Introduction

The salinisation of soil has become a major global resource concern and environmental issue. The salt content in soil affects the geographical distribution of plants in nature, whilst also limiting the productivity of plants (Zhu 2016; Agarwal et al. 2020). Increased concentrations of Na+ and Cl− affect plant metabolism and reduce the rate of photosynthesis, and consequently, slow down plant growth (Waqas et al. 2021; Zhao et al. 2020). In the process of long-term evolution and environmental adaptation, plants have formed a series of physiological and molecular mechanisms in response to salt stress, which can be divided into two categories: the first is salt tolerance, whereby effector molecules directly participate in changes to the physiological cellular metabolism, and therefore, contribute to salt tolerance, and the second is regulatory molecules, which are located upstream of effector molecules and are involved in signal transduction, primarily including various kinases and transcription factors comprising the signal transduction system (Chen et al. 2021).

Mei is a traditional, well-known flower in China, which has been cultivated for over 3000 years (Zhang et al. 2012). Soil salinity and alkali content are important factors that can limit the distribution of P. mume (Yang et al. 2018; Li 2009). The role of core binding factor (CBF) (NF-Y), phospholipase D (PLD), late embryogenesis abundant protein (LEA), basic leucine zipper (bZIP), and small heat shock proteins (HSP) has been established in abiotic stress (Du et al. 2012; Yang et al. 2016; Peng et al. 2016; Wan et al. 2016). Overexpression of PmLEA can improve cold and drought resistance of tobacco (Bao et al. 2017). Moreover, CBF is also associated with cold resistance, salt tolerance, and antioxidant capacity of P. mume (Yang et al. 2016). However, the salt tolerance mechanism of P. mume has yet to be systematically studied, and the signal transduction pathway involved in salt tolerance remains unclear. To expose the mechanism of salt tolerance of P. mume and screen the key genes of salt tolerance, we aim to provide a theoretical basis for the breeding of new salt tolerant varieties and to promote the cultivation and distribution of P. mume in saline alkali soil.

Materials and methods

Plant material and salt treatment

The experimental aspect of this study was carried out in the nursery greenhouse of Beijing Forestry University (116.3°E, 40.0°N), between June–August. The annual branches of P. mume ‘Meiren’ were collected from one mother plant and grafted onto the same growing rootstock of P. mume ‘Meiren’. Seedlings with relatively uniform growth were planted in plastic flowerpots with a size of 35 cm × 30 cm × 35 cm in March, with trays arranged underneath the flowerpots; the planting density was 1 plant/pot. The culture medium was mixed by volume ratio of peat soil: perlite: river sand = 2:1:1, and the mass of each pot was 7.0 kg.

When each test seedling entered the vigorous growth period, the pot soil was controlled for 7 days to ensure the soil was dry and to facilitate the even distribution of salt during salt application. Once the pot soil was dry, 1500 ml of distilled water was poured into the control pots, whilst 1500 ml of salt solution was poured into the test pots, and the exuded water from the pots was poured into the pots again. The soil NaCl content (w / W) of control group and experimental group were 0.0% and 0.3% respectively, 3 replicates of each treatment were conducted. The soil moisture content was maintained at 60–80% by adding clean water every 5 days. During treatment 0d, 3d, 10d, and 30d, samples were taken in the morning for index determination, and 2 or 3 new middle and upper functional leaves were taken from different directions of each plant. At 0 h, 6 h, 24 h, and 72 h of salt treatment, new middle and upper functional leaves were taken for transcriptome sequencing and gene expression. The transcripts were marked as T0, T6, T24, and T72, and each treatment was repeated 3 times, resulting in a total of 12 samples. The samples were immediately frozen in liquid nitrogen after collection and were stored at -80 °C until use.

