NO supplementation decreases cadmium accumulation
To investigate the efficacy of different treatments used in the study, NO contents were measured (Fig. 1, Table 1, Table S1, S2 and S3). By using two-way Analysis of Variance and LSD test, there was a significant difference in NO contents between roots treated with or without Cd (Cd level). A significant difference was also found at NO level among groups with addition of NO donor (SNP and T1 treatment), NO production inhibitor and NO scavenger (L-NAME + c-PTIO and T2 treatment) and without addition of NO donor, inhibitor or scavenger (Control and Cd treatment). However, there was no significant difference at Cd × NO level (Table 1). NO contents in the roots treated by L-NAME plus c-PTIO were significantly lower than those of control (Table S1, Table S3). By contrast, SNP alone treatment significantly increased NO content in roots compared to control. Moreover, NO contents in the roots were significantly increased in plants under Cd treatment compared to control. NO accumulation was further increased when plants were treated with Cd plus NO donor SNP (T1), while it was obviously reduced in roots exposed to Cd, L-NAME (an inhibitor of NO synthase) plus c-PTIO (NO scavenger) (T2), compared to Cd treatment, respectively. These results suggested that endogenous NO contents can be effectively increased in tall fescue roots when treated by SNP, but significantly decreased by L-NAME together with c-PTIO addition, under both control and Cd-treated conditions. To further investigate NO-regulated cadmium stress adaptation in tall fescue, we focused on the four treatment groups including Control, Cd, T1 and T2. The Cd content was reduced in tall fescue roots by 11% in T1 regime relative to Cd sole treatment (Fig. 2). On the contrary, the Cd content was markedly increased by 24.2% when plants were treated with c-PTIO and L-NAME. Besides, there was no significant difference in root dry weight and root length among Cd, T1 and T2 treatment group that may be due to short time treatment (48 h) (data not shown). These results indicated that NO plays a significant role in effectively reducing the Cd accumulation in the tall fescue roots.
Table 1
The result of two-way Analysis of Variance about relative fluorescence intensity in tall fescue roots except L-NAME treatment and c-PTIO treatment.
Source of variation | Sum of Squares | df | Mean Square | F | P-value | F0.05 |
NO level | 10.5757 | 2 | 5.2879 | 72.2668 | 0.0000 | 3.5546 |
Cd level | 1.5019 | 1 | 1.5019 | 20.5265 | 0.0003 | 4.4139 |
Cd × NO level | 0.0704 | 2 | 0.0352 | 0.4811 | 0.6258 | 3.5546 |
Error | 1.3171 | 18 | 0.0732 | | | |
Total | 13.4652 | 23 | | | | |
Transcriptome Analysis
To investigate the molecular mechanisms of Cd detoxification by NO in tall fescue, transcriptome analysis was performed. A total of twelve RNA-seq libraries were constructed and sequenced to identify the DEGs that were responsive to Cd stress in the tall fescue roots with or without NO treatment. Table S4 shows an overview of the RNA-Seq reads obtained from the 12 libraries. A total of 2,005,577 transcripts were retrieved from the clean reads. The length of the transcripts ranged from 201 bp to 16,784 bp, and the mean length was 680 bp. A total of 968,924 unigenes were obtained from these transcripts, and the mean length was475 bp with an N50 length of 560 bp (Table S5). Based on de novo assembly, a total of 904 DEGs, of which 414 were up-regulated and 490 down-regulated genes were identified in the T1 as compared to Cd, but only 118 DEGs (74 up- and 44 down-regulated) were identified in the T2 vs Cd comparison (Fig. S1). Additionally, we found that 1482 DEGs (591 up-regulated and 891 down-regulated genes) were differentially expressed in the Cd vs Con comparison. Meanwhile, to confirm the reliability of the Illumina RNA-Seq, some DEGs involved in different biological processes were selected and detected by quantitative reverse transcription-PCR (qRT-PCR) analysis. The strong correlation between the result from qRT-PCR and RNA-Seq (r = 0.8935) indicated that the RNA-Seq was accurate and effective as shown in Fig. S8. GO functional and enrichment analyses were performed to classify the functions of DEGs (Fig. S2). The results demonstrated that biological processes (BP) were the most enriched among the GO categories, and metabolic processes formed the dominant group in this category. The molecular function (MF) was the second most enriched, and the highly represented GO terms were oxidoreductase activity, metal ion binding and cation binding. The cellular components (CC) were the least enriched category in the GO classification. Furthermore, the number of DEGs was significantly enriched using the KEGG database compared to the background number (q < 0.05). The enriched KEGG pathways were presented using a scatterplot method (Fig. 3). According to their enrichment factor, the top-eight enriched pathways were related to “stilbenoid, diarylheptanoid and gingerol biosynthesis”, “taurine and hypotaurine metabolism”, “flavonoid biosynthesis”, “alpha-linolenic acid metabolism”, “tyrosine metabolism”, “nitrogen metabolism”, “lysine biosynthesis” and “sulfur metabolism” from largest to smallest.
