Scientific Papers

Non-targeted metabolomics analysis of metabolite changes in two quinoa genotypes under drought stress | BMC Plant Biology


Effects of PEG stress on osmotic adjustment substances and MDA in quinoa leaves

As shown in Fig. 1A-F, the cent of soluble protein and MDA increased continuously with the prolongation of drought time in HZ1. The content of soluble protein increased by 118.05% compared with the control at 9 days after stress, and the content of MDA increased by 38.15% compared with the control. The content of soluble sugar and proline increased first and then decreased after PEG stress. The content of soluble sugar increased by 91.74% compared with the control at 9 days, and the content of proline increased by 60.84% compared with the control. In L1, with the continuous drought time, the contents of soluble protein, soluble sugar and proline increased first and then decreased, but they were higher than those of the control. The content of soluble protein increased by 131.05% compared with the control at 9 days after stress, and the content of soluble sugar increased by 43.40% compared with the control at 9 days after stress. The content of MDA increased continuously during the stress period, and increased by 68.92% compared with the control at 9 days after stress. In addition, with the duration of drought stress, the relative water content of quinoa seedling leaves showed a decreasing pattern. The RWC in HZ1 decreased by 6.73%, 32.55%, 33.63% and 40.62% at 3,6,9 and 12 days after stress, respectively. The RWC in L1 decreased by 4.44%, 3.99%, 12.31% and 16.16%, respectively. In addition, this study found that with the prolongation of drought stress days, the root activity of quinoa decreased continuously in both materials. On the third day of drought stress, the root activity of the two materials began to decrease significantly, and reached the minimum on the 12 th day after stress. The comparison between materials showed that the root activity of L1 was always higher than that of HZ1 after stress, and was 7.17% (day 3), 33.46% (day 6), 49.96% (day 9) and 54.23% (day 12) higher than that of HZ1, respectively.

Fig. 1
figure 1

Effects of drought stress on osmotic adjustment substances, root activity and relative water content in leaves of Quinoa

Effects of Drought stress on Leaf Anatomical structure of Quinoa

Combined with early physiological indicators, the 9th day quinoa leaves were selected to observe the effects of drought stress on leaf anatomical structure (Table 1). It can be seen from Figure S1 that the anatomical structure of the leaves in the control group was filled with cells, the structure was clear, the intercellular space was small, and the arrangement was neat. The thickness of the upper and lower epidermis of L1-CK was 14.38 μm and 12.51 μm, and the thickness of the upper and lower epidermis of L1-T was 14.54 μm and 9.52 μm. The thickness of the lower epidermis of L1-T was significantly lower than that of L1-CK, and decreased by 23.86% compared with L1-CK. The thickness of palisade tissue of L1 increased by 52.43% compared with the control after drought stress. In addition, the ratio of grid to sea under drought stress of L1 was significantly higher than that of the control, with a value of 1.11. The thickness of sponge tissue under drought stress in HZ1 was significantly higher than that of the control, and increased by 16.07% compared with the control, and the ratio of palisade tissue to spongy tissue under drought stress was lower than that of the control.

Table 1 Leaf mesophyll structure parameters of quinoa under drought stress

Effects of drought stress on stomatal characteristics of quinoa leaf lower epidermis

Figure S2 shows the stomatal distribution and stomatal characteristics of the lower epidermis of quinoa leaves under drought stress. We found that the stomatal length, stomatal width, stomatal area, stomatal density and stomatal opening number per unit area of the two materials under drought stress were lower than those of the control (Table 2). The stomatal length, stomatal width, stomatal area and stomatal density in L1 decreased by 39.78%, 52.54%, 48.22% and 12.22% respectively compared with the control, while the number of stomatal opening per unit area increased by 50% compared with the control. Stomatal length, stomatal width, stomatal area, stomatal density and stomatal opening number per unit area in HZ1 decreased by 31.42%, 34.44%, 61.60%, 15.56% and 15.39% respectively compared with the control.

Table 2 Leaf stomatal structure eigenvalue of lower epidermis of quinoa under drought stress

Metabolite sample quality control analysis

PLS-DA analysis was performed on drought and control conditions of drought-tolerant and drought-sensitive genotype materials at two different time points (Fig. 2). The first PLS component (PC1) explained 46.5% of the total variation, while the second component (PC2) explained 12.5% of the variation for the entire dataset. The fractional plot between PC1 and PC2 shows two different groups associated with drought and control samples. It shows that there are obvious differences in metabolite accumulation under two conditions. The sensitive genotype material (HZ1) and the tolerant genotype material (L1) samples were separated from each other under drought and control conditions, especially under drought conditions.