Physiological index determination

The direct leaching method was used to determine chlorophyll content, whilst the conductivity meter method was used to determine relative conductivity, and the determination of malondialdehyde (MDA) was achieved through the thiobarbituric acid method. The determination of soluble sugar was attained by the anthrone colorimetric method, acid ninhydrin colorimetry was used to determine free proline (Pro), and the Coomassie Brilliant Blue G250 staining method was used to determine soluble protein. In addition, the nitrogen blue tetrazolium (NBT) photoreduction method was implemented to determine superoxide dismutase (SOD) activity, and the guaiacol method was used to determine peroxidase (POD) activity (Li 2003).

The photosynthetic gas exchange parameters, such as net photosynthesis (Pn), transpiration (Tr), stomatal conductance (Gs), and intercellular CO2 concentration (Ci), were measured by LI-6400 portable photosynthetic apparatus. The temperature was maintained at 30 °C, the light intensity was 1,000 μmol·m−2·s−1, and the CO2 concentration was 400 μmol·mol−1.

RNA extraction, library construction and data processing

RNA was extracted from the plant samples using RNAprep Pure Plant Kit (Polysaccharides and Polyphenolics-rich). Agarose gel electrophoresis was used to determine the extent of RNA degradation and contamination, whilst Nanodrop was used to detect the purity of the extracted RNA (OD260/280 ratio). In addition, a qubit accurately quantified the RNA concentration, and Agilent 2100 accurately detected RNA integrity. The cDNA library was sequenced using Illumina HiSeq 4000.

The raw image data files obtained by high-throughput sequencing were converted into original sequenced reads by CASAVA base calling analysis; the raw reads were then filtered to obtain clean reads to be used as the basis of subsequent analysis. The data processing steps were as follows: 1) reads with adapters were removed; 2) reads containing N (undetermined bases) greater than 10% were removed; and 3) low quality reads were removed (Qphred ≤ 20 bases that accounted for more than 50% of the read length). HISAT software was selected to perform genomic localisation analysis on the filtered sequencing sequence. The raw reads were uploaded to NCBI, reference: PRJNA678761.

Screening and enrichment analysis of differentially expressed Unigene

RNA-seq expression quantification was based on the fragments per kilo base per million mapped reads (FPKM) values. Also, the negative binomial distribution model was used to calculate the hypothesis test probability (pvalue). The false discovery rate (FDR) was calculated to correct for multiple testing, as applicable for multi-hypothesis testing. In addition, the p value adjusted (padj) < 0.05 was used as a differential gene screening standard.

The Gene Ontology (http://www.geneontology.org/) was divided into three parts: molecular function, biological process, and cellular component (Young et al. 2010). Pathway enrichment analysis was based on the pathway in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Also, the hypergeometric test was used to identify the pathways that were significantly enriched in the differentially expressed genes compared with the entire genome background (Kanehisa 2008). Furthermore, KOBAS (2.0) was used for pathway enrichment analysis.

Transcript screening and statistics

Plant transcription factor prediction was carried out using iTAK software. The basic principle of this process relied on the use of the TF family and rules defined in the database to identify transcription factors through hmmscan in the HMMER suite. The identification and classification of transcription factors correspond to the method of Perez-Rodriguez et al. (2010).

qRT-PCR verification

cDNA was synthesised using Primer Script™ RT Master Mix. PCR was then performed based on the CFX Connect Real-time System (Bio-Rad, USA) using SYBR Primix Ex TaqTM II (Tli RNaseH Plus). The 25μL reaction system contained 2μL of template, 1μL of upstream and downstream primers (10 μM) (Table 1), 8.5μL RNase Free dH2O, and 12.5μL of SYBR Premix Ex TaqII (Takara). All experiments were performed as 3 biological replicates. The ubiquitin C (UBC) gene was used as the internal reference gene (Wang et al. 2014). The relative expression levels of the genes were calculated using the double standard curve method.

