Identification of Alkaline Salt Tolerance Genes in Brassica napus L. by Transcriptome Analysis

Soil salt alkalization is one major abiotic factor reducing the productivity of crops, including rapeseed, an indispensable oil crop and vegetable. The mechanism studies of alkali salt tolerance can help breed highly resistant varieties. In the current study, rapeseed (B. napus) line 2205 exhibited more tolerance to alkaline salt than line 1423 did. In line 2205, the lesser plasma membrane damage index, the accumulated osmotic solute, and higher antioxidant enzyme activities contributed to alkaline tolerance. A more integrated mesophyll-cell structure was revealed under alkali salt stress by ultrastructure observation in line 2205, which also implied a lesser injury. Transcriptome analysis showed that more genes responded to alkaline salt in line 2205. The expression of specific-response genes in line 1423 was lower than in line 2205. However, most of the specific-response genes in line 2205 had higher expression, which was mainly enriched in carbohydrate metabolism, photosynthetic processes, ROS regulating, and response to salt stress. It can be seen that the tolerance to alkaline salt is attributed to the high expression of some genes in these pathways. Based on these, twelve cross-differentially expressed genes were proposed as candidates. They provide clues for further analysis of the resistance mechanism of rapeseed.


Introduction
Salinity is among the most detrimental stresses in plants, as salt hampers plant growth, development, yield, and quality [1]. Soil salinization is intensifying because of unsuitable irrigation practices and other causes of climate change and environmental pollution [2]. The Global Map of Salt-affected Soils displays that 85% of topsoils (0-30 cm) and 62% of subsoils (30-100 cm) are salines, 34% and 19% of which are sodic and saline, respectively. Salt-affected soils are distributed in over 100 countries, especially in the arid and semi-arid regions, including China. However, the basic national conditions of China's large population make it difficult to grow crops in non-saline land. Therefore, rational exploitation and improvement of the salinized land can relieve the pressure on cultivated land resources. According to reports, the total area of available saline-alkali soil is 3.67 × 107 ha in China, 30% of which shows great agricultural potential [3,4].
Salt-affected soils are mainly categorized into two groups: saline and sodic (alkaline). Generally, there is the presence of excess sodium on exchange sites in both soils, while the anions are different. The dominant anions are chloride and sulfate in saline soils. However, a high concentration of carbonate/bicarbonate anions in sodic soils causes high pH, a high sodium absorption ratio, and low electrical conductivities compared with those in saline soils [5]. In salt-stress plants, the imbalance of osmotic potential induces water stress, a high concentration of Na + causes severe ion toxicity, and the interaction between salts and tolerance-related genes, which will lay a solid foundation for the breeding and mechanism of alkaline salt tolerance in B. napus.

Plant Materials and Growth Conditions
The salt-tolerant line 2205 (T) and salt-sensitive line 1423 (S) were grown in a greenhouse at 25°C and 60% relative humidity under a 16 h light/8 h dark photoperiod. Firstly, the seeds were germinated on filter papers with distilled water for three days until the cotyledons unfolded. Secondly, the buds were transplanted into the matrix (Mineral/vermiculite/Perlite = 3:1:1). When the seedlings reached the two-leaf stage after ten days, the uniform seedings were placed to the Hoagland nutrient solution (pH 6.56) [31] for hydroponics. After growth for 6 days, alkaline salt stress was performed in the Hoagland nutrient solution containing 75 mmol/L NaHCO 3 (pH 8.12), while the control group contained only the nutrient solution. After treatment with 75 mmol/L NaHCO 3 for 0, 12, and 24 h, the young leaves of line 2205 (T0, T12, and T24) and 1423 (S0, S12, and S24) were collected for RNA-seq. The leaves from 10 plants were pooled to form one biological replicate and three replicates were generated for each treatment. In total, 18 samples were collected and stored at −80 • C as independent biological samples.

Measurement of Physiological Indexes
To further evaluate and analyze resistance differences to alkali stress between lines 2205 and 1423, six conventional physiological indicators related to stress response, including relative electric conductivity (REC), soluble sugar (SS), soluble protein (SP), proline (Pro), superoxide dismutase (SOD), and peroxidase (POD), were examined using the young leaves after being treated for 0, 12, and 24 h. REC was detected using the method described by Lutts et al. [32], SS and Pro were measured by the method of Zhao et al. [33], the examination of SP was performed by the description of Woolhouse et al. [34], SOD was determined by the method of Beauchamp and Fridovich [35], and POD was detected based on the procedure provided by Zhao et al. [36]. These experiments were repeated three times with three biological replicates. For each biological replicate, three plants were collected.