Metabolomics
GC-TOF-MS was performed to identify the differentially expressed metabolites modulated by exogenous NO in tall fescue under Cd stress. A total, 823 metabolites were detected in T1 vs Cd comparison. Ninety-nine metabolites, mainly comprising amino acids and their derivatives (14), flavones (18), anthocyanins (11), phenol amides (7), organic acids and their derivatives (4), sugars (3) and others (42), showed a significant response to the T1 treatment compared with the Cd treatment (VIP ≥ 1, fold change ≥ 2 or fold change ≤ 0.5) (Table S6). For these differentially expressed metabolites above, 65 metabolites were up-regulated and 34 metabolites were down-regulated. Besides, 131 differentially expressed metabolites (45 up- and 86 down-regulated) were identified in the T2 vs Cd group and 197 differentially expressed metabolites (161 up- and 36 down-regulated) were found in the Cd vs Con group. Moreover, the majority of the compounds detected were altered in response to the NO treatment (Fig. S3), and the most prominent group was the secondary metabolites, particularly anthocyanin and flavone. In the T1 vs Cd comparison, the metabolites peonidin (Fes0607) and rhoifolin (Fes1007) were up-regulated at levels 360.7- and 99.9-fold higher, respectively. However, the peonidin down-regulation was 2525-fold higher in the Cd vs Con group. On the other hand, the flavone metabolites acacetin (Fes1079) and sakuranetin (Fes1078) were largely down-regulated in the T1 vs Cd comparison (Table S7). Principal component analysis (PCA) revealed that the first and second principal components accounted 27.63% and 14.27% of the total variance, respectively (Fig. S4). Furthermore, the first principal component showed a separation of the Con treatment and the treatments containing Cd (including Cd, T1 and T2). In addition, the second principal component implied that the sample was treated by exogenous NO or inhibited to produce NO. The analysis of the metabolite profiles by PCA revealed a clear separation of all treated samples. Simultaneously, the results of the hierarchical cluster analysis (HCA) suggested that all the control samples (Con) clustered as a single distinct group (Fig. 4). In contrast, the other treatment lines did not form a single cluster. It is important that the two Cd lines and the two T2 lines clustered in the same manner, suggesting that NO level may be lower and closer to the Cd treatment in the large category.
Overview of the correlations in Metabolomics and Transcriptomics
Correlation analysis was performed between metabolomics and transcriptomics data to investigate new insights into the role of NO in plant Cd stress response. By comprehensive integrated analysis of all data with or without NO treatment, emphasis was given to related genes and metabolites between T1 and Cd treatment. The integrated analysis of the transcriptome and metabolome between T1 and Cd treatment demonstrated that 81 out of 904 DEGs showed enriched correlations with metabolites and mainly included GSTs, nitrate reductase (NAD(P)H), trans-cinnamate 4-monooxygenase, and ABC transporters. In addition, 255 of the metabolites detected were enriched, but only 15 metabolites were differently expressed. The data for the annotated metabolites were presented as a heat-map (Fig. S5). Three clusters were informative with regard to the differential NO response under Cd stress. The metabolites in the middle cluster were mainly involved in amino acids, especially L-pyroglutamic acid (Fes0791), and they showed higher levels under the T1 treatment than the Cd treatment. The differently expressed genes and metabolites were enriched with the same KEGG pathways, including phenylpropanoid biosynthesis, nitrogen metabolism, flavone and flavonol biosynthesis, and ABC transporters (Fig. 5 and Table S8). The DEGs were mainly involved in antioxidant systems, secondary metabolic pathways, nitrogen metabolism and metal ion transport mechanism, suggesting that NO alleviated Cd stress by a wide series of defense mechanisms.