Table 1 Primer sequences for quantitative expression
Full size table

Results

Salt response: physiological characteristics

The photosynthetic gas exchange parameters, relative conductivity, MDA level, osmotic regulators, and antioxidant enzyme activities were measured at early-stage stress (3d), mid-stage stress (10d), and late-stage stress (30d) (Fig. 1a–d). The results of Pn, Gs, and SOD enzyme showed a downward trend over time; contrastingly, relative conductivity, free Pro, and POD enzyme activities showed an upward trend. MDA, soluble sugar, soluble protein, and Ci showed an initial increasing trend, followed by decreasing levels. On the other hand, Tr initially decreased, and then increased, but the value remained lower post-stress than pre-stress. Moreover, the physiological characteristics of P. mume in response to salt stress varied in different stress periods. In the early period of stress, the photosynthetic gas exchange parameters, relative conductivity, MDA, and soluble sugar content changed significantly (P ≤ 0.05); in the mid-stage of stress, the indicators all changed with the exception of Ci, which returned to a normal value; in the late stage of stress, soluble sugar and Ci returned to normal levels, while Pn, Tr, and SOD showed values lower than those of control group, and relative conductivity, MDA, free proline, and soluble protein were higher than normal levels. Under salt stress, photosynthetic gas exchange, membrane system activity, osmotic adjustment substance, and antioxidant enzyme activity showed significant positive correlation (Fig. 1e).

Fig.1
figure 1

a Effects of salt stress on photosynthetic gas exchange parameters. b Effects of salt stress on malondialdehyde and relative conductivity. c Effects of salt stress on osmotic adjustment substances. d Effects of salt stress on antioxidant enzyme activity. e The correlation heat map of four physiological indexes: photosynthetic gas exchange parameters, membrane system related indexes, osmotic adjustment substances and antioxidant enzyme activity

Full size image

The seedlings with relatively consistent growth were planted in individual flowerpots, and the NaCl content (w/w) of the soil for the control and test plants was 0.0% and 0.3%, respectively. The samples were collected at 0d, 3d, 10d, and 30d to determine each index. According to Duncan’s test, different letters represent a different level of significance (P ≤ 0.05). The correlation heat map was drawn using GraphPad Prism 8, whereby A represents photosynthetic gas exchange parameters, B represents membrane system related indexes, C represents osmotic adjustment substances, and D represents antioxidant enzyme activity.

Transcriptome sequencing analysis

To investigate the expression patterns of the P. mume leaves at 0 h, 6 h, 24 h, and 72 h of salt stress, four samples (T0, T6, T24, and T72) were transcribed using Illumina HiSeq 4,000. After removing the splice sequence, repetitive sequence, unknown reading segment, and low-quality reading segment, 66,150,124, 51,282,181, 54,613,020, and 52,907,542 clean reads were obtained, which covered 85.53%, 84.57%, 83.05%, and 83.59% of the genome sequence of P. mume, respectively (Table 2).

Table 2 Statistical results of Illumina HiSeq 4000 sequencing
Full size table

Gene expression analysis

A total of 64,916 transcripts were obtained from the 4 libraries constructed based on sequencing. The DEGs were filtered with expression levels log2 fold change > 1 or < –1, q value < 0.005 in each pairwise comparison, and 5,161 genes were differentially expressed under salt stress (T6 vs. T0, T24 vs. T0, T72 vs. T0). Among these three groups, the greatest number of differentially expressed genes was 3140 at T72. In T6 vs. T0, the differentially expressed genes included more down-regulated genes than up-regulated genes, while at T24 vs. T0 and T72 vs. T0 the differentially expressed genes presented more up-regulation than down-regulation (Fig. 2a). The differential expression of genes varied at different stress time points. Compared with 0 h stress, a total of 527 genes were significantly differentially expressed at 6 h, 24 h, and 72 h salt stress. It is possible that these co-expressed genes are related to the salt response (Fig. 2b).

Fig. 2
figure 2

a Histogram and b Venn diagram of differentially expressed genes. A total of 5161 genes were differentially expressed under salt stress; the histogram depicts the number of genes up-regulated or down-regulated in T6 vs. T0, T24 vs. T0, and T72 vs. T0. The Venn diagram shows the number of differentially expressed genes overlapping from the three time points after salt stress

Full size image

Hierarchical clustering based on gene expression patterns (Fig. 3a) identified T24 and T72 as being more likely to be clustered together. As the stress time increased, the gene expression patterns became closer. Hierarchical clustering was used to cluster the relative expression levels of differential genes log2 (ratios) (Fig. 3b). The two most significant enrichment groups were Subcluster 61,371 genes and Subcluster 31,092 genes. At 6 h, a smaller fluctuation was identified than at 0 h, which then increased or decreased significantly at 24 h. T72 followed essentially the same pattern as T24, as the expression of a considerable number of genes changed significantly at 24 h, and this trend of change was maintained at 72 h of stress.