Electron Microscopic Observation of Seedling Cells under NaHCO 3 Stress
At the three-leaf stage, the third true leaf was collected at 0 h (CK) and 7 d after alkaline salt stress and washed with distilled water. Then, the tissue block with a width of 1 mm and a length of 5 mm along the middle of the main leaf vein was cut. The same part of each leaf was taken once, and the process was repeated using three plants. The cut tissue was quickly placed into a 4% glutaraldehyde solution (pH = 6.8) for 12 h after vacuuming. Then, the tissue was washed with a 0.1 mM phosphate buffer saline (PBS) buffer (pH = 6.8) for 10 min four times. The leaves were transferred to 0.1 mM osmium acid for 1.5-2.0 h, washed with a 0.1 mM PBS buffer (pH = 6.8) for 10 min five times, and then dehydrated in ethanol solutions with gradient concentrations (30%, 50%, 70%, 80%, and 90%) for 15 min. The samples were then treated with 100% ethanol for 30 min and 100% acetone for 30 min twice, infiltrated with acetone, permeated with pure glue for 24 h, and then embedded with resin. The half-thin section was sliced by a microtome, stained by uranyl acetate and lead citrate, and observed by a transmission electron microscope (HT7700, Hitachi, Tokyo, Japan).

RNA Extraction, Library Construction, and Sequencing
The total RNA of these samples was extracted with an RNAprep Pure Plant Kit (TIANGEN, DP441) and subjected to quality control with a Bioanalyzer 2100 system (Agilent Technologies, Palo Alto, CA, USA). The RNA concentration was measured by NanoDrop 2000 (Thermo). A total amount of 1 µg of RNA per sample was used as the input material for RNA sample preparations. The libraries were sequenced on an Illumina HiSeq X Ten platform at Beijing BMK Biotechnology Co., Ltd. (Beijing, China).

RNA-Seq Analysis
Quality control was performed by removing reads containing an adapter, ploy-N, and low quality from raw data. The clean paired-end reads were mapped to the designated reference genome of B. napus (http://www.genoscope.cns.fr/brassicanapus/data/, accessed on 20 September 2021) by the Hisat2 tools soft [37]. Only reads with a perfect match or one mismatch were further analyzed and annotated based on the reference genome. Gene expression levels were estimated by fragments per kilobase of transcript per million fragments mapped (FPKM) [38]. Pairwise differential expression analyses were performed using the DEseq [39]. Comparison schemes for these two lines, the salt-tolerant line 2205 (T) and salt-sensitive line 1423 (S), at three-time points (0, 12, and 24 h) were set up to analyze the gene expression, including S0 vs. T0, S12 vs. T12, S24 vs. T24, S0 vs. S12, S0 vs. S24, T0 vs. T12, and T0 vs. T24. The resulting p-values were adjusted using the method by Benjamini and Hochberg for controlling the false discovery rate (FDR). Genes with an FDR < 0.01 and |log 2 fold change| ≥ 1 were considered to be differentially expressed.
Further analyses of DEGs were performed using the tools on the platform BMKCloud (www.biocloud.net, accessed on 27 February 2022). Venn and clustering heatmap plots were generated by R packages "VennDiagram" and "heatmap". The box plots of DEG sets were drawn using GraphPad Prism v9.0 software (San Diego, CA, USA).

Functional Enrichment Analysis of DEGs
Gene Ontology (GO) enrichment analysis of DEGs was implemented by the GOseq R packages with the significant threshold of p < 0.05 [40], and the significantly enriched terms were graphed using GraphPad Prism v9.0 software. In addition, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were enriched using the KOBAS online analysis database [41], and the top 20 pathways were displayed.

Verified DEGs by qRT-PCR
To verify the differential expression detected by the RNA-seq, the remaining RNAs of the 18 samples were reverse-transcribed in a 20 µL reaction mixture with a FastKing RT Kit (With gDNase, TIANGEN, KR116) following the manufacturer's instructions. The transcriptional expression of six randomly selected DEGs was explored by quantitative real-time PCR (qRT-PCR). The Actin gene was used as the internal control for qRT-PCR data analysis. The primes for these genes are listed in Table S1. The reaction was performed in 8-tube strips on a QuantStudio™ 7 Flex (Thermo Fisher Scientific, Waltham, MA, USA) using SYBR Green Master ROX (TaKaRa, Dalian, China) in accordance with the manufacturer's instructions. Three technical replicates for each biological-replicate sample were analyzed. Relative fold differences of the selected DEGs among two lines or experiments were calculated using the 2 −∆∆Ct method [42]. Pearson's correlation coefficient was used to express the correlation of log 2 (fold change) between samples. The results were shown by GraphPad Prism 9.0.

