The seed germination rate of LC and LR under salinity stress was also tested. As Fig. 1b and 1c show, germination rate was more severely impaired in LC than LR; the germination in LR was higher and occurred sooner than LC under the control condition and 150 mM NaCl concentration. Furthermore, relative to the control, more sown seeds of LR (86%) germinated than those of LC (18%), indicating the ability to germinate of LC is severely hindered by salinity stress, but not LR.
Hormone changes in wolfberry in response to salinity stress
To further explore the differences between LC and LR in tolerance to salinity, the abscisic acid (ABA), jasmonic acid (JA), and salicylic acid (SA) content in the leaves of LC and LR under control and 150 mM NaCl conditions were tested. As Fig. 2 shows, for the control group, the ABA content in LR was significantly lower than that in LC. Compared with the control condition, 150 mM NaCl treatment induced the ABA content slightly increase in LC leaves, but this was not a significant difference, whereas a significantly large amount of ABA accumulated in the leaves of LR (Fig. 2a). Further, the JA content was significantly higher in leaves of LR than LC under the control condition, though JA accumulation was significantly reduced by salinity stress, more in LR than in LC (Fig. 2b). The SA content in the leaves of LR was lower than that in LC under control condition, but was not significantly affected by the salinity treatment in both species (Fig. 2c). These results indicated that the tolerance to high salinity stress in LR was driven by accumulating ABA while reducing its JA content.
Overview of the transcriptomic responses of LC and LR to salt stress
To better understand the molecular basis of salinity stress responses in LC and LR, we carried out transcriptome sequencing and analyzed different expressed genes (DEGs) between LC and LR under control or salinity conditions. A total of 2836 DEGS were detected in LC under salinity stress compared with the control group, in which 1337 genes were up-regulated and 1499 genes down-regulated. For LR, however, only 141 genes were differentially expressed when treated with high salinity,in which 80 genes were up-regulated and 61 genes were down-regulated (Fig. 3a). To identify the key determinate factors of the transcriptome, PCA was performed on the genes of the two species under control or salinity treatment conditions. The first two principal components (PC1, PC2) were able completely distinguish our combinations of species and treatment (i.e., 2 species x 2 treatment levels [mock and 150mM salinity concentration]). The PCA shows a clear separation between the wolfberry species along PC1 and the separation of treatment can be observed along PC2. In addition, the three biological replicates were projected closely in the ordination space, which suggested a good correlation between replicates (Fig. 3b). A Venn diagram was used to analyze and display the differences between variation genes of LC and LR under salinity stress respectively. As depicted in Fig. 3c, group LC-mock vs. LC-NaCl and group LR-mock vs. LR-NaCl only shared two changed genes in total under salinity stress.
KEGG enrichment analysis of DEGs in LC and LR under salinity stress
Evidently, as shown in Fig. 4a, for LC-mock vs. LC-NaCl, the differential genes in LC between the control and salinity condition are mainly enriched in metabolic pathways (48.57%), biosynthesis of secondary metabolites (22.81%), plant–pathogen interaction pathway (10.26%), MAPK signaling pathway (6.16%), amino sugar and nucleotide sugar metabolism pathway (6.16%), carbon metabolism (5.47%), and plant hormone signal transduction (5.02%). For the group LR-mock vs. LR-NaCl (Fig. 4b), the DEGs in LR between control and salinity condition are mainly enriched in metabolic pathways (60%), biosynthesis of secondary metabolites (38.18%), carbon metabolism, protein processing in endoplasmic reticulum, spliceosome, tryptophan metabolism, and lysine degradation..
Dynamic transcriptome analysis between LC and LR in response to salinity stress
To study the genes expression patterns in LC and LR under background and salinity conditions, a K-means cluster analysis was performed, in which the expression patterns of genes of the LC-mock, LC-NaCl, LR-mock, and LR-NaCl groups of wolfberry plants were classified into 10 subclasses; these were then roughly divided into six categories (Fig. 5). The first category was class of genes that showed no regulation change in LC when subjected to salinity stress compared with the background condition, yet they showed a trend of up-regulation in LR (subclass1, subclass9). The second category, by contrast, was a class of genes whose regulation levels also went unchanged under salinity stress (compared with background condition) in LC but whose tendency was to become down-regulated in LR under salinity stress (subclass7, subclass8). The third category of genes featured an up-regulated expression trend under salinity stress in LC, which remained unchanged under salinity stress in LR (subclass2, subclass3). The genes of the fourth category were instead down-regulated in LC under salinity stress while their expression levels were mostly unchanged in LR (subclass6, subclass10). Concerning the fifth category genes, under salinity stress, they were up-regulated in LC yet down-regulated in LR when exposed to high salinity stress(subclass4). The sixth category of genes had expression levels not induced by salinity stress in either LC or LR (subclass5).
