Research ResourcePlant biology

Genetics and Phosphoproteomics Reveal a Protein Phosphorylation Network in the Abscisic Acid Signaling Pathway in Arabidopsis thaliana

See allHide authors and affiliations

Science Signaling  09 Apr 2013:
Vol. 6, Issue 270, pp. rs8
DOI: 10.1126/scisignal.2003509


Abscisic acid (ABA) is a phytohormone that regulates diverse plant processes, including seed germination and the response to dehydration. In Arabidopsis thaliana, protein kinases of the SNF1-related protein kinase 2 (SnRK2) family are believed to transmit ABA- or dehydration-induced signals through phosphorylation of downstream substrates. By mass spectrometry, we identified proteins that were phosphorylated in Arabidopsis wild-type plants, but not in mutants lacking all three members of the SnRK2 family (srk2dei), treated with ABA or subjected to dehydration stress. The number of differentially phosphorylated peptides was greater in srk2dei plants treated with ABA than in the ones subjected to dehydration, suggesting that SnRK2 was mainly involved in ABA signaling rather than dehydration. We identified 35 peptides that were differentially phosphorylated in wild-type but not in srk2dei plants treated with ABA. Biochemical and genetic studies of candidate SnRK2-regulated phosphoproteins showed that SnRK2 promoted the ABA-induced activation of the mitogen-activated protein kinases AtMPK1 and AtMPK2; that SnRK2 mediated phosphorylation of Ser45 in a bZIP transcription factor, AREB1 (ABA-responsive element binding protein 1), and stimulated ABA-responsive gene expression; and that a previously unknown protein, SnRK2-substrate 1 (SNS1), was phosphorylated in vivo by ABA-activated SnRK2s. Reverse genetic analysis revealed that SNS1 inhibited ABA responses in Arabidopsis. Thus, by integrating genetics with phosphoproteomics, we identified multiple components of the ABA-responsive protein phosphorylation network.


Abscisic acid (ABA) is a phytohormone that plays an integral role in many aspects of the plant life cycle. For example, ABA is essential for properly regulating seed maturation and stomatal opening and closing, as well as abiotic and biotic stress responses (15). Therefore, understanding the regulatory system of ABA responses is critical for improving many agricultural issues related to stress tolerance and seed quality control. A simple ABA signaling model consists of the soluble ABA receptors, which are members of the pyrabactin resistance 1 (PYR1) and PYR1-like (PYL) proteins, also known as regulatory component of ABA receptor (RCAR) family and collectively referred to as PYR/PYL/RCAR, a subgroup of type 2C protein phosphatases (PP2Cs), and the SNF1-related protein kinase 2 (SnRK2) family (3, 69). In the absence of ABA, PP2Cs dephosphorylate and inactivate SnRK2s, and ABA-dependent responses are repressed (10, 11). When ABA concentrations increase in response to abiotic stresses or developmental cues, ABA binds to a receptor of the PYR/PYL/RCAR family, which inhibits PP2C activity, causing SnRK2 activation to initiate ABA signaling (10, 1214). This signaling model is supported by biochemical and structural studies that provided the molecular basis for the interactions involved in the formation of the PYR/PYL/RCAR-ABA-PP2C complex (15, 16).

The current model of ABA perception and signaling indicates the importance of protein phosphorylation mediated by SnRK2 (3, 7). SnRK2 is a plant-specific protein kinase family, and there are 10 members in Arabidopsis genome, which are classified into subclasses I to III (1719). Because SnRK2 members are activated by ABA or osmotic stress pathways (18, 2022), they might contribute to either or both of the ABA-dependent and ABA-independent pathways that mediate dehydration stress signaling (1, 23). ABA activates three members of subclass III: SRK2D (also known as SnRK2.2), SRK2E (also known as OST1 or SnRK2.6), and SRK2I (also known as SnRK2.3). A triple knockout mutant (srk2dei or snrk2.2/2.3/2.6) is largely ABA-insensitive (10, 2426), indicating that these kinases play a central role in ABA signaling. For example, srk2dei seeds germinate even in the presence of high concentrations of ABA, which normally inhibit germination (10, 24, 25), and srk2dei plants show a severe wilting phenotype due to fully opened stomata (24, 26).

Given the evidence that subclass III SnRK2s are involved in the regulation of ABA or osmotic stress responses and the biochemical evidence that these SnRK2s are active protein kinases (18, 19, 21), identifying the downstream targets (that is, phosphorylation substrates) of these kinases is critical for further understanding of ABA signal transduction (3, 5, 7). Several SnRK2 substrates have been identified or proposed. For example, a family of basic leucine zipper–type proteins, ABA-responsive element binding proteins [AREBs; also known as ABA-responsive element–binding factors (ABFs)], requires phosphorylation by SnRK2s to activate ABA-responsive gene expression (27, 28). In addition, SnRK2s can phosphorylate the anion channel SLAC1 to stimulate stomatal closure (29, 30), and other Arabidopsis proteins, such as the NADPH oxidase RbohF and the potassium channel KAT1, have been proposed as SnRK2 substrates (31, 32).

Phosphoproteomics enables the analysis of thousands of phosphoproteins in vivo (33, 34). Therefore, we used phosphoproteomic analyses to identify proteins phosphorylated in Arabidopsis thaliana in a SnRK2-dependent manner and to analyze protein phosphorylation networks involved in ABA signaling or dehydration stress signaling. We performed comparative analyses between wild-type and srk2dei triple mutant plants treated under different conditions to identify SnRK2-dependent phosphoproteins. This integration of genetics and phosphoproteomics identified SnRK2 substrate candidates, and in vitro or in planta analyses validated SnRK2-dependent regulation or the molecular functions of several candidates.


Overview of phosphoproteome analysis in Arabidopsis seedlings treated with ABA or dehydration stress

We applied a liquid chromatography–tandem mass spectrometry (LC-MS/MS) approach to identify the SnRK2-regulated phosphoproteome in plants either subjected to dehydration stress or treated with ABA (fig. S1). We confirmed the ABA insensitivity of srk2dei by demonstrating that srk2dei germinated and expanded normal cotyledons in the presence of 100 μM ABA (fig. S1). We also performed microarray analysis with the Affymetrix GeneChip ATH1 to collect data regarding ABA- or dehydration-responsive gene expression in wild-type plants and srk2dei (figs. S2 and S3 and table S1). Consistent with previous reports (25, 26), a majority of ABA-responsive changes (73.7%) and a part of dehydration-responsive changes (42.2%) in gene expression were impaired in srk2dei (fig. S3), suggesting that some other pathways could be involved.

