Abstract Through crosstalk, FLAGELLIN SENSITIVE 2 (FLS2) and RESPIRATORY BURST OXIDASE HOMOLOG D (RBOHD) are involved in regulating the homeostasis of cellular reactive oxygen species (ROS) and are linked to the metabolic response of plants toward both biotic and abiotic stress. In the present study, we examined the metabolome of Arabidopsis seedlings under drought and salt conditions to better understand the potential role of FLS2 and RBOHD‐dependent signaling in the regulation of abiotic stress response. We identified common metabolites and genes that are regulated by FLS2 and RBOHD, and are involved in the response to drought and salt stress. Under drought conditions, D‐aspartic acid and the expression of associated genes, such as ASPARAGINE SYNTHASE 2 (ASN2), increased in both fls2 and robed/f double mutants. The accumulation of amino acids, carbohydrates, and hormones, such as L‐proline, D‐ribose, and indoleacetaldehyde increased in both fls2 and rbohd/f double mutants under salt conditions, as did the expression of related genes, such as PROLINE IMINOPEPTIDASE, PHOSPHORIBOSYL PYROPHOSPHATE SYNTHASE 5, and NITRILASE 3. Collectively, these results indicate that the FLS2‐RBOHD module regulates plant response to drought and salt stress through ROS signaling by adjusting the accumulation of metabolites and expression of genes related to metabolite synthesis. Keywords: drought, FLS2, metabolome, RBOHD, salt, stress, transcriptome __________________________________________________________________ To explore the potential signaling pathway that is required for the quick induction of stress‐related metabolites in plants in response to drought and salt stress, in this study, we studied the changes of differentially abundant metabolites in the fls2 mutant and the rbohd rbohf (rbohd/f) double mutant and revealed a new metabolomic perspective on the role of FLAGELLIN SENSITIVE 2 and RESPIRATORY BURST OXIDASE HOMOLOG D in regulating metabolites accumulation in response to drought and salt stress conditions graphic file with name PEI3-4-36-g007.jpg 1. INTRODUCTION During growth and development, plants continuously interact with the environment during their growth and development, inducing the synthesis of many different primary and secondary metabolites (Yang et al., [50]2018). Different metabolites are specifically or non‐specifically expressed in different developmental stages and tissues of plants, as well as in response to a variety of biotic and abiotic stresses (Isah, [51]2019). In this regard, drought and salt stress can have a significant negative impact on plant growth and development, and also induce dynamic changes in the level of different metabolites (Arif et al., [52]2020; Kumar et al., [53]2021). This includes metabolites such as amino acids, carbohydrates, and also secondary metabolites, such as phenolic acids and flavonoids, terpenes and steroids, and alkaloids (Isah, [54]2019; Singh et al., [55]2020). Notably, studies have indicated that these metabolites play a key role in plant adaptation to biotic and abiotic stresses (Fàbregas & Fernie, [56]2019). A recent study reported a higher accumulation of L‐aspartic acid in leaves of tolerant varieties of chickpea (Cicer arietinum L.) under drought stress conditions, indicating that L‐aspartic acid may serve as a marker metabolite for drought response (Khan et al., [57]2019). The concentration of soluble protein and proline, as well as superoxide dismutase and catalase activity, was also found to be elevated in Apiaceae (Bupleurum chinense DC.) in response to drought stress (Yang et al., [58]2020). Collectively, studies have shown that the synthesis of a series of metabolites is induced as part of plant response to adverse conditions. FLAGELLIN SENSITIVE 2 (FLS2) and RESPIRATORY BURST OXIDASE HOMOLOG D (RBOHD) function in regulating cytosolic calcium and the production of reactive oxygen species (ROS) during the defense response in plants (Chi et al., [59]2021; Li et al., [60]2014; Melotto et al., [61]2006). Receptor‐like cytoplasmic kinase BIK1 (BOTRYTIS‐INDUCED KINASE 1), one of the components of the FLS2 immunoreceptor complex, positively regulates the accumulation of ROS by directly phosphorylating RBOHD (Li et al., [62]2014). FLS2 forms a functional complex with the BRASSINOSTEROID INSENSITIVE 1 (BRI1)‐associated kinase receptor 1 (BAK1), resulting in Ca^2+ influx and ROS production (Chinchilla et al., [63]2007). Abiotic stresses, such as salt stress, can also enhance the accumulation of ROS in plants by activating RBOHD (Luo et al., [64]2021). ROS accumulation and the activation of the Ca^2+ signaling pathway are known to be involved in abiotic stress response in plants (Bush, [65]1995; Yu et al., [66]2002). Notably, our recent study revealed the essential roles of the FLS2‐RBOHD‐PIF4 module in regulating the adaptive response of plants to drought and salt stress (Liu et al., [67]2022). Although FLS2‐RBOHD module has been studied on how plants respond to abiotic stress, the effects of FLS2‐RBOHD model on metabolite synthesis and level changes in plants under abiotic stress conditions have been less studied. Therefore, we used liquid chromatograph‐mass spectrometry (LC–MS) to analyze the dynamic changes in metabolite levels in an fls2 mutant and an rbohd rbohf (rbohd/f) double mutant under drought and salt stress conditions. The transcriptomes of the fls2 mutant and rbohd/f double mutant under drought and salt stress conditions were also characterized. Differentially abundant metabolites (DAMs) in the fls2 mutant and rbohd/f double mutant were identified and these data were combined with the transcriptome data and analyzed. Key metabolites and their corresponding regulatory genes that may be involved in regulating plant response to drought and salt stress conditions were identified. Our results provide novel insight that contributes to a comprehensive understanding of the mechanisms by which FLS2 and RBOHD regulate plant response to abiotic stress through ROS signaling by adjusting the accumulation of metabolites. 2. EXPERIMENTAL PROCEDURES 2.1. Plant material and growth conditions The WT is Columbia (Col‐0). The fls2 (SALK_141277) and rbohd/f double mutants (CS9558) in Col‐0 background were obtained from the Arabidopsis Biological Resource Center (Table [68]S1). Homozygous T‐DNA insertion lines were confirmed by polymerase chain reaction (PCR) using gene‐specific and T‐DNA‐specific primers (Table [69]S2; Figure [70]S3). For NaCl treatments, 1‐week‐old seedlings were first transplanted into the soil to grow for 1 week under normal growth conditions. Subsequently, these were watered with an aqueous solution containing 100 mM NaCl and allowed to grow further for 1 week. For drought treatment, the seedlings were first transplanted into normal watered soil; afterward, watering was stopped after transplantation. After 1 week, the soil water content decreased to about 10%, and the seedlings were further grown for 1 week. For control, seedlings of the same batch were transplanted into the soil and grown under normal watering conditions (watering once a week) for 3 weeks. 