KEGG pathway enrichment analysis was carried out for the 10 subclasses. These results demonstrated that these DEGs were mainly involved in metabolic pathways and biosynthesis of secondary metabolites pathways. For the first category, the first five pathways in both subclass1 and subclass9 are plant–pathogen interaction, plant hormone signal transduction, carbon metabolism, MAPK signaling pathway, and RNA transport. For the second category, the first five pathways in subclass7 are protein processing in endoplasmic reticulum, carbon metabolism, starch and sucrose metabolism, plant–pathogen interaction, and phenylpropanoid biosynthesis; the first five pathways in subclass8 are plant–pathogen interaction, ribosome, carbon metabolism, plant hormone signal transduction, and biosynthesis of amino acids. For the third category, the first five pathways in subclass2 and subclass3 are plant–pathogen interaction, plant hormone signal transduction, protein processing in endoplasmic reticulum, MAPK signaling pathway, and carbon metabolism. For the fourth category, the first five pathways in subclass6 and subclass10 are carbon metabolism, plant hormone signal transduction, biosynthesis of amino acids, starch and sucrose metabolism, and MAPK signaling pathway. For the fifth category, the main enriched pathways in subclass4 are plant–pathogen interaction, and protein processing in endoplasmic reticulum. For the sixth category, the main enriched pathways in subclass5 are plant–pathogen interaction, carbon metabolism, ribosome, biosynthesis of amino acids, and RNA transport.
Metabolomic analysis of LC and LR responses to salinity stress
Next, the metabolites of Chinese wolfberry and black wolfberry plants under salinity stress were detected, and the difference in the metabolites between species or conditions was analyzed. As Fig. 6a shows, for 80 metabolites in LC, their expression levels were changed under salinity stress, in which 57 were up-regulated and 23 were down-regulated. The expression levels of 69 metabolites in LR were changed under salinity stress, in which 34 were up-regulated and 35 were down-regulated. Compared with LC, 207 metabolites were differentially expressed in the leaves of LR under background condition,in which 151 were up-regulated and 56 were down-regulated. In all, 234 metabolites were differentially expressed in the leaves between LC and LR under salinity stress, of which 146 were up-regulated and 88 were down-regulated. The PCA of the metabolites in the control group and the salt treatment group of LC and LR showed that PC1 and PC2 could completely distinguish the four combinations of species and treatment (Fig. 6b). In Fig. 6c, the difference in metabolites’ change between different comparative groups is summarized (using a Venn diagram). Groups LC-mock vs. LC-NaCl and LR-mock vs. LR-NaCl shared eight metabolites with common changes,in which seven were up-regulated, and one was down-regulated. Compared with LC-mock vs. LR-mock, the group LC-NaCl vs. LR-NaCl had 161 metabolites featuring the same change tendency, 113 of which were up-regulated and 48 down-regulated. In the other comparisons, LC-mock vs. LR-mock and LC-mock vs. LC-NaCl, 25 metabolites had the same trend in variation, all of which were up-regulated.
KEGG enrichment analysis of different metabolites in LC and LR under salinity stress.
The metabolites in the four comparison groups (LC-mock vs. LC-NaCl, LR-mock vs. LR-NaCl, LC-mock vs. LR-mock, LC-NaCl vs. LR-NaCl) were enriched by KEGG, with the results summarized in Fig. 7. All the metabolites were mainly enriched in metabolic pathways and biosynthesis of secondary metabolites pathways, followed by a detailed analysis of other enrichment pathways. As seen in Fig. 7a, in the group LC-mock vs. LC-NaCl, the changes in metabolites induced by salinity stress in LC mainly concerned these pathways: microbial metabolism in diverse environments, biosynthesis of alkaloids derived from shikimate pathway, biosynthesis of phenylpropanoids, arginine and proline metabolism, and flavone and flavonol biosynthesis. In Fig. 7b, for the group LR-mock vs. LR-NaCl, the metabolites variation induced by salinity stress in the leaves of LR were mainly enriched in the following pathways: biosynthesis of amino acids, purine metabolism, cysteine and methionine metabolism, biosynthesis of amino acids, and protein digestion and absorption. For the group LC-mock vs. LR-mock (Fig. 7c), the different metabolites between leaves of LC and LR at the background condition were mainly concentrated in five pathways: microbial metabolism in diverse environments, pyrimidine metabolism, purine metabolism, flavone and flavonol biosynthesis, and biosynthesis of phenylpropanoids. Finally, in the LC-NaCl vs. LR-NaCl group (Fig. 7d) the disparity in leaf metabolites between LC and LR under salinity pressure mainly arose in these pathways: microbial metabolism in diverse environments, purine metabolism, tryptophan metabolism, phenylpropanoid biosynthesis, and flavone and flavonol biosynthesis.
