Initiation and amplification of SnRK2 activation in abscisic acid signaling

The phytohormone abscisic acid (ABA) is crucial for plant responses to environmental challenges. The SNF1-regulated protein kinase 2s (SnRK2s) are key components in ABA-receptor coupled core signaling, and are rapidly phosphorylated and activated by ABA. Recent studies have suggested that Raf-like protein kinases (RAFs) participate in ABA-triggered SnRK2 activation. In vitro kinase assays also suggest the existence of autophosphorylation of SnRK2s. Thus, how SnRK2 kinases are quickly activated during ABA signaling still needs to be clarified. Here, we show that both B2 and B3 RAFs directly phosphorylate SnRK2.6 in the kinase activation loop. This transphosphorylation by RAFs is essential for SnRK2 activation. The activated SnRK2s then intermolecularly trans-phosphorylate other SnRK2s that are not yet activated to amplify the response. High-order Arabidopsis mutants lacking multiple B2 and B3 RAFs show ABA hyposensitivity. Our findings reveal a unique initiation and amplification mechanism of SnRK2 activation in ABA signaling in higher plants.

ABA-insensitivity is much weaker in m3kδ1/δ6/δ7 or OK-quatdec mutants 27,28 . Thus, the role of RAFs in ABA signaling still needs further investigation.
Here, we show that the B2 and B3 subgroup RAFs phosphorylate Ser171 and Ser175 in SnRK2. 6 with different specificity and that transphosphorylation is essential for initiating SnRK2.6 phosphorylation and activation. After phosphorylation by RAFs, the activated SnRK2.6 can quickly autophosphorylate (intermolecularly) and activate more SnRK2 proteins. We also generate a series of high-order mutants carrying null mutations in the B2, B3, or both B2 and B3 subgroup RAFs. From phenotypic assays of these high-order mutants, we find that both B2 and B3 subgroup RAFs are essential for ABA signaling. ABA-induced activation of SnRK2.2/2.3/2.6 and ABA-induced gene expression are strongly impaired in OK 100 -oct and OK 100 -nonu mutants lacking 8 and 9 members, respectively, of the B2 and B3 subgroups. OK 100 -oct and OK 100 -nonu also exhibit ABA hyposensitivity and can germinate and grow under extremely high ABA concentrations. We find that ABA does not activate B2 and B3 RAFs; instead, the basal level of RAF kinase activity is essential for SnRK2 activation and initiation of ABA signaling. Our results reveal a crucial RAF-SnRK2 cascade in ABA receptor-coupled core signaling and unique activation machinery for initiating and amplifying stress signaling in higher plants.
We further evaluated the role of B2 and B3 subgroup RAFs in ABA signaling by assaying seed germination in response to ABA. All tested mutants carrying high-order mutations in B2 and B3 subgroup RAFs, including OK 100 -oct and OK 100 -nonu, showed insensitivity to ABA in seed germination and post-germination seedling growth (Fig. 4d, e, Supplementary Fig. 7d-h). The order of insensitivity to ABA in seed germination was OK 100 -quin < OK 100 -oct = OK-quatdec < OK 100 -nonu < snrk2-triple, pyl112458, and pyl-duodec (Fig. 4d, Supplementary Fig. 7d). Higher-order RAF mutants were clearly more insensitive than lower-order RAF mutants to ABA in seed germination, which further suggests that the RAF members in the B2 and B3 subgroups have essential and partially redundant roles in ABA signaling. OK 100 -oct and OK 100nonu also exhibited hyposensitivity to ABA in seedling growth as indicated by root growth, fresh weight measurements, and leaf yellowing ( Supplementary Fig. 7g-j).
We then measured ABA-induced SnRK2.2/2.3/2.6 activation in these mutants by in-gel kinase assay. The ABA-induced SnRK2.2/ 2.3/2.6 activation was almost completely abolished in the OK 100oct and OK 100 -nonu mutants, which resembles that in the snrk2-triple and pyl112458 mutants (Fig. 4f). Interestingly, we observed a weak OK 100 band (indicated by arrows) in wild type, snrk2triple and pyl112458 even without ABA or mannitol treatment, and this band was not induced by ABA, when compared to the strong induction by mannitol treatment (Fig. 4f, rightmost lane). Consistently, the immunoblot result showed that the ABAinduced phosphorylation of conserved serine residues corresponding to Ser171 and Ser175 in SnRK2.6 was markedly reduced in the OK 100 -oct and OK 100 -nonu mutants (Fig. 4g). Interestingly, perhaps due to the abolishment of ABA signaling, the ABAinduced rapid SnRK2 degradation was also abolished in the OK 100 -oct and OK 100 -nonu mutants (Fig. 4g). These results strongly indicate that the B2 and B3 subgroup RAFs are essential for ABA-induced SnRK2.2/2.3/2.6 activation.
