Seedlings of Arabidopsis thaliana pER8::cMYC-MKK7 Col-0 were grown in 50 mL MES-buffered, half-strength Murashige and Skoog medium with 4.70 mM potassium nitrate and 5.15 mM ammonium nitrate. Nitrogen salts were supplied either in normal ¹⁴N form or in ¹⁵N-enriched form (>98% ¹⁵N atom % K¹⁵NO₃, ¹⁵NH₄¹⁵NO₃). Medium was supplemented with 2.5 g/L sucrose. Plants were grown at 22°C in continuous light (70 μE/m²/sec). Twelve-day-old seedlings were treated with β-estradiol (1 μM) in ethanol or with ethanol as the control and collected 6 h after treatment.

Prior to protein extraction and fractionation, ground tissue powder of ¹⁴N and ¹⁵N-labeled was mixed in a 1:1 ratio. Phenolic total protein extraction and enrichment of phosphorylated proteins was performed as described before (Thomas et al., 2015). In brief, ground tissue powder was washed twice with ice-cold acetone before resuspension of the tissue pellet in 10% (w/v) trichloroacetic acid (TCA) in acetone and sonication for 10 min in a sonication bath. Pellets were washed twice with 10% (w/v) TCA in acetone, twice in 10% (w/v) TCA, and twice in 80% (v/v) acetone. Air-dried pellets were resuspended in dense SDS buffer (30% (w/v) sucrose, 2% (w/v) SDS, 0.1 M Tris-HCl pH 8.0, 5% (v/v) 2-mercaptoethanol) before adding Tris-buffered phenol and vigorous vortexing. The upper phenol phase was separated by centrifugation (15 min at 4,000 × g) and proteins were recovered by precipitation at -20°C (60 min) with 5 volumes of 0.1 M ammonium acetate in methanol. After centrifugation the protein pellets were washed twice with 0.1 M ammonium acetate in methanol and twice with 80% (v/v) acetone.

For Al(OH)₃-based MOAC enrichment of phosphoproteins, protein pellets were re-solubilized in buffer A (30 mM MES pH 6.1, 8 M urea, 150 mM sodium L-glutamic acid, potassium L-aspartic acid, 20 mM imidazole, 0.25% (w/v) CHAPS). Phosphoprotein binding to Al(OH)₃ (prewashed twice with buffer A) was achieved by incubation on a rotator at 10°C for 60 min. Al(OH)₃-bound phosphoproteins were washed 6 times with buffer B (similar to buffer A, except that the concentration of amino acids was increased to 200 mM). Phosphoproteins were eluted in buffer C (200 mM potassium pyrophosphate pH 9.0, 8 M urea) by head-over-head incubation for 30 min at room temperature. After centrifugation (3,000 × g for 5 min) the phosphoprotein containing supernatant was loaded onto a Amicon Ultra-15 centrifugal filter unit (Merck, Darmstadt, Germany) to concentrate the phosphoprotein sample. The retentate was diluted with 4 volumes of ddH₂O before addition of 0.01 volume of 2% (w/v) sodium deoxycholate. After vortexing and incubation for 5 min, 0.1 volume of 100% (w/v) TCA was added to precipitate phosphoproteins for 120 min on ice. Precipitated phosphoproteins were collected by centrifugation (14,000 × g for 10 min at 4°C) and subsequently washed once with 25% TCA in ddH₂O, once with 80% acetone in 50 mM Tris-HCl pH 7.5, and once with 100% acetone, before air-dried pellets were stored at -20°C.

Protein samples were subjected to SDS-PAGE, transferred to nitrocellulose, and used for immunodetection as described (Beckers et al., 2009). Strep-Tactin^® HRP conjugate was purchased from IBA Lifesciences (Göttingen, Germany); primary rabbit antibodies against the thiophosphatester (α-TPE) from Abcam (Cambridge, United Kingdom) all other antibodies were purchased from Cell Signaling Technology (Danvers, MA, United States). Chemiluminescence detection of antigen-antibody complexes was done with Immobilon^TM Western substrate (Merck, Darmstadt, Germany).