2.2. RNA‐seq analysis Total RNA was extracted using the mirVana miRNA isolation kit (Ambion) following the manufacturer's protocol. RNA integrity was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies). The samples with RNA Integrity Number (RIN) ≥ 7 were subjected to subsequent RNA‐seq analysis. The libraries were constructed using TruSeq Stranded mRNA LTSample Prep Kit (Illumina) according to the manufacturer's instructions. Libraries were sequenced on the Illumina sequencing platform (HiSeqTM 2500 or Illumina HiSeq X Ten), and 125‐bp/150‐bp paired‐end reads were generated. Then, raw data (raw reads) were processed using Trimmomatic. The reads containing ploy‐N and the low‐quality reads were removed to obtain clean reads. Then, the clean reads were mapped to the reference genome using hisat2. After that, the FPKM value of each gene was calculated using cufflinks, and the read counts of each gene were obtained by htseq‐count. DEGs were identified using the DESeq (2012) R package functions estimateSizeFactors and nbinomTest. p‐value <.05 and fold change >2 or fold change <.5 were set as the threshold for significantly differential expression. Hierarchical cluster analysis of DEGs was performed to explore gene expression patterns. RNA sequence data are available at [71]https://dataview.ncbi.nlm.nih.gov/?search=SUB8234436 (Liu et al., [72]2022). 2.3. LC–MS analysis All chemicals and solvents were analytical or HPLC grade. Water, methanol, acetonitrile, and formic acid were purchased from CNW Technologies GmbH. L‐2‐chlorophenylalanine was purchased from Shanghai Hengchuang Bio‐technology Co., Ltd. 2.4. Sample preparation Transfer 80 mg of accurately weighed sample to a 1.5 mL Eppendorf tube. Two small steel balls were added to the tube. 20 μL of 2‐chloro‐l‐phenylalanine (0.3 mg/mL) dissolved in methanol as internal standard and a 1 mL mixture of methanol and water (7/3, vol/vol) were added to each sample, samples were placed at −80°C for 2 min. Then grinded at 60 HZ for 2 min, and ultrasonicated at ambient temperature for 30 min after vortexed, then placed at 4°C for 10 min. Samples were centrifuged at 13,000 rpm, 4°C for 15 min. Supernatant in a brown and glass vial was dried in a freeze concentration centrifugal dryer. A mixture of methanol and water (1/4, vol/vol) were added to each sample, samples vortexed for 30 s, then placed at 4°C for 2 min. Samples were centrifuged at 13,000 rpm, 4°C for 5 min. The supernatants (150 μL) from each tube were collected using crystal syringes, filtered through 0.22 μm microfilters, and transferred to LC vials. The vials were stored at −80°C until LC–MS analysis. QC samples were prepared by mixing aliquots of all samples to be a pooled sample. 2.5. Sample on‐board processing An ACQUITY UHPLC system (Waters Corporation) coupled with an AB SCIEX Triple TOF 5600 System (AB SCIEX) was used to analyze the metabolic profiling in both ESI‐positive and ESI‐negative ion modes. An ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm) was employed in both positive and negative modes. The binary gradient elution system consisted of (A) water (containing 0.1% formic acid, v/v) and (B) acetonitrile (containing 0.1% formic acid, v/v) and separation was achieved using the following gradient: 0 min, 5% B; 2 min, 20% B; 4 min, 25% B; 9 min, 60% B; 14 min, 100% B; 18 min, 100% B; 18.1 min, 5% B and 19.5 min, 5% B The flow rate was 0.4 mL/min and column temperature was 45°C. All the samples were kept at 4°C during the analysis. The injection volume was 2 μL. Data acquisition was performed in full scan mode (m/z ranges from 70 to 1000) combined with IDA mode. The QCs were injected at regular intervals (every 10 samples) throughout the analytical run to provide a set of data from which repeatability can be assessed. Mass spectrum conditions: ESI was used as ion source. Positive and negative ion scanning modes were used to collect the sample quality spectrum signals. The mass spectrum parameters are shown in Table [73]1. TABLE 1. Mass spectrum parameter Parameter Positive ion Negative ion Spray voltage (V) 3800 3200 Capillary temperature (°C) 320 320 Aux gas heater temperature (°C) 350 350 Sheath gas flow rate (Arb) 35 30 Aux gas flow rate (Arb) 8 0 S‐lens radio frequency level 50 50 Mass range (m/z) 100–1000 100–1000 Full ms resolution 70,000 70,000 MS (Mass)/MS resolution 17,500 17,500 NCE (Normalized Collisional Energy)/stepped NCE 10, 20, 40 10, 20, 40 [74]Open in a new tab 2.6. LC–MS data preprocessing and statistical analysis The acquired LC–MS raw data were analyzed by the progenesis QI software (Waters Corporation) using the following parameters. The resulting matrix was further reduced by removing any peaks with a missing value (ion intensity = 0) in more than 50% samples. The internal standard was used for data QC (quality control) (reproducibility). Metabolites were identified by progenesis QI (Waters Corporation) Data Processing Software, based on public databases such as [75]http://www.hmdb.ca/; [76]http://www.lipidmaps.org/ and self‐built databases. The positive and negative data were combined to get a combined data which were imported into R ropls package. Principal component analysis (PCA) and (orthogonal) partial least‐squares‐discriminant analysis (O) PLS‐DA were carried out to visualize the metabolic alterations among experimental groups, after mean centering (Ctr) and Pareto variance (Par) scaling, respectively. PCA is to explore the degree of correlation among multiple possible correlation variables, find the maximum or minimum correlation direction, and achieve the purpose of data compression or noise reduction (dimension reduction). OPLS‐DA analysis combines orthogonal signal correction and PLS‐DA methods to decompose the X matrix into two types of information related and unrelated to Y, and screen the differential variables by removing the unrelated differences. The abscissa and ordinate of the PCA graph represent the projected score values of each sample on the PC1 (principal components 1) and PC2, respectively (Nicholson et al., [77]1999). The