and REC for the P-J treatment were increased by 18.85%, 30.76% (P<0.05), 22.66%, and 2.94%, respectively, and the MDA content was decreased by 0.01% compared with those for the CK-J treatment (Fig. 1B). Meanwhile, the SOD activity for the P-J treatment was increased by 19.72%, and the POD and CAT activities and REC were decreased by 23.72% (P<0.05), 31.22% (P<0.05), and 18.38%, respectively, compared with those for the P-Y treatment (Fig. 1B).

Overview of the transcriptomic responses
Transcriptome of each sample was sequenced on Illumina paired-end sequencing platform. The number of generated reads ranged from 39 to 48 million, with a mean of 44 million reads for each sample. The reads were mapped onto the cotton reference transcriptome. The mapping ratio varied from 53.60% to 67.40%, with a mean of 64.11%. The counts of mapped reads were summarized at gene level (Additional le 1: Table S1, Additional le 2: Fig. S1). The principal component analyses (PCA) was performed based on the gene counts (Additional le 3: Fig. S2). The results showed that samples from saline and alkaline treatments were clearly separated on the PC2 dimension, whereas the modi ed and unmodi ed samples were separated by at PC1 dimension. To verify the accuracy of RNA-seq data, six genes were randomly selected for quantitative RT-PCR (qRT-PCR) analysis. The expression abundances estimated by qRT-PCR and RNA-seq were highly correlated (R 2 =0.80, Additional le 4: Fig. S3), indicating that the RNA-seq results were robust and suitable for the subsequent analysis.

Identi cation of Differentially Expressed Genes
To determine the differences of transcriptional response to the treatments, differentially expressed genes (DEGs) were identi ed by pair-wise comparisons of the samples. Compared with the CK-Y treatment, 386 genes were up-regulated and 275 genes were down-regulated for CK-J treatment ( Fig. 2A). Compared with the corresponding controls, a total of 1937 and 2365 DEGs were identi ed for the NaCO 3 treatments (CK-J and P-J treatments) and NaCl treatments (CK-Y and P-Y treatments), respectively ( Fig. 2A). These results indicated that expression patterns of more genes were altered by NaCl treatment compared with the NaCO 3 treatments. Compared with the CK-Y treatment, NaCl treatments resulted in that 1424 genes were upregulated and 941 genes were downregulated after applied compound material. Compared with the CK-J treatment, 1448 genes were upregulated and 489 genes were downregulated for NaCO 3 treatments after the application of compound material ( Fig. 2A). After the application of compound material, compared with the NaCl treatment (P-Y), the NaCO 3 treatment (P-J) resulted in that 1184 genes were upregulated and 373 genes were downregulated. Venn diagram was draw to identify the common and speci c DEGs. The results showed that there were 7 common genes differentially expressed for the four treatments (Fig. 2B). To further understand the molecular interactions among the DEGs, KEGG enrichment analysis was carried out. The results showed phenylpropanoid biosynthesis, pertussis, and brassinosteroid biosynthesis pathway were signi cantly enriched in NaCl treatments (CK-Y and P-Y treatments) (Fig. 4A); systemic lupus erythematosus and alcoholism pathway were signi cantly enriched in NaCO 3 treatments (CK-J and P-J treatments) (Fig. 4B); phenylpropanoid biosynthesis, glycerolipid metabolism, amino sugar and nucleotide sugar metabolism, and pentose and glucuronate interconversions pathway were signi cantly enriched in controls (CK-J and CK-Y treatments) (Fig. 4C); phenylpropanoid biosynthesis and alpha-Linolenic acid metabolism pathway were signi cantly enriched in compound material treatments (P-J and P-Y treatments) (Fig. 4D).

Response of the salt ion transporter in cotton leaves
The transporter-mediated salt ion balance is an important component in cotton leaves when the compound material regulates saline stress and alkaline stress. In NaCl treatments (CK-Y and P-Y treatments), K + transporter and K + channel genes were signi cantly regulated by the compound material, and four up-regulated K + channel genes (GH_A13G1568, GH_D01G0882, GH_D13G1517, and GH_A01G0868) were identi ed. Besides, one K + transporter gene (GH_D08G2294) was up-regulated, one sodium/hydrogen exchanger gene (GH_A09G0801) was down-regulated (Table 1). For NaCO 3 treatments (CK-J and P-J treatments), the expression levels of genes related to K + transporter (GH_D05G2808) and K + channel (GH_A01G0868, GH_D01G0882, and GH_D13G1517) changed markedly, and all were upregulated (Table 1). For compound material treatments (P-J and P-Y treatments), one K + channel gene (GH_A09G0801) was down-regulated (Table 1).