Statistical Analysis
Data processing was conducted using Microsoft Excel 2019, and figures were generated using GraphPad Prism 9. All triplicate data were subjected to one-way analysis of variance (ANOVA) using IBM SPSS Statistics 21. The data are presented as the mean (±SD) of three replications. Means were separated by the least significant difference test at 5% level of significance.

Differences of Alkaline Salt Tolerance and Physiological Indicators between Lines 2205 and 1423 under Alkaline Salt Stress
The growth of lines 2205 and 1423 showed no difference under normal conditions ( Figure 1A uble sugars (SS), soluble protein (SP), proline (Pro), and the activity of POD and SOD in both cultivars, and the tendency increased over the stress time ( Figure 1). The content of SP was higher in line 2205 than in line 1423, and the gap was increased after 24 h of stress ( Figure 1E). Under normal and alkaline salt conditions, the concentration of SS and REC was significantly lower in line 2205 compared with that in line 1423, which was opposite to SOD activity ( Figure 1F-H). There was no significant difference in Pro content and POD activity between line 2205 and line 1423 under normal growth conditions and 12 h of stress, which were significantly elevated in line 2205 after 24 h of stress compared with that in line 1423 ( Figure 1I-J). These results implied that the physiological responses to alkaline salt were different between line 2205 and line 1423.

Ultrastructural-Feature Differences between Line 2205 and Line 1423 under Alkaline Salt Stress
Observation of microstructures showed that the mesophyll cells of lines 2205 and 1423 exhibited typical features under normal conditions. The chloroplasts were spindleshaped structures, the basal particles were arranged in an orderly manner, the thylakoids To further evaluate the difference in alkaline salt tolerance between these two cultivars, the physiological indexes related to stress response were measured. The results showed NaHCO 3 stress increased relative electric conductivity (REC), the content of soluble sugars (SS), soluble protein (SP), proline (Pro), and the activity of POD and SOD in both cultivars, and the tendency increased over the stress time ( Figure 1). The content of SP was higher in line 2205 than in line 1423, and the gap was increased after 24 h of stress ( Figure 1E). Under normal and alkaline salt conditions, the concentration of SS and REC was significantly lower in line 2205 compared with that in line 1423, which was opposite to SOD activity ( Figure 1F-H). There was no significant difference in Pro content and POD activity between line 2205 and line 1423 under normal growth conditions and 12 h of stress, which were significantly elevated in line 2205 after 24 h of stress compared with that in line 1423 ( Figure 1I-J). These results implied that the physiological responses to alkaline salt were different between line 2205 and line 1423.

Ultrastructural-Feature Differences between Line 2205 and Line 1423 under Alkaline Salt Stress
Observation of microstructures showed that the mesophyll cells of lines 2205 and 1423 exhibited typical features under normal conditions. The chloroplasts were spindleshaped structures, the basal particles were arranged in an orderly manner, the thylakoids were arranged compactly, and starch grains were present in the chloroplast of both lines (Figure 2A,B,G,H). However, the cells of line 2205 were approximatively round, and more and tighter chloroplasts were found, which showed a long oval shape in line 1423 ( Figure 2C,I). After alkaline salt stress for 7 days, line 1423 suffered a greater disruption, obvious plasmolysis occurred, and the organelles in the cells decreased or dissolved. However, many intact organelle structures could still be observed in line 2205 ( Figure 2F,L). In contrast to the shrunken manifestation occurring on the thylakoid and starch grains of chloroplasts in the salt-sensitive line 1423, the thylakoid of line 2205 swelled, and the starch grains in the chloroplasts became larger and more transparent. Moreover, the basal particles of line 2205 still maintained the regular arrangement to some extent, while those of line 1423 were arranged loosely and distorted. In addition, line 2205 had fewer osmiophilic granules compared with line 1423 ( Figure 2D,E,J,K). From these results, we speculated the stronger resistance of line 2205 was related to the more complete chloroplast structure under NaHCO 3 stress. many intact organelle structures could still be observed in line 2205 ( Figure 2F,L). In contrast to the shrunken manifestation occurring on the thylakoid and starch grains of chloroplasts in the salt-sensitive line 1423, the thylakoid of line 2205 swelled, and the starch grains in the chloroplasts became larger and more transparent. Moreover, the basal particles of line 2205 still maintained the regular arrangement to some extent, while those of line 1423 were arranged loosely and distorted. In addition, line 2205 had fewer osmiophilic granules compared with line 1423 ( Figure 2D,E,J,K). From these results, we speculated the stronger resistance of line 2205 was related to the more complete chloroplast structure under NaHCO3 stress.