Analysis of top 20 differentially expressed metabolites in LC and LR under normal and salinity stress conditions
The top 20 differentially expressed metabolites with more significant log2FC whose expression pattern matched their related genes in the four comparison groups (LC-mock vs. LC-NaCl, LR-mock vs. LR-NaCl, LC-mock vs. LR-mock, LC-NaCl vs. LR-NaCl) are shown in Fig. 8. In the group LC-mock vs. LC-NaCl, these highly ranked metabolites were linked to flavonol, anthocyanins, polyamine, nucleotide and its derivatives, organic acids and quinate. More specifically, most flavonols, including hyperoside, hyperin, avicularin, and biorobin, were down-regulated in LC when exposed to salinity. In the anthocyanins classification, malvidin-3-O-rutinoside-5-O-glucosides were up-regulated, while both delphinidin 3-galaactoside chloride and procyanidin B2 were down-regulated. The polyamines were up-regulated, while the nucleotide and its derivatives were down-regulated (Fig. 8a). In the group LR-mock vs. LR-NaCl, the primarily changed metabolites were associated with amino acids derivatives, nucleotide and its derivatives, polyamine, vitamins, anthocyanin, coumarins, nicotinic acid derivatives, hydroxycinnamoyl derivatives, and organic acids. Examined in greater detail, the amino acids derivatives, 3-hydroxykynurenine and L-(-)-cystine were all down-regulated, whereas S-(5’-adenosy)-L- homocysteine and L-cysteine were both up-regulated during salinity stress. Nucleotide and its derivatives, such as adenosine 5’-monophosphate, adenine, and iP7G, along with the polyamines, such as N-sinapoyl cadaverine, diCaf-put, and N-sinapoyl putrescine, in addition to the organic acids like D-erythronolactone, were all up-regulated in LR when exposed to high salinity. Furthermore, some vitamins, namely nicotinamide-N-oxide and (-)-riboflavin, were down-regulated in LR during salinity stress (Fig. 8b). In the group LC-mock vs. LR-mock, under background conditions, the main differential metabolites found were flavonoid, anthocyanins, and polyamine. Some flavonoids were up-regulated in LR compared with LC, like C-hexosyl-apigenin O-caffeoylhexoside, C-hexosyl-tricetin O-pentoside and isorhamnetin rutinose, but others were evidently down-regulated, such as hesperetin C-hexosyl-O-hexosyl-O-hexoside, kaempferol-3-O-glucoside-7-O-soph, luteolin O-hexosyl-O-hexoside and quercetin-3-O-glucose-7-O-soph (Fig. 8c). In the group LC-NaCl _vs._ LR-NaCl, the major differential metabolites in LR compared with LC under salinity stress were related to flavonoid anthocyanins and polyamine. Most of the gathered anthocyanins and polyamines were down-regulated in LR compared with LC under high salinity. In the flavonoid classification (Fig. 8d), some were up-regulated in LR compared with LC—e.g., C-hexosyl-tricetin O-pentoside, quercetin-O-glucoside, isoquercitroside, physcion-8-O-β-D-glucoside, biorobin and C-hexosyl-apigenin O-caffeoylhexoside—while several other flavonoids were down-regulated in LR compared with LC under salinity stress (such as kaempferol-3-O-glucoside-7-O-soph, luteolin O-hexosyl-O-hexosyl-O-hexoside, and quercetin-3-O-glucose-7-O-soph).
KEGG pathway enrichment in DEGs and different metabolites in LC and LR
According to the above KEGG enrichment analysis, a histogram was drawn to show the common pathways in which DEGs and differential expressed metabolites were highly enriched in. As Fig. 9 shows, in which a taller ordinate column corresponds to greater enrichment. In group LC-mock vs. LC-NaCl (Fig. 9a), the metabolic pathways distinguished by a simultaneously higher enrichment of DEGs and differential metabolites are arginine and protein metabolism, benzoxazinoid biosynthesis, and riboflavin metabolism. In group LR-mock vs. LR-NaCl (Fig. 9b), the corresponding metabolic pathways are lysine degradation pathway, nitrogen metabolism, and purine metabolism.