Twelve genes were selected to verify the transcriptome sequencing results (Enclose Table 1), whereby strong positive correlation was determined between the results of RNA-seq and qRT-PCR (Fig. 4); the correlation coefficients R2 in T6 vs. T0, T24 vs. T0, and T72 vs. T0 were 0.89, 0.90, and 0.88, respectively. This signified the reliability of the transcriptome sequencing data.

Fig. 3
figure 3

. a Hierarchical cluster analysis, using the FPKM value of the differential gene as the expression level; the areas with different colours represent different clustering grouping information. b Differential genetic clustering of the H-cluster method.

Full size image

GO and KEGG analysis of differential genes

GO and KEGG pathway enrichment analyses were conducted for the differential genes. Significant enrichment was denoted when FDR ≤ 0.05. The results found a lack of significant enrichment in T6 vs. T0. However, 17 functional groups were significantly enriched in T24 vs. T0, and the highest proportion of “organonitrogen compounds” in the biological processes class was 16.93%. In T72 vs. T0, a total of 62 functional groups were significantly enriched. The highest proportion of “cellular metabolic processes” in the biological processes class was 19.07%, the highest proportion of “thylakoids” in the cell composition class was 16.22%, and the highest proportion of “small molecule binding” in the molecular function class was 6.10% (Fig. S1 and Fig. S2). The analysis showed that up-regulated transcripts in biological processes primarily derived from the pathways of “phosphate-containing compound metabolic process”, “protein metabolic process”, “protein modification process”, and “macromolecule modification”. In addition, up-regulated transcripts in cell composition were mostly related to photosynthesis, including “photosystem”, “photosynthetic membrane”, and “thylakoid”. In terms of molecular function, many up-regulated transcripts were related to enzyme activity, including “kinase activity”, “transferase activity”, “protein kinase activity”, and “phosphotransferase activity”, whilst others were related to substance synthesis, including “ATP binding”, “small molecule binding”, “nucleotide binding”, and “anion binding”.

The result of KEGG enrichment analysis showed 1150 differential genes belonging to 113 KEGG pathways in the T6 vs. T0 group, but a significant enrichment pathway was not identified. The T24 vs. T0 group had 1138 differential genes belonging to 106 KEGG pathways, of which the “photosynthesis-antenna protein” pathway was significantly enriched. Furthermore, the T72 vs. T0 group showed 1553 differential genes belonging to 116 KEGG pathways, and the “photosynthesis-antenna protein” pathway was also significantly enriched at 72 h (Fig. S3).

Salt response-related transcription factor

A total of 362 transcripts were identified from the 4 transcripts (T0, T6, T24, and T72), which belonged to 56 gene families. Of these families, 13 had more than 10 genes (Fig. 5a). In the T6 vs. T0 group, differentially expressed transcription factors belonged to 29 families. The gene families with more than 5 genes included the basic helix–loop–helix (bHLH) gene family (11 entries), WRKY gene family (8 entries), MYB gene family (7 entries), and ethylene response factor (ERF) (7 entries). In the T24 vs. T0 group, differentially expressed transcripts belonged to 25 families; those that comprised more than 5 genes included the ERF gene family (17 entries), WRKY gene family (14 entries), bHLH gene family (8 entries), and the MYB gene family (7 entries). The T72 vs. T0 group had differentially expressed transcripts belonging to 31 gene families, and those that included more than 5 genes were the WRKY gene family (18 entries), bHLH gene family (11 entries), ERF (10 entries), DIVARICATA gene family (6 entries), and MYB gene family (5 entries). The T6 vs. T0, T24 vs. T0, and T72 vs. T0 ranges co-expressed transcripts belonging to 20 gene families, which are possibly related to salt tolerance. Among these families, bHLH, WRPK, ERF, and MYB included the largest number of genes. Four of the gene families presented gene expression at 6 h, 24 h, and 72 h of stress (Fig. 5b). Moreover, the transcription factors up-regulated at 24 h and 72 h stress were more than that at 6 h stress, with the exception of the MYB gene family whereby the up-regulated genes at 72 h were lower than 6 h. However, the transcription factors down-regulated at 6 h of stress were more than that at 24 h and 72 h.