Illumina Sequencing and Differentially Expressed Genes between Line 2205 and Line 1423
In order to explore the discrepant genes responding to alkaline salt between line 2205 and line 1423, RNA-seq was performed after 0 h (control), 12 h, and 24 h NaHCO3 stresses on three independent biological replicates. In total, 116.40 GB of clean data from 18 RNA libraries were obtained after filtering. The clean data of each sample reached 4.90 GB, the GC content was between 45.85-46.88%, and the percentage of Q30 base was more than 91.71%. The efficiency of sequence comparison between the clean reads of each sample and the reference genome ranged from 83.09% to 88.39% (Table S2).
To identify changes in transcriptional expression under alkali stress, the differentially expressed genes (DEGs) were identified by comparing the expression of genes between

Illumina Sequencing and Differentially Expressed Genes between Line 2205 and Line 1423
In order to explore the discrepant genes responding to alkaline salt between line 2205 and line 1423, RNA-seq was performed after 0 h (control), 12 h, and 24 h NaHCO 3 stresses on three independent biological replicates. In total, 116.40 GB of clean data from 18 RNA libraries were obtained after filtering. The clean data of each sample reached 4.90 GB, the GC content was between 45.85-46.88%, and the percentage of Q30 base was more than 91.71%. The efficiency of sequence comparison between the clean reads of each sample and the reference genome ranged from 83.09% to 88.39% (Table S2).
To identify changes in transcriptional expression under alkali stress, the differentially expressed genes (DEGs) were identified by comparing the expression of genes between 12 and 24 h and those at 0 h in both cultivars. In the sensitive line S, 5220 (S0 vs. S12) and 7049 (S0 vs. S24) genes were altered at the transcriptional level, whereas 8395 M and 12960 (T0 vs. T24) were differentially expressed in the tolerant line T ( Figure 3A and Table S3). It was found that line T had nearly twice as many DEGs as line S. Among these DEGs, 9401 and 2802 genes specifically respond for alkaline salt in line T (T. spe.) and line S (S. spe.), respectively ( Figure 3A). Futher, 5239 DEGs had significantly different expression between T and S in T. spe, as subset I. Similarly, 1666 DEGs in S. spe differently expressed between T and S, as subset II ( Figure 3B). Additionally, 10,284 DEGs responded simultaneously to alkaline salt in the two lines, and their expression exhibited significant differences. These genes were defined as subset III ( Figure 3B). and 12960 (T0 vs. T24) were differentially expressed in the tolerant line T ( Figure 3A and Table S3). It was found that line T had nearly twice as many DEGs as line S. Among these DEGs, 9401 and 2802 genes specifically respond for alkaline salt in line T (T. spe.) and line S (S. spe.), respectively ( Figure 3A). Futher, 5239 DEGs had significantly different expression between T and S in T. spe, as subset Ⅰ. Similarly, 1666 DEGs in S. spe differently expressed between T and S, as subset Ⅱ ( Figure 3B). Additionally, 10,284 DEGs responded simultaneously to alkaline salt in the two lines, and their expression exhibited significant differences. These genes were defined as subset Ⅲ (Figure 3B). , S0 versus T0 (yellow circle), S12 versus T12 (blue four-sided arc), S24 versus T24 (red dumbbell arc). The purple, orange, and gray filling DEGs with underlines mark the concrete genes in subsets Ⅰ, Ⅱ, and Ⅲ for further analysis, respectively.

Co-expression Clustering of Filtered DEGs
To excavate the potential alkali-salt-regulating genes, the co-expression clustering analyses of filtered DEGs (subsets Ⅰ, Ⅱ, and Ⅲ) were performed, and the values of Log10 (FPKM + 1) were used for clustering. Each subset could be divided into three clusters (Figure 4). On subset Ⅰ, 72.25% of DEGs (clusters 2 and 3) had higher expression in T0, T12, and T24 ( Figure 4A,B). However, the higher expressed DEGs accounted for just 28.36% in subset Ⅱ, 73.37% of which (clusters 2 and 3) was expressed at a lower level in the tolerant line T ( Figure 4C,D). On subset Ⅲ, the top two clusters were downregulated (cluster 1) and upregulated (cluster 2) in S0, 12, 24 vs. T0, 12, 24 ( Figure 4E,F). Taken together, the 3785 DEGs in subset Ⅰ and the 3508 DEGs in subset Ⅲ with an upregulated expression in line T might be responsible for the salt tolerance, and the other 1189 (in subset Ⅱ) and 4553 (in subset Ⅲ) genes with an upregulated expression in line S might have negative effects on salt resistance.
To analyze the functional differences of the DEGs in these two groups, the synchronously up-and downregulated genes (FDR ≤ 0.01, |log2 (fold change)| ≥ 1) in groups S0 vs. T0, S12 vs. T12 and S24 vs. T24 were further investigated. As a result, 2236 of 7293 (up-R. in T., higher expression in line T) and 2240 of 5742 DEGs (down-R. in T., lower expression in line T) were selected as the important DEGs (Table S4).

