Changes in the respiration rates of intact PDS and the PDS with cracked seed coats
After seed coats were cracked, the respiratory rate of PDS significantly (P < 0.05) increased to 0.012 μmol CO2 g-1minute-1 on the fourth day of incubation compared with the first day of incubation (Fig. 1a). Furthermore, from the sixth day of incubation, the respiration rate continually increased and reached significantly (P < 0.05) a maximum (0.022 μmol CO2 g-1minute-1) on the fourteenth day of incubation when the germination is completed.
The respiratory rate of intact primary dormant seeds significantly (P < 0.05) increased to 0.012 μmol CO2 g-1minute-1 after 3 days of incubation and then gradually reduced to 0.006 μmol CO2 g-1minute-1 after 10 days of incubation (Fig. 1b). During the rest of time of incubation period, the respiratory rate varied between 0.004 and 0.012 μmol CO2 g-1minute-1.
Changes in the masses of PDS
The masses of PDS rapidly significantly (P < 0.05) increased 23% on the fourth day of incubation compared with the first day of incubation and then slowly increased, varying between 22 and 28% during the rest of time of incubation period (Fig. 1c).
Germination of PDRS and PDS
The germination percentage of PDRS was 93% (Fig. 1d), which was significantly higher than that of PDS (5%, P < 0.05) (Fig. 1e).
Principal component analysis of metabolites in the embryos
PCA score plot is shown in Fig. 2, respectively. According to the PCA analysis, 72.4% of the total variance was explained by PC1 (40.6%) and PC2 (31.8%) (Fig. 2). The parameters of the model were as follows: R2X=0.994,Q2=0.972. A clear formation of six distinct groups (PDRS, PDRS5, PDRS11, PDS, PDS5 and PDS11) was observed in the PCA score plot (Fig. 2). Both the PDRS5 and PDRS11 samples were discriminated from the PDRS along the first principal component. The maximum separation occurred between the PDRS5 and PDRS11 groups along the first principal component. The PDS samples were clearly grouped away from the PDS5 and PDS11 samples based on the second principal component. The PDS5 and PDS11 samples were situated relatively close to one another.
Clustering analysis of metabolites in the embryos
The HCA of identified metabolites was constructed to visualize the clustering of 98 differential components (depicted on the heat map, Fig. 3). Pearson correlation coefficients among differential metabolites were considered as a metric method to visualize the distance between metabolites. To determine the relationships and trends among the differential metabolites, a HCA plot were divided into four groups. In zone A, the contents of metabolites in group A (e.g., glycerol, mannitol, xylitol and arabitol) were the lowest in the embryos of PDS. Most of metabolites in group B exhibited the highest contents in the embryos of PDS, PDS5 and PDS11, which mainly contained polyols (e.g., ethylene glycol, erythritol, threitol, myo-inositol, inositol-3-phosphate, hexyl alcohol and 2,3-butanediol) and fatty acids (e.g., trans-9-octadecenoic acid, oleic acid, linoleic acid, stearic acid, palmitic acid, heptadecanoic acid and arachidic acid). A large proportion of amino acids, TCA cycle intermediates (e.g., malic acid, pyruvic acid, citric acid and oxoglutaric acid) and glycolysis cycle intermediates (e.g., D-glucose-6-phosphate, fructose-6-phosphate and 3-phosphoglyceric) clustered in group C. These metabolites had the highest contents in the embryos of PDRS5. The contents of the corresponding metabolites from group D were the highest in the embryos of PDRS11, including monosaccharides (e.g., mannose, trehalose, sucrose, fructose, glucopyranose, glucose and xylose). A relatively small change in the contents of these metabolites clustered in groups C and D were observed between the embryos of PDS, PDS5 and PDS11.
Partial least squares discriminant analysis of metabolites in PDRS, PDRS5 and PDRS11 embryos
The differences between PDRS5 and PDRS were mostly explained by PC1 (77.6%) (Fig. 4a). The R2 and Q2 values of the PLS-DA model were 0.99 and 0.98, respectively (Fig. 4d). Of the identified metabolites, ninety exhibited significant differences between PDRS5 and PDRS groups (P < 0.05). Sixty-five metabolites with VIP > 1 significantly contributed to the separation of PDRS5 from PDRS (Fig. 4b,c).
The variance of component 1 (88.6%) explained the differences between PDRS11 and PDS5 (Fig. 4e), which had a very high accuracy (100%) (Fig. 4h). Ninety-two exhibited significant differences between PDRS11 and PDRS5 groups (P < 0.05). Seventy-five metabolites with VIP > 1 were significantly contributed to the separation of PDRS11 from PDRS5 (Fig. 4f,g).