We compared the phosphoproteome profiles from wild-type or srk2dei Arabidopsis seedlings treated with ABA (50 μM) or subjected to dehydration for 15, 30, and 90 min. Three independent biological replicates were performed for each treatment condition. After isolation of total protein and tryptic digestion, phosphopeptides were enriched by hydroxy acid–modified metal oxide chromatography and analyzed with a LTQ-Orbitrap LC-MS/MS instrument (35). MS and MS/MS spectra were assigned to specific peptide sequences by the MASCOT search engine. We detected 5288 unique phosphopeptides belonging to 2204 proteins (table S2). The majority of identified phosphopeptides contained either one or two phosphorylated sites (Fig. 1A), and the relative distribution of phosphorylated sites among serine, threonine, and tyrosine (Fig. 1B) is consistent with the distribution of phosphorylated residues reported in other plant studies (36, 37).

Fig. 1 An overview of the results of the phosphoproteomic analysis.

(A) Distribution of the number of phosphosites per peptide. (B) Distribution of phosphorylated residues in each peptide.

ABA- or dehydration stress–responsive phosphoproteome in Arabidopsis

To gain an overview of the extent of unique or conserved responses to exogenous ABA treatment or dehydration stress, we used label-free quantitation to compare responses in Col wild-type seedlings. We quantified the relative abundance of each phosphopeptide using the peak area of the extracted ion current chromatogram derived from MS analysis (table S2). We also analyzed the same samples using proteolytic isotope labeling with 16O/18O and quantified phosphopeptides. Because this labeling method was less efficient and provided less coverage of the proteome than the label-free method, these data only partially support the results obtained by the label-free quantitation. After normalization, phosphopeptides that showed a more than threefold increase or a decrease to less than one-third of the value in the wild-type samples in response to either ABA treatment or dehydration stress in at least two of three replicates were assigned to the “up-regulation” or “down-regulation” group, respectively. Fifty-three peptides were found to be up-regulated by ABA, and about half of these ABA-responsive peptides (27 of 53) were also regulated by dehydration (Fig. 2A and table S3). Dehydration stress resulted in the up-regulation of 125 phosphopeptides that were not altered in response to ABA. This result indicates that dehydration induces a large number of ABA-independent changes in protein phosphorylation. The same trend was observed in the down-regulated groups (Fig. 2B and table S3). These experiments indicated that short-term phosphoproteomic changes to ABA and dehydration stress are mostly distinct.

Fig. 2 Overlap of phosphopeptides regulated by ABA treatment and dehydration stress.

(A) Venn diagram showing the number of phosphopeptides that met the criteria for up-regulation in Col wild-type (WT) plants exposed to ABA or dehydration (DH). (B) Venn diagram showing the number of phosphopeptides that met the criteria for down-regulation in Col (WT) plants exposed to ABA or DH. (C) An example of comparative analysis with the phosphoproteomic data. Graphs show the extracted ion current chromatogram data of a phosphopeptide, [SpTVGTPAYIAPEVLLR], in WT and srk2dei Arabidopsis seedlings. This peptide corresponds to an activation loop of SRK2I (also known as SnRK2.3). N.D., not detectable.

Comparative analysis of phosphoproteome profiles to identify SnRK2 substrates

SnRK2s are protein kinases that are activated in response to ABA or dehydration (18, 19, 38, 39). Therefore, we hypothesized that phosphorylation of substrates downstream of SnRK2 in response to ABA or dehydration should be strongly reduced or absent in srk2dei. We performed a comparative analysis between Col wild-type plants and srk2dei plants to identify proteins that were differentially phosphorylated after stress or hormone treatment. We screened phosphopeptides with the following criteria: (i) an increase in phosphorylation in response to ABA or dehydration (a more than threefold change in wild-type plants untreated compared to treated), (ii) a decrease in phosphorylation to less than 1/5-fold of the value occurred in the srk2dei mutant compared to wild-type, and (iii) the same trend in phosphorylation status occurred in at least two of the three biological replicates. The first criterion represented the 178 peptides identified as up-regulated by ABA or dehydration in wild type (Fig. 2A), and thus, those peptides were used for further analysis with second and third criteria. Thirty-five phosphopeptides met these criteria and were considered dependent on SnRK2 (Table 1) and these corresponded to proteins in many functional classes, including transcription factors, RNA binding proteins, and enzymes. Three of the phosphopeptides corresponded to the activation loop of subclass III SnRK2s and were eliminated as potential substrates; thus, 32 phosphopeptides were final candidates for SnRK2 substrates. The phosphorylated sites identified in SnRK2 were consistent with a previous report (10), suggesting that this region in SnRK2 was phosphorylated in response to ABA and dehydration (Fig. 2C). Phosphopeptides corresponding to the activation loop of SRK2D and SRK2I were not detected in any srk2dei samples (Fig. 2C), confirming that the mutant does not have kinase activity (10).

Table 1 List of phosphopeptides down-regulated in the srk2dei mutant.

Accession number represents the proteins in which the peptide occurs. Protein descriptions were annotated from MASCOT. Peptide(seq+mod) shows the amino acid sequence and modification of each peptide (p, phosphorylation; o, oxidation; c, carbamidomethylation). Abbreviations for the amino acids are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr. In the Motif column, each number represents motif patterns around phosphorylation site: Motif 1, [-(K/R)-x-x-(pS/pT)-]; Motif 2, [-(pS/pT)-x-x-x-x-(D/E)-]; Motif 3, [-(pS/pT)-P-]; and Motif 4, [all others]. In the Stress column, “A” and “D” mean that each peptide was up-regulated in wild-type plants by ABA treatment and dehydration stress, respectively.

View this table:

Motif analysis of ABA- and dehydration stress–responsive phosphopeptides

We next performed a phosphorylation motif analysis of the differentially phosphorylated proteins using the Motif-X algorithm (40) to gain insights into the possible number and variety of kinases involved in the responses based on the assumption that different kinases will have different preferences toward primary sequence recognition. Using the data set from Col wild type (Fig. 2, A and B), we first extracted ABA- or dehydration-responsive motifs. Seven motifs were detected in the up-regulation group ([-R-S-x-pS-], [-pS-P-], [-S-x-x-pS-], [-S-x-x-x-pS-], [-K-x-x-pS-], [-R-x-x-pS-], and [-pS-x-x-x-x-E-]) (fig. S4). In the down-regulated group, Motif-X detected three motifs ([-pS-P-], [-pS-x-x-E-], and [-pS-F-]). The minimal mitogen-activated protein kinase (MAPK) target motif, [-(pS/pT)-P-], was detected in both groups, possibly indicating the inverse regulation of different MAPK family members (41, 42). The three motifs [-R-S-x-pS-], [-K-x-x-pS-], and [-R-x-x-pS-] can be assigned to the more general motif, [-(K/R)-x-x-(pS/pT)-], which is a recognition motif for SnRK2 or calcium-dependent protein kinase (CDPK) in plants (27, 28, 43, 44). The motifs [-S-x-x-pS-] and [-S-x-x-x-pS-] represent phosphorylation motifs targeted by casein kinase II or glycogen synthase kinase-3 (GSK-3), respectively (42). This motif analysis suggested that both direct targets of SnRK2s, as well as downstream targets of MAPKs and GSKs, are represented in this subgroup of the phosphoproteomic data.