To evaluate which members of the B2 and B3 subgroups have predominant roles in ABA-regulated seed germination and seedling establishment, we backcrossed OK 100 -nonu with Col-0 wild type and screened F2 populations on 1/2 MS medium containing 10 µM ABA and sucrose. By genotyping 103 individual F2 seedlings with strong ABA insensitivity, we found that each RAF might contribute to ABA hyposensitivity (Chisquare test, p < 0.05). RAF3, RAF4, RAF5, and RAF7-9 (closed linked together) might have predominant roles (p < 0.0001) in ABA-regulated germination and seedling establishment (Supplementary Table 1).
Among the DEGs between wild type and the mutant, 67 genes showed significantly higher expression (>= 3-fold, p < 0.05) in the OK 100 -oct mutant under control conditions. Gene Ontology (GO) analysis indicated that these genes are enriched in plant response to chitin, fungus, bacterium, and oxidative stress, suggesting a potential role of the B2 and B3 RAFs in these biotic stress responses (Supplementary Data 4).
To further evaluate the role of B2 and B3 RAFs in ABAinduced gene expression, we used transient activation assays. We used the LUCIFERASE (LUC) reporter gene driven by the ABAresponsive RD29B promoter as an indicator of ABA response 8 . In the wild type, ABA clearly induced the expression of RD29B-LUC in the mesophyll cell protoplasts, while ABA-induced RD29B-LUC expression was completely abolished in the protoplasts of OK 100 -oct, snrk2-triple, pyl112458, and abi1-1 mutants (Fig. 5d). Co-expression of RAF3, RAF5, or RAF11 fully rescued the ABAinduced expression of RD29B-LUC, and co-expression of RAF4, RAF6, RAF7, RAF9, or RAF10 partially rescued the ABA-induced expression of RD29B-LUC in the protoplasts of OK 100 -oct mutant (Fig. 5e). No rescue was seen with co-expression of RAF1, RAF2, RAF8, or RAF12 (Fig. 5e).
ABA does not activate B2 and B3 RAFs in plants. Unlike the strong activation of RAFs by hyperosmolarity, the kinase activity of RAFs was not enhanced by ABA treatment (Fig. 4f). Consistently, the phosphorylation of pSTAGTPEWMAPEVLR, a conserved peptide located in the activation loop of RAF2/EDR1 and RAF3, was not affected by ABA treatment but highly induced by osmotic stress caused by mannitol treatment (Fig. 6a, b,  Supplementary Fig. 8a). Multiple phosphosites in this region also could be detected without ABA treatment 27,36 (Fig. 6b, Supplementary Fig. 8a, b, highlighted in Supplementary Data 5). SnRK2.6 showed clearly induced phosphorylation in the peptide containing the phosphorylation site Ser175 by both ABA and mannitol treatments (Supplementary Fig. 8c). We further tested whether RAF3 phosphorylation is required for its activity on SnRK2.6 by generating a non-phosphorylatable mutation of RAF3. Co-transfection of RAF3 S763AS766AT770A , with Ser to Ala substitutions in the activation loop, did not rescue the ABAinduced RD29B-LUC expression in the protoplasts of OK 100 -oct (Fig. 6c). RAF3 S763AS766AT770A completely lost its ability to phosphorylate SnRK2.6 in vitro (Fig. 6d, left panel). However, mutating these conserved residues of RAF10, producing RAF10 T706AT709AT713A , hardly affected RAF10 activity in in vitro kinase or transient expression assays (Fig. 6d, right panel, Supplementary Fig. 8d). Therefore, the activation mechanism of RAF10 in the B2 RAF subgroup may differ from that of RAF3 in c Expression of the ABA-inducible marker genes in wild type, OK 100 -oct, and OK 100 -nonu seedlings after 6 h of ABA treatment. Error bars, SEM (n = 3 biological replicates). Two-tailed paired t-tests, *p < 0.05, **p < 0.01, ***p < 0.001. d The activation assay of the ABA-responsive RD29B-LUC reporter gene in wild type and mutants. The protoplasts were transformed with the reporter plasmid and incubated with or without 5 µM ABA for 5 h under light. Error bars, SEM (n = 3 individual transfections). e Add-back assay testing RAF1 to RAF12 in activating the reporter gene in the protoplasts of OK 100 -oct. Error bars, SEM (n = 4 individual transfections). f Activation of the reporter gene by the combinations of RAFs with SnRK2.2, SnRK2.3, or SnRK2.6 in the protoplasts of OK 100 -oct. The ratio of RD29B-LUC expression in the protoplasts with 5 µM ABA relative to that without ABA treatment was used to indicate the activation activity of RAF-SnRK2 pairs. Error bars, SEM (n = 4 individual transfections). g Activation of the reporter gene by RAF3 or RAF7 in the protoplasts of wild type, pyl112458, or snrk2-triple. Error bars, SEM (n = 3 individual transfections). Source data are provided as Source Data files. the B3 RAF subgroup. Taken together, these results indicate that B2 and B3 RAFs have basal levels of phosphorylation and activity under normal conditions and that application of ABA does not increase their phosphorylation.