Enrichment of phosphopeptides by MOAC was essentially done as described (Hoehenwarter et al., 2013). Briefly, 500 μg MOAC-enriched phosphoproteins were trypsin digested [Poroszyme immobilized trypsin (1/100 v/w) (Thermo Fisher Scientific, Waltham, MA, United States)] overnight before protein digests were desalted using a self-packed graphite (500 mg) spin column with a polyethylene filter of 10 μM pore size (MoBiTec, Göttingen, Germany). Meanwhile, graphite was equilibrated by washing twice with 500 μL of 1 M ammonia, once with 500 μL acetonitrile (ACN) and twice with 500 μL 1% (v/v) trifluoroacetic acid (TFA). Meanwhile phosphoprotein digestion mixture was acidified by adding 10% (v/v) TFA to bring the final concentration to 1.25% (v/v) TFA and centrifuged for 10 min at 16,000 × g before loading on the column. Graphite was washed twice with 500 μL 1% (v/v) TFA before peptides were eluted twice with 200 μL 0.1% (v/v) formic acid (FA) in 50% (v/v) ACN. Eluted peptides were pooled and dried in a vacuum concentrator. After dissolving peptides in 100 μL buffer A (phthalic acid-saturated 50% (v/v) ACN, 2.5% (v/v) TFA) they were loaded onto a TiO₂ column (12.5 mg) pre-equilibrated with buffer A (25 mg TiO₂ per 1 mg peptides). After phosphopeptide binding, TiO₂ was washed twice with 250 μL buffer A, twice with 250 μL 50% (v/v) ACN, 0.1% (v/v) TFA, and twice with 250 μL 0.1% (v/v) TFA. Finally, phosphopeptides were eluted three times with 100 μL 5% (v/v) ammonia, eluates were pooled, dried in a vacuum concentrator and stored at -20°C until MS-analysis.

Liquid chromatography and mass spectrometry (LC-MS) of phosphorylated peptides was described in detail (Thomas et al., 2015). In brief, peptides were separated using C18 reverse phase chemistry employing a pre-column (EASY column SC001, length 2 cm, inner diameter (ID) 100 μm, particle size 5 μm) in line with an EASY column SC200 with a length of 10 cm, ID of 75 μm and a particle size of 3 μm (both from Thermo Fisher Scientific). Peptides were eluted into a Nanospray Flex ion source (Thermo Fisher Scientific) with a 180 min gradient increasing from 5% to 35% (v/v) ACN in ddH₂O and electrosprayed into an Orbitrap Velos Pro mass spectrometer (Thermo Fisher Scientific). The source voltage was set to 1.9 kV, the S Lens RF level to 50%. The delta multipole offset was –7.00.

Measurements employed a data dependent acquisition (DDA) scan strategy wherein up to 20 of the most abundant ions with a minimum signal of 1,000 recorded in an MS survey scan were isolated and subjected to collision induced dissociation (CID). The AGC target value was set to 1e06 and the maximum injection time (max IT) to 500 ms in the Orbitrap. The parameters were set to 1e04 and 100 ms in the LTQ with an isolation width of 2 Da for precursor isolation and MS/MS scanning. Multi stage activation (MSA) was applied to further dissociate fragment ion peaks resulting from neutral loss of the phosphate moiety by dissociation of the high-energy phosphate bond to generate b- and y-fragment ion series rich in peptide sequence information. Neutral loss masses were 24.49, 32.66, 49, 65.3, 98, 147, 195.94, and 293.91. The repeat count was set to 1 and the repeat duration to 30 s. The exclusion duration was set to 40 s and the exclusion width to 10 ppm.