We extended the motif analysis using the quantitative data of phosphoproteins in srk2dei to determine which motifs were dependent on SnRK2s. On the basis of the Motif-X results, motifs were consolidated into groups: Motif 1, [-(K/R)-x-x-(pS/pT)-]; Motif 2, [-(pS/pT)-x-x-x-x-(D/E)-]; Motif 3, [-(pS/pT)-P-]; and Motif 4, [all others]. Phosphopeptides of the up-regulated group in the wild-type plants (Col) were classified to each motif group. The quantitative data of phosphopeptides from srk2dei and Col were sorted according to motif groups, and the average fold change for each group was displayed as a heat map (Fig. 3). The heat map showed that phosphopeptides from all motif groups were increased in response to ABA and dehydration in wild-type seedlings. However, the motif groups showed distinct categories for SnRK2 dependence in ABA responses or dehydration stress responses. All motif groups were SnRK2-dependent (decreased) during ABA treatment, with group 4 motifs showing the least reduction. The dehydration stress responses showed little to no SnRK2 dependence for each motif group; however, there were clearly some phosphopeptides within each motif group that were SnRK2-dependent during dehydration stress (Table 1). This analysis indicated that in plants subjected to dehydration, there were a large number of changes in phosphopeptides containing these motifs that did not require SnRK2 for their phosphorylation. Therefore, as a whole, the phosphorylation motif analysis defined a subset of SnRK2-dependent phosphorylation groups regulated in response to ABA. Because Motifs 1, 2, and 3 showed the strongest dependence on SnRK2, we focused on candidates from these groups for functional analysis of SnRK2 downstream factors.

Fig. 3 Motif analysis of phosphopeptides in Col (WT) and srk2dei.

Sequences of 13 amino acids (−6 to +6) in which phosphorylated residue is the center (0) were extracted from phosphopeptides meeting the criteria of up-regulation by ABA or dehydration in Col (WT). The conserved residues around the phosphorylated Ser are shown. Dots represent nonconserved residues. The percentages of phosphopeptides of the total in the ABA- and dehydration-responsive groups are shown. The heat map represents relative difference in the abundance of the phosphorylated peptides, with up-regulation shown in yellow and down-regulation shown in blue in response to ABA or dehydration in Col (WT) and srk2dei.

MAPK pathway(s) involved in SnRK2-mediated ABA signaling

Motif analysis revealed that phosphorylation of peptides containing [-(pS/pT)-P-] (Motif 3) was reduced in srk2dei plants in response to ABA (Fig. 3), and the abundance of some phosphopeptides containing this motif was similarly reduced during dehydration stress (Table 1). Because this motif is a known MAPK target motif, we hypothesized that SnRK2 plays a role in ABA- and dehydration-induced activation of MAPK pathway(s). Indeed, a peptide from AtMPK1 or AtMPK2 containing the canonical TEY motif found in the activation loops of MAPKs was phosphorylated in response to ABA and dehydration (Table 1 and Fig. 4, A and B). The sequence of this peptide is identical to both AtMPK1 and AtMPK2 (Fig. 4C), which prevented exact assignment of this peptide to one protein or the other, and thus we refer to these together as AtMPK1/2. Microarray analysis indicated that the transcript abundance of both AtMPK1 and AtMPK2 was not affected by ABA treatment or dehydration stress (Fig. 4, D and E). ABA-responsive phosphorylation of AtMPK1/2 in srk2dei was reduced (Fig. 4B), therefore indicating that subclass III SnRK2 is required for AtMPK1/2 activation. We tested AtMPK1 enriched from Arabidopsis seedlings by immunoprecipitation with an antibody recognizing AtMPK1 in an in vitro phosphorylation assay. In general, we found that ABA-responsive activation of AtMPK1 was consistently reduced in srk2dei compared to wild type, but that dehydration-responsive activation was more variable (Fig. 4F and fig. S5). These results suggest that subclass III SnRK2s stimulate MAPK cascade(s) involving AtMPK1/AtMPK2 in response to ABA.

Fig. 4 Phosphorylation of AtMPK1 and AtMPK2 in response to ABA or dehydration is reduced in srk2dei.

(A) MS/MS spectrum of a phosphopeptide derived from the activation loop of AtMPK1/AtMPK2. (B) Label-free nano-LC-MS/MS quantitation of GQFMpTEpYVVTR abundance in Col wild-type (empty) and srk2dei (solid) plants treated with 50 μM ABA or dehydration for the times indicated. Extracted ion chromatogram from MS analysis was used for quantitation of the phosphopeptide. (C) Sequence alignment of the Arabidopsis MAPK family, organized by subgroup (A, B, C, and D). The residues in AtMPK1 and AtMPK2 corresponding to the phosphorylated activation loop peptide are outlined in yellow. Phosphorylated residues are indicated by asterisks and colored green. (D) Relative transcript abundance of AtMPK1 in response to 50 μM ABA or dehydration stress for 90 min. (E) Gene expression pattern of AtMPK2 in response to 50 μM ABA or dehydration stress. Data in (D) and (E) are means ± SE (n = 3) of fold change relative to control (Col) from microarray analysis of Col (empty) and srk2dei (solid). (F) In vitro kinase assay of AtMPK1 immunoprecipitated from Col and srk2dei (dei) seedlings exposed to 50 μM ABA or dehydration for indicated times. The substrate MBP was phosphorylated in the presence of 32P. Arrows indicate MBP labeled with 32P (solid) or stained with Coomassie brilliant blue (empty) to show the loading of MBP. Phosphorylation of MBP was detected by autoradiography, quantified by gel densitometry, and normalized to the time 0 Col control. Data are representative of four experiments (see fig. S5 for all other experiments).

In vivo phosphorylation of AREB-type transcription factors by SnRK2

The AREB family of bZIP transcription factors promotes a wide range of ABA-responsive gene expression by binding to ABA-responsive promoter elements (1). Furihata et al. (28) predicted multiple [-R-x-x-S-] phosphorylation sites in four conserved regions (boxes I to IV) of AREBs, and mutagenesis of these predicted phosphorylation sites indicated that AREBs may be regulated by phosphorylation. A phosphoproteomic study confirmed that some of [-R-x-x-S-] motifs in AREBs were phosphorylated in vivo (45), and it was suggested that SnRK2 can phosphorylate [-R-x-x-S-] motifs in vitro (27, 28).