Discussion
Subgroup B RAFs belong to the MITOGEN-ACTIVATED PROTEIN KINASE KINASE KINASE (MAPKKK) family due to their similarity with animal B-Raf protein kinases [37][38][39] . In a canonical MAPK cascade, MAPKKKs are activated by extracellular signals and phosphorylate and activate MAPK KINASEs (MAPKKs), which then phosphorylate and activate MAPKs to regulate various cellular processes. Instead of phosphorylating MAPKKs, after rapid activation by osmotic stresses, B2, B3, and B4 RAFs phosphorylate and activate SnRK2s 27-31 . Plants use this noncanonical RAF-SnRK2 cascade to relay early osmotic stress signaling. Here we show a unique initiation-amplification mechanism of RAF-SnRK2 in ABA signaling to ensure rapid activation of SnRK2s with a basal level of Raf kinase activity (Fig. 7). Unlike the B4 subgroup RAFs, which are quickly activated by hyperosmolality 27 , the application of exogenous ABA does not enhance the phosphorylation or activity of B2 and B3 subgroup RAFs, as indicated by in-gel kinase assays and phosphoproteomics (Figs. 4 and 6). The phosphoproteomics and the in-gel kinase assay result also revealed the existence of basallevel phosphorylation and activity of B2 and B3 RAFs even without ABA (Figs. 4f and 6a, Supplementary Data 5). This basal level activation of RAFs might be necessary and sufficient for maintaining the SnRK2 activity and ABA signaling required for normal growth and development 6 . The upstream kinases or other mechanisms for this basal activity of RAFs need to be determined in the future. In the presence of ABA, the ABA and PYR1/PYL/ RCAR complex releases SnRK2s from PP2C-mediated inhibition, resulting in the accumulation of uninhibited forms of SnRK2s. The RAFs quickly trans-phosphorylate uninhibited SnRK2s to initiate SnRK2 activation. The activated SnRK2s then intermolecularly transphosphorylate and activate other SnRK2 molecules not yet activated by RAFs, to amplify the ABA signaling (Fig. 7). By this activation-amplification mechanism, the basal level activity of RAF kinases is sufficient to quickly activate SnRK2s to transduce the ABA signal. It would be interesting to apply the ATP analog-based method to evaluate whether this activation-amplification mechanism also exists in other kinase cascades in plants or animal cells, as autophosphorylation is a general feature of many protein kinases.
Several B3 subgroup Raf-like kinases, M3Kδ6/SIS8/RAF5, M3Kδ7/RAF4, M3Kδ1/RAF3, and RAF6, phosphorylate SnRK2.6, and are essential for ABA-induced SnRK2 activation 27,28 . In this study, we show that the B2 subgroup RAFs, together with B3 subgroup, have an essential role in the ABA core signaling pathway. OK 100 -nonu and OK 100 -oct show strong ABAinsensitivity in germination, leaf yellowing, and stomatal closure. The OK 100 -nonu seeds even germinate at an ABA concentration of  (Fig. 4). To our knowledge, OK 100 -oct and OK 100nonu are among the few mutants, including snrk2-triple, pyl112458 and pyl-duodec, that can germinate on such an extremely high concentration of ABA, further supporting the critical role of the B2 and B3 RAFs in ABA responses. However, ABAinsensitivity of OK 100 -nonu is still not identical with that of snrk2triple and pyl high-order mutants. This suggests that, besides B2 and B3 RAFs, additional protein kinases also participate in ABAinduced SnRK2 activation. At least two protein kinases, BRASSINOSTEROID-INSENSITIVE2 (BIN2) and BRASSINOS-TEROID INSENSITIVE 1-ASSOCIATED RECEPTOR KINASE 1 (BAK1), can phosphorylate SnRK2s and are involved in ABA signaling 40,41 . BIN2 likely phosphorylates only SnRK2.2 and SnRK2.3, but not SnRK2.6 41 , at the conserved threonine corresponding to Thr179 in SnRK2.6 ( Supplementary Fig. 2). Thus, whether BIN2, or other members of the GSK family, cooperate with RAFs in amplifying ABA-triggered SnRK2 activation, needs to be further studied.