Fig. 2
figure 2

Partial least square discriminant analysis and 2D scores loading plot for the quinoa HZ1 and L1 under control and drought treatments at two time points (3 and 9 days). Samples at control and drought treatments did not overlap with each other, indicating an altered state of metabolite levels in the quinoa leaves. HZ1, sensitive variety; L1, tolerant variety

Metabolite KEGG annotation

The KEGG annotation analysis showed that 846 metabolites (Figure S3) were involved in 13 primary pathways, including Global and overview maps (804), Biosynthesis of other secondary metabolites (615), Metabolism of terpenoids and polyketides (152), Amino acid Metabolism, metabolism of cofactors and vitamins (94), Lipid Metabolism (79), Carbohydrate Metabolism (29), Membrane transport (21), Glycan biosynthesis and Metabolism (20), Metabolism of other amino acids (18), Energy Metabolism (10), Translation (8) and Nucleotide Metabolism (7). Secondly, these 13 primary pathways contain a total of 153 secondary pathways, the 13 primary pathways contain 14,30,10,8,9,20,12,3,11,9,21,2 and 4 secondary pathways respectively. Furthermore, 113 of 846 metabolites were found to be annotated into 64 pathways by analysis, with the Metabolic pathways (KO01100) having the largest number of metabolites (75) ; They are Biosynthesis of metabolites secondary (KO01110) pathway (46), ABC transporters (KO02010) pathway (7), Phenylalanine metabolism (Ko00360) pathway (7), Tropane, piperidine and pyridine alkaloid Biosynthesis (KO00960) pathway (7), Tryptophan metabolism (KO00380) pathway (6), alpha-Linolenic acid metabolism (KO00592) pathway (5), Biosynthesis of amino acids (KO01230) pathway (5), Glucosinolate Biosynthesis (KO00966) pathway (5), Phenylpropanoid Biosynthesis (KO00940) pathway (5), the other 54 metabolic pathways had less than 5 metabolites.

PCA and OPLS-DA were used to analyze the changes of metabolites in leaves of different quinoa cultivars under drought stress

As shown in Fig. 3, significant segregation was observed in all four treatments under drought stress, with 95% confidence intervals for each group. The cumulative values of R2Y and q 2 in OPLS-DA plots were 0.993 and 0.960(a), respectively, on Day 3 of L 1 drought stress, and 0.993 and 0.960(a), respectively, on Day 9 of L 1 drought stress, the cumulative values of R2Y and Q2 in OPLS-DA plots were 0.987 and 0.914, respectively, and the cumulative values of R2Y and Q2 in OPLS-DA plots were 0.993 and 0.947, respectively, on the third day of HZ1 drought stress The cumulative values of R2Y and Q2 in OPLS-DA plots were 0.986 and 0.943(a), respectively, on the 9th day of HZ1 drought stress. The results show that the OPLS-DA model is not over-fitting and has high reliability and repeatability, which can be used in the follow-up analysis.

Fig. 3
figure 3

OPLS-DA scatter plot of L1 and HZ1 leaves under drought stress

Identification of different metabolites in leaves of different quinoa cultivars under drought stress

Based on the OPLS-DA model, the differential metabolites (DEMS) were screened by VIP value (VIP > 1) and P < 0.05. In L1,523 DEMs were identified on the 3rd day under drought stress, of which 102 DEMs were up-regulated and 421 DEMs were down-regulated, and 406 DEMs were identified on the 9th day under drought stress (Table 3, Table S1-Table S2), of which 140 DEMs were up-regulated, 266 DEMs downgraded. In HZ1,301 DEMs were identified, 177 of which were up-regulated and 124 were down-regulated by drought stress on Day 3, and 272 were identified by drought stress on Day 9, and 136 of which were up-regulated by drought stress on Day 9, 136 DEMs were downgraded. At the same time, it is obvious that the number of up-regulated DEMs in L 1 is always less than the number of down-regulated DEMs.