Here, we did not identify any phosphorylated [-R-x-x-S-] motifs in AREBs in either the ABA- or dehydration-responsive groups, which may reflect selective coverage of phosphopeptides dependent on the technology used. Instead, we detected a phosphopeptide, 42DFGpSMNMDELLK53, of AREB1 that was phosphorylated in response to both ABA and dehydration (Table 1 and Fig. 5, A to C). This site is in the context of Motif 2, [-(pS/pT)-x-x-x-x-(D/E)-], and is conserved in the AREB family (Fig. 5B). Furthermore, our analysis revealed that Ser45 phosphorylation was reduced in the srk2dei mutant in response to both ABA and dehydration (Fig. 5C). This label-free result was cross-supported by 18O labeling analysis (Fig. 5D), indicating that Ser45 of AREB1 is an in vivo target of SnRK2-dependent phosphorylation.

Fig. 5 Ser45 of AREB1 is phosphorylated by subclass III SnRK2s.

(A) MS/MS spectrum of a phosphopeptide matching a conserved sequence in several AREB family members. (B) Sequence alignment of the box I region, one of four conserved boxes in Arabidopsis AREB family. The sequence corresponding to the phosphorylated peptide is outlined by the yellow box. Phosphorylated residues are green and indicated with asterisks. Alternative names for the proteins are shown with slashes. The two phosphorylation motifs in this region conform to the following motifs: [-R-x-x-pS-] and [-pS-x-x-x-x-E-]. (C) Label-free quantitation of DFGpSMNMDELLK. Col (empty) and srk2dei (solid) were treated with 50 μM ABA or dehydration for 0, 15, 30, and 90 min, and then analyzed by nano-LC-MS. Extracted ion chromatogram was used for quantitation of the phosphopeptide. (D) Quantitation of DFGpSMNMDELLK by proteolytic isotope labeling. Two samples labeled with 16O/18O were analyzed simultaneously for each run of nano-LC-MS. In this case, ABA-treated or dehydration-stressed samples were labeled by 18O, and then analyzed one by one together with control sample (16O). (E) Transactivation assay of AREB1 in Arabidopsis mesophyll protoplasts. Protoplasts expressing a vector control, AREB1WT, AREB1S45A, or AREB1S45D were analyzed for transactivation of RD29B-GUS with (solid) or without (empty) 50 μM ABA. Data are means ± SE (n = 3).

We investigated the biological importance of Ser45 phosphorylation by transactivation assay in Arabidopsis protoplasts. We prepared green fluorescent protein (GFP) fusion constructs of AREB1 with or without Ala (S45A) (phosphorylation-deficient) or Asp (S45D) (phosphorylation-mimetic) substitutions, transiently expressed these constructs in Arabidopsis mesophyll protoplasts, and monitored transcriptional activity by GUS expression driven by tandem ABA-responsive elements (ABREs), a well-defined cis-element of AREBs (1). Consistent with previous reports, expression of wild-type AREB1 enhanced ABRE-driven GUS activity in response to ABA compared to vector control (Fig. 5E). In contrast, expression of AREB1 (S45A) failed to promote ABA-induced GUS activity above that induced in the vector control. The phosphorylation-mimetic AREB1 (S45D) also exhibited enhanced GUS activity in response to ABA treatment compared to the ABA-treated vector control and also exhibited an increase in basal GUS activity (Fig. 5E). These data do not take into account potential differences in the stability of the expressed proteins or other potential differences in posttranslational modifications; however, the results suggest that phosphorylation of Ser45 contributes to the regulation of AREB1 activity by ABA.

Identification and characterization of SnRK2-substrate 1

Among the SnRK2-dependent phosphopeptides that accumulated in response to both ABA and dehydration treatments was a peptide belonging to At1g26470, which we named SnRK2-substrate 1 (SNS1) (Fig. 6A). This is a relatively small protein (133 amino acids) of unknown function and is conserved among higher plants (Fig. 6B). Label-free quantitation indicated that this protein was phosphorylated in response to both ABA and dehydration, and its ABA-dependent phosphorylation was abolished and its dehydration-induced phosphorylation was reduced in srk2dei (Fig. 6C). In an in vitro kinase assay with a synthetic peptide, containing the phosphorylation site from At1g26470, and SRK2E immunoprecipitated from Arabidopsis cells, the wild-type peptide from At1g26470 was phosphorylated, whereas a peptide containing an alanine substitution at Ser43 was not (Fig. 6D). Thus, SnRK2s can phosphorylate Ser43 of SNS1.

Fig. 6 SNS1 is phosphorylated by SnRK2 in vivo and in vitro.

(A) Mass spectrum matching a phosphopeptide from SNS1. (B) Sequence alignment of full-length SNS1 protein and its orthologs in castor bean, poplar, rice, maize, and sorghum. The residues matching the phosphorylated peptide are outlined with a yellow box. Phosphorylation motif is shown under the sequence and includes the red and blue residues. The asterisk indicates Ser43 of SNS1, and dotted line indicates the region annotated as Eaf7 superfamily. (C) Label-free quantitation of EQpSQVELELR. Col (empty) and srk2dei (solid) were treated with 50 μM ABA or dehydration for 0, 15, 30, and 90 min, and then analyzed by nano-LC-MS. Extracted ion chromatogram was used for quantitation of the phosphopeptide. (D) In vitro phosphorylation of SNS1 peptides by SRK2E (also known as OST1). Synthetic peptides SNS1 residues 38–49 (WT) and SNS1 residues 38–49 with an S43A mutation (S43A) were used as substrates of SRK2E-GFP purified from cultured Arabidopsis cells treated with (solid) or without (empty) 50 μM ABA. Phosphorylation of peptides was analyzed by measurement of 32P radioactivity incorporated. Data are means ± SE (n = 3). ND, not detected.