Although both B2 and B3 RAFs are essential for ABA-induced SnRK2 activation, they might have distinct roles in some ABAregulated biological processes. OK 100 -B3, but not OK 100 -B2, has arrested growth in soil, suggesting a unique role of B3 subgroup RAFs in growth regulation (Fig. 3). OK 100 -B2 shows similar ABAinsensitivity on medium with or without sucrose, whereas OK 100 -B3 only shows strong ABA-insensitivity with exogenous sucrose (Fig. 3). Supporting this notion, the miRNA-M3K was screened from the medium containing sucrose 28,42 . RAF3-5 and RAF7-9 might have more dominant roles in germination and seedling establishment, while RAF3, RAF5, RAF7, and RAF11 rescue the ABA-induced RD29B-LUC expression more robustly. Several key regulators in ABA signaling and synthesis are also involved in sugar responses 43,44 . Mutants raf1/ctr1/sis1, and raf5/sis8, are resistant to high concentrations of sugar 45,46 . Together with these findings, our results suggest a crucial role of B3 subgroup RAFs in sucrose signaling and/or in seed development (e.g., in the accumulation of energy reserve in the seeds). It is notable that the role of RAFs in rescuing ABA-induced RD29B-LUC transcription is not identical to their contribution in gemination and seedling establishment, further indicating that different RAFs have various contributions to different ABA-mediated biological processes. Such functional diversity is also observed in the 14 PYR1/PYL/ RCAR ABA receptors. Only pyl112458, but not 3791112 (pyl3/7/ 9/11/12) shows arrested growth under normal condition 33 . By contrast, pyl112458 has more predominant roles in ABAmediated regulation of germination, stomatal movement, etc. 33 . In guard cell, PYL2 is sufficient for guard cell ABA-induced responses, whereas in the responses to CO 2 , PYL4 and PYL5 are essential 47 . PYL8 directly binds to the transcription factor MYB77 to regulate auxin responsive gene expression 48 . PYR1 especially participates in cross-talk between salicylic acid and ethylene, thereby redirecting defense disease resistance towards fungal Plectosphaerella cucumerina 49 . In Arabidopsis, 14 PYLs, at least 8 PP2Cs, three SnRK2s, and 12 members of the B2 and B3 RAF subgroups comprise a complex network in ABA sensing and signaling, which may ensure that plants precisely respond to everchanging environments. The engineering of ABA receptors is an efficient way to improve stress resistance in both Arabidopsis and crops [50][51][52] . Our findings regarding B2 and B3 RAFs in stress signaling provide additional targets (e.g., ectopic expression of stress-inducible or constitutively activated forms of RAFs in guard cells) for engineering crops resistant to harsh environmental conditions.
Besides involvement in sugar and ABA signaling, CTR1/RAF1 is a crucial component in ethylene signaling. We excluded RAF1/ CTR1 from the OK 100 high-order mutants because the ctr1 mutant displays severe growth inhibition under normal conditions 35 . However, although the KD of RAF1/CTR1 strongly phosphorylates SnRK2.6 in vitro, neither the full-length RAF1/ CTR1 nor RAF1/CTR1-KD rescued the ABA-induced expression of RD29B-LUC in the protoplasts of OK 100 -oct ( Supplementary  Fig. 8e). Thus, additional mechanisms may determine RAF1 specificity in vivo. Similarly, RAF2, RAF8, and RAF12 only show weak activities on the induction of RD29B-LUC expression in the protoplasts. The roles of these RAFs in ABA signaling therefore need to be further investigated.
