Phenotype comparison and chromosome identification of diploid and tetraploid rice
Chromosome numbers in root tips of diploid and tetraploid plants were counted. Chromosomes of Balilla-2x and Yangdao 6-2x were 2n = 2x = 24, and of Balilla-4x and Yangdao 6-4x were 2n = 4x = 48 (Fig. 1C, D). The phenotypes of diploid and tetraploid brown rice significantly differed. Tetraploid brown rice was larger (Fig. 1A, B), with longer and wider grains (Fig. 1E). The 1000-grain weight of tetraploid brown rice was 25.56%–28.94% higher than that of diploid brown rice (Fig. 1F).
Widely targeted metabolic profiling of diploid and tetraploid brown rice based on LC-MS/MS
Two groups of diploid–tetraploid brown rice (Fig. 1A, B), with three biological replicates, making a total of 12 samples were used to portray the metabolic profiles employing the widely targeted metabolomics approach. A typical total ion current plot of one QC sample is shown in Fig. 2A, which is the spectrum obtained by continuously summing the intensity of all ions in the mass spectrum at different time points. The multi-substance extracted ion chromatogram is usually used to determine the ion flux spectrum of each extracted substance in MRM mode. The multi-peak detection plot of metabolites in MRM mode is shown in Fig. 2B. Based on the MVDB and KEGG databases, and MRM, the qualitative and quantitative mass spectrometry analyses were performed on the metabolites in the samples. In total, 401 metabolites were identified, comprising 70 lipids, 68 amino acids and derivatives, 50 phenolic acids, 2 anthocyanins, 30 flavonoids, 9 flavonols, 22 flavonoid carbonosides, 2 isoflavones, 7 phenolamines, 22 alkaloids, 6 plumeranes, 33 nucleotides and derivatives, 29 organic acids, 27 saccharides and alcohols, 7 vitamins, and 17 others (Supplementary Table S1).
PCA for diploid vs tetraploid groups
The PCA score scatter plots for all samples are shown in Fig. 2C, where the abscissa and the ordinate represent the PC1 and PC2 scores, respectively. The distinction between the diploid–tetraploid Balilla (japonica) group, the diploid–tetraploid Yangdao 6 (indica) group, and the mix group was significant based on the top-ranking PCs, all the samples were within 95% confidence intervals (Hotelling’s T-squared ellipse). The PCA results suggested significant differences in metabolic phenotypes between each sample. The values for the diploid and tetraploid brown rice were separated in the PCA score plot of metabolites, and were clearly divided into two categories. The results indicated that after polyploidization of different rice lines, the change trend of their metabolites was similar, which might lead to the same changes in rice nutrients.
Orthogonal projections to latent structures-discriminant analysis (OPLS-DA) for diploid vs tetraploid groups
Compared with PCA, OPLS-DA can maximize the distinction between groups, and is more conducive to finding differential metabolites. The scatter score plots inferred from the inter-group comparison of the diploid–tetraploid Balilla group and the diploid–tetraploid Yangdao 6 group in OPLS-DA are shown in Fig. 3A and B, respectively. The R2Y and Q2Y scores were all greater than 0.99 in the B-2x vs B-4x(Fig. 3A) and Y-2x vs Y-4x (Fig. 3B), demonstrating that the ploidy difference led to the differential metabolism. The OPLS-DA model was established using many (n = 200) alignment experiments (Trygg & Wold, 2002). The horizontal line corresponds to the R2 and Q2 of the original model, and the black and gray points represent the R2 and Q2 of the model after Y replacement, respectively (Fig. 3C, D). The stable and reproducible model provided a satisfactory explanation of the difference between the two groups of samples. The OPLS-DA results showed that the differential metabolites could be screened according to VIP value in the subsequent analysis.
HCA and volcano plot of differential metabolites for diploid vs tetraploid groups
The HCA can classify metabolites with the same characteristics and identify the differences between groups. So, it can be used to evaluate the characteristic difference of metabolite accumulation caused by ploidy differences. The HCA plot of the differential metabolites identified in comparing the diploid with the tetraploid groups is shown in Fig. 2D. The HCA results showed a clear grouping pattern of different species.
Differential metabolites were also analyzed using volcano plots (Fig. 4F, G). The points in the volcano plot represent the metabolites, the abscissa indicates the fold change (FC) (log2) of each substance in the group, and the ordinate indicates the p-values (log10) of the Student’s t-test. Metabolites with FC of ≥ 2 or FC ≤ 0.5, and p < 0.05 were selected. In the japonica group, there were 182 up-regulated and 56 down-regulated metabolites in the tetraploid compared with diploid, there were 86 up-regulated and 120 down-regulated metabolites in the indica group.
The FC can describe the changes from initial to final values. In the current study, log2 FC was used to analyze the relative expression changes of metabolites between diploid and tetraploid brown rice. If log2 FC > 0, this indicates that the relative content of the metabolite was up-regulated, if log2 FC < 0, the relative content was down-regulated. Of the 401 differential expression metabolites of diploid–tetraploid, among japonica and indica groups, 180 metabolites showed opposite expression trends but 221 showed the same trends (141 up-regulated vs 69 down-regulated) (Fig. 4A, B). Moreover, the numbers of up-regulated metabolites of lipids, amino acids and derivatives, and phenolic acids in tetraploid rice were significantly increased in both the japonica and indica groups (Fig. 4A, C–E). In particular, among the 70 lipid differential metabolites, 53 showed the same trends for the japonica and indica rice groups, of which, levels of 77.36% (41 out of 53) were up-regulated (Fig. 4E).