Regulation of antioxidative defense in cotton leaves
Many DEGs in cotton leaves were signi cantly enriched in oxidoreductase activity Go term. Eight peroxidase genes (GH_A06G1119, GH_D11G2319, GH_D10G1060, GH_A12G2651, GH_A05G0628, GH_D10G1977, GH_D06G1268, and GH_A06G1247) were up-regulated for NaCl treatments (CK-Y and P-Y treatments). Three peroxidase genes (GH_A05G4223, GH_A06G1247, and GH_A05G0628) were upregulated for NaCO 3 treatments (CK-J and P-J treatments). Besides, one peroxidase gene (GH_A05G1582) was down-regulated, ve peroxidase gene (GH_A03G1283, GH_D03G1634, GH_D04G0154, GH_D03G1633, and GH_D08G2611) was up-regulated for compound material treatments (P-J and P-Y treatments). One peroxidase gene (GH_D05G1612) was down-regulated, and one peroxidase gene (GH_D10G1977) was up-regulated for controls (CK-J and CK-Y treatments) ( Table 2). DEGs involved in the phenylpropanoid biosynthesis pathway The expression of the genes involved in phenylpropanoid biosynthesis of cotton leaves in response to the application of compound material regulating saline stress and alkaline stress was showed in Fig.5 were down-regulated (Fig. 5C).