Table 3 basic information of differential metabolites

Analysis of common differential metabolites

In 4 group comparisons (HZ1-C3-VS-T3, L1-C3-vs-T3, HZ1-C9-vs-T9 versus L1-C9-VS-T9), we found 59 differentially expressed metabolites that were co-expressed; Among them, 47 differential metabolites were all down-regulated in the comparison of 4 groups after drought stress, and 5 differential metabolites were all up-regulated in the comparison of 4 groups after drought stress, 5(S)-hpete, Theasapogenol a, 8(R)-HETE, planagonine and His Gly Val; 3-phenyl-1-propanol were up-regulated in HZ1 and down-regulated in L1. Among them, 18 metabolites showed significant differences between treatment and control (Fig. 4). Compared with the control, the high accumulation of metabolites in plants under drought conditions included organic acids (5(S)-hpete, 8(R)-HETE and planagonine) and amino acids (His Gly Val). On the other hand, metabolites that show reduced levels under drought include quinoline, borax alcohol B, amino acids (Tyr His Leu Cys, Gln Lys Cys Phe, Tyr Phe Tyr Phe), L-Phenylalanine, 1-(14-methyl-pentadecyl) -2-(8- [3]-gradient alkane-octyl)-tin-glycerol, Chenopodium, 1-(4-phenyl-1-yl) ethyl ](prop-2-en-1-yl) amine, 2, 5-dimethoxy-4-(1-phenylpropyl-2-enyl) phenol. The model of metabolite clustering clearly shows the metabolic changes under different water conditions.

Fig. 4
figure 4

Heatmap of expression of common 18 differential metabolites in four contrast conditions (HZ1-C3-VS-T3, L1-C3-vs-T3, HZ1-C9-vs-T9 and L1-C9-vs-T9).

Unique expression of differential metabolites

At the same time, the differentially expressed metabolites in HZ1 and L1 on the 3rd and 9th day of drought stress were analyzed (Fig. 5). 103 unique differentially expressed metabolites (Table S3) were found in L1 on day 3 of drought stress, of which 93 differentially expressed metabolites were downregulated and 10 differentially expressed metabolites were upregulated. Based on the LOG2FC value, the contents of Erucamide, Lys Ser, 6-Ketoprostaglandin e 1,7-Oxo-11-dodecenoic acid, p-tert-Amylphenol and Glycodeoxycholic acid were significantly down-regulated under drought stress, the contents of 8-Dimethyl-2-phenyl-4H, 8H-benzo [1,2-b: 3,4-b’ ] dipyran-4-one, a kind of flavonoid, 3,3-Dimethylacrylic acid, 4-hydroxy-4-(3-pyridyl)-butanoic acid, 3-Acetylnerbowdine and polypeptide (Trp Met Trp His) were significantly down-regulated under drought stress.

Fig. 5
figure 5

Differentially expressed metabolites in L1 on day 3 of drought stress

There were 53 differentially expressed unique metabolites (Table S3) in L1 on day 9 of drought stress, of which 21 differentially expressed metabolites were downregulated and 32 differentially expressed metabolites were upregulated (Fig. 6). Based on LOG2FC values, the contents of polypeptides (Ile His Asp His, His Met Tyr Val, Cys Trp Arg His), trans, trans-Farnesol, 3,4,7-Trihydroxy-5-methoxy-8-prenylflavan, 4-o-(beta-d-xylopyranosyl-(1-& GT; 6) -beta-d-glucopyranoside) and Methylcyclopentane were significantly decreased under drought stress, the contents of Aplotaxene, Tyrosyl-Proline, 1,3-Diisopropylbenzene, 3’-methoxy- [6]-gingerdiol 3,5-diacetate were significantly up-regulated under drought stress.

Fig. 6
figure 6

Differentially expressed metabolites in L1 on day 9 of drought stress

There were 60 uniquely differentially expressed metabolites (Table S3) in HZ1 on day 3 of drought stress, of which 11 differentially expressed metabolites were downregulated and 49 differentially expressed metabolites were upregulated (Fig. 7). Based on the LOG2FC values, the contents of Tapentadol, 5,6,7,8-tetrahydro-2-Naphthoic Acid and Longifolenaldehyde were significantly decreased under drought stress, the contents of 5-Naphthalenetriol and (2-(5-Methyl-2-furanyl) -3-piperidinol were significantly increased under drought stress.

Fig. 7
figure 7

Differentially expressed metabolites of HZ1 on day 3 of drought stress

There were 34 uniquely differentially expressed metabolites (Table S3) in HZ1 on Day 9 under drought stress, of which 13 differentially expressed metabolites were downregulated and 21 differentially expressed metabolites were upregulated (Fig. 8). Based on the LOG2FC value, the contents of 4’-methyl-α-pyrrolidinohexanophenone, Coroglaucigenin-3-o-alpha-l-rhamnopyranoside and Ser Leu Ala were significantly down-regulated under drought stress, while the contents of Sabinol, alpha-Phellandrene dimer and soladulcidine were significantly up-regulated under drought stress.