projected score value of each sample on the plane is composed of the first principal component and the second principal component is the spatial coordinate, which can intuitively reflect the similarity or difference between the samples (Okada et al., [78]2010). The closer the distance between different samples indicates the closer the composition and concentration of the molecules they contain (Jolliffe & Cadima, [79]2016). On the OPLS‐DA graph, there are two principal components, the predicted PC1 and the orthogonal principal component (PCo1) (Trygg & Wold, [80]2002). OPLS‐DA maximizes the differences between groups and reflects the PC1, so the differences between groups can be directly distinguished from the PC1, while the PCo1 reflects the intra‐group differences (Trygg & Wold, [81]2002). The Hotelling's T2 region, shown as an ellipse in score plots of the models, defines the 95% confidence interval of the modeled variation. Variable importance in the projection (VIP) ranks the overall contribution of each variable to the OPLS‐DA model, and those variables with VIP >1 are considered relevant for group discrimination. In this study, the default 7‐round cross‐validation was applied with one‐seventh of the samples being excluded from the mathematical model in each round, in order to guard against overfitting. The differential metabolites were selected based on the combination of a statistically significant threshold of VIP values obtained from the OPLS‐DA model and p values from a two‐tailed Student's t‐test on the normalized peak areas, where metabolites with VIP values larger than 1.0 and p values less than .05 were considered as differential metabolites (Tables [82]S3–S24). Metabolites involved in metabolic pathway analysis using MetaboAnalyst5.0: [83]https://www.metaboanalyst.ca/ (Pang et al., [84]2022). 2.7. Association analysis of metabolome and transcriptome Metabolite synthesis‐related genes were found according to the metabolic pathways involved in metabolites in the KEGG: [85]https://www.kegg.jp/. The genes involved in metabolite‐related metabolic pathways were compared with the transcriptome data obtained by RNA‐seq to obtain the expression levels of genes in different samples. The heat map of gene expression was drawn using the website of Oebiotech: [86]https://cloud.oebiotech.com/task/detail/heatmap/. And the GO analyses were performed on Metascape: [87]http://metascape.org (Zhou et al., [88]2019). Firstly, the metabolite synthesis‐related genes list was uploaded on Metascape according to the operating manual. After uploading the data, GO analyses were performed on Metascape automatically. After the analysis was completed, the results of GO were downloaded from Metascape. 3. RESULTS 3.1. LC–MS analysis of fls2 mutant and rbohd/f double mutant under no‐stress (CK), salt, and drought stress conditions We previously demonstrated that an Arabidopsis rbohd/f double mutant exhibited sensitivity to drought and salt stress, relative to wild‐type (WT) and fls2 mutant plants which exhibited tolerance to both stresses (Liu et al., [89]2022). Therefore, we analyzed the metabolome of WT, fls2 mutant, and rbohd/f double mutant under drought and salt stress conditions to determine if FLS2 and RBOHD/F regulate the response of plants to drought and salt stress through their ability to mediate the accumulation of specific metabolites. An LC–MS full scan was first used to obtain data on the ionic strength of different molecules (Figure [90]S1a). We determined the intensity of the strongest ions at each retention time point in fls2, rbohd/f, and WT plants under salt stress, drought stress, and CK conditions (Figure [91]S1a). Differences in the number of peaks and ion intensity in each of the graphs derived from the different samples indicate that the types and amounts of metabolites differed in each sample under the different treatment conditions (Figure [92]S1a). A PCA diagram illustrates that the samples clustered together, and therefore indicates that there was little difference between samples (Figure [93]S2). The metabolomic data obtained in the experiment were multidimensional and some of the variables were highly correlated. Therefore, we utilized a multivariate statistical analysis to identify DAMs between different comparison groups. Principal component analysis was first used to observe the overall distribution of DAMs in different samples (Figure [94]S1b). Then, orthogonal projections to latent structures discriminant analysis (OPLS‐DA) was used to determine differences between samples (Figure [95]S1c). DAMs with biological significance were thus identified through the OPLS‐DA analysis. In addition, we also used fold change to further confirm the significance of the DAMs between comparison groups (Figures [96]2a, [97]4a, and [98]6a; Figure [99]S5a). FIGURE 2. FIGURE 2 [100]Open in a new tab Identification and analysis of the differentially abundant metabolites (DAMs) in the comparison between “fls2_NaCl versus WT_NaCl” samples. (a) Volcano plots of DAMs were identified in the various comparisons of fls2 and WT under CK and salt stress conditions. DAMs were identified based on a p‐value <.05 and a |log2FC| > .58. Red dots indicate significantly higher abundant metabolites in the comparison groups, green dots indicate significantly lower abundant metabolites, and gray dots indicate metabolites with insignificant changes in abundance. (b) Bar chart of the number of metabolites in different categories of metabolites in “fls2_NaCl versus WT_NaCl” samples. The height of the blue bars represents the number of metabolites with increased abundance in the different comparison groups. The gray bars represent the number of metabolites with decreased abundance. (c) Heat map of metabolite‐related gene expression in “fls2_NaCl versus WT_NaCl” comparisons. The heat map indicates the expression of genes related to L‐aspartic acid, L‐proline, L‐glutamic acid, L‐methionine and sucrose, maltose, epicatechin, indoleacetaldehyde, guanine, and cytosine in fls2_NaCl, fls2_CK, WT_CK, WT_NaCl samples. (d) Heat map of gene expression of metabolite synthesis‐related genes identified in GO enrichment analysis. Genes presented are related to (c). FIGURE 4. FIGURE 4 [101]Open in a new tab Identification and analysis of the differentially abundant metabolites (DAMs) in different comparison groups under CK and salt stress conditions. (a) Volcano plots of DAMs were identified in the various comparisons of rbohd/f and WT samples under CK and salt stress conditions. DAMs were determined based on p‐value and fold change (p‐value < .05, |log2FC| > .58). Red dots indicate metabolites with significantly increased abundance, green dots represent metabolites with significantly decreased abundance, and gray dots indicate a change in metabolites that were nonsignificant. (b) Bar chart of the number of metabolites in different categories of metabolites in the “rbohd/f_NaCl versus WT_NaCl,” “rbohd/f_CK versus WT_CK,” “WT_NaCl versus WT_CK,” and “rbohd/f_NaCl versus rbohd/f _CK” comparison groups. The height of the blue bars represents the number of metabolites with increased abundance in the different comparison groups. The gray bars represent the number of metabolites with decreased abundance. (c) Heat map of metabolite‐related gene expression in the different rbohd/f_NaCl versus WT_NaCl comparisons. Heat map of metabolite‐related gene expression in rbohd/f_NaCl, rbohd/f_CK, WT_CK, and WT_NaCl samples. Heat map indicates gene expression levels for genes related to L‐proline, N2‐acetyl‐L‐ornithine, D‐ribose, sucrose, indoleacetaldehyde, epicatechin, dAMP, and UDP‐D‐galactose. (d) Heat map of gene expression of metabolite synthesis‐related genes identified in GO enrichment analysis. Genes presented are related to (c). FIGURE 6. FIGURE 6 [102]Open in a new tab Identification and analysis of differentially abundant metabolites (DAMs) in different comparison groups of fls2 and WT plants under CK and drought stress conditions. (a) Volcano plots of DAMs in fls2 and WT samples under CK and drought stress conditions. DAMs were identified based on a p‐value < .05 and |log2FC| > .58. Red dots indicate metabolites with significantly increased abundance, green dots indicate metabolites with significantly decreased abundance, and gray dots indicate DAMs whose change in abundance was nonsignificant. (b) Bar chart of the number of metabolites in different categories of metabolites in each of the comparison groups. (c) Heat map of metabolite‐related gene expression in fls2 and WT samples under CK and drought stress conditions. The heat map illustrates the expression level of genes related to D‐aspartic acid, L‐glutamic acid, D‐glucose 1‐phosphate, maltose, epicatechin, and guanine and uracil synthesis in fls2_Drought, fls2_CK, WT_CK, and WT_Drought samples. (d) Heat map of gene expression of metabolite synthesis‐related genes identified in GO enrichment analysis. Genes presented are related to (c). 3.2. FLS2 is involved in sensing salt stress FLAGELLIN SENSITIVE 2 is involved in sensing the bacterial flagellin flg22, which in turn leads to a transient increase in cytosolic calcium ions and ROS (Chinchilla et al., [103]2006; Gómez‐Gómez & Boller, [104]2000). In the present study, we characterized changes in the metabolome and transcriptome of fls2 mutant and WT plants under CK and NaCl conditions to determine the potential involvement of FLS2‐mediated ROS signaling in response to salt stress. Principal component analysis indicated that the distance between fls2 mutant and WT plants under salt stress conditions is relatively far, suggesting that the types and abundance of metabolites in fls2 mutant plants were different from WT plants under salt stress conditions (Figure [105]S1b). We then identified the DAMs in fls2 mutant samples and WT samples under CK and NaCl conditions (VIP >1, p < .05) via LC–MS analysis. The metabolic pathway matching analysis of the identified DAMs was conducted utilizing the exact mass number, MetaboAnalyst ([106]www.metaboanalyst.ca/), and Kyoto Encyclopedia of Genes and Genomes (KEGG) (Figure [107]1a,b). We focused on metabolites that were annotated in the KEGG database, especially those known to be development‐related metabolites and pathways (Figure [108]1a,b). Results identified 48 DAMs in the comparison between “fls2_NaCl versus WT_NaCl” that were assigned to 26 metabolic pathways mainly associated with plant development, including: “arginine biosynthesis,” “galactose metabolism,” and “citrate cycle (TCA cycle)” (Figure [109]1a,b). FIGURE 1. FIGURE 1 [110]Open in a new tab Metabolic network diagram of differentially abundant metabolites (DAMs) in fls2_NaCl and WT_NaCl samples. (a) Metabolic network diagram of DAMs in the comparison between fls2_NaCl and WT_NaCl samples. Solid lines indicate one‐step reactions; dashed lines indicate two‐ or more‐step reactions. (b) Impact of DAMs in fls2_NaCl versus WT_NaCl comparisons on different metabolic pathways. All matching pathways for metabolites are shown based on p‐values from pathway enrichment analysis and pathway impact values are derived from pathway topology analysis. The color of the circle is based on the p‐values, from light to dark, and the p‐value from large to small. The radius of the circle is based on its pathway impact value, from small to large, and the impact value from small to large. The pathway impact value is the ratio of the sum of the importance measures of each matched metabolite to the importance measures of all metabolites in each pathway. Next, we analyzed changes in the abundance of these metabolites. Results indicated that the number of amino acids and carbohydrates with increased abundance were greater in fls2_NaCl than in WT_NaCl, including L‐proline, L‐glutamic acid, and maltose (Figure [111]2a,b). Under normal growth conditions in the comparison group of “fls2_ CK versus WT_CK,” the number of flavonoids with increased expression in fls2 mutant is higher than WT. In contrast, under NaCl treatment conditions, in the comparison group of “fls2_Nacl versus WT_NaCl,” the number of flavonoids with increased expression in fls2 mutant is lower than WT (Figure [112]2b). The high abundance of the identified metabolites in the fls2 mutant indicates that the fls2 mutant was sensitive to salt stress. Notably, we also found that the abundance of indole acetaldehyde was significantly higher in the “fls2_NaCl versus WT_NaCl” comparison. An association analysis of the transcriptome and metabolome data was then performed to determine if changes in the expression of genes related to the DAMs were following each other. As shown in Figure [113]2c, we focused on DAMs such as L‐aspartic acid, L‐proline, L‐glutamic acid, L‐methionine, sucrose, maltose, epicatechin, indoleacetaldehyde, guanine, and cytosine in the correlation analysis (Figure [114]2c). The expression data (transcriptome) of metabolite‐related genes were used to construct a heatmap and GO process chart (Figure [115]2c,d). The results indicated that in “fls2_NaCl versus WT_NaCl,” “fls2_CK versus WT_CK,” “fls2_NaCl versus fls2_CK,” and “WT_NaCl versus WT_CK” comparisons, changes in the expression of metabolite‐related genes