Discussion
Under saline stress and alkaline stress, excessive Na + will accumulate in plant leaves, inhibiting the transport of K + and causing K + /Na + imbalance in plant cells [22]. However, the regulation mechanisms of ion balance under saline stress and alkaline stress are different. Wang, et al. [9] showed that the Na + content under alkaline stress was greater than that under saline stress in pot experiments. In this study, the Na + and K + contents of cotton leaves under alkaline stress were signi cantly higher than those under saline stress, and there was no signi cant difference in K + /Na + . In order to regulate the saline stress and alkaline stress on cotton, the modi er was applied in eld trials. Zhang, et al. [23] found that the sodium/hydrogen exchanger 4 of sesame aerial parts was up-regulated under saline stress through hydroponic culture. Zhao, et al. [24] found 17 Na + /H + antiporters in the root of chrysanthemum in response to saline stress. Niu, et al. [25] culture experiments found that salinity signi cantly decreased the expression of NHX4 in leaf veins. In the study, for the treatments treated with the modi er, the Na + content was decreased, which was because the stress signal of Na + could be quickly inhibited by the down-regulation of a Na + /H + exchanger 3 gene when the modi er was applied to salinized soil. Na + /H + exchanges (NHXs) reduced the accumulation of Na + by xing Na + and storing it in vacuoles [26,27]. Besides, Huang, et al. [28] found that the potassium channel KAT1 of the aboveground part of barley was down-regulated under saline stress. In this study, under saline stress and alkaline stress, the K + content and K + /Na + ratio for the modi er treatments were increased, which was because the application of the modi er increased the transcription level of certain genes encoding K transporters and K channels in cotton leaves. In particular, under saline stress, several genes of two-pore K + channel 1 and K + transporter 1 were obviously up-regulated for saline treatments, and several genes of K + transporter 2 and two-pore K + channel 1 were also obviously up-regulated for alkaline treatments. Moreover, a K + channel SKOR gene was down-regulated for the alkaline treatments compared with that for the saline treatments, indicating that the application of the modi er had a better effect on the recovery of K + of cottons under saline stress.
The differences in K + and Na + contents between saline treatments and alkaline treatments were due to the different physiological damages suffered by cotton. In this study, it was found that alkaline stress caused more serious physiological damage to cotton leaves than saline stress. Gong, et al. [12] found that the application of foreign substances could promote the antioxidant system to remove excess free radicals and regulate physiological damage. In this study, we also found that the application of the modi er could regulate the physiological damage suffered by cotton. It has the same effect on the regulation of antioxidant enzymes under saline stress and alkaline stress, but the degrees of the effects are different. For example, no matter under saline stress or alkaline stress, the modi er could increase the SOD, POD, and CAT activity of cotton leaves, which was because a large number of DEGs were related to oxidoreductase activity for CK-Y, P-Y, CK-J and P-J treatments. There were 51 DEGs of oxidoreductase activity Go term were regulated for NaCl treatments (CK-Y and P-Y treatments), and 29 DEGs were regulated for NaCO 3 treatments (CK-J and PJ treatments); moreover, the modi er activated the oxidative stress response under saline stress and alkaline stress. Among them, the modi er signi cantly increased the activities of SOD and CAT under saline stress, while the modi er did not signi cantly increase the activities of SOD and CAT under alkaline stress, indicating that the modi er had less effect on the activities of SOD and CAT under alkaline stress, and SOD activity may not be necessary for saline and alkaline tolerances [29]. Studies have shown that POD is the main detoxi cation enzyme of plants under saline stress and alkaline stress [30]. This study found that the modi er signi cantly increased the POD activity of cotton leaves under saline stress and alkaline stress, which was because the genes related to peroxidase were up-regulated for CK-Y, P-Y, CK-J, and P-J treatments, and the expression of antioxidant enzyme gene was also increased, leading to the improvement of the tolerance of cotton to saline stress and alkaline stress after applying the modi er; besides, the application of modi er under saline stress increased the expression of a great number of antioxidant enzyme genes. Luo, et al. [31] showed that SOD1 and CAT1 genes were involved in the cottons' response to saline stress. G Geng, et al. [32] found that the POD7 and SOD [Cu-Zn] genes of the salt-tolerant varieties of sugar beet were signi cantly upregulated, while in our study, the regulation of saline stress and alkaline stress by the application of modi er only signi cantly regulated the POD A2/50/5/46/29/12/P7 in cotton leaves. This might be because the permeability of the soil in the eld enhanced root vitality and promoted plant resistance, so only peroxidase-related genes responded to the reactions to saline stress and alkaline stress. We also noticed that under saline stress and alkaline stress, the REC and MDA in leaves were affected, resulting in membrane damage. The difference of REC under saline stress and alkaline stress was not signi cant, but the MDA contents under alkali stress was much higher than that under saline stress. Cui, et al. [33] found that the RCE of peanut leaves was increased during saline stress recovery. Gong, et al. [12] found that the leaf MDA content of Malus hupehensis Rehd. was decreased by applying melatonin under alkali stress.
Our study also found that the application of the modi er to cottons under saline stress and alkaline stress increased the REC content of cotton leaves but decreased the MDA content. Among them, only the increases in REC of cottons under saline stress were signi cant, indicating that the modi er alleviated the damaged of saline stress and alkaline stress to the stability of the cell membrane and maintained the integrity of the cell membrane; moreover, the effect of the modi er on the cell membrane of cottons under saline stress was better than that under alkaline stress.
The genes for lignin biosynthesis are dynamically regulated at different levels to protect plant cell metabolism from oxidative damage [34]. In the transcription of data, functional analysis of DEGs was performed through KEGG and GO enrichment analysis, and it was found that a large number of genes were involved in the phenylpropanoid biosynthesis pathway. The phenylpropanoid biosynthesis pathway is one of the most important secondary metabolite pathways in plants, and is related to the plant's response to saline stress and alkaline stress. The lignin metabolites produced in this pathway are of great signi cance for plants to resist abiotic stress [35,36]. Besides, four lignins (p-hydroxyphenyl lignin, guaiacyl lignin, 5-hydroxy-guaiacyl lignin and syringyl lignin) were aggregated by four monomers (pcoumaryl alcohol, coniferyl alcohol, 5-hydroxy-coniferyl alcohol, and sinapyl alcohol), while four alcohols were catalyzed by peroxidase (EC: 1.11.1.7), leading to the formation of these lignins (Fig. 5). Shen, et al. [37] found that seven genes related to lignin biosynthesis in Arabidopsis thaliana were up-regulated under saline stress. This study found that 4CL, HCT, COMT, TOGT1, F5H, CAD, and POD were down-regulated for P-Y treatment compared with those for the CK-Y treatment, indicating that these enzymes might play a vital role in the decrease of lignin synthesis and the protection of cotton from the damage caused by saline stress by modi er. Moreover, 4CL enzyme changes the accumulation of lignin [38], HCT enzyme modi es H lignin [39], COMT enzyme participates in the biosynthesis of S lignin [40], F5H enzyme regulates the composition of S/G lignin (syringyl (S)/guaiacyl (G) lignin) in plants [41], CAD enzymes change the lignin content and structure [42] and POD enzymes participate in lignin biosynthesis and affect plant growth and development [43]. The expression levels of HCT and POD for P-J treatment were down-regulated compared with those for the CK-J treatment, indicating that 5-O-Caffeoylshikimic acid and caffeoyl quinic acid would not be converted into caffeoyl-CoA. However, caffeoyl-CoA is an essential intermediate for lignin biosynthesis [44]. The above indicates that under both saline stress and alkaline stress, the application of modi er will down-regulate the peroxidase (EC: 1.11.1.7), which was because that the modi er may reduce the lignin biosynthesis caused by saline stress and alkaline stress.