Co-expression Clustering of Filtered DEGs
To excavate the potential alkali-salt-regulating genes, the co-expression clustering analyses of filtered DEGs (subsets I, II, and III) were performed, and the values of Log 10 (FPKM + 1) were used for clustering. Each subset could be divided into three clusters ( Figure 4). On subset I, 72.25% of DEGs (clusters 2 and 3) had higher expression in T0, T12, and T24 ( Figure 4A,B). However, the higher expressed DEGs accounted for just 28.36% in subset II, 73.37% of which (clusters 2 and 3) was expressed at a lower level in the tolerant line T (Figure 4C,D). On subset III, the top two clusters were downregulated (cluster 1) and upregulated (cluster 2) in S0, 12, 24 vs. T0, 12, 24 ( Figure 4E,F). Taken together, the 3785 DEGs in subset I and the 3508 DEGs in subset III with an upregulated expression in line T might be responsible for the salt tolerance, and the other 1189 (in subset II) and 4553 (in subset III) genes with an upregulated expression in line S might have negative effects on salt resistance.
To analyze the functional differences of the DEGs in these two groups, the synchronously up-and downregulated genes (FDR ≤ 0.01, |log 2 (fold change)| ≥ 1) in groups S0 vs. T0, S12 vs. T12 and S24 vs. T24 were further investigated. As a result, 2236 of 7293 (up-R. in T., higher expression in line T) and 2240 of 5742 DEGs (down-R. in T., lower expression in line T) were selected as the important DEGs (Table S4).
Genes 2022, 13, x FOR PEER REVIEW 9 of 18 "photosynthesis", "PSII associated light-harvesting complex II catabolic process", "photosystem II oxygen evolving complex assembly"; "glycolytic process", "starch biosynthetic process", "gluconeogenesis", and "glucose catabolic process". In addition to the cellular component related to the chloroplast, apoplast, mitochondria, and peroxisome were also significantly enriched. However, only three GO terms, "cytosol" in CC, "snRNA binding" and "succinate-semialdehyde dehydrogenase (NAD + ) activity" in MF, were significantly enriched among the downregulated 2240 DEGs in line 2205 ( Figure 5A). Several studies have shown that the photosynthetic mechanisms of different salt-tolerant germplasms have different trends under salt stress [7,43]. Therefore, we hypothesized the higher expression of multiple genes participating in photosynthetic systems delayed the attenuation of photosynthesis in the tolerant line 2205. Chloroplasts are not only the site of photosynthesis, but also ROS production [44]. In addition, mitochondria and peroxisome significantly enriched in CC are also important sites for ROS releasing. Many researchers have indicated that ROS plays a dual role under abiotic stresses [13]. Thereby, we speculated that ROS homeostasis was stronger in line 2205 than line 1423, which increased the tolerance to alkaline salt.  Several studies have shown that the photosynthetic mechanisms of different salttolerant germplasms have different trends under salt stress [7,43]. Therefore, we hypothesized the higher expression of multiple genes participating in photosynthetic systems delayed the attenuation of photosynthesis in the tolerant line 2205. Chloroplasts are not only the site of photosynthesis, but also ROS production [44]. In addition, mitochondria and peroxisome significantly enriched in CC are also important sites for ROS releasing. Many researchers have indicated that ROS plays a dual role under abiotic stresses [13]. Thereby, we speculated that ROS homeostasis was stronger in line 2205 than line 1423, which increased the tolerance to alkaline salt.  Figure 5. Similar to GO enrichment, the two-group DEGs were enriched in different pathways. The four most significantly enriched pathways were "glyoxylate and dicarboxylate metabolism", "carbon metabolism", "peroxisome", and "photosynthesis" for the 2236 upregulated genes in line 2205 ( Figure 5B). The other four pathways, "biosynthesis of unsaturated fatty acids", "N-Glycan biosynthesis", "caffeine metabolism", and "SNARE interactions in vesicular transport", were the top pathways significantly enriched for these 2240 upregulated genes in line 1423 ( Figure 5C). To sum up, GO and KEGG analyses both showed that carbohydrate metabolism, photosynthetic processes, and ROS regulating were enriched for these 2236 DEGs, which were not enriched for those 2240 DEGs It is surmised that the higher expression of the genes in these pathways enhanced the tolerance to alkali salt stress of line 2205.

Validation of the RNA-Seq Data by qRT-PCR
To verify the differential expression detected by the Illumina RNA-Seq data, qRT-PCR was performed for six random genes using 18 RNA libraries. Similar expression trends were shown for the selected genes, and a high positive correlation (0.83) was obtained between RNA-seq and qRT-PCR data at different time points (Figure 7), suggesting that the RNA-Seq data were reliable.