exhibited similar trends to changes in the abundance of metabolites. For example, the expression level of the L‐glutamic acid‐related genes gamma‐glutamyl transpeptidase 1 (GGT1) and glutathione s‐transferase tau 18 (GSTU18) were upregulated in the “WT_NaCl versus WT_CK” and “fls2_NaCl versus WT_NaCl” (log[2]FC >0) comparison groups, which is consistent with the increase of L‐glutamic acid in “fls2_NaCl versus WT_NaCl” comparison group (Figure [116]2a,c). Additionally, the guanine‐related gene GUANYLATE KINASE—(AGK2 or GK‐2) was downregulated in “fls2_NaCl versus WT_NaCl” and “fls2_NaCl versus fls2_CK” comparison groups, which is consistent with the changes of observed in guanine abundance (Figure [117]2a,c). Furthermore, as shown in Figure [118]2c, we also found that genes related to metabolites such as sucrose and indole acetaldehyde were significantly upregulated in the “fls2_NaCl versus fls2_CK” comparison group (Figure [119]2c), indicating that the fls2 mutant is either highly sensitive to salt stress or that FLS2 is involved in sensing salt stress signals. 3.3. Increased salt stress sensitivity of rbohd/f double mutant plants Excessive soil salinity induces an increase in ROS that is regulated by RBOHD/F (Xie et al., [120]2011). To protect cells from the harmful effects of excessive levels of ROS produced under salt stress conditions, plants have developed a variety of antioxidant defense mechanisms to maintain the homeostasis of the intracellular redox state (Acosta‐Motos et al., [121]2017). Therefore, we conducted an analysis of the metabolome and transcriptome of the rbohd/f double mutant and WT plants under normal and salt stress conditions to examine the effect of the ROS signal generated by RBOHD/F on the growth of plants under salt stress conditions. Principal component analysis indicated that the distance between rbohd/f double mutant and WT samples was considerable in the “rbohd/f_NaCl versus WT_NaCl” comparison, suggesting that the level of metabolites in rbohd/f plants was altered in response to salt stress, relative to WT plants (Figure [122]S1b). A total of 45 DAMs were identified in the “rbohd/f_NaCl versus WT_NaCl” comparison group that were assigned to 18 metabolic pathways, including “arginine and proline metabolism,” “alanine, aspartate and glutamate metabolism,” and “riboflavin metabolism” (Figure [123]3a,b). FIGURE 3. FIGURE 3 [124]Open in a new tab Metabolic network diagram of differentially abundant metabolites (DAMs) in the “rbohd/f_NaCl versus WT_NaCl” comparison. (a) Metabolic network diagram of DAMs in the “rbohd/f_NaCl versus WT_NaCl” comparison group. Solid lines indicate one‐step reactions; dashed lines indicate two‐ or more‐step reactions. (b) Impact of DAMs in the “rbohd/f_NaCl versus WT_NaCl” comparison group on different metabolic pathways. Matching pathways for metabolites were selected based on p‐values in a pathway enrichment analysis and pathway impact values were derived from pathway topology analysis. The number of amino acids, carbohydrates, and flavonoids with increased abundance was greater in rbohd/f_NaCl samples, relative to rbohd/f_CK samples, including metabolites such as N2‐acetyl‐L‐ornithine, sucrose, and epicatechin. The number of amino acids, carbohydrates, and flavonoids with increased abundance in the “rbohd/f_NaCl versus rbohd/f_CK” comparison group also increased, relative to the “WT_NaCl versus WT_CK” comparison group (Figures [125]3a and [126]4a,b). Moreover, rbohd/f_NaCl samples exhibited an increase in the numbers of carbohydrates and flavonoids, compared to WT_NaCl samples, as was well as an increase in the abundance of auxin‐related metabolites such as indole acetaldehyde and indoleacetic acid (Figure [127]4a,b). The number of amino acids with increased abundance also increased in the “rbohd/f_NaCl versus WT_NaCl” comparison group, relative to the “rbohd/f_CK versus WT_CK” comparison group (Figure [128]4b). Most of the metabolites with increased abundance were involved in “galactose metabolism” and “arginine and proline metabolism,” which contribute to the energy supply available for plant growth, as well as plant defense (Abedi et al., [129]2018; Majumdar et al., [130]2016) (Figure [131]3a,b). In contrast, the number of pyrimidine and purine metabolites decreased in rbohd/f_NaCl samples in the “rbohd/f_NaCl versus WT_NaCl” comparison group, including cGMP and UDP‐D‐galactose (Figure [132]4b). We also found that the number of carbohydrate metabolites that decreased in abundance was higher than the number of carbohydrate metabolites with increased abundance in the “rbohd/f_CK versus WT_CK” comparison group (Figure [133]4b). Next, we performed an association analysis between the transcriptome and metabolome data obtained for the “rbohd/f_CK versus WT_CK,” “rbohd/f_NaCl versus WT_NaCl,” “WT_NaCl versus WT_CK,” and “rbohd/f_NaCl versus rbohd/f _CK” comparison groups. Association analysis focused on the expression of genes related to L‐proline, N2‐acetyl‐L‐ornithine, D‐ribose, sucrose and indoleacetaldehyde, epicatechin, dAMP, and UDP‐D‐galactose (Figure [134]4c,d). The analysis indicated that the expression of genes related to L‐proline, sucrose, indoleacetaldehyde, N2‐acetyl‐l‐ornithine, and D‐ribose, epicatechin synthesis in rbohd/f_NaCl samples was mostly higher than in other samples (Figure [135]4c). The expression of N‐ACETYL‐L‐GLUTAMATE SYNTHASE 1 (NAGS1), SUCROSE SYNTHASE 3 (SUS3) was also significantly higher in rbohd/f_NaCl samples than in other samples. The level of expression of genes related to dAMP and UDP‐D‐galactose in rbohd/f_NaCl samples, however, was lower than in WT_NaCl samples (Figure [136]4d). These results suggest that RBOHD/F are involved in regulating the expression of genes related to metabolites induced by salt stress. 