A correlation between the transcripts and metabolites was analyzed for the transcripts related to the secondary metabolic pathways, such as those for phenylpropanoid biosynthesis (trans-cinnamic acid), flavone biosynthesis (isotrifoliin, acacetin) and organic acids (2, 5-dihydroxybenzoic acid). On the other hand, correlations were found for pathways related to nitrogen metabolism and some amino acid biosynthesis (L-citrulline, L-pyroglutamic acid and L-alanine).
To understand the interaction between differently expressed genes and metabolites more clearly, we selected some related DEGs mapped into the related metabolic pathways. In the T1 vs Cd comparison, correlations were observed for pathways, such as phenylpropanoid biosynthesis and flavone and flavonol biosynthesis (Fig. 6). The transcripts related to phenylpropanoid biosynthesis showed the highest negative correlations with the metabolite trans-cinnamic acid. In the region from trans-cinnamic acid to 4-Coumarate, trans-cinnamic acid and the genes encoding CYP73A, located in the downstream of trans-cinnamic acid, were up-regulated. It is possible that the up-regulated content of trans-cinnamic acid may contribute to the increased transcription of CYP73A. It has been reported that trans-cinnamic acid improved the growth of plants and showed an inverse effect compared to the plants treated with heavy metals (Hojati et al., 2016). Additionally, some interrelated DEGs (TRINITY_DN367754_c1_g1, TRINITY_DN359843_c1_g1, TRINITY_DN359843_c2_g1, TRINITY_DN380055_c1_g4, TRINITY_DN380055_c1_g2) were up-regulated not only in the T1 vs Cd condition but also in the T2 vs Cd condition. Strikingly, the level of isotrifoliin increased more than 4-fold in the T2 vs Cd comparison, whereas the opposite trend was observed in the T1 vs Cd comparison. Meanwhile, the gene encoding CYP75A, located in the upstream of isotrifoliin, was down-regulated in the T1 vs Cd comparison. However, the gene encoding CYP75A was not changed in the T2 vs Cd comparison. Therefore, we speculated that the down-regulated expression of CYP75A gene caused the decline of isotrifoliin indirectly. Furthermore, the organic acid 2, 5-dihydroxybenzoic acid showed similar expression patterns like isotrifoliin. All the related genes regarded ADH1, an alcohol dehydrogenase, which catalyzes the conversion of pyruvate to ethanol. Additionally, all of the ADH1-related genes were up-regulated, but only the gene encoding the flavonoid 3', 5'-hydroxylase (CYP75A) was down-regulated in the T1 treatment.
There is no doubt that SNP application will be involved in nitrogen metabolism. Additionally, it has been reported that SNP application could protect the plants against abiotic stress through its impact on nitrogen metabolism. Therefore, it was obvious to find a correlation between metabolome and transcriptome for nitrogen metabolism in T1 treatment. In this case, it was obvious that nitrate reductase (TRINITY_DN380554_c0_g2, TRINITY_DN348032_c3_g1, TRINITY_DN367739_c0_g2), as the first identified NO biosynthetic enzyme, was up-regulated in the T1 vs T2 condition, and the concentration of L-citrulline (the NO-accompanying product) was increased compared with that of the Cd treatment. Some genes mainly enzymes involved in NO-generating mechanisms were also changed. The gene encoding arginase was down-regulated, and some nitrate or nitrite transporter (NRT) were apparently up-regulated. Similarly, two transporter members of the ABC chaperone family (ABCB1 and ABCC10) were significantly up-regulated.