Validation of the RNA-Seq Data by qRT-PCR
To verify the differential expression detected by the Illumina RNA-Seq data, qRT-PCR was performed for six random genes using 18 RNA libraries. Similar expression trends were shown for the selected genes, and a high positive correlation (0.83) was obtained between RNA-seq and qRT-PCR data at different time points (Figure 7), suggesting that the RNA-Seq data were reliable.

Discussion
Salt stress is a severe hazard that adversely affects all stages of plant growth, such as inhibiting seed germination, slowing plant development, decreasing biomass yield, and decreasing crop yield [1,29]. Therefore, it is a great advantage to breed salt-tolerant rapeseed cultivars. Excavating the physiological and biochemical response mechanisms would be helpful to understanding the molecular tolerance traits and develop salt-tolerant genotypes. However, our present knowledge on the mechanisms regulating salt tolerance (especial alkaline salt stress) in rapeseed is still in its infancy. Hence, the selection and identification of high alkaline salt tolerance cultivars are crucial to expanding rapeseed production in saline-alkali soil.

Alkali Salt Tolerance of Line 2205 Is Reflected at Multiple Levels
Compared with line 1423, the 2205 cultivar exhibits strong tolerance not only to the neutral salt NaCl [45], but also to the alkaline salt NaHCO3. The discrepancy between lines 2205 and 1423 is reflected not only in the visual phenotype but also in physiological responses and ultramicrostructures of mesophyll cells. REC is an indicator used to quantify the cell membrane stability during adversary stress [46]. Reports have shown that stress (such as heat, drought, and salt) can damage the plasma membrane, inducing an increase in REC [46]. In the current study, REC was significantly higher in line 1423, leading to severe damage and weak selectivity of the membrane compared with those of line 2205 (Figure 1). Ultrastructural observation also discovered cell structures, and organelles of line 1423 appeared to dissolve when subjected to 7-day alkali salt stress (Figure 2), which may be correlated with the serious injury of the plasma membrane. Additionally, as important osmotic regulators, salinity-induced accumulation of SS and soluble protein SP have been reported [47,48]. In the current study, the increasing trend in SS and SP was detected after NaHCO3 stress, which is consistent with previous reports. However, the tolerant line 2205 presented a lower SS content but an opposite pattern of SP (Figure 1). Whether SP and SS play different roles in osmotic regulation in response to alkaline salt needs further investigation. Many studies have reported that Pro plays an important and positive role in plant growth as an osmoprotectant and antioxidative defense molecule

Discussion
Salt stress is a severe hazard that adversely affects all stages of plant growth, such as inhibiting seed germination, slowing plant development, decreasing biomass yield, and decreasing crop yield [1,29]. Therefore, it is a great advantage to breed salt-tolerant rapeseed cultivars. Excavating the physiological and biochemical response mechanisms would be helpful to understanding the molecular tolerance traits and develop salt-tolerant genotypes. However, our present knowledge on the mechanisms regulating salt tolerance (especial alkaline salt stress) in rapeseed is still in its infancy. Hence, the selection and identification of high alkaline salt tolerance cultivars are crucial to expanding rapeseed production in saline-alkali soil.

Alkali Salt Tolerance of Line 2205 Is Reflected at Multiple Levels
Compared with line 1423, the 2205 cultivar exhibits strong tolerance not only to the neutral salt NaCl [45], but also to the alkaline salt NaHCO 3 . The discrepancy between lines 2205 and 1423 is reflected not only in the visual phenotype but also in physiological responses and ultramicrostructures of mesophyll cells. REC is an indicator used to quantify the cell membrane stability during adversary stress [46]. Reports have shown that stress (such as heat, drought, and salt) can damage the plasma membrane, inducing an increase in REC [46]. In the current study, REC was significantly higher in line 1423, leading to severe damage and weak selectivity of the membrane compared with those of line 2205 (Figure 1). Ultrastructural observation also discovered cell structures, and organelles of line 1423 appeared to dissolve when subjected to 7-day alkali salt stress ( Figure 2), which may be correlated with the serious injury of the plasma membrane. Additionally, as important osmotic regulators, salinity-induced accumulation of SS and soluble protein SP have been reported [47,48]. In the current study, the increasing trend in SS and SP was detected after NaHCO 3 stress, which is consistent with previous reports. However, the tolerant line 2205 presented a lower SS content but an opposite pattern of SP (Figure 1). Whether SP and SS play different roles in osmotic regulation in response to alkaline salt needs further investigation. Many studies have reported that Pro plays an important and positive role in plant growth as an osmoprotectant and antioxidative defense molecule under abiotic stress [49]. We also found it was accompanied by a higher proline content in the more tolerant line 2205 under alkaline salt (Figure 1). Salt stress triggers the excess generation of ROS, which causes oxidative damage to cells and even results in plant death. The antioxidant defense system constituted of non-enzymatic and enzymatic components saves plants from oxidative damage by detoxifying the excessive ROS and maintaining the balance of ROS [50]. Meanwhile, SOD and peroxidase POD are two important constituents of the antioxidant defense system [6]. In this study, higher POD and SOD activities in line 2205 endowed it with a higher scavenging ability for ROS (Figure 1). We speculated this might partially contribute to the tolerance to alkaline salt stress.
It was found that the integrity of mesophyll cells of line 2205 was better than that of line 1423 under NaHCO 3 stress, and no plasmolysis was observed by microstructure observation (Figure 2). Previous studies found that the cell structure can be changed to minimize the damage when plants are subjected to saline-alkali stress [51]. The intact cell wall can be used as a barrier between the cell and the external environment to restrict ions from entering the mesophyll cells, thus protecting the cells from toxicity. The responses of mesophyll cells to saline-alkali stress in salt-tolerant and salt-sensitive materials are significantly different [52]. The mesophyll cells are important sites for photosynthesis; damage to mesophyll cells disturbs photosynthesis under stress [1]. In this study, the mesophyll-cell structure of line 2205 showed less damage compared with that in line 1423, indicating there may be more effective photosynthesis.