Clustering, pathway, and enrichment analyses of lipid metabolites for diploid vs tetraploid groups
The previous analysis showed that most (levels of 77.36%) lipid metabolites were up-regulated. The heatmap of lipid metabolite changes for the diploid vs tetraploid groups is shown in Fig. 5A. The results showed a clear grouping pattern of different species. The most differential lipid metabolites identified in the study showed similar positively or negatively regulated trends between japonica and indica groups, consistent with the previous analysis. The results of these lipid metabolite annotations were classified according to the type of pathway in the KEGG database (http://www.genome.jp/kegg/). A total of 13 metabolic pathways were involved (Fig. 5B). Among these pathways, “Biosynthesis of unsaturated fatty acids” and “Metabolic pathways” were mainly involved. This indicates that the increase in ploidy mainly changed the content of unsaturated fatty acids of brown rice.
Studies have shown that the main free fatty acids in rice are unsaturated fatty acids (Zhou et al., 2003). In order to verify this previous conclusion, we determined the content of free fatty acids in brown rice (Fig. 5C). The content of free fatty acids in Balilla-4x was 28.36 nmol/g, which was 45.66% higher than in Balilla-2x (19.47 nmol/g), and the content in Yangdao 6-4x was 11.82 nmol/g, which was 52.91% higher than that in Yangdao 6-2x (7.73 nmol/g). Thus, content of free fatty acids mainly composed of unsaturated fatty acids increased significantly in tetraploid rice, consistent with previous analysis.
Statistical analysis of significant differential lipid metabolites for diploid vs tetraploid groups
In the current study, the VIP and p-value were used to analyze the significant differential lipid metabolites. If VIP > 1 and p < 0.01, a significant difference in the metabolite exists between the diploid and tetraploid groups. There were 11 metabolites with significant differences (Table 1) : two free fatty acids (γ-linolenic and punicic acids), five lysophosphatidylcholines (LysoPC 15:0, LysoPC 16:1, LysoPC 18:1, LysoPC 18:3, and LysoPC 18:3 2n isomer), two glycerol esters (MAG 18:2 and MAG 18:3 isomer1), one sphingolipid (4-hydroxysphinganine), and one phosphatidylcholine (choline alfoscerate). Most of these metabolites were significantly up-regulated in tetraploid rice (Table 1). Among them, all the lysophosphatidylcholines increased significantly, and may be important contributors to the increase of PL content in tetraploid rice. Among the annotated free fatty acid metabolites, the level of γ-linolenic increased significantly and indicated that up-regulation of γ-linolenic acid may play an important role in the increase of unsaturated fatty acids in tetraploid rice.
Table 1. Identification of significantly different lipid metabolites in diploid and tetraploid groups based on the criteria of VIP > 1 and p < 0.01.
Lipid compounds
|
B-2x vs B-4x
|
Y-2x vs Y-4x
|
VIP
|
p
|
Trend
|
VIP
|
p
|
Trend
|
γ-Linolenic acid
|
2.13
|
1.85×10-3
|
↑
|
1.11
|
8.38×10-4
|
↑
|
Punicic acid
|
2.22
|
4.29×10-3
|
↓
|
1.55
|
4.28×10-4
|
↓
|
LysoPC (15:0)
|
1.28
|
3.73×10-4
|
↑
|
1.25
|
3.94×10-4
|
↑
|
LysoPC (16:1)
|
1.67
|
8.71×10-4
|
↑
|
1.76
|
3.73×10-4
|
↑
|
LysoPC (18:1)
|
2.50
|
9.17×10-4
|
↑
|
2.31
|
1.15×10-3
|
↑
|
LysoPC (18:3)
|
5.62
|
4.16×10-5
|
↑
|
2.82
|
1.97×10-3
|
↑
|
LysoPC (18:3) (2n isomer)
|
5.48
|
2.69×10-4
|
↑
|
3.02
|
2.54×10-3
|
↑
|
4-Hydroxysphinganine
|
1.07
|
6.91×10-4
|
↓
|
1.82
|
1.74×10-3
|
↑
|
Choline alfoscerate
|
1.34
|
06.37×10-3
|
↑
|
1.39
|
7.18×10-4
|
↑
|
MAG (18:2)
|
5.34
|
2.33×10-4
|
↑
|
4.83
|
9.12×10-5
|
↑
|
MAG (18:3) isomer1
|
2.26
|
1.89×10-4
|
↑
|
1.08
|
1.13×10-4
|
↑
|
Note: the ↑ and ↓ represent increased and reduced levels of lipid metabolites between the two groups, respectively.
The γ-linolenic acid is formed by the elongation and desaturation of palmitic acid carbon chain. The pathways for synthesis of certain unsaturated fatty acids as affected by rice polyploidization are shown in Fig. 6. Heatmap and log2 FC value of lipid metabolite changes between diploid and tetraploid brown rice indicated that the level of palmitoleic (16:1, ∆9), stearic (18:0), linoleic (18:2, ∆9, 12), α-linolenic (18:3, ∆9, 12, 15), and γ-linolenic (18:3, ∆6, 9, 12) acids were increased. These unsaturated fatty acids are the main components of free fatty acids in rice. It indicated that the expression level of palmitoleic acid, stearic acid, linoleic acid, α-linolenic acid, and especially γ-linolenic acid were up-regulated after polyploidization. Finally, the content of free fatty acids in tetraploid rice was increased significantly.