Fig. 8
figure 8

Differentially expressed metabolites of HZ1 on day 9 under drought stress

Screening of main DEMs

Based on the log2FC value, we screened the DEMS which were different from the control and drought treatments (Table 4). In the comparison of L1-C3-vs-T3, we found that organic acids (5(S)-hpete), volatile compounds (E) -4,8-dimethyl-1,3, the contents of 7-nonyltriene, lovastatin acid, 3-methylene indole and amino acids (glutamic acid, glycine, L-Aspartic Acid and tryptophan) increased in leaves under drought stress. And the contents of Ovalitenin B, Anabsine, p-cresol and 3-phenyl-1-propanol decreased significantly under drought conditions. On the 9th day of drought stress, the contents of Aplotaxene, L-Histidinol, Zidovudine in L1 showed an increasing trend in the leaves under drought stress compared with those in L1-C9-vs-T9. The contents of alkaloids such as pseudoequine, wool phenol (Pubescenol), 3-phenyl-1-propanol, Alkyl cycloalkane (Methylcyclopentane) and phenylalanine decreased significantly under drought conditions.

Table 4 Important metabolites with L1-C3-vs -T3 and L1-C3-vs -T3

On the 3rd day of drought stress of HZ1 (Table 5), in the comparison of HZ1-c3-vs-T3, we found that the contents of organic acid (5(S)-Hpete), 15(R)-tablet thromboxane A2, one of the Phosphatidylinositol (PI), L-Asparagine-arginine and Peptides (histidine, glycine and leucine) in HZ1 tended to increase under drought stress. And the contents of quercetin 3-β-d-glucoside, phenylalanine, Ovalitenin B, Tapentadol were decreased. On the 9th day of drought stress, we found that the contents of hcl (soladucidine), APIRENE (alpha-Phellandrene dimer), 5-acetyl-3,4-dihydro-2 h-pyrrole, l-histidine (L-Histidinol) and 15(R)-tablet thromboxane a 2 in HZ1 tended to increase under drought stress. However, the contents of fern lactam (Pterolactam), NO inhibitor (V-PYRRO/No), 2,5-dimethoxy-4-(1-phenylpropyl-2-enyl) phenol, pyroglutamic acid and secondary amide (Myrtine) decreased significantly under drought condition.

Table 5 Important metabolites with HZ1-C3-vs -T3 and HZ1-C9-vs -T9

KEGG metabolic analysis of differential metabolites

In order to further identify the key metabolic pathways of two quinoa materials under drought stress, KEGG enrichment analysis of identified DEMS was carried out. Further analysis showed that 43(50 DEMs) and 42(44 DEMs) metabolic pathways were enriched in L 1 on Days 3 and 9 under drought stress, respectively, the top 20 metabolic pathways were showed in Fig. 9. On Day 3 of drought stress, 50 DEMs in L 1 were assigned to 43 metabolic pathways, with alpha-Linolenic acid metabolism (KO00592) being associated with ABC transporters; (KO02010) pathway having the most DEMs (5,10%) ; followed by Biosynthesis of amino acids (KO01230) pathway (4,8%), Tropane, piperidine and pyridine alkaloid biosynthesis (KO00960) pathway (4,8%), Tryptophan metabolism (KO00380) pathway (4,8%), Glucosinolate biosynthesis (KO00966) pathway (3,6%), Glycerophospholipid metabolism; KO00564 pathway (3,6%), Arachidonic acid metabolism (KO00590) pathway (3,6%), Biosynthesis of unsaturated fatty acids (KO01040) pathway (3,6%) and 2-Oxocarboxylic acid metabolism (KO01210) pathway (3,6%). In addition, 10 metabolic pathways contained 2 DEMs and 23 metabolic pathways contained 1 DEMs respectively. Meanwhile, alpha-Linolenic acid metabolism (KO00592) and Glucosinolate biosynthesis (KO00966) pathways were significantly enriched. On the 9th day of drought stress, 44 DEMs were distributed to 42 metabolic pathways in L 1, and the Arachidonic acid metabolism (KO00590), Tropane, piperidine and pyridine alkaloid biosynthesis (KO00960) and ABC transporters (KO02010) pathways had the most DEMs (4,9.09%) ; Secondly, Glucosinolate Biosynthesis (KO00966) pathway (3,6.82%), 2-Oxocarboxylic acid metabolism (KO01210) pathway (3,6.82%), Biosynthesis of amino acids (KO01230) pathway (3,6.82%), Tryptophan metabolism (KO00380) pathway (3,6.82%), alpha-Linolenic acid metabolism (KO00592) pathway (3,6.82%), in addition, 5 metabolic pathways contained 2 DEMS and 29 metabolic pathways contained 1 DEMs. Meanwhile, Arachidonic acid metabolism (KO00590) and Glucosinolate biosynthesis (KO00966) pathways were significantly enriched. 42(37 DEMs) and 37(34 DEMS) metabolic pathways were enriched in HZ1 on Day 3 and Day 9 under drought stress, respectively.