The phosphorylation of Ser1029 of the ABA AND ABIOTIC STRESS-RESPONSIVE RAF-LIKE KINASE (PpARK)/PpCTR1 in Physcomitrella patens is induced by exogenous ABA in P. patens [53][54][55] , which is inconsistent with our observations on RAF3 and RAF10 (Fig. 6). Therefore, P. patens and higher plants may adopt different machinery to relay ABA signaling. In addition, non-phosphorylatable mutations at Ser1029 in PpARK/PpCTR1, or Ser763Ser766AThr770 in RAF3, abolished their kinase activities, suggesting phosphorylation-dependent activation of PpARK/PpCTR1 and RAF3. By contrast, the activation of RAF10 might be independent of phosphorylation. In animal cells, RAF kinases can be activated through phosphorylation, dimerization, or by binding of small GTPases, scaffold protein, 14-3-3 proteins, etc. [56][57][58] . Future work will investigate the phosphorylation or other activation mechanisms of RAF-SnRK2 cascades in different plant species and their roles in plant adaptive plasticity.

Methods
Seed germination and plant growth assay. Seeds were surface-sterilized in 70% ethanol for 10 min, followed by four times washing with sterile-deionized water. For the germination assay, seeds were sown on 1/2 Murashige and Skoog (MS) medium (0.75% agar, pH 5.7) with or without the indicated concentrations of ABA and 1% sucrose. Plates were kept at 4°C for 3 days in darkness for stratification and then shifted to a plant growth chamber set at 23°C and a 16 h light/8 h dark photoperiod. After 72 h of transfer, radical emergence was examined, and photographs of seedlings were taken at the times indicated. For growth assays, seeds were placed on 1/2 MS medium (0.75% agar, pH 5.7) and plates were placed vertically in a plant growth chamber after 3 days of stratification. After 3-4 days, the seedlings were transferred to medium with or without the indicated concentrations of ABA. Root length and fresh weight were measured at the indicated days. For seed dormancy assays, fresh seeds were harvested and sown on 1/2 MS medium (0.75% agar, pH 5.7) and plates were placed in a plant growth chamber. Radical emergence was measured 48 h after transferring.
Generation of OK 100 high-order mutants. The clustered regularly interspaced short palindromic repeats/CRISPR-associated 9 (CRISPR-Cas9) and guide RNA fragment from pCAMBIA-2300-11RAFs 27 was cloned into pCAMBIA-1300. The resulting vectors containing sgRNAs targeting B2, B3, or B2/B3 RAFs were used to transform wild type to generate OK 100 -B2, OK 100 -B3, OK 100 -oct, and OK 100 -nonu. The transgenic plants were screened for hygromycin resistance. The T1 transformants were identified by sequencing the fragments with the RAF target regions, which were amplified by PCR using primer pairs listed in Supplementary Data 6.
In-gel kinase assay. For in-gel kinase assays, 20 µg extract of total proteins were electrophoresed on 10% SDS/PAGE embedded with histone in the separating gel as a substrate for kinase. The gel was then washed three times at room temperature for 30 min each with washing buffer (25 mM Tris-Cl, pH 7.5, 0.5 mM Dithiothreitol (DTT), 0.1 mM Na 3 VO 4 , 5 mM NaF, 0.5 mg/mL BSA, and 0.1% Triton X-100). The kinase was allowed to renature in renaturing buffer (25 mM Tris-HCl, pH 7.5, 1 mM DTT, 0.1 mM Na 3 VO 4 , and 5 mM NaF) and incubated at 4°C overnight with three changes of renaturing buffer. The gel was further incubated at room temperature in 30 mL reaction buffer (25 mM Tris-Cl, pH 7.5, 2 mM EGTA, 12 mM MgCl 2 , 1 mM DTT, and 0.1 mM Na 3 VO 4 ) with 200 nM ATP plus 50 µCi of [γ-32 P]ATP for 90 min. The reaction was stopped by transferring the gel into 5% (w/v) trichloroacetic acid and 1% (w/v) sodium pyrophosphate. The gel was then washed to remove unincorporated [γ-32 P]ATP in the same solution for at least 5 h with five changes. Radioactivity was detected with a Personal Molecular Imager (Bio-Rad).