We also characterized the in planta functions of SNS1. Microarray analysis revealed that SNS1 was constitutively expressed in low amounts in Arabidopsis seedlings, both in the presence and in the absence of exogenous ABA (Fig. 7A). This is consistent with the expression data of SNS1 in the public eFP browser database in which SNS1 expression does not change in response to any phytohormones or abiotic or biotic stresses and shows ubiquitous expression from seeds to vegetative tissues (fig. S6). We isolated a SNS1-knockout mutant (sns1; SALK_008996) from the Arabidopsis Biological Resource Center for loss-of-function analyses. To assess the role of SNS1 in ABA responses, we tested the growth of the SNS1-knockout mutant in the presence or absence of exogenous ABA. In the absence of ABA, sns1 grew normally (Fig. 7B); however, 1 μM ABA severely inhibited the postgermination growth of sns1 (Fig. 7, C and D). This ABA-hypersensitive phenotype of sns1 was recovered by complementation of SNS1 gene (Fig. 7D). We assayed stress-responsive gene expression in wild-type and sns1 seedlings treated with ABA or exposed to salt stress by monitoring the expression of two ABA-responsive genes, Rab18 and RD29B, and one ABA- and salt-responsive gene, RD29A (1). There was a tendency for expression of Rab18 and RD29B, but not of RD29A, to be increased in sns1 in response to ABA, but no difference in Rab18 or RD29B expression was observed between wild type and sns1 in response to salt stress (Fig. 7E and fig. S7). Thus, our results suggest that SNS1 functions as a negative regulator of ABA signaling in Arabidopsis at the postgermination stage.

Fig. 7 Functional analysis of SNS1 in ABA signaling.

(A) Relative transcript abundance from microarray analysis of SNS1 in plants subjected to 50 μM ABA or dehydration stress. Data are means ± SE (n = 3), normalized to the WT control (Cont.). Col (WT) (empty); srk2dei (solid). (B) Col (WT) and transferred DNA knockout mutant (sns1) after 1 month from germination. (C) Effect of ABA on postgermination growth of Col (WT) and sns1. Seeds of the indicated genotype were germinated and grown in the presence of 1 μM ABA for 4 weeks. (D) Complementation test of sns1 mutant. Col (WT), sns1, and a complementation line, sns1/SNS1, were germinated and grown on GM agar plates containing the indicated concentrations of ABA for 2 weeks. (E) RNA gel blot analysis of ABA-responsive genes. Col and sns1 were treated with ABA or NaCl. Total RNA was subjected to RNA gel blot analysis with 32P-labeled complementary DNA (cDNA) probes of Rab18, RD29B, and RD29A. Ethidium bromide (EtBr) staining shows ribosomal RNA in each lane as an indication of loading. Expression of each gene was detected by autoradiography and quantified by gel densitometry as shown under the gel images. Data are representative of four experiments (see fig. S7 for all other experiments).


Protein phosphorylation and dephosphorylation play a central role in ABA signaling. The core component system for ABA perception consists of PYR/PYL/RCAR receptors, group A PP2Cs, and subclass III SnRK2s, which phosphorylate downstream factors to transmit the signal (3, 69). Other protein kinases, such as CDPKs, are also involved in ABA signaling (43, 44). Although several studies have described the ABA-responsive phosphoproteome in plants, identifying 50 (45) and 82 (46) phosphoproteins differentially phosphorylated in response to ABA, these analyses did not delineate the direct relationship between phosphorylation of the proteins and the predicted core receptor complex, nor did they investigate the extent of overlap between ABA- and dehydration-induced signaling. Previously, ABA-dependent or ABA-independent pathways in dehydration stress signaling had been analyzed mainly in terms of transcriptomics or metabolomics (1, 47). Here, we used a phosphoproteomic analysis that integrated genetic mutants to clearly identify components of SnRK2-dependent, ABA-induced signaling pathways. Here, multiple sets of large-scale phosphoproteomic data from Arabidopsis seedlings were compared in two dimensions: a comparative analysis between ABA treatment and dehydration stress and a mutant-oriented analysis between wild type and srk2dei mutant.

The number of phosphopeptides detected in this study (5288) was comparable to those found in other previous phosphoproteomic studies in Arabidopsis (36, 37, 45, 46). Comparative analysis between ABA treatment and dehydration stress revealed that dehydration stress induced multiple protein phosphorylation pathways in addition to the ABA-dependent pathway, supporting that multiple protein kinases are involved in dehydration stress signaling, including SnRK2s, MAPKs, and CDPKs (1821, 48, 49). Further studies will be required for understanding how those multiple kinases mediate dehydration stress signaling.

We identified candidates of SnRK2-dependent signaling (Table 1) and subclassified these candidates into potentially distinct regulatory pathways by motif analysis. We found that phosphorylation of the three specific motifs was SnRK2-dependent during ABA responses but less so during dehydration stress. Although subclass III SnRK2s are activated by both ABA and dehydration, probably due to different mechanisms (50), our results suggest that subclass III SnRK2s may be uniquely used during ABA responses but be partially redundant during dehydration responses. In contrast, subclass II SnRK2s are the predominant subclass that regulates dehydration stress responses, although they are also activated by both ABA and dehydration (51). Furthermore, it had been reported that in rice, subclass II SnRK2s were not activated in response to ABA (18). Subclass II and III members have different types of C-terminal region called as “ABA-box,” which is important for SnRK2-PP2C interaction (10, 18, 52, 53), and this region determines the different functions of subclass II and III SnRK2.

Among three motifs, Motif 3, [-(pS/pT)-P-], is a well-known, minimal MAPK target motif, and we found that phosphopeptides containing Motif 3 did not accumulate in srk2dei plants (Table 1). It is not immediately obvious why Motif 3 phosphorylation was affected by the loss of SnRK2 because there is no evidence that SnRK2 can directly phosphorylate this motif [-(pS/pT)-P-]. Instead, we found phosphorylation of the activation loop of AtMPK1/AtMPK2 in our phosphoproteomic candidate list, and we demonstrated that ABA-responsive activation of AtMPK1 was suppressed in srk2dei. One hypothesis is that SnRK2 somehow affects AtMPK1/2 activity, which in turn phosphorylates Motif 3–containing proteins; this hypothesis is consistent with AtMPK1/2 being implicated in ABA signaling (54, 55). However, the mechanistic connection between AtMPK1/2 regulation and SnRK2 remains to be elucidated.

Previous studies identified Motif 1, [-(K/R)-x-x-(pS/pT)-], as a phosphorylation motif of SnRK2 (56) and to be present in AREB transcription factors (27, 28, 57). The AREB family is a major group of transcription factors that regulate ABA-responsive gene expression, and its transcriptional activity is regulated by phosphorylation of multiple [-R-x-x-pS-] motifs in four conserved boxes (28). In our study, phosphopeptides from AREBs were phosphorylated in response to both ABA and dehydration, and this phosphorylation was suppressed in srk2dei, suggesting that this peptide is an in vivo substrate of SnRK2. However, ABA-responsive phosphorylation occurred at Ser45, which is within another motif, Motif 2: [-(pS/pT)-x-x-x-x-(D/E)-]. Previously, studies suggested that this site in AREB2 was phosphorylated in response to ABA (46) and showed that a truncated ABF3 protein containing this site was phosphorylated by recombinant SnRK2 proteins (57). Supporting these previous studies, we found that ABA-activated SnRK2 phosphorylates this conserved site in plant cells. In addition, mutational analysis of Ser45 of AREB1 suggested that its transcriptional activity was partly regulated by phosphorylation of this site. Therefore, we propose that multiple phosphorylations of Motif 1 and Motif 2 in AREBs orchestrate their transcriptional activity. Other modifications, such as SUMOylation, or proteasome-dependent protein degradation could also be involved (58, 59) in the regulation of AREB activity. It will be important to connect multiple phosphorylation events and other posttranslational regulatory mechanisms of AREB to fully understand its regulatory mechanism.