Procedures were described in detail previously and performed with Mascot Distiller v.2.5.1.0 linked to an in-house Mascot server v.2.5.0 (Thomas et al., 2015). To determine the level of ¹⁵N incorporation into peptides, MS data was used to search the TAIR10 database with common contaminants amended (35,394 sequences and 14,486,974 residues) with the enzyme set to trypsin/P. The tolerated precursor and fragment ion mass errors were 7 ppm and 0.8 Da, respectively. Oxidation of methionine and phosphorylation of serine and threonine residues were set as variable modifications. The quantitation option was set to ¹⁵N metabolic. Peptides and proteins were identified and extracted ion currents (XIC) of naturally occurring and heavy isotope incorporated peptide pairs were fit to calculated ion currents of the respective peptide pairs with the corrections option iteratively set to impurity ¹⁵N 96%, 97%, 98%, 98.5%, and 99%. R² values of all XIC to calculated ion current fits for all identified phosphopeptides for the set incorporation levels were extracted. The R² distribution with the highest mode was used as an estimate of phosphopeptide ¹⁵N incorporation. The results of the search, XIC extraction and fitting obtained with the respective incorporation level was then used to identify and quantify phosphopeptides by calculating the XIC peak area ratio of the naturally occurring and ¹⁵N incorporated peptides of all peptide pairs (light to heavy ratio, L/H).

Only the highest ranking phosphopeptide annotation for each MS/MS spectrum was kept and used for peptide quantification. Phosphopeptide L/H ratios that were more than 100-fold different from the median across the replicates were considered outliers and excluded from further analysis with the Perseus software platform version 1.5.5.3 (Tyanova et al., 2016). Phosphopeptide quantification ratios were log₂-transformed and subsequently median-normalized before a one-sample t-test (p < 0.05) was performed on all phosphopeptides that were quantified in at least four sample replicates (Yang et al., 2013).

StrepII-tagged eYFP, VQ4 (AT1G28280), HON5 (AT1G48620), DRP2B (AT1G59610), WRKY6 (AT1G62300), CLMP1 (AT1G62390), MKK2 (AT4G29810), and NOT2/3/5 (AT5G18230) were expressed in a coupled cell-free transcription–translation system based on tobacco BY-2 cell lysates as described (Buntru et al., 2015).

Immunocomplex kinase assays were essentially performed as described previously (Leissing et al., 2016). StrepII-tagged substrates were purified from tobacco BY-2 cell lysates using Strep-Tactin^® Macroprep^® affinity beads (IBA Lifesciences) according to manufacturer’s instructions.

To illuminate the identity of downstream targets of Arabidopsis MKK7, we used transgenic pER8::cMYC-MKK7 plants conditionally expressing a cMYC-tagged form of AtMKK7 from the β-estradiol-inducible promoter (pER8) (Zuo et al., 2000). Treatment of these transgenic conditional overexpression plants with the steroid hormone ß-estradiol leads to accumulation of cMYC-MKK7 that specifically phosphorylates endogenous MPK3 and MPK6 (Figure 1A) (Jia et al., 2016). Next, we embarked on a reciprocal ¹⁵N-stable isotope metabolic labeling-based quantitative phosphoproteomic analysis to accurately quantify differentially regulated tandemMOAC-extracted phosphopeptides. Transgenic pER8::cMYC-MKK7 seedlings were grown in medium containing either the natural abundance of nitrogen isotope (¹⁴N) salts, or stable “heavy” nitrogen isotope (¹⁵N, >98%) enriched salts, as the sole nitrogen source (Figure 1B). This metabolic labeling approach led to an ¹⁵N incorporation of nearly 98% as determined by quality of fit analysis of measured and calculated XICs taking into account different levels of possible ¹⁵N incorporation (Supplementary Figure S1). After 12 days, seedlings were treated with ß-estradiol (or ethanol as the solvent control) to induce the transgene and activate the MKK7-MPK3/6 module. At the 6-h time point post induction seedlings were harvested and ground in liquid nitrogen before ¹⁴N and ¹⁵N-labeled tissue powder was combined in a 1:1 ratio. In total, we processed eight combined ^14/15N-labeled tandemMOAC-enriched phosphopeptide samples by LC-MS/MS; this set comprised four biological replicate experiments. Each replicate included two reciprocal ^14/15N-labeled samples. One of which was inversely labeled with heavy nitrogen (label switch) (Figure 1B). In a first step, phosphoproteins were enriched from denatured total protein samples by Al(OH)₃-based MOAC before enriched phosphoprotein fractions were digested with trypsin. In a second step, the resulting peptide mixture was desalted via reverse-phase chromatography using graphite columns before phosphopeptides were enriched by TiO₂-MOAC. Subsequent LC-MS/MS analysis of all tandemMOAC samples together yielded 9,324 MS/MS spectra annotated as phosphopeptides (pPSMs) using Mascot, resulting in the identification of 1,650 nonredundant phosphopeptide sequences with high confidence scores (Mascot identity threshold <0.05). These phosphopeptides mapped to 764 identified phosphoproteins.