RNA sequencing and data analysis. Total RNA was isolated from two-week-old seedlings of Col-0 and OK 100 -oct mutant, with and without ABA treatment, using RNeasy Plant Mini Kit (Qiagen). Total RNA (1 µg) was used for library preparation with NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (New England BioLabs, E7765) following the manufacturer's instructions. Prepared libraries were assessed for fragment size using NGS High-Sensitivity kit on a Fragment Analyzer (AATI), and for quantity using Qubit 2.0 fluorometer (Thermo Fisher Scientific) and KAPA Library Quantification Kit (Kapa, KK4824). All libraries were sequenced in paired-end 150 bases protocol (PE150) on an Illumina Nova sequencer.
The paired-end reads were cleaned by Trimmomatic 59 (version 0.39). After trimming the adapter sequence, removing low quality bases, and filtering short reads, clear read pairs were retained for further analysis. The Arabidopsis thaliana reference genome sequence was downloaded from TAIR10. Clean reads were mapped to the genome sequence by HISAT (2.1.0) 60 with default parameters. Number of reads that were mapped to each gene was calculated with the htseqcount script in HTSeq (0.11.2) 61 . EdgeR 62 was used to identify genes that were differentially expressed. Genes with at least three-fold change in expression and with an FDR < 0.05 were considered differentially expressed genes (DEGs).
Analysis of gene expression by qRT-PCR. Total RNA was extracted from twoweek-old wild-type, OK 100 -oct, andOK 100 -nonu seedlings with or without 50 µM ABA treatment for 6 h. Total RNA was isolated using the RNeasy Plant Mini Kit (Qiagen) according to the manufacturer's instructions. Genomic DNA was removed using RNase-free DNase and subsequently, 1 µg of total RNA was reverse transcribed using the iScript TM gDNA Clear cDNA Synthesis Kit (Bio-Rad) following the manufacturer's instructions. The actin gene was used as an internal control. Quantification was performed using three independent biological replicates.
Water loss measurement. The water loss was estimated on detached rosette leaves of 4-week-old plants by weighing using a weighing dish. Leaves were then kept on the laboratory bench for at least 30 min. Fresh weight was monitored before and after the procedure and at the times indicated. Water loss was expressed as a percentage of initial fresh weight.
Stomatal bioassay. For stomatal aperture assay, rosette leaves of 4-week-old Arabidopsis seedlings were taken. Epidermal strips were peeled out and incubated in buffer containing 50 mM KCl, 10 mM MES, pH 6.15, in a plant growth chamber for 3 h before ABA treatment. Stomatal apertures were measured 2 h after the addition of 5 μM ABA. The apertures of about 60 stomata per sample were measured by quantifying the pore width of stomata using Image J software (1.51 K). All the experiments were repeated at least three times.
Protoplast isolation and transactivation assay. Protoplasts were isolated from leaves of 4-week-old plants grown under a short photoperiod (10 h light at 23°C/14 h dark at 20°C). Leaf strips were excised from the middle parts of young rosette leaves, dipped in enzyme solution containing cellulase R10 (Yakult Pharmaceutical Industry) and macerozyme R10 (Yakult Pharmaceutical Industry) and incubated at room temperature in the dark. The protoplast solution was diluted with an equal volume of W5 solution (2 mM MES, pH 5.7, 154 mM NaCl, 125 mM CaCl 2 , and 5 mM KCl) and filtered through a nylon mesh. The flow-through was centrifuged at 100 g for 2 min to pellet the protoplasts. Protoplasts were resuspended in W5 solution and incubated for 30 min. 100 μL of protoplasts suspended in MMG solution (4 mM MES, pH 5.7, 0.4 M mannitol, and 15 mM MgCl 2 ) were mixed with the plasmid mix and added to 110 μL PEG solution (40% w/v PEG-4000, 0.2 M mannitol, and 100 mM CaCl 2 ). The transfection mixture was mixed completely by gently tapping the tube followed by incubation at room temperature for 5 min. The protoplasts were washed twice with 1 mL W5 solution. After transfection, protoplasts were left for incubation for a further 5 h under light in washing and incubation solution (0.5 M mannitol, 20 mM KCl, 4 mM MES, pH 5.7) with or without 5 μM ABA. The RD29B-LUC (7 μg of plasmid per transfection) and ZmUBQ-GUS (1 μg per transfection) were used as an ABA-responsive reporter gene and as an internal control, respectively. For wildtype and mutated RAF, SnRK2 plasmids, 3 μg per transfection were used. After transfection, protoplasts were incubated for 5 h under light in washing and incubation solution (0.5 M mannitol, 20 mM KCl, 4 mM MES, pH 5.7) with or without 5 μM ABA. The mutations were introduced into wild-type RAFs using the primers listed in Supplementary Data 6.