Synergistic Upregulation of Genes in Carbohydrates Metabolism, Photosynthetic Process, and ROS Equilibrium under NaHCO 3 Stress
In this study, RNA-seq analysis identified more DEGs in the groups of T12 vs. T0 and T24 vs. T0 compared with the groups of S12 vs. S0 and S24 vs. S0, implying that more regulatory pathways had altered in line 2205 under 75 mM NaHCO 3 stress ( Figure 3A). This is consistent with the study of Mohamed et al. [29], which shows that the tolerant rapeseed YY9 has higher DEG regulation than the salt-sensitive one ZS11 on germination under salt stress. The higher number of DEGs in the tolerant line 2205 indicates that more pathways are involved in accommodating the alkali salt stress.
Further analysis discovered that 72.25% of the alkali-salt-specific response DEGs in line 2205 had higher expression than those in line 1423. However, 73.37% of line 1423-specific response DEGs showed lower expression in line 2205. Among these DEGs simultaneous responding to alkaline salt in both cultivars, 44.27% and 34.11% DEGs possessed higher expression in line 1423 and line 2205, respectively (Figures 3 and 4). From the differentially up-and downregulated DEGs in these subsets, we speculated that their differences in alkali salt resistance were due to the diverse coordinative capabilities of positive and negative regulation of alkaline salt tolerance.
In this paper, in order to excavate the functions and pathways regulating alkaline salt in rapeseed, the functional enrichments were performed for the target-group DEGs in line 2205. A number of significant terms and pathways were annotated by GO and KEGG analysis. The genes with lower expression in line 2205 were mainly enriched in cytosol, spliceosome, and ubiquitin-mediated proteolysis ( Figure 5). Numerous studies have revealed that the splicing of stress-responsive genes and spliceosome components can be altered by salt stress, which influences the salt tolerance of plants by adjusting the homeostasis of ROS and osmosis. For example, serine/arginine-rich (SR)like protein, as a component of spliceosome, have been reported to negatively regulate salt stress [53,54]. Among the twelve serine/arginine-rich splicing factors among the selected DEGs in this study, nine genes (BnaA03g00590D, BnaA06g21030D, BnaA07g06170D, BnaC02g27300D, BnaC07g07660D, BnaC08g16960D, BnaC08g38300D, BnaCnng19170D, Bras-sica_napus_newGene_614) were from the set "down-R. in T." (Table S6). It is speculated that the weak alkali salt tolerance of line 1423 is correlated with the high expression of these genes.
However, the screened DEGs from "up-R. in T." were mainly enriched in carbohydrate metabolism, photosynthetic processes, ROS regulating, and response to salt stress ( Figure 5). A number of studies confirmed that there are complicated mechanisms in regulating salt tolerance [1,45]. However, these extensive studies mainly focus on neutral salt NaCl, but not alkaline salt, especially for the oil crop, B. napus. Our analysis showed that the higher expression of some genes in carbohydrate metabolism, photosynthetic processes, and ROS adjustment improved the alkalinity tolerance of rapeseed. The results were different from the annotated pathways (fatty acid and amino acids metabolisms) in response to Na 2 CO 3 in canola [55].
Considering the intricacy of quantitative traits, the intersection genes in target pathways were identified as candidate genes in regulating alkaline salt tolerance (Table S6). Malate dehydrogenase 2 (MDH2), a mitochondrial tricarboxylic acid cycle gene, has been proven the mRNA level of MDH2 is higher in the tolerant rye than the sensitive one in response to aluminum stress [56], consistent with the candidate gene BnMDH2 (BnaA06g08600D) in this study. Various studies have testified cytosolic glyceraldehyde-3phosphate dehydrogenase (GAPC) is participated in plant response to stresses, including heat [57], clod [58], oxidative stress [59], and cadmium [60], but no to alkali salt stress. Two homologous genes, BnGAPC1 (BnaC05g47450D) and BnGAPC2 (BnaC01g40210D), were screened out in this study (Table S6). Both genes maintained high levels of expression in line 2205 even under the NaHCO 3 treatment, but not in line 1423 ( Figure 6). 1-Aminocyclopropane-1-carboxylate oxidase (ACO), which participates in the final step of ethylene synthesis, is induced by heat, cold stresses, wounding of leaves, soil-flooding, etc. in potato [61]. In our study, two BnACO1 (BnaA01g01700D and BnaC01g02790D) showed high expressions in line 2205 ( Figure 6). Plants will burst ROS under stress. Excessive production of ROS will lead to damage for plants. However, when these molecules are at the right levels, they play an active role in signal transduction [13]. Glycolate oxidase (GOX), a typical enzyme of peroxisomes, can trigger. In plants, pathogen-induced expression of GOX activates the expression of several defense genes and improves the resistance to pathogens by increasing H 2 O 2 [62,63]. In our study, three BnGOXs (BnaC01g36940D, BnaA05g25050D, and BnaA01g29410D) with significantly higher expression in line 2205 were identified, which may result in more glucose digestion and less SS, and simultaneously boost the POD and SOD activities in line 2205 (Figures 1 and 6). Glutamate:glyoxylate aminotransferase (GGAT) plays an important role in photorespiratory carbon cycles and amino acid metabolism. The growth is strongly affected in aoat1-1 seedlings with a weak GGAT activity [64]. In our results, two BnGGTA1 (BnaA07g10020D and BnaC07g13140D) showed no expression in line 1423, and the FPKM of two BnGGTA2 (BnaAnng22050D and BnaC06g43710D) were lower in line 1423 compared with line 2205 (Figure 6).
At present, there are few studies on the salt-tolerance mechanism of rapeseed. Only a handful of neutral salt-regulated genes such as alternative oxidases (BnaAOXs) [65], calcineurin B-like proteins (BnaCBL), and CBL-interacting protein kinases (BnaCIPK) [66] were cloned, and even fewer alkali-salt-tolerance genes have been reported in B. napus. Our study proposed firstly candidate genes for alkaline salt tolerance in Brassica napus by RNA-seq, including twelve target genes in metabolic pathways (such as carbohydrate metabolism, photosynthetic processes, and ROS regulating), which positively responded to alkaline stress in rapeseed. Our study provides new insight into the mechanism of alkali salt tolerance in B. napus. Cloning and functional analyses of these candidate genes have been arranged in follow-up projects to more accurately interpret these results. Subsequent application of molecular biotechnology to interest candidate genes can facilitate the breeding production of alkali salt resistance in rapeseed. For example, gene editing technology can be used to knock out alkaline salt-sensitive genes in sensitive materials to enhance their resistance. Genes positively regulating alkaline salt resistance can be overexpressed in resistant cultivars to further enhance their resistance.