We performed reciprocal labeling within each biological quadruplicate experiment to allow treatment effects, such as labeling effects, biological variation or quantitation errors to be distinguished from experimental bias (Schulze and Usadel, 2010). In each sample, light and heavy labeled forms of a specific phosphopeptide were recorded as distinct, mass shifted XIC peaks. The ratio of peak areas of the light and heavy forms of each peptide was therefore calculated based on the XICs of monoisotopic peaks and the isotopologue distribution. Peptide quantification for exemplarily forward and reciprocal labeled samples is shown for two representative phosphopeptides in Figure 2. Mascot analysis of all tandemMOAC samples together yielded peptide abundance ratios for all 1,650 unique phosphopeptides. To confidently quantify phosphopeptides we only evaluated phosphopeptides recorded in at least four of the replicates. From this set of 454 phosphopeptides we extracted overrepresented phosphorylation patterns using Motif-X (Figure 3A) (Schwartz and Gygi, 2005; Chou and Schwartz, 2011). We discovered consensus phosphorylation motifs that resemble three main categories of S/T kinase recognition sequences: acidophilic-, basophilic-, and proline-directed phosphorylation sites (Douglass et al., 2012). This result indicates that tandemMOAC-enrichment of phosphopeptides is not biased toward enrichment of phosphorylated peptides that belong to either of these categories.

To confidently identify phosphorylation targets of the MKK7-MPK3/6 cascade, a one-sample Student’s t-test with p < 0.05 was performed with the above mentioned 454 phosphopeptides that were quantified in at least half of the sample replicates. Statistical analyses revealed that in total 27 phosphorylated peptides differed in their abundance at least 1.5-fold upon in planta induction of MKK7 (Supplementary Table S1). Graphical representation of these quantitative phosphoproteomics data in a volcano plot revealed that the abundance of 20 phosphopeptides was significantly enhanced ≥1.5-fold, whereas the abundance of only seven phosphopeptides was reduced (Figure 4). Within this set of 27 regulated phosphopeptides the Motif-X algorithm discovered a single predominant consensus phosphorylation motif reminiscent of the minimal motif required for MPKs (Figure 3B) (Alvarez et al., 1991). The latter suggests that most phosphoproteins that differ in their abundance upon ß-estradiol-induced accumulation of MKK7 are targeted by its downstream proline-directed MPKs. Accordingly, a phosphorylated MPK substrate motif was found in the sequence of 16 of the 20 accumulating phosphopeptides (Table 1). In contrast, the abundance of only a single phosphorylated MPK-targeted phosphopeptide was reduced. Together, these results indicate that ß-estradiol treatment of transgenic pER8::cMYC-MKK7 seedlings specifically induces the phosphorylation of downstream targets of the MKK7-MPK3/6 cascade. The MPK motif-containing peptides that are phosphorylated upon activation of the MKK7-MPK3/6 cascade map to 14 full-length phosphoproteins (Table 1). Except the CCCH-type zinc finger protein 56 (AT5G12850) all substrates contain a kinase-interaction motif characterized by a cluster of basic residues positioned amino-terminal to a (L/I)X(L/I) motif (Supplementary Table S2) (Barsyte-Lovejoy et al., 2002). This corroborates our assertion that these proteins are likely direct in vivo substrates of MPK3/6. Although mostly novel, some of the proteins have been reported MPK3/6 substrates (Hoehenwarter et al., 2013; Pecher et al., 2014). When examining this set of direct phosphorylation targets of MKK7-MPK3/6 we found that many of them are DNA/RNA-binding proteins that predominantly localize to the nucleus (Supplementary Figures S2A,B). Furthermore, the enrichment of proteins in the GO biological process category ‘response to biotic and abiotic stimuli’ supports the notion that MKK7 is involved in stress-responsive signal transduction pathways in Arabidopsis (Supplementary Figure S2C).