3.4. Fls2 is involved in sensing drought stress signals Plants subjected to drought stress accumulate a variety of organic and inorganic substances, including sugars, amino acids, and inorganic ions, to increase the osmotic potential of their cells and enhance their water retention capacity (Rhodes & Samaras, [137]1994). Therefore, we analyzed the metabolome and transcriptome of fls2 mutant and WT plants under CK and drought conditions to examine the function of FLS2 in drought response. Principal component analysis indicated that the fls2 mutant subjected to drought stress did not significantly differ in their accumulation of metabolites compared to WT plants (Figure [138]S1b). A total of 16 DAMs were identified in the “fls2_Drought versus WT_Drought” comparison that were assigned to 14 metabolic pathways. These included “glyoxylate and dicarboxylate metabolism,” “alanine, aspartate, and glutamate metabolism,” and the “citrate cycle (TCA cycle)” (Figure [139]5a,b). FIGURE 5. FIGURE 5 [140]Open in a new tab Metabolic network diagram of differentially abundant metabolites (DAMs) in the “fls2_Drought versus WT_Drought” comparison group. (a) Metabolic network diagrams of DAMs in the “fls2_Drought versus WT_Drought” comparison group. Solid lines indicate one‐step reactions; dashed lines indicate two‐or more‐step reactions. (b) The impact of DAMs in the “fls2_Drought versus WT_Drought” comparison on different metabolic pathways. All matching pathways for metabolites were determined based on p‐values obtained from a pathway enrichment analysis and pathway impact values derived from a pathway topology analysis. Analysis of the “fls2_CK versus WT_CK” and “fls2_Drought versus WT_Drought” comparison groups indicated that the number of amino acids with increased abundance was lower in fls2_Drought plants, while the number of flavonoid metabolites and carbohydrates with higher abundance increased in the “fls2_Drought versus WT_Drought” comparison group (Figure [141]6a,b), including L‐lyxonate and epicatechin. Additionally, the number of amino acids and carbohydrates with increased abundance was lower in the “fls2_Drought versus fls2_CK” comparison group than it was in the “WT_Drought versus WT_CK” comparison group (Figure [142]6b). Unlike fls2_NaCl samples, no increase in the accumulation of indoleacetaldehyde was observed in fls2_Drought samples (Figures [143]2a and [144]6a). An association analysis of metabolome and transcriptome data was performed for the DAMs of D‐aspartic acid (DAA), L‐glutamic acid, D‐glucose 1‐phosphate, maltose, epicatechin, guanine, and uracil in “fls2_Drought versus fls2_CK,” “fls2_CK versus WT_CK,” “fls2_Drought versus WT_Drought,” and “WT_Drought versus WT_CK” sample groups. Results indicated that the expression of genes related to most of the DAMs showed similar trends to their corresponding DAMs (Figure [145]6a,c). For example, the level of the DAA‐related gene ASPARTATE AMINOTRANSFERASE 2 (ASP2) was higher in fls2_Drought and WT_CK samples than in fls2_CK and WT_Drought samples (Figure [146]6c). Furthermore, the expression of genes related to maltose synthesis was significantly upregulated in fls2_Drought samples (Figure [147]6c,d). The expression of genes related to epicatechin, D‐glucose 1‐phosphate, and DAA, such as ASPARTATE AMINOTRANSFERASE 1 (ASP1) and PHOSPHOGLUCOMUTASE 2 (PGM2), were also upregulated in WT_CK samples (Figure [148]6c). 3.5. rbohd/f double mutant exhibits enhanced sensitivity to drought stress Electron transfer during photorespiration is inhibited when plants are subjected to drought stress, resulting in the production of a large amount of ROS (Noctor et al., [149]2014). Therefore, we performed an analysis of the metabolome and transcriptome of rbohd/f double mutant and WT plants under CK and drought conditions to examine the role of RBOHD/F‐mediated ROS signaling in plant response to drought stress. A total of 37 DAMs were identified in the “rbohd/f_Drought versus WT_Drought” comparison group that were assigned to 23 metabolic pathways known to be involved in plant development. Most of the DAMs with increased abundance in rbohd/f_Drought samples are involved in “arginine biosynthesis,” “phenylpropanoid biosynthesis,” and as well as other pathways. The major metabolic pathways identified are related to response to stress and removal of ROS in plants (Amarowicz & Weidner, [150]2009) (Figure [151]S4a,b). The analysis of “rbohd/f_Drought versus rbohd/f_CK” and “WT_Drought versus WT_CK” comparison groups indicated that the number of amino acids, carbohydrates, and flavonoids with increased abundance significantly increased in rbohd/f_Drought samples (Figure [152]S5b). In contrast, the number of purines and pyrimidines with increased abundance were observed to decrease in rbohd/f_Drought samples (Figure [153]S5a,b). These results indicate that the abundance of a great number of metabolites in the rbohd/f double mutant changes under drought stress conditions. Additionally, the analysis of the “rbohd/f_CK versus WT_CK” and “rbohd/f_Drought versus WT_Drought” comparison groups indicated that the number of amino acids, flavonoids, and carbohydrate metabolites with increased abundance was significantly higher in rbohd/f_Drought samples (Figure [154]S5b). An association analysis of metabolome and transcriptome data for DAMs was conducted that include the metabolites L‐proline, L‐glutamic acid, sucrose, indoleacetic acid, AMP, and UDP‐D‐galactose. As shown in Figure [155]S5c, results indicated that the expression of genes related to the synthesis of L‐glutamic acid, sucrose, L‐proline, and indoleacetic acid was higher in rbohd/f samples than in WT samples under drought stress conditions. For example, the level of the indoleacetic acid‐related genes NITRILASE 2 (NIT2) and SULFOTRANSFERASE 16 (SOT16) were significantly higher in rbohd/f_CK and rbohd/f_Drought samples. The level of expression of genes related to AMP and UDP‐D‐galactose synthesis was higher in WT_CK samples than in rbohd/f_CK samples (Figure [156]S5a,c). 3.6. Expression analysis of antioxidant defense‐related genes in fls2 and rbohd/f mutants The activity and abundance of several antioxidant proteins in plants are enhanced when they are exposed to abiotic stress to provide protection from oxidative injury to cells (Gill & Tuteja, [157]2010). Thus, FLS2 and RBOHD/F‐mediated ROS signaling are a critical aspect of plant response to abiotic stress as demonstrated by the increased sensitivity of the Arabidopsis rbohd/f double mutant to abiotic stresses (Liu et al., [158]2022). Plants have evolved a complex enzymatic and non‐enzymatic antioxidant system to maintain the homeostasis of their intracellular redox state (Nadarajah, [159]2020). Protective enzymes include SUPEROXIDE DISMUTASE (SOD), CATALASE (CAT), ASCORBATE PEROXIDASE (APX), and ALTERNATIVE OXIDASE (AOX) (Mittler et al., [160]2004). The non‐enzymatic antioxidant system consists of various reducing substances such as ascorbic acid (AsA), vitamin E, cytochrome f, and anthocyanins (Noctor, [161]2005). We also analyzed the expression level and GO enrichment of genes in the fls2 mutant, rbohd/f double mutant, and WT samples that were related to antioxidant defense metabolites (Figure [162]S6a,b). The level of many genes related to antioxidant defense in rbohd/f double mutant samples was mostly higher than they were in WT under both stress and CK conditions. For example, the level of expression of GLUTATHIONE S‐TRANSFERASE F3 (GSTF3) and MONODEHYDROASCORBATE REDUCTASE 1 (MDAR1) was higher in rbohd/f_Drought and rbohd/f_NaCl samples than in WT_Drought and WT_NaCl samples (Figure [163]S6a). Interestingly, the level of expression of genes related to antioxidative stress in the fls2 mutant did exhibit significant differences from the WT under either NaCl or drought conditions (Figure [164]S6a). 