Fig. 4
figure 4

Correlation between RNA-seq and qRT-PCR results. Twelve genes were randomly selected to verify the transcriptome sequencing results

Full size image

Identification of DEGs in different treatment periods

The genes that showed a statistically significant difference in expression levels at T6, T24, or T72, compared with those at T0, were enriched to reveal the specific genes and their functions that were significantly different in the early stage of salt tolerance in P. mume. The functional categories of the differentially expressed genes (DEGs) significantly up-regulated are shown in Fig. 6a. The results showed that a large number of transcription factors and hormone signal transduction genes were up-regulated at T6, T24, and T72. Sucrase synthase (SUSY) and photosystem I reaction centre (PS I) subunit genes of the metabolic signal transduction pathway were significantly up-regulated, indicating their considerable role in signal transduction in the early stage of salt tolerance. Several genes related to ion transport and reconstruction of ion balance, as well as genes related to the biosynthesis of osmotic protective substances, reactive oxygen species scavenging, and stress resistance-related protein genes were also significantly up-regulated. The functional categories of DEGs with significantly down-regulated expression are shown in Fig. 6b. Numerous transcription factors and aquaporins (AQPs) related to dehydration protection and water transport were significantly down-regulated in T6, T24, and T72. These findings suggest that regulatory networks involving hormone signal transduction, ion transport and reconstruction of ion balance-related genes, biosynthesis of osmotic protective substances-, reactive oxygen species scavenging, and stress resistance-related genes regulated the early salt tolerance activities of P. mume.

Fig. 5
figure 5

a 13 gene families with more than 10 transcription factors. b Differential expression of four transcription factor families under salt stress

Full size image

Based on the enrichment of GO and KEGG, protein–protein interaction analysis of DEGs and TFs was undertaken to identify the key genes conveying salt tolerance in P. mume. It was possible that these structural genes were regulated by bZIP family transcription factors. In addition, cyclic-nucleotide gated channels (CNGCs), POD, and SUSY may contribute to salt tolerance of P. mume (Fig. 7).

Fig.6
figure 6

Heatmap comparison of a significantly up-regulated DEGs and b significantly down-regulated DEG, in three treatment periods. The functional categories of significantly over-represented DEGs were identified in T0 and T72

Full size image
Fig.7
figure 7

Construction of protein–protein interaction network model of salt-tolerant candidate genes in P. mume using Cytoscape software

Full size image

Discussion

Genes related to ion transport and reconstruction of ion balance

Plants must undergo osmotic adjustments to rebuild the ion balance when under certain environmental adversity, like salt stress. Many genes were found to be involved in the reconstruction of ion balance in the gene expression profile of P. mume. The K+ ion channel (AKT) was up-regulated at 24 h of salt stress, while the Mg2+ transporter was up-regulated at 6 h, 24 h, and 72 h of stress. In addition, the K+ transporter was also up-regulated at 6 h, 24 h, and 72 h (Table 3a). Previous studies have shown that cation/H + exchangers (CHX) may also participate in sodium efflux (Manohar et al. 2011). In the present study, a CHX18 gene was found to be up-regulated at 6 h and 72 h in the differential expression profile of P. mume, which may be involved in ion metabolism regulated by CHX. A comparison of the differential expression profiles of salt stress at 6 h, 24 h, and 72 h identified greater down-regulation of genes than up-regulation in the genes corresponding to ion balance at 6 h stress, whilst greater up-regulation than down-regulation was established of the genes at 24 h and 72 h of stress. Ion metabolism and transport are significant areas in relation to the salt tolerance of P. mume because salt stress affects the ion balance, and the plant uses the process of ion transport to regain such balance.