Consistent with previous studies, our results suggested that subclass III SnRK2s might recognize, in addition to Motif 1, [-(K/R)-x-x-(pS/pT)], a second motif, Motif 2, [-(pS/pT)-x-x-x-x-(D/E)-]. SnRK2 phosphorylated motifs matching Motif 1 in SLAC1 [Ser120 (-K-x-x-pS-)] (29, 30) and AtrbohF [Ser174 (-R-x-x-pS-)] (31). Another channel, the potassium channel KAT1, was phosphorylated at Thr306 [-pT-x-x-x-x-D-], which resembles Motif 2 (32). A comprehensive survey of SnRK2 phosphorylation motifs performed with a combinatorial peptide matrix found only Motif 1 to be phosphorylated (56). The peptide library covered only −4 to +4 residues from Ser; thus, residues in the +5 position, such as [-(pS/pT)-x-x-x-x-(D/E)-], would have been missed. A study of the phosphorylation of spinach sucrose phosphate synthase by CDPK or SnRK1 found that acidic residues (D/E) at +5 position are conserved among dicots, suggesting that this position was a recognition element of CDPK or SnRK1 (60). Our data show that one-third of the peptides exhibiting SnRK2-dependent phosphorylation contained Motif 2, and thus support that the +5 residue is important for SnRK2 recognition.

We identified a previously unknown Motif 2–containing protein, SNS1, which we determined to function as an negative regulator of ABA signaling in plants, participating in postgermination growth and ABA-responsive gene expression in Arabidopsis seedlings. Furthermore, the function of SNS1 may be specific to ABA responses because salt-responsive gene expression was not changed in sns1. SNS1 is conserved from mosses to angiosperms (fig. S8) and may have been evolutionarily conserved along with the core components of ABA signaling. Although the mechanism of SNS1 action is unknown, domain searches detected a single unique domain that was found in the Eaf7 superfamily (Fig. 6B). Eaf7 is a subunit of the only essential histone acetyltransferase (HAT) complex in yeast (61). Therefore, it will be interesting to examine whether SNS1 is part of a HAT complex. In addition, further analysis will be required to elucidate the relationship between SNS1 function and SnRK2-dependent phosphorylation.

Through integration of genetics with phosphoproteomic analyses, we connected protein kinases and their in vivo signaling pathways. In particular, this study provided insight into the ABA signaling pathway in plants, identifying and classifying targets regulated by subclass III SnRK2s (Fig. 8). In yeast, several studies with a similar concept, but using small-molecule kinase inhibitors, identified substrates of protein kinases (62, 63). However, such approaches require high specificity inhibitors, which are not always available, particularly in plants. Therefore, the mutant-oriented approach in this study may be more broadly suitable, especially for plant research. In addition, integrative analysis of proteome and transcriptome data or motif analysis may be helpful for interpreting such large-scale data from phosphoproteomics.

Fig. 8 A proposed model of the regulatory network downstream of SnRK2.

Subclass III SnRK2 is involved in the core component system of ABA signaling, and it is regulated by PYR/PYL/RCAR ABA receptor and group A PP2Cs. Upon ABA-dependent SnRK2 activation, SnRK2 phosphorylates at least two motifs: Motif 1, [-(K/R)-x-x-(pS/pT)-], and Motif 2, [-(pS/pT)-x-x-x-x-(D/E)-]. Proteins written in red text were identified as substrates in this study. SnRK2 also stimulates the MAPK cascade(s) involving AtMPK1 and AtMPK2 by unknown mechanism(s). Motif 3, [-(pS/pT)-P-], is the consensus motif recognized by AtMPK1/2. Phosphorylation of the other motifs is likely to be indirect to SnRK2 or mediated by other mechanisms.

Materials and Methods

Plant materials and protein extraction

The srk2dei triple mutant was generated as described previously (10). Col wild-type and srk2dei seeds were sown on germination medium (GM) agar plates. After stratification for 4 days at 4°C, the plates were placed at 22°C under 16-hour light/8-hour dark conditions for 2 weeks. After preconditioning plants in 10 ml of water in a 9-cm plastic plate overnight, 30 to 40 plants were exposed to ABA treatment or dehydration stress as follows. For ABA treatment, plants were placed in 10 ml of 50 μM (±)-ABA (Sigma) for 0, 15, 30, and 90 min. For dehydration stress, plants were placed on a 9-cm plastic plate without any solution for 0, 15, 30, and 90 min. Plant samples were frozen in liquid nitrogen and stored at −80°C.

For protein extraction, plant samples were ground in liquid nitrogen with a mortar and pestle and then homogenized in extraction buffer H containing 50 mM Hepes (pH 7.5), 5 mM EDTA, 5 mM EGTA, 1 mM Na3VO4, 25 mM NaF, 50 mM β-glycerophosphate, 10% glycerol, 2 mM dithiothreitol (DTT), and proteinase inhibitor cocktail (Sigma). The homogenate was filtrated through triple layers of Miracloth and centrifuged at 17,000g for 20 min, and then the supernatant was collected and protein concentration was measured.

Enrichment of phosphopeptides with metal oxide chromatography

Protein concentration of the crude extracts was measured by bicinchoninic acid assay, and aliquots of each crude extracts containing 100 μg of protein were mixed with urea to a final concentration of 8 M. The solution was reduced with DTT, alkylated with iodoacetamide, and digested with Lys-C, followed by dilution and tryptic digestion, as described (64). These digested samples were desalted using StageTips with SDB-XC Empore disk membranes (3M) (65). Metal oxide chromatography (MOC) with titania was performed as described in the literature (35) with slight modifications. Custom-made MOC tips were prepared using C8-StageTips and titania bulk beads (Titansphere, GL Sciences; 0.5 mg of beads per 10-ml pipette tip) as described for SCX(beads)-C18 tips (66). Before loading samples, the MOC tips were equilibrated with lactic acid (300 mg/ml) in 0.1% trifluoroacetic acid (TFA), 80% acetonitrile (solution A). The digested sample was diluted with 100 μl of solution A and then loaded to the MOC tip. After successive washing with solution A and solution B (0.1% TFA, 80% acetonitrile), 0.5% piperidine was used for elution (67). The eluted fraction was acidified with TFA and desalted using SDB-XC StageTips as described above. The desalted sample was concentrated in a vacuum evaporator and then dissolved in solution A for a subsequent nano-LC-MS/MS analysis.