Conclusions
Field test results show that saline stress and alkaline stress are two different stresses. Under salt stress, the contents of Na + and MDA in cotton leaves are high, the activities of POD and CAT are low, and the inhibitory effect of alkaline stress are greater than saline stress. The application of the modi er is mainly to increase cotton K + /Na + and POD activity to regulate cotton damages from saline stress and alkaline stress. Through transcription analysis, it was further found that K + transporter genes and peroxidaserelated genes were up-regulated during the regulation of saline stress and alkaline stress by the modi er, which improved the saline and alkaline tolerances of cotton; and the enzymes involved in lignin biosynthesis were down-regulated, which protected cotton from the damage caused by saline stress and alkaline stress (Fig.6). Among them, these up-regulated genes and down-regulated enzymes are abundant in cotton leaves during the regulation of saline stress by modi er. Moreover, these differentially expressed genes obtained in eld trials have high stability, which deserves to be popularized.

Experiment Site
Experiments were conducted at the Experimental Station of Grape Research Institute in Shihezi City, Xinjiang Province, China (44°20' N 86°03' E). The soil is a desert grey soil. Soil basic characteristics are shown in Table 3.

Experimental materials and experimental design
The cotton (Xinluzao 62) and a compound material were used as the experimental material in this study.
The experiment was conducted from April 20th to September 20th, 2018. There were four treatments in total. Two of which were compound material treatments: (1) P-Y treatment (compound material was applied in salinized soil), (2) P-J (compound material was applied in salinized soil); the other two treatments were the controls (CK-Y: no modi er was applied in salinized soil; CK-J: no modi er was applied in alkalized soil), with three repetitions per treatment and a randomized block design. On April 20th, soils were putted into the plastic barrels (0.5 m in diameter and 0.6 m in height) according to the soil layer of the eld, and then barrels were buried back to the eld. NaCl and Na 2 CO 3 was added into the barrels and mixed fully in the plough layer. NaCl of 8 g kg -1 were applied into the NaCl treatments (CK-Y and P-Y treatments), and the pH and EC of the soil were 8.24 and 1.84 s m -1 , respectively; Na 2 CO 3 of 8 g kg -1 were applied into the NaCO 3 treatments (CK-J and P-J treatments), and the pH and EC of the soil were 9.78 and 1.03 s m -1 , respectively. On April 29th, for all treatments, 360 kg hm −2 urea and 795 kg hm −2 synthetic fertilizer (formulated for drip irrigation; N: 135 kg hm −2 ; P: 135 kg hm −2 ; K: 135 kg hm −2 ) were applied. On May 4th, cotton was sown; and after emergence, six seedlings were retained in each barrel.
On May 6th, compound material was applied after diluting with water. Compound material of 300 kg hm -2 was applied into the compound material treatments (P-Y and P-J treatments). Both the fertilizer and compound material were applied to soils with drip irrigation at once. The seedlings were irrigated for the rst time on June 25th. The irrigation cycle was 3 days. At the owering and boll-forming stage (August 19th), new leaves were collected for transcriptome sequencing (three replicates per treatment). All samples were immediately placed in liquid nitrogen and stored at −80 °C until use.