Table 3 Key genes conveying salt tolerance in P. mume
Full size table

Genes related to biosynthesis of osmotic protective substances

The differential expression profile of P. mume showed the up-regulation of SUSY genes after stress. Sugars, polyols, amino acids (such as Pro), and betaine are all non-toxic soluble substances, which accumulate in the cytoplasm in an external hypertonic environment and have no effect on cellular metabolism (Chen et al. 2002). Pyrroline-5-carboxylate synthase (P5CS) is a key enzyme for Pro synthesis and degradation; up-regulation of the P5CS gene was identified at 6 h of stress. The P5CS gene has been successfully overexpressed in many plants such as arabidopsis and tobacco, and can effectively improve the salt tolerance of transgenic plants (Cao et al. 2011; Huang et al. 2013). The differential expression profile showed only one mannitol dehydrogenase gene, which was down-regulated at 6 h and 24 h of stress (Table 3b). Subsequent data analysis found that most of the key genes encoding osmotic protective substances lacked significant expression changes. It is possible that these genes had not yet accumulated during the stress response process at the selected time ranges, and that they primarily assist the process of stress resistance.

Genes related to active reactive oxygen species scavenging and anti-stress proteins

Under salt stress, plant cells produce large amounts of reactive oxygen species due to hindered metabolism. These species include oxygen free radicals (O2−·), hydrogen peroxide (H2O2), and hydroxyl free radicals (·OH). Although a normal product of cellular reactions, an excessive amount of reactive oxygen species is harmful to cells as they can cause lipid peroxidation of cell membranes and DNA damage that leads to cell death (Luo et al. 2020). To overcome this potential issue, plants synthesise antioxidants, which are renowned for their free radical scavenging ability. Many antioxidant genes, such as POD, SOD, metallothionein protein (MT), glutathione sulfur transferase, aldehyde dehydrogenase (ALDH), and peroxiredoxin (PrxR), were detected in the differential expression profile of P. mume (Table 3c). In high-salt and arid environments, the ALDH gene family was up-regulated, indicating their correlation to the stress response in cotton (Guo et al. 2017). Data analysis of the antioxidant genes identified greater up-regulation of POD, MT, glutathione sulfur transferase, and ALDH than down-regulation as the stress time increased;

In addition, regarding physiological response, SOD enzyme activity and expression of the SOD gene decreased after 24 h of stress, which implies that SOD was not the key gene involved in salt tolerance of the P. mume in this study. This finding differs from previous results studying the salt stress tolerance of barley and arabidopsis, whereby the SOD gene was depicted as being highly involved in salt tolerance (Zhang et al. 2021; Zheleznova et al. 2020). This difference may be due to differences in salinity tolerance of species or treatment conditions. Therefore, the role of the SOD gene in the salt stress tolerance of P. mume needs further exploration and verification.

Genes related to dehydration protection and water transport

Two AQP genes were up-regulated after stress exposure, while 9 AQP genes were shown to be down-regulated after stress. The results of the study established an increase in the expression of LEA and DHN genes of P. mume at 24 h and 72 h of stress. LEA proteins possess high hydrophilicity and thermal stability, which enables stablisation of the cell membrane structure, and can be used as molecular chaperones. Furthermore, LEA can be used to bind metal ions and prevent oxidation in cells, and induce expression under abiotic stresses including salt, drought, and cold temperatures (Cheng et al. 2021). DHN is a member of the LEA family; studies have found that overexpression of the DHN gene can enhance the response of Arabidopsis seeds to salt stress (Williams et al. 2015). In terms of P. mume, Du et al. (2012) identified 6members of the DHN gene family in the P. mume genome. With the exception of PmLEA29, the expression levels of the remaining 5 DHN genes increased under salt stress (Table 3d). Therefore, dehydration protection is an important approach for salt tolerance of P. mume.