Nano-LC-MS system

Nano-LC-MS/MS analyses were conducted with LTQ-Orbitrap (Thermo Fisher Science), Ultimate3000 RS nano-LC system (Dionex), and HTC-PAL autosampler (CTC Analytics) instrumentation. ReproSil C18 materials (3 μm, Dr. Maisch) were packed into a self-pulled needle (150 mm length × 100 μm inside diameter, 6 μm opening) with a nitrogen-pressurized column loader cell (Nikkyo Technos) to prepare an analytical column needle with “stone-arch” frit (68). The injection volume was 5 μl, and the flow rate was 500 nl/min. The mobile phases consisted of 0.5% acetic acid (A) and 0.5% acetic acid and 80% acetonitrile (B). A three-step linear gradient of 5% to 10% B in 5 min, 10% to 40% B in 60 min, 40% to 100% B in 5 min, and 100% B for 10 min was used throughout this study. A spray voltage of 2400 V was applied via the metal connector as described (68). The MS scan range was mass/charge ratio (m/z) 300 to 1400, and the top 10 precursor ions were selected for subsequent MS/MS scans. A lock mass function was used for the LTQ-Orbitrap to obtain constant mass accuracy during gradient analysis (69).

Proteolytic isotope labeling with 16O/18O-water

Crude extracts were treated with trypsin in the presence of 16O- or 18O-labeled water according to a previous report (70) with a slight modification before phosphopeptide enrichment. Briefly, the desalted tryptic digest of samples (100 μg of proteins) was first concentrated in a vacuum evaporator, then dissolved in 50 μl of 50 mM citrate/NaOH buffer (pH 6.0), prepared in 16O- or 18O-water. After incubation with 1 μg of Lys-C at 37°C for 3 hours, the samples were treated with 1 μg of trypsin overnight. Finally, the reaction mixtures were boiled for 10 min, acidified with formic acid, and desalted with SDB-XC StageTips as mentioned above. After phosphopeptide enrichment, the samples were analyzed by a nano-LC-MS system as described earlier.

Database searching

Mass Navigator v1.2 (Mitsui Knowledge Industry) was used to create peak lists on the basis of the recorded fragmentation spectra. Mass Navigator v1.2 discarded all peaks with an absolute intensity of less than 10 and with an intensity of less than 0.1% of the most intense peak in MS/MS spectra, and an in-house Perl script converted the m/z values of the isotope peaks to the corresponding monoisotopic peaks when the isotope peaks were selected as the precursor ions (71). Peptides and proteins were identified by means of automated database searching using Mascot v2.2.04 (Matrix Science) against TAIR version 8 (02 April 2008) with a precursor mass tolerance of 3 parts per million (ppm), a fragment ion mass tolerance of 0.8 dalton, and strict trypsin specificity (72), allowing for up to two missed cleavages. Carbamidomethylation of cysteine was set as a fixed modification, and methionine oxidation and serine, threonine, and tyrosine phosphorylation were allowed as variable modifications. For the samples enzymatically labeled with 16O/18O-water, [18O2]-C terminus was also considered as a variable modification. Note that neutral loss products from both precursor and fragment ions were considered for MASCOT scoring in this phosphorylation modification setting, although no assignment was indicated for precursor-origin neutral loss peaks in the output results according to the supplier. Peptides were considered identified if the Mascot score was over the 95% confidence limit based on “identity” score of each peptide and was less than the threshold indicating 1.17% false-positive rate estimated with a randomized decoy database created by a Mascot Perl program. Phosphorylation sites were unambiguously determined when b- or y-ions, which were between the existing phosphorylated residues, were observed in the peak list of fragment ions (36).

Peptide quantitation

We applied two approaches to quantify phosphoproteome changes, that is, label-free (7375) and proteolytic 16O/18O labeling methods (70, 76, 77). In both of the methods, the LC-MS peak area of each peptide and the corresponding isotope pair in all samples were determined by integration of ion intensity in a survey MS scan based on an observed m/z of monoisotopic ion and a retention time. The peak integration was performed by Gaussian approximation of an extracted ion chromatogram within 5 mD of the observed m/z using Mass Navigator v1.2.

Motif analysis

We classified phosphopeptides detected in this study by phosphorylation motifs. On the basis of peptide sequences and modification sites from MS/MS analysis, we extracted peptide fragments (13 amino acids) that have a phosphorylated residue at the central position. Motif enrichment analysis was then performed with Motif-X algorithm (40). TAIR9 data set was used as an analytical background in Motif-X analysis.

Microarray analysis

Microarray analysis was performed as described previously (51). Total RNA was isolated from Col and srk2dei plants treated with ABA or dehydration stress for 0 or 90 min. We used the Affymetrix ATH1 genome array (Affymetrix Inc.) according to the manufacturer’s instructions. Each analysis was repeated in triplicate, and the microarray data were analyzed using affylmGUI package within the statistical program R.

Site-directed mutagenesis of AREB1

Full-length AREB1 cDNA was amplified by reverse transcription polymerase chain reaction from Arabidopsis RNA and cloned into pENTR/D-TOPO vector according to the manufacturer’s instructions (Invitrogen). To convert Ser45 to Ala or Asp, we performed site-directed mutagenesis using QuikChange Site-Directed Mutagenesis Kit (Agilent Technologies Inc.) with the following oligonucleotide pairs: 5′-CAGTGTAGGGAAAGATTTTGGGGCTATGAACATGGATGAGTTGTTAAAG-3′ and 5′-CTTTAACAACTCATCCATGTTCATAGCCCCAAAATCTTTCCCTACACTG-3′ for S45A; 5′-CAGTGTAGGGAAAGATTTTGGGGATATGAACATGGATGAGTTGTTAAAG-3′ and 5′-CTTTAACAACTCATCCATGTTCATATCCCCAAAATCTTTCCCTACACTG-3′ for S45D.