Conclusions
In summary, alkaline salts can affect membrane permeability by stimulating many strong physiological responses to alkaline salts, and the damage to mesophyll cells in tolerant canola is attenuated and delayed. The responses include enhanced osmoregulation through the addition of soluble protein and proline, and more favorable ROS scavenging through higher SOD and POD activities. RNA-seq analysis demonstrated that high expression genes in lines 2205 and 1423 participated in different metabolic pathways. In line 2205, high-expression genes participated in positive pathways such as carbohydrate metabolism, photosynthetic processes, and ROS regulating, which enhanced its tolerance. Twelve candidate genes were identified in multiple target pathways which might be the key genes responsible for alkaline salt responses in rapeseed.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13081493/s1. Table S1: Primes for qRT-PCR verification; Table S2: The sequencing results of 18 RNA libraries; Table S3: Differentially expressed genes on different groups; Table S4: The differentially expressed genes for further functional analysis; Table S5: Target pathways and annotated DEGs; Table S6: Fifty-five DEGs overlapped on multiple pathways.  . RNA-seq analyses were performed using BMKCloud (www.biocloud.net). The reference genome of B. napus was downloaded from http://www.genoscope. cns.fr/brassicanapus/data/ (accessed on 20 September 2021). The RNA-seq data have been submitted to NCBI BioProject with BioProject ID: PRJNA682529. The datasets supporting the conclusions of this article are included within the article and its additional files.

Conflicts of Interest:
The authors declare no conflict of interest.