Genes related to stress signal transduction

In the calcineurin B-like protein-CBL-interacting protein kinase (CBL-CIPK) pathway, the CBL, CIPK, and calmodulin-binding transcription activator (CAMTA) gene families all showed changes in their expression levels. CBLs and CIPKs jointly participate in the process of ion stress signal transduction. Low K+ content and high Mg2+ content generates an ion stress signal, which subsequently causes the interaction of transmembrane proteins with their respective target proteins through Ca2+ flow. This initiates cascade phosphorylation (CBLs and CIPKs) (Boudsocq et al. 2013), which results in corresponding changes to the transcription factors of downstream related genes. Alternatively, cascade phosphorylation results in direct interaction with Ca2+ transcription activators, such as CAMTAs and GT element-binding-like proteins (GTLs) (Pandey et al. 2013; Hua et al. 2012) to regulate the salt stress ionic balance genes and the expression of transcription factors related to stress resistance.

In the signal transduction pathways of mitogen-activated protein kinase (MAPK) and calcium-dependent protein kinase (CDPK), these protein kinases were up-regulated at 6 h, 24 h, and 72 h of stress. It could be speculated that the MAPK cascade achieves osmotic balance by adjusting the osmotic solute, including the functionally interconnected protein kinases of MAPK, MAPK kinase (MAPKK), and MAPK kinase kinase (Table 3e). In the differential expression profile of P. mume, only changes in the expression of phospholipase were detected; yet the corresponding gene of its hydrolysate was not found. Overall, establishing whether salt stress initiates the phospholipase transduction pathway requires more in-depth research.

Analysis of the DEGs of P. mume showed more down-regulated genes of PP2C than up-regulated genes at 6 h of stress, whilst more up-regulated genes were detected at 24 h and 72 h than down-regulated genes. The gene expression of PYL, PYR, and SnRKs increased post-stress, which subsequently positively regulated the abscisic acid (ABA) signalling pathway (Soon et al. 2012). In pathways involving protein synthesis, salt stress signals were recognised by ABA and were then transmitted to downstream transcription factors, such as MYB and MYC. These intermediate proteins recognise each other due to cis-regulatory elements, like MYBRS and MYCRS which are drought responsive elements, on the promoters of osmotic stress-induced genes. This enables the regulation of gene expression in response to salt stress (Yong et al. 2019; Li et al. 2010). The expression of bZIP, MYB, and MYC transcription factors in P. mume varies significantly; however, these factors all participate in the ABA-dependent signal transduction pathway.

Analysis of the internal mechanism of physiological response

The internal mechanism of the physiological response of P. mume to salt stress was analysed according to the physiological response results alongside the expression of key genes (Table 3). In the early stage of stress, the photosynthetic gas exchange parameters of Pn, Gs, and Tr all decreased, while Ci increased; the initial decline of Pn to Ci reflects the efficiency of carbon carboxylation, and correlation with RuBisCO activity was identified to an extent (Nik et al. 2020). Two RuBisCO-related genes were identified in the differential gene spectrum of P. mume; these were down-regulated at 24 h and 72 h of salt stress, which was consistent with the physiological results, as salt stress inhibits the activity of Rubisco, which is the key photosynthetic enzyme of P. mume. Consequently, the photosynthetic rate was found to decrease. In addition, P5CS and SUSY genes related to Pro synthesis were detected in the gene expression profile of P. mume; these genes regulate the synthesis of Pro and sucrose, promote the accumulation of Pro and soluble sugar under salt stress, and alleviate osmotic stress (Funck et al. 2020).

The results also portrayed the important role played by POD in the removal of reactive oxygen species from plant cells, and the regulation of POD enzyme synthesis by POD genes. POD enzyme activity increased after 10 days of stress. In terms of transcription level, POD genes were up-regulated at 6 h and 24 h of stress, and up-regulated genes were more than down-regulated genes. Therefore, gene expression can be viewed as related to protein synthesis, but not completely unified.