3.7. Fls2‐ RBOHD module co‐regulates the expression of key metabolites We further analyzed the expression of identical metabolites and related synthetic genes produced in WT, fls2, and rbohd/f mutants under drought and salt stress. The results showed that DAA and L‐lyxonate were accumulated in WT, fls2, and rbohd/f double mutants under drought conditions. The expressions of ASP2, DEHYDROASCORBATE REDUCTASE 1 (DHAR1), and ASCORBATE PEROXIDASE 3 (APX3) involved in the synthesis of DAA and L‐lyxonate were also increased correspondingly (Figure [165]7a). Under salt stress, the contents of L‐proline, sucrose, D‐ribose, and indolealdehyde were increased in WT, fls2, and rbohd/f double mutants, and the expressions of related genes, such as P5CR, sucrose synthetase 4 (SUS4), AT4G34880, and 6‐phosphogluconate dehydrogenase 1 (PGD1), were also increased (Figure [166]7b). FIGURE 7. FIGURE 7 [167]Open in a new tab Analysis of the expression of genes related to common differentially abundant metabolites in WT, fls2 mutant, and rbohd/f double mutants under drought and salt stress conditions. (a) Heat map of metabolite‐related gene expression in fls2_Drought/CK, rbohd/f_Drought/CK, and WT_Drought/CK samples. (b) Heat map of metabolite‐related gene expression in fls2_NaCl/CK, rbohd/f_NaCl/CK, and WT_NaCl/CK samples. 4. DISCUSSION 4.1. FLS2 regulates the response of plants to drought and salt stress by modulating metabolite levels in plants Plants are inevitably affected by abiotic stresses such as drought, salinity, or high and low temperatures, which greatly limits their growth and development (Nadarajah, [168]2020). FLS2 is a kinase that functions as a receptor of flg22, a conserved 22 amino acid peptide of plant–bacterial pathogens (Zipfel et al., [169]2004). Our previous study revealed the important role that FLS2 also plays in regulating the response of plants to abiotic stresses (Liu et al., [170]2022). In the present study, we further examined the potential effect of FLS2 on the accumulation of metabolites in plants subjected to salt and drought stress. Our results revealed an increased abundance of amino acids and carbohydrates, such as L‐proline, L‐isoleucine, D‐ribose, and others in the “fls2_NaCl versus WT_NaCl” comparison group (Figure [171]2a,b). L‐proline contributes to maintaining the homeostasis of the intracellular redox environment and intracellular ROS content, both of which have been shown to be associated with abiotic stress tolerance (Krishnan et al., [172]2008; Szabados & Savouré, [173]2010). Carbohydrates are important carbon energy reserves and are also involved in intracellular redox homeostasis (Poltronieri et al., [174]2011). The substantial accumulation of metabolites in fls2_NaCl mutant plants is due to the sensitivity of fls2 mutant to salt stress. FLS2 is a receptor kinase, which suggests that it may be involved in the perception of salt stress, a premise indicated in our previous study (Liu et al., [175]2022). The sensitivity of fls2 mutant to salt stress may be due to the defective receptor kinases in fls2 mutant that sense abiotic stress. Notably, in the analysis of the “fls2_Drought versus fls2_CK” comparison group, no significant accumulation of amino acids or carbohydrates was observed in fls2_Drought samples (Figure [176]6a,b). Based on these data, we speculate that this may indicate that fls2 mutant does not have a strong sensitivity to drought stress but rather are likely to be tolerant to drought conditions. 4.2. rbohd/f double mutant exhibits hypersensitive changes in the level of metabolites under drought and salt stress conditions RESPIRATORY BURST OXIDASE HOMOLOG D functions as an essential regulator of ROS production in plants. In plants, the activation of RBOHD activity in response to salt stress results in the accumulation of ROS; thus, RBOHD participates in plant response to salt stress (Liu et al., [177]2022; Luo et al., [178]2021). The abundance of metabolites in the rbohd/f double mutant exhibited a high level of sensitivity to salt stress conditions and LC–MS analysis revealed that the abundance of DAMs such as L‐proline, sucrose, epicatechin, indoleacetic acid, and others, increased in rbohd/f_NaCl samples, relative to rbohd/f_CK samples (Figure [179]4a). Indoleacetic acid has been reported to regulate root development in response to many abiotic stresses, including salt stress (Korver et al., [180]2018). Our results indicated that the abundance of DAMs such as dAMP and UDP‐D‐galactose and others decreased in rbohd/f_NaCl samples, relative to WT_NaCl samples (Figure [181]4a,b). In this regard, ROS accumulation has been reported to have a negative impact on sugar and base moieties and results in oxidative damage to DNA (Boiteux et al., [182]2017). Consequently, this may lead to a decrease in the abundance of metabolites such as dAMP under salt stress conditions. Our results suggest that ROS signaling mediated by RBOHD/F induces dynamic changes in metabolite production in plants under salt stress. In the analysis of the “rbohd/f_Drought versus rbohd/f_CK” and “WT_Drought versus WT_CK” comparison groups, the abundance of DAMs such as indoleacetic acid, L‐glutamic acid, L‐aspartic acid, sucrose, and others was significantly greater in rbohd/f_Drought samples than in the other samples (Figures [183]S4a and [184]S5a,b). Extensive amino acid accumulation has often been observed in plants growing under abiotic stress conditions, including maize, cotton, Arabidopsis thaliana, and others (Huang & Jander, [185]2017; Ranieri et al., [186]1989; Showler, [187]2002). Additionally, the accumulation of indoleacetic acid is known to have a significant effect on plant growth and stress tolerance (Yemelyanov et al., [188]2020). In the analysis of the “rbohd/f_CK versus WT_CK” comparison group, the number of carbohydrates that decreased in abundance was high (Figure [189]4b), which may indicate that the absence of RBOHD/F under CK conditions results in a decrease in the abundance of metabolites related to synthetic energy. The changes in the number of DAMs in rbohd/f_Drought samples under drought stress conditions indicate that the rbohd/f double mutant has an increased sensitivity to oxidative stress response, relative to rbohd/f_CK samples. These initial results indicate that the potential mechanism by which FLS2 and RBOHD participate in plant response to drought and salt stress is by regulating changes in the abundance of metabolites. 