Plant Physiological Analysis
Leaf samples (0.5 g) were homogenized in 5 mL of potassium phosphate buffer (pH 7.8) in ice. The homogenate was centrifuged at 10,000×g for 20 min at 4 °C, and the supernatant was used for the determination of the enzyme activities with the spectrophotometric method. Superoxide dismutase (SOD) activity was determined following the method of Paoletti, et al. [45] based on the photochemical reduction in the NBT. The reaction mixture (3 mL) was comprised of 0.05 M potassium phosphate buffer (pH 7.8), 130 mM methionine, 750 μM NBT, 20 μM ribo avin, 100 μM EDTA-Na2, and 200 μL of enzyme extract. The SOD activity was measured at 560 nm (one unit of SOD activity was denoted by the amount of enzyme required to cause 50% inhibition of the NBT photochemical reduction). The peroxidase (POD) activity was determined as described by Zhou, et al. [46]. The reaction mixture contained 0.  [47].
The contents of thiobarbituric acid reactive substances (TBARS) of leaf samples were determined to calculate the malondialdehyde (MDA) content (nmol/g; extinction coe cient: 155 mM cm −1 ) [48]. Fresh cotyledons (0.5 g) were homogenized and extracted in 10 mL of 0.6% TBA (made in 10% trichloroacetic acid). After that, the extract was heated at 100 °C for 30 min and then cooled on ice. Then, the samples were centrifuged at 5000×g for 10 min; the absorbance was measured at 532 nm. Non-speci c turbidity was corrected by subtracting the absorbance value at 600 nm. For relative electrical conductivity (REC) [49], 0.1 g of fresh leaves were cut into 1 cm slices, placed in 10 mL of deionized water, and shaking for 24 h at room temperature on a rotary shaker (QL200H, Shanghai, China). Then, electrical conductivity of the solution (L1) was measured using a conductivity meter (EM38, ICT international, Armidale, NSW, Australia). The solution was boiled for 15 min and cooled to room temperature, and electrical conductivity (L2) was again measured. REC was calculated as REC = L1/L2.
The contents of Na + and K + in the leaf samples were determined according to the method of Bao [50]. Leaf samples were immersed in 98% H 2 SO 4 and 30% H 2 O 2 , and a ame spectrophotometer (AP1200 type, Shanghai, China) was used for the determination.

Transcriptome Sequencing and Data Analysis
In this study, Oligo (dT) magnetic beads were used to enrich the mRNA with polyA structure in the total RNA, and the RNA was interrupted (300 bp) by ion interruption. After the RNA-Seq library was constructed, library fragments were enriched by PCR ampli cation, and then library selection was performed based on the fragment size (450 bp). Then, the quality of the library was tested by Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, Calif.). After RNA extraction, puri cation, and library building, the samples were paired-end (PE) based on the Illumina platform using Next-Generation Sequencing (NGS).
Quality of the generated reads was checked using FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Fastp was used to remove the adapter and low-quality sequences in the reads [51]. Cotton genome sequence of (Gossypium hirsutum, ZJU) were downloaded from Hu, et al. [52] and used as the reference genome. The clean reads were qausimapped on to all cotton annotated transcripts using Salmon [53]. Expression abundance at the unit of transcript per million (TPM) was calculated at gene level. DESeq2 was used to identify the differentially expressed genes (DEGs) between samples with the thresholds of adjusted p-value less than 1 and absolute value of log2(fold change) larger than 1 [54]. Principal component analysis (PCA) was performed to display the transcriptomic similarity among the samples based on the counts of top 1000 genes. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted using clusterPro ler [55].

Quantitative real-time PCR validation
To validate the RNA-seq data, 11 DEGs from the pathway enrichment analysis were selected for qRT-PCR analysis. Samples of RNA-Seq were reverse transcribed into cDNA for real-time qPCR validation using the PrimeScript TM 1st stand cDNA Synthesis Kit and SYBR Green Master Mixes (Vazyme Biotech, Nanjing, China). qRT-PCR was performed on a uorescence quantitative system TIB8600 (Taipu, Biotech, Xiamen, China). Each sample was measured with three biological and three technical replicates, and the relative expression levels were calculated using the 2 -⊿⊿Ct method. The endogenous reference gene used was GhEF1α. The gene-speci c primers are listed in Table S2 (Additional le 1).

Figure 1
Page 23/28 Effect of the application of compound material on the contents of K+ and Na+, K+/Na+ (A), antioxidative enzymes activity, and the contents of MDA and REC (B) in leaves.     Proposed model for the function of the compound material in regulating saline stress and alkaline stress of cotton leaves. The up-pointing red arrows mean that the candidate genes are up-regulated; the downpointing green arrows mean that the candidate genes are down-regulated.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.