On Day 3 of drought stress, 37 DEMs in HZ1 were assigned to 42 metabolic pathways, with the most DEMs (4,11.11%) in the Glycerophospholipid metabolism (KO00564) versus ABC transporters (KO02010) pathway; The next were Arachidonic acid metabolism (KO00590) pathway (3,8.33%), Tryptophan metabolism (KO00380) pathway (3,8.33%), Vitamin B 6metabolism (KO00750) pathway (5.5.56%), Isoquinoline alkaloid biosynthesis (KO00950) pathway (5.5.56%), Aminoacyl-tRNA biosynthesis (KO00970) pathway (5.5.56%), Nicotinate and nicotinamide metabolism (KO00760) pathway (5.5.56%), Glucosinolate biosynthesis (KO00966) pathway (5.5.56%), Glycine, Glycine, Glycine, Glycine and nicotinamide, serine and threonine metabolism (Ko00260) pathway (5.5.56%), 2-Oxocarboxylic acid metabolism (KO01210) pathway (5.5.56%), Phenylpropanoid Biosynthesis (KO00940) pathway (5.5.56%), Biosynthesis of amino acids (KO01230) pathway (5.5.56%), alpha-Linolenic acid metabolism (KO00592) pathway (5.5.56%) and Tropane, piperidine and pyridine alkaloid biosynthesis (KO00960) pathway (5.5.56%). In addition, 27 metabolic pathways each contained 1 DEMs, and the significantly enriched metabolic pathway was Glycerophospholipid metabolism (KO00564). On the 9th day of drought stress, 34 DEMs in HZ1 were assigned to 37 metabolic pathways. Arachidonic acid metabolism (KO00590) pathway (3,9.09%), Phenylpropanoid biosynthesis (KO00940) pathway (3,9.09%), and ABC transporters (KO02010) pathway had the most DEMs (3,9.09%) Glucosinolate Biosynthesis (KO00966) pathway (2,6.06%), Vitamin B 6metabolism (KO00750) pathway (2,6.06%), Zeatin Biosynthesis (KO00908) pathway (2,6.06%), Linoleic acid metabolism (KO00591) pathway (2,6.06%), Glycine, serine and threonine metabolism (Ko00260) pathway (2,6.06%), Biosynthesis of amino acids (KO01230) pathway (2,6.06%), Tropane, piperidine and pyridine alkaloid biosynthesis (KO00960) pathway (2,6.06%). In addition, there were 27 metabolic pathways with 1 dems, and the Vitamin B 6 metabolism (KO00750) pathway and Zeatin biosynthesis (KO00908) pathway were significantly enriched.

Fig. 9
figure 9

Bubble plots of KEGG enrichment pathway under four contrast conditions of top 20 (HZ1-C3-VS-T3, L1-C3-vs-T3, HZ1-C9-vs-T9 and L1-C9-vs-T9). Note: We have obtained the rights holder’s copyright permission from KEGG officials, who allowed us to modify the KEGG images for this study as appropriate.

α-Linolenic acid metabolism pathway analysis

Based on the KEGG metabolic pathway, we found that the α-Linolenic acid metabolism pathway was enriched in many different metabolites (Fig. 10). We found that the contents of Stearidonic Acid, 13(s)-HPOTRE, Heptadecatrienal, 12-OPDA and Methyl jasmonate were significantly down-regulated under L1-C3-VS-T3 treatment, and the contents of Stearidonic Acid, 12-OPDA and Methyl jasmonate were also significantly down-regulated under L1-C9-VS-T9 treatment. However, the contents of Heptadecatrienal and Methyl jasmonate were significantly up-regulated under HZ1-C3-VS-T3 treatment. Interestingly, these differential metabolites were found to be down-regulated in the drought-resistant genotype L1 and up-regulated in the drought-sensitive genotype HZ1, these results suggested that this change of different metabolites may be different strategies for quinoa materials to cope with drought stress.

Fig. 10
figure 10

The α-Linolenic acid metabolism pathway based on the KEGG pathway



Source link