Background

Flooding is among the most severe abiotic stresses that occur during plant growth and development [1]. Flooding is a general term referring to excessively wet conditions, that is where excess water replaces gas-spaces surrounding roots and/or shoots. It mainly includes four aspects: (1) Waterlogging or soil flooding: only the root-zone is flooded, (2) Partial waterlogging or soil flooding: partial flooding of the root-zone, (3) Submergence refers to the entire plant being underwater (4) Partial submergence: the entire root system and part of above-ground organs are under water [2]. Flooding directly affects the diffusion of oxygen in plant tissues/soil, resulting in hypoxia. Hypoxia greatly disrupts respiration and photosynthesis, which leads to a reduction in the ATP supply and has deleterious effects on normal life activities of plants [3]. When plants are completely submerged, they are deprived of oxygen. Most plant species cannot survive prolonged submergence, but they can temporarily adapt to submergence stress through the Low Oxygen Quiescence Syndrome (LOQS) or the Low Oxygen Escape Syndrome (LOES) [4,5,6]. Heavy rainfall and flood disasters have become frequent in recent years, and there is an urgent need to study the submergence tolerance of plants and their mechanisms to maintain the effective adaptation of plants to climate change [7, 8].

Reactive oxygen species (ROS) are a normal product of plant cell metabolism. Reactive oxygen can be used as a signal molecule to respond to stress, and excessive ROS is harmful to plant cells. Under prolonged hypoxia condition, excessive ROS can be accumulated, causing membrane lipid peroxidation and altering the structure of proteins and nucleic acids [9]. Malondialdehyde (MDA) is one of the most important products of membrane lipid peroxidation and constitutes a common parameter of membrane damage. Plants have evolved a complex set of enzymatic and non-enzymatic detoxification mechanisms to eliminate oxidative damage caused by ROS [Observation of the internal structure of leaf tissue

Paraffin sections were used to observe the changes in leaf tissue in ZNL2067 and ZL100 from three treatments. After stress treatment, the leaves were quickly cut into 5 mm × 10 mm pieces, and then fixed with FAA fixing solution. The slices were made by paraffin sectioning method and sliced with a microtome (RM2016) with a thickness of 5 μm. Then stained with safranine and solid green, and sealed with neutral gum. Finally, a light microscope (NIKON ECLIPSE E100) was used to observe and take photos. We used CaseViewer software to observe longitudinal Sects. (200 ×) of paraffin sections of leaves. Palisade tissue thickness, spongy tissue thickness, and leaf thickness were measured, and palisade tissue/spongy tissue ratio, CTR and SR were calculated. Three replicates were measured for each treatment, and 5 readings were taken for each replicate.

Dry matter determination

10 representative cotton plants were randomly selected from three treatments, separated by roots, stems, and leaves, placed in an oven. First, they were dried at 105℃ for 30 min and then at a temperature of 75 ℃ to a constant weight. Finally, the dry matter weight was calculated.

Net photosynthetic rate determination

The photosynthetic parameters of the main stem functional leaves of ZNL2067 were measured by a Li-6400 portable photosynthesis meter (produced by LI-COR, USA) on the day of submergence for 3 days and reoxygenation for 3 days. Three replicates were measured for each treatment, and 10 readings were taken for each replicate.

MDA and POD measurements

0.1 g of leaf tissue was weighed in three treatments, and 1 mL of extract was added to homogenize in an ice bath. After centrifugation at 8000 g for 10 min at 4 ℃, the supernatant was removed and placed on ice with three biological replicates for each sample. Samples were taken to determine the activity of POD and MDA, and a POD assay kit and MDA assay kit (#A084-3–1, #A003-3–1, Jiancheng Bioengineering Institute, Nan**g, Jiangsu, China) were used to measure the enzyme activity.

cDNA library construction and sequencing

Representative samples of ZNL2067 were randomly selected from the Nor, Sub and Reo treatments. RNA was separately extracted from roots, stems, and leaves, and then, equal amounts of RNA from roots, stems, and leaves were mixed. Each treatment was repeated three times independently.

We isolated and purified total RNA using TRIzol (Invitrogen, Carlsbad, CA, USA). RNA was quantified using NanoDrop ND-1000 (NanoDrop, Wilmington, DE, USA), and the RNA integrity was assessed using a Bioanalyzer 2100 (Agilent, CA, USA) with a RIN of > 7.0. The cDNA library was built utilizing the methods of Fan et al. [68]. The average insert length in the cDNA library was 300 ± 50 bp. Finally, we performed 2 × 150 bp paired-end sequencing (PE150) on an Illumina NovaSeq™ 6000 (LC-Bio Technology CO., Ltd., Hangzhou, China).

Identification of differentially expressed genes (DEGs)

We used StringTie (version: stringtie-1.3.4d.Linux_x86_64) to assemble the mapped reads for each sample [69]. All transcriptomes from the samples were merged to reconstruct a comprehensive transcriptome using the gffcompare software (version: gffcompare-0.9.8.Linux_x86_ 64). We estimated the expression levels of all transcripts using StringTie and Ballgown and determined mRNA expression levels by calculating FPKM values. The differentially expressed mRNAs and genes were selected with log2fold change (FC) > 1 or log2 (FC) < -1 and p value < 0.05 by R package edge R [70]. We used TBtools software to display heatmaps [71].

Gene Ontology (GO) and KEGG pathway enrichment analyses

We performed GO enrichment analysis of DEGs using the GOseq R package [72], and the length bias of DEGs was corrected. GO terms (p value < 0.05) were considered significantly enriched by DEGs. KOBAS77 software was used to test the statistical enrichment of DEGs in KEGG pathways. All DEGs were compared against the GO and KEGG [73,74,75].

Real-time quantitative PCR (RT-qPCR) validation and analysis

We selected thirty DEGs to validate the reliability of the transcriptome database. Thirty pairs of primers were designed using the Primer 6.0 software (Table S1), and RT-qPCR was performed [68]. The Actin gene was used as a reference.

Data Processing

SPPS(Ver.21) and EXCEL software were used for statistical analysis. One-way analysis of variance (ANOVA) or Duncan's method was used to compare the significant levels of differences between different treatments (α = 0.05).