Recently, it has been shown that MKK7 phosphorylates MPK3 and MPK6 in response to pathogen infection and/or during plant growth and development (Jia et al., 2016). To confirm the phosphorylation of targets of MKK7 listed in Table 1, we tested the ability of MPK3/6 to directly phosphorylate a selected set of these proteins. In total, six strepII-tagged proteins (HON5, DRP2B, WRKY6, CLMP1, MKK2, and NOT2/3/5) were purified with Strep-Tactin^® Macroprep resin after successful cell-free protein synthesis in tobacco BY2 cell lysates (Supplementary Figure S3) (Buntru et al., 2015). Subsequent phosphorylation assays were performed with pre-activated, wild-type, and engineered analog-sensitive AS-MPK3 and AS-MPK6 in the presence of the bulky ATP analog N⁶-benzyl-ATPγS (Bn-ATPγS) as a co-factor (Figure 5 and Supplementary Figure S4) (Leissing et al., 2016). Mutation of the so-called gatekeeper amino acid which enlarges the ATP-binding pocket of the kinase allows the AS kinase to catalyze thiophosphorylation of its substrate proteins. VQ4 served as a positive control for MPK3/6, as it was shown by us and others that this VQ-motif containing protein is specifically phosphorylated by MPK3/6 in vivo (Hoehenwarter et al., 2013; Pecher et al., 2014). In addition, in vitro synthesized eYFP served as a negative control protein in the kinase assays. All the phosphorylation targets of the MKK7-MPK3/6 module we tested were directly phosphorylated by AS-MPK6, whereas AS-MPK3 was unable to phosphorylate HON5 and WRKY6. Thus, most of the targets identified by the tandemMOAC approach are direct MPK3/6 substrates, at least in vitro (Figure 5).

Remarkably, in this study we found that upon in planta expression of MKK7 two functionally redundant Dynamin proteins of Arabidopsis, DRP2A, and DRP2B, are phosphorylated by MPK3/6 at a highly conserved serine residue (Table 1 and Figure 5). Although their precise mode of action has yet to be disclosed, these proteins have been shown to play a key role in different membrane scission events (Taylor, 2011). For instance, DRP2 has been shown to localize to the Golgi apparatus where it enables trafficking of cargo molecules from the trans-Golgi network to the central vacuole (Jin et al., 2001; Huang et al., 2015). In addition, DRP2 colocalizes with clathrin light chain at the plasma membrane where dynamin proteins are thought to act in endocytosis by assembling as spirals around the neck of invaginating endocytic vesicles. Recently, DRP2B was also shown to play an important role in plant immunity (Smith et al., 2014). Mutant drp2b plants are impaired in ligand-induced endocytosis of the flagellin receptor FLS2 and are more susceptible to bacterial pathogens. Although the present study identified DRP2 as a downstream target of the MKK7-MPK3/6 phosphorylation module, the role of MPK-dependent phosphorylation of DRP2 in endocytosis remains unclear. Since the MKK7-MPK3/6 module is involved in basal and systemic plant immunity, it is tempting to speculate that the module-mediated DRP2 phosphorylation contributes to the overall plant immune response.