Fold changes of important metabolites with a VIP value > 1 in PDRS, PDRS5 and PDRS11 embryos
83% of the metabolites with VIP >1 (including the majority of sugars, organic acids and amino acids) accumulated from days 0 to 5 (Fig. 5). 73% of the metabolites with VIP >1 (including the majority of organic acids and amino acids) reduced substantially from days 5 to 11. Fructose 6-phosphate, inositol-3-phosphate, 3-phosphoglyceric and D-glucose-6-phosphate contents showed the most decrease, with decreasing 409-, 75-, 58- and 41-fold, respectively (Fig. 5). There was a 2-6 fold increase in the levels of xylose, glucose, glucopyranose, fructose and trehalose from days 0 to 5 and a further 1-3 fold continual increase from days 5 to 11 (Fig. 5).
Partial least squares discriminant analysis of metabolites in PDS, PDS5 and PDS11 embryos
The data were clearly separated into two groups, resulting score plot showing a strong separation between PDS5 and PDS (Fig. 4i). Contents of seventy-seven metabolites exhibited significant differences between PDS5 and PDS (P < 0.05). Sixty-four metabolites with VIP > 1 significantly contributed to the separation of PDS5 from PDS (Fig. 4j,k). The flexibility of the model was evaluated by explained variance R2 = 0.99 and accuracy in prediction Q2 = 0.98 (Fig. 4l).
The variance of component 1 (59.4%) explained the differences between PDS11 and PDS5 (Fig. 4m). Contents of sixty-eight metabolites of the identified metabolites exhibited significant differences between PDS11 and PDS5 (P < 0.05). Fifty-two metabolites with VIP > 1 significantly contributed to the separation of PDS11 from PDS5 (Fig. 4n,o). The model R2 and Q2 scores were 0.97 and 0.95, respectively (Fig. 4p).
Fold changes of important metabolites with a VIP value > 1 in PDS, PDS5 and PDS11 embryos
The contents of 70% metabolites with VIP >1 (including the majority of sugars, organic acids and amino acids) accumulated from days 0 to 5. Contents of D-glucose-6-phosphate, fructose 6-phosphate, 3-phosphoglyceric acid, oxoglutaric acid and fumaric acid showed the apparent changes, increasing 104-, 55-, 30-, 21- and 16-fold, respectively (Fig. 6). The levels of these metabolites with VIP >1 displayed relatively small changes from days 5 to 11.
The main altered metabolic pathways in PDRS embryos
Those pathways located on the top right corner of the ‘metabolome view’ are the main altered metabolic pathways (Fig. 7). From days 0 to 5, the main altered metabolic pathways included amino acid metabolisms (glycine, serine and threonine metabolism, alanine, aspartate and glutamate metabolism, arginine and proline metabolism and tyrosine metabolism), carbohydrate metabolism (glyoxylate and dicarboxylate metabolism, pyruvate metabolism, TCA cycle and glycolysis) and the metabolism of cofactors and vitamins (pantothenate and CoA biosynthesis) and glycerophospholipid metabolism (Fig. 7a).
From days 5 to 11, the following metabolic pathways were mainly altered, including amino acid metabolism (alanine, aspartate and glutamate metabolism, beta-alanine metabolism, glycine, serine and threonine metabolism and arginine and proline metabolism), carbohydrate metabolism (citrate cycle, pyruvate metabolism, glycolysis and pentose phosphate pathway), lipid metabolism (sphingolipid metabolism) and the metabolism of cofactors and vitamins (pantothenate and CoA biosynthesis) (Fig. 7b).
The main altered metabolic pathways in PDS embryos
With the exception of glycerophospholipid metabolism, those metabolic pathways significantly changed in PDRS embryos also displayed significant alternation in PDS embryos from days 0 to 5 (Fig. 7c). In addition, pentose phosphate pathway also showed major alternation. The pronouncedly altered metabolic pathways from days 0 to 5 significantly changed from days 5 to 11, except for glycine, serine and threonine metabolism and pyruvate metabolism (Fig. 7d).
Comparative metabolic pathway analysis between PDRS and PDS
From days 0 to 5, there was an increased activity of carbohydrate metabolism and amino acid metabolism in PDRS embryos and a similar but less intense increase in PDS embryos (Fig. 8). From days 5 to 11, most intermediates of carbohydrate metabolism and amino acid metabolism in PDRS embryos decreased substantially. In contrast, there were minor changes in these metabolites (including sugars, organic acids and amino acids) in PDS embryos.