4.3. Combined metabolome and transcriptome analysis of the effect of FLS2 and RBOHD on regulating the abundance of stress‐induced metabolites and the expression of associated genes We conducted a combined metabolome and transcriptome analysis to examine the relationship between DAMs and the expression of their related genes in fls2 and rbohd/f double mutants under drought and NaCl stress conditions (Figures [190]2c, [191]4c, and [192]6c; Figure [193]S5c). The analysis revealed that many genes, including GSTF2, ASN2, YUCCA 8 (YUC8), GLUCOSIDE GLUCOHYDROLASE 2 (TGG2), and others, that were related to the identified DAMs were significantly upregulated in fls2_NaCl samples, relative to WT_NaCl samples, suggesting that the fls2 mutant was more sensitive to NaCl than WT plants (Figure [194]2c). Glutathione S‐transferase has been reported plays an important role in maintaining redox homeostasis and reducing oxidative damage (Chen et al., [195]2012), and GSTF2 is also closely related to the regulation of oxidative stress response in Arabidopsis (Lee et al., [196]2014). ASN2 has been shown to play an important role in the response of Arabidopsis to salt stress (Maaroufi‐Dguimi et al., [197]2011). TGG2 can decrease the accumulation of ROS in Arabidopsis through its antioxidant properties (Zhao et al., [198]2015). The upregulation of these genes, however, may indicate the increased sensitivity of fls2 mutant to NaCl (Figure [199]2c). Although maltose‐related genes exhibited increased expression in fls2_Drought samples (Figure [200]6c), other genes such as GSTF7 were also highly expressed in WT_CK samples, and thus, did not exhibit a significant increase in fls2_Drought samples (Figure [201]6b). This is consistent with the metabolome data for fls2 mutant indicating that they exhibited tolerance to drought stress (Figure [202]6b). The expression level of genes such as PYRROLINE‐5‐CARBOXYLATE REDUCTASE (P5CR), flavonol synthase 1 (FLS1), NIT2, and others that are related to the synthesis of L‐proline, epicatechin, and indoleacetic acid was higher in rbohd/f_NaCl and rbohd/f_Drought samples than in WT_NaCl and WT_Drought samples (Figure [203]4c; Figure [204]S5c). These results are consistent with the metabolome data (Figure [205]4a; Figure [206]S5a). P5CR is a key enzyme involved in proline biosynthesis and is associated with enhanced drought tolerance in Arabidopsis (Chen et al., [207]2021). FLS2 also plays a key role in plant tolerance to abiotic stress by controlling flavonol accumulation (Zhang et al., [208]2020). Our results suggest that the rbohd/f double mutant is more sensitive to drought and salt stress, and that both RBOHD and FLS2 are required for plant response and adaptation to abiotic stress. The expression of genes related to antioxidant defense, such as CATALASE 2 (CAT2), GLUTATHIONE PEROXIDASE 1 (GPX1), CAROTENOID CLEAVAGE DIOXYGENASE 7 (CCD7), and others, was greater in rbohd/f double mutant than in WT plants (Figure [209]S6). CAT2 has been reported to play an important role in ROS scavenging in plants under abiotic stress conditions (Ono et al., [210]2021) and GPX1 is known to be involved in the detoxification of H[2]O[2] (Avsian‐Kretchmer et al., [211]2004). These genes, however, did not exhibit a specific increase in expression in the fls2 mutant (Figure [212]S6). These data indicate that both RBOHD and FLS2 are essential for regulating gene expression related to antioxidant defense under drought and salt stress conditions. The present study identified common metabolites and genes regulated by FLS2 and RBOHD under drought and salt stress conditions (Figure [213]7). DAA and L‐lyxonate accumulated in both fls2 mutant and rbohd/f double mutant under drought conditions, and L‐proline, sucrose, D‐ribose, and indoleacetaldehyde increased in both fls2 and rbohd/f double mutants under salt conditions, and the expression of their related genes, including P5CR, SUS4, AT4G34880, PGD1, and others, also increased (Figure [214]7b). Notably, SUS4 has been reported to be strongly upregulated in wheat in response to drought stress and other abiotic stresses (Wang et al., [215]2019). Changes in ROS and Ca^2+ signaling induced by FLS2 and RBOHD result in the production of the same metabolites and enhancement in the expression of their related genes in response to drought and salt stress conditions. The results of our study provide new insights into plant metabolic processes involved in ROS and Ca^2+ signaling in response to exposure to abiotic stress. Both biotic and abiotic stresses can lead to the accumulation of intracellular ROS. Thus, plants respond to both types of stresses by regulating changes in intracellular ROS (Fichman & Mittler, [216]2020). FLS2 and RBOHD are closely related to the ability of RBOHD to induce ROS and Ca^2+ signaling, which play an essential role in regulating plant response to abiotic stresses (Noirot et al., [217]2014). In conclusion, our study revealed the potential role of FLS2 and RBOHD in regulating the abundance of DAMs in response to drought and salt stress conditions and the expression of DAM‐related genes. We provide evidence that the metabolites and genes controlled by both FLS2 and RBOHD regulate ROS and Ca^2+ signaling in plants subjected to abiotic stress. Our results reveal the potential role of FLS2 and RBOHD in the regulation of the response of a higher plants to abiotic stresses and also provide new strategies for the combined analysis of metabolomic and transcriptomic data (Figure [218]8). FIGURE 8. FIGURE 8 [219]Open in a new tab Proposed model of FLAGELLIN SENSITIVE 2 (FLS2) and RESPIRATORY BURST OXIDASE HOMOLOG D (RBOHD) signaling in response to abiotic stress conditions. FLS2 and RBOHD regulate the synthesis of differentially abundant metabolitesthrough reactive oxygen species signaling when plants are subjected to drought and salt stress, which includes the participation of related enzymes. RBOHD also regulates the synthesis of antioxidant enzymes in plants and the expression of their related genes, enabling plants to respond to abiotic stresses. CONFLICT OF INTEREST STATEMENT The authors declare no conflict of interest. Supporting information Figure S1. [220]Click here for additional data file.^ (2.5MB, docx) Table S1. [221]Click here for additional data file.^ (609.5KB, xlsx) ACKNOWLEDGMENTS