Morphological and physiological changes of wheat under nitrogen deficiency
The morphological and physiological changes of wheat were shown in Fig. 1. The crop height of N0 was 0.75 times significantly lower than that of N1 (Fig. 1a); the leaf area per plant of N0 was 0.70 times significantly smaller than that of N1; the specific leaf area of N0 and N1 had no significant difference; the net photosynthetic rate (Pn) of N0 was 0.47 times significantly lower than that of N1 (Fig. 1d); the shoot fresh wight of N0 was significantly 0.61 times less than that of N1. In brief, the nitrogen deficiency may cause a lower crop height, smaller leaf area per plant, low Pn, and less fresh weight of wheat shoot.
The root length per plant of N0 was 1.61 times significantly longer than that of N1 (Fig. 1b); the root volume per plant of N0 was 0.61 times lower than that of N1; the root surface area per plant of N0 was 1.04 times bigger than that of N1; the root fresh weight of N0 was 0.82 times more than that of N1. The root shoot ratio of N0 was 1.36 times higher than that of N1 (Fig. 1c). In brief, the nitrogen deficiency may cause a longer root length, lower root volume per plant, bigger root surface area per plant, more root fresh weight; moreover, the nitrogen deficiency increased the root shoot ratio.
Global analysis of RNA-seq data resulting from nitrogen deficiency
The number of genes expressed in different regions were calculated, and stacked histogram was drawn (Fig. 2a). There were 72487-78729 genes expressed in wheat shoots, and 17116-22418 genes had Fragments Per Kilobase of transcript per Million fragments mapped (FPKM) values greater than 1. There were 63273-64413 genes expressed in wheat roots, and 27785-29233 genes had FPKM values greater than 1.
Principal component analysis (PCA) was applied to explore the relationship between samples by locating the samples at different dimensions (Fig. 2b). The closer the clustering distance was, the more similar the samples were. The results of PCA analysis showed that PCA1 reflected the difference of root and shoot, accounting for 99.41% of the total variation; PCA2 reflected shoot transcription difference under N0 and N1, accounting for 0.21% of the total variation; PCA3 reflected root transcription difference under N0 and N1, accounting for 0.11% of the total variation.
The volcanogram (Fig. 3ab) and cluster map (Fig. 3cd) of P value and log2FC were applied to screen the differentially expressed genes (DEGs) under nitrogen deficiency treatment (N0) as compared to control (N1). There were 3949 DEGs in shoot, 1535 of them were up-regulated, 2414 of them were down-regulated. There were 3911 DEGs were screened in roots, 1236 of them were up-regulated, 2675 were down-regulated (Fig. 3e). The venn map (Fig. 3f) revealed that 1535 DEGs were up-regulated and 2414 were down regulated in both shoot and root. There were 372 DEGs differentially expressed in roots and shoot.
Functional analysis of DEGs identified under nitrogen deficiency
The gene ontology classification (Fig. 4) enriched 1205 up-regulated genes and 1888 down-regulated genes in shoots, while 961 up-regulated genes and 1883 down-regulated genes in roots. The enriched genes were classified into 3 major classes and 64 sub-classes. Some genes belongs to the two or more classifications. The largest four classifications (more than 980 DEGs) were cellular process, metabolic process, binding, and catalytic activity (Table 1).
The KEGG classification (Fig.5a, b) analyse of wheat under nitrogen deficiency showed the followings. In shoot, the pathway of gene information processing-translation had the most down-regulated genes (142 genes), whereas the pathway of metabolism-biosynthesis of other secondary metabolites had the most up-regulated genes (54 genes). In root, the pathway of metabolism-carbohydrate metabolism had the most down-regulated genes (118 genes), whereas the pathway of metabolism-biosynthesis of other secondary metabolites had the most up-regulated genes (78 genes). The highest enrichment score in shoot was monobactam biosynthesis (Fig.5c), whereas the highest enrichment score in shoot was nitrogen metabolism (Fig.5d).
Table 1 The number of differentially expressed genes (DEGs) in the four pathways with the largest number of genes under nitrogen deficiency
|
GO_classify 1
|
GO_classify 2
|
DEGs-Up
|
DEGs-Down
|
Shoot
|
Biological process
|
Cellular process
|
385
|
947
|
Biological process
|
Metabolic process
|
498
|
1029
|
Molecular function
|
Binding
|
672
|
1031
|
Molecular function
|
Catalytic activity
|
560
|
854
|
Root
|
Biological process
|
Cellular process
|
312
|
669
|
Biological process
|
Metabolic process
|
390
|
891
|
Molecular function
|
Binding
|
527
|
893
|
Molecular function
|
Catalytic activity
|
429
|
936
|
Analysis of gene families associated with cellular process
Expansin family members were mainly belonged to the cellular process in gene ontology (GO) classification. Under nitrogen deficiency, there were 3 DEGs (Fig. 6) of wheat Expansin family in shoot, including TreasCS2B02G411700 (up-regulated), TreasCS1A02G30020 (down-regulated) and TreasCS1B02G310300 (down-regulated); there were 6 down-regulated wheat Expansin family members (TreasCS6A02G307900 and so on) and 24 up-regulated wheat Expansin family members (TreasCS5B02G528400 and so on) in root.
Analysis of gene families associated with metabolic process
Pet and Psb family members were important proteins in photosystem in wheat shoot, which mainly belonged to metabolic process in gene ontology (GO) classification. Under nitrogen deficiency (N0), there were 3 down-regulated DEGs (Fig. 7) of wheat Pet family (TreasCS7A02G325500 and so on); there were 8 down-regulated wheat Psb family members (TreasCS3D02G523300 and so on) and 1 up-regulated wheat Psb family members (TreasCS6B02G412100).
Nar and Nrt family members had functions related to nitrogen metabolism, and mainly belonged to metabolic process in gene ontology (GO) classification. Under nitrogen deficiency, there were 3 down-regulated DEGs (Fig. 8) of wheat Nar family (TreasCS6A02G326200, TreasCS6B02G356800, and TreasCS6D02G306000) in both shoot and root of wheat; but there were other 2 up-regulated wheat Nar family members (TreasCS6A02G210000 and TreasCS6D02G193100) in root; there were 9 up-regulated DEGs of wheat Nrt family in root (TreasCS6A02G031100).
Validation of transcriptome data
In the 100 candidate genes (Fig. 9a), there were 94 genes whose RT-qPCR results were consistent with FPKM results in transcriptome data, in which there were 46 shoot data and 48 root data. This indicated that nearly 94% of the transcriptome data were reliable. The coefficients of X of regression lines were 0.93 and 1.05 for shoot and root, respectively, which indicated that the transcriptome data had high accuracy. The RT-qPCR data of 50 candidate genes in root and 50 candidate genes in shoot were shown in Fig.9b and Fig.9c, respectively. The comparison between RT-qPCR and transcriptome data of each genes can be queried in Table S1 and Table S2.