Recombinant proteins and synthetic peptides

Partial AREB1 cDNA fragments (206 to 277 nucleotides) were amplified from AREB1 cDNA clone of wild type and S45A with the primers 5′-TCTTTCAGAGCAGTGTAGGGAAAG-3′ and 5′-TCAAGCACTCCATATATTCTTTAAC-3′ and subcloned into pGEX-4T-3 (GE Healthcare) to produce glutathione S-transferase (GST)–fused truncated AREB1 protein (Phe35 to Ala58) in BL21(DE3). Recombinant GST-AREB1:35–58WT and GST-AREB1:35–58S45A were expressed and isolated using glutathione Sepharose resin according to the manufacturer’s instructions (GE Healthcare). Two synthetic peptides matching either the wild-type SNS1 sequence 35ASSLSKEQSQVELELRLL52 or the S43A-substituted SNS1 sequence 35ASSLSKEQAQVELELRLL52 were synthesized by Hokkaido System Science Co. Ltd.

In vitro kinase reactions

Col wild-type and srk2dei seedlings were treated with 50 μM ABA or dehydration stress for 0, 15, 30, and 90 min. Crude extracts were prepared using extraction buffer H and AtMPK1 proteins immunoprecipitated with polyclonal anti-AtMPK1 antibody and protein G–Sepharose (GE Healthcare). The AtMPK1-specific antibody was produced against synthetic peptides corresponding to the N terminus of AtMPK1 (MATLVDPPNGIRNEGC). This peptide was conjugated with keyhole limpet hemocyanin carrier. Polyclonal antisera were raised in rabbits and purified by affinity chromatography (Sawady Technology Inc.). In vitro kinase reaction of immunoprecipitated AtMPK1 and myelin basic protein (MBP) was performed with [32P]ATP (adenosine 5′-triphosphate) as described (48). Phosphorylation of MBP was detected by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography.

For SnRK2 reactions, Col wild-type and srk2dei seedlings were treated with 50 μM ABA for 0 and 30 min, and SRK2E proteins were immunoprecipitated with protein G–Sepharose (GE Healthcare) and polyclonal anti-SRK2E antibody as described previously (19). In vitro kinase reactions of recombinant GST-AREB1(35–58WT) or GST-AREB1(35–58 S45A) by purified SRK2E were performed in 50 mM tris-HCl (pH 7.5), 1 mM DTT, 10 mM MgCl2, 10 mM MnCl2, 50 mM ATP, and 0.037 MBq of [γ-32P]ATP (60 Ci/mmol) as previously described (78). Phosphorylation of AREB1 was detected by SDS-PAGE and autoradiography. In vitro phosphorylation of SNS1 synthetic peptides SNS1:35–52WT or SNS1:35–52S43A by SRK2E was performed under identical reaction conditions. To detect peptide phosphorylation, SRK2E-bound protein G–Sepharose beads (GE Healthcare) were removed by centrifugation and SNS1 peptides trapped by filtration of supernatant through nitrocellulose membrane filters (Millipore). The filters were washed three times with extraction buffer, and the radioactivity was measured with a Tri-Carb scintillation counter (PerkinElmer).

Transactivation assay of AREB1

AREB1 cDNA was subcloned into a modified pGreen0029 binary vector (35SΩ-GFP) (79) to make 35SΩ-AREB1WT-GFP. 35SΩ-AREB1S45A-GFP and 35SΩ-AREB1S45D-GFP were generated by site-directed mutagenesis as described above. A reporter plasmid, RD29B-GUS, was generated from pGreen0029 by replacement of CaMV35S promoter upstream of GUS gene with 1.0 kb of RD29B promoter. pGreen0029 harboring luciferase gene, 35SΩ-LUC, was used for transfection control. Preparation of Arabidopsis mesophyll protoplasts, transfection, and transactivation assay were performed according to the literature (80) with slight modifications (51).

Functional analysis of SNS1

SNS1 knockout mutant, sns1 (SALK_008996), was obtained from Arabidopsis Biological Resource Center. For the first germination test, Col wild-type and sns1 were placed on GM agar plates containing 1 μM ABA. Then, SNS1 cDNA was cloned to pGreen0029 to make a 35S:SNS1 construct, and it was introduced to sns1 mutants to make a complementation line, sns1/SNS1. Seeds of Col wild-type, sns1, and sns1/SNS1 were sown on GM agar plates containing 0, 0.1, and 0.5 μM ABA. After 4 days of stratification, plates were placed at 22°C under 16-hour light/8-hour dark conditions. For gene expression analysis, 2-week-old seedlings of Col wild-type and sns1 were treated with 50 μM ABA or 100 mM NaCl solution as described above. RNA gel blot analysis was performed as described previously (81). Three cDNA probes—Rab18, RD29A, and RD29B—were labeled with 32P and used for hybridization on RNA blots. After washing, radioactivity was detected by autoradiography, and quantitative data were obtained by gel densitometry of ImageJ software.

Supplementary Materials

Fig. S1. Experimental workflow of the phosphoproteomic analysis.

Fig. S2. Venn diagram of the ABA-responsive genes and dehydration-responsive genes in wild-type plants.

Fig. S3. Comparative analysis of ABA- or dehydration-responsive gene expression data from wild type and srk2dei.

Fig. S4. Motif-X analysis of the ABA- or dehydration-responsive phosphopeptides.

Fig. S5. In vitro kinase assay of AtMPK1 immunoprecipitated from wild-type or srk2dei.

Fig. S6. Expression patterns of SNS1.

Fig. S7. RNA gel blot analysis of ABA-responsive genes in wild type and sns1.

Fig. S8. SNS1 orthologs in plants.

Table S1. Microarray data from wild type and srk2dei.

Table S2. List of phosphopeptides detected in seedlings exposed to ABA or dehydration.

Table S3. List of ABA- or dehydration stress–responsive phosphopeptides.

References and Notes

Acknowledgments: We thank H. Kobayashi, S. Mizukado (RIKEN), S. Onuma (Keio University), K. Ishizuka, and R. Terao (Tokyo University of Agriculture and Technology) for technical assistance. We are grateful to Arabidopsis Biological Resource Center (ABRC) and GABI-Kat project for providing Arabidopsis transferred DNA–tagged mutants. Funding: This work was partly supported by a grant from the Bio-Oriented Technology Research Advancement Institution of Japan to K.S., Grants-in-Aid for Scientific Research for Young Scientists (B) to T.U. from MEXT, research funds from Yamagata prefecture and Tsuruoka City to Keio University, and NSF grant IOS-1025837 to S.C.P. Author contributions: T.U. designed the study. T.U., N.S., and F.T. performed the experiments and analyzed the data. N.S. and Y.I. developed the experimental system for mass spectrometry. T.U., N.S., J.C.A., S.C.P., and K.S. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All MS data were deposited in Pepbase ( Microarray data have been submitted to ArrayExpress (, accession number: E-MEXP-3713). The genetically modified plants require a materials transfer agreement from Tokyo University of Agriculture and Technology or RIKEN.
View Abstract

Navigate This Article