Flavonoid metabolism in LC and LR during salinity stress
Flavonoid metabolism plays an important role in protecting plants against adverse effects of salinity stress. Figure 10a illustrates the flavonoid biosynthesis pathway, for which the marked genes that were analyzed appear in Fig. 10b. The expression profiles of almost all these marked genes in the flavonoid metabolism pathway had a pattern of lower abundance in LR than LC, either under the background condition or salinity stress. Specifically, the genes encoding chalcone synthase (Cluster-40571.102907), flavone synthase II, 2-hydroxyisoflavanone synthase-like (Cluster-40571.125750), and flavonol synthase, and flavonol synthase/flavanone 3-hydroxylase (Cluster-40571.25710) were apparently not in LR but up-regulated in LC when wolfberry plants were exposed to salinity stress. However, the genes for flavone synthase II, 2-hydroxyisoflavanone synthase-like (Cluster-40571.123809), flavone synthase II, 2-hydroxyisoflavanone synthase-like (Cluster-40571.199168), and flavonoid 3’-monooxygenase, flavonoid 3’-monooxygenase (Cluster-40571.294286) were all no-regulated in LR yet down-regulated in LC during salinity stress. Moreover, the genes encoding naringenin 3-dioxygenase and naringenin 2-oxoglutarate 3-dioxygenase (Cluster-40571.120883) were up-regulated in both LC and LR under salinity stress, while that for naringenin 3-dioxygenase (Cluster-40571.135119) was up-regulated in LR but non-regulated in LC. Furthermore, the genes for flavone synthase II, 2-hydroxyisoflavanone synthase-like (Cluster-40571.303908) were down-regulated both in LR and LC in salinity stress conditions (Fig. 10b). Interestingly, most of the changed metabolites in the flavonoid biosynthesis pathway during salinity stress persisted in higher abundance in LR but at a lower level in LC. Including butin, catechin, neohesperidin, naringenin, and afzelecin; in which, butin was down-regulated both in LR and LC, catechin was non-regulated in LR but up-regulated in LC, neohesperidin was up-regulated in LR but down-regulated in LC, naringenin was down-regulated in LR but up-regulated in LC, and afzelechin was non-regulated in LR but up-regulated in LC. Apart from those metabolites, eriodictyol remained at a lower level in LR than LC, yet it was down-regulated in LR though not regulated in LC. While pinocembrin also occurred at a low level in LR than LC, it was up-regulated in both LR and LC. In contrast, chlorogenic acid was generally higher LR than LC, and it was up-regulated in LR but down-regulated in LC (Fig. 10c).'
Alterations in flavone and flavonol metabolism in LC and LR after exposure to high salinity
It is noteworthy that the flavone and flavonol biosynthesis pathway was enriched significantly in both transcriptomic and metabolomic data of wolfberry plants. The flavone and flavonol biosynthesis pathway appears in Fig. 11a, and the marked expression pattern of relative genes are detailed in Fig. 11b. This revealed that most of the genes involved in flavone and flavonol biosynthesis pathway stay constitutively expressed at low level in LR but at a higher level in LC, such as flavonol-3-O-glucoside glucosyltransferase (Cluster-40571.113183) and galactoside glucosyltransferase (Cluster-40571.291876), although both their expression levels went unchanged during salinity stress in both species. For glucosyltransferase (Cluster-40571.113184), flavonol-3-O-glucoside galactoside glucosyltransferase (Cluster-40571.249135), and flavonoid 3’-monooxygenase (Cluster-40571.294286), their transcription levels were high in LC and low in LR, and down-regulated in LC but not-regulated in LR, when exposed to salinity stress. Regarding flavonol-3-O-glucoside L-rhamnosyltransferase (Cluster-40571.163208) and kaempferol 3-O-beta-D-galactosyltransferase (Cluster-40571. 188476), their abundance of transcripts were higher in LC than LR, with expression up-regulated in LC yet non-regulated in LR during salinity stress. Besides, some other genes—including the novel plant SNARE (Cluster-40571.242780), flavonoid 3’-monooxygenase (Cluster-40571.294284), and flavonoid 3’-monooxygenase-like (Cluster-40571.294288), showed consistently greater expression in LR than LC, with transcription levels up-regulated in LR but down-regulated or not regulated in LC when the plants were exposed to salinity stress (Fig. 11b). Further, for most of the changed metabolites in flavone and flavonol biosynthesis pathway during salinity stress in wolfberry, their content stayed at a high level in LR but a lower level in LC. As seen in Fig. 11c, the 3,7-di-O-methylquercetin continued to have a lower content under the background condition, but this was up-regulated to greater extent in LC and down-regulated in LR when the plants were exposed to salinity stress. Moreover, isovitexin, astragalin, and cosmosiin had contents that stayed at higher level in LR than LC under the background condition, yet under salinity stress both of them were down-regulated in LR and LC. The rutin content was found to be higher in LR than LC, but this did not change during salinity stress. This was not so for cynaroside, whose background content remained low in LC but high in LR, while it was down-regulated in LC and up-regulated in LR under high salinity conditions (Fig. 11c).