Our previous study on rice PD-proteome analysis identified several members of the RLK family resided in the PD region, including an unpublished Os02g20400 (Jo, 2011; Jo et al., 2011). By using bioinformatic analysis, we identified an Arabidopsis RLK protein family member (AT2G28250, named NCRK) with the highest identity compared with Os02g20400 (named OsNCRK) ( Figure 1A ). This protein was formerly reported as an NCRK, a member of the RLK family with an unknown function ( Figure 1B ) (Molendijk et al., 2008). In Arabidopsis, NCRK could be hyperphosphorylated and partially co-localized with the small GTPase ARABIDOPSIS RAB HOMOLOG F2A (RapF2a) in endosomes (Molendijk et al., 2008). The amino acid alignment between RLK orthologs reveals that the extracellular receptor domain encoded by NCRK protein contains two repeats of about 40 amino acids sharing a unique and highly conserved Cys-rich motif, WXCXCX13–18CX3CXC across the species, suggesting their specific function ( Figure 1B ).

First, we examined the subcellular localization of OsNCRK and NCRK in the transient Nicotiana benthamiana plant by transiently expressing green fluorescent protein (GFP) fusion to the C-terminal of both proteins, followed by fluorescence imaging using confocal microscopy. OsNCRK and Arabidopsis NCRK showed punctate spots at the cell periphery; a typical pattern was observed resembling PD localization. To confirm if the spots are co-localized with PD, we co-expressed NCRK-GFP with PD markers such as Turnip vein-clearing virus (TVCV) movement protein, callose staining with Aniline Blue dye, PDLP1 and PDLP5 (Thomas et al., 2008; Jo et al., 2011; Lee et al., 2011). NCRK-GFP fluorescence spots co-localized well with the four PD markers ( Figure 1C and Supplementary Figure 1A ). To quantify the percentage of NCRK co-localization at the PD area, we measured the PD index by calculating the ratio of mean fluorescent intensity of the PD enrichment area and its PM neighbor area ( Supplementary Figure 1C ). NCRK was highly localized in the PD area, similarly with PDLP1 or PDLP5, but not with RECEPTOR DEAD KINASE1 (RDK1) used as PM markers (Kumar et al., 2017) ( Supplementary Figure 1B ). Analyses of transgenic seedlings expressing a GUS gene driven by Arabidopsis NCRK promoter (a) showed that the underlying gene (gus) is broadly expressed throughout the whole seedling and more prominently in the root, hypocotyl, and cotyledon. In cotyledon, relatively strong expression was observed in vascular tissues but to less extent in mesophyll cells ( Supplementary Figure 2B ). Analyses of pNCRK::NCRK-mCherry transgenic lines showed that the fluorescence was co-localized or localized to the vicinity of the aniline blue-stained callose ( Figure 1D ). To investigate the potential function of NCRK in PD regulation, we obtained T-DNA insertion lines (ncrk-1, ncrk-2) ( Supplementary Figure 2A ) and performed the PD callose staining experiment. ncrk-1 and ncrk-2 mutants showed less basal callose accumulation ( Figure 1D and Supplementary Figure 3C ). ncrk-1 mutant plants exhibited increased PD permeability and decreased PD callose compared with wild-type plants in the leave and hypocotyl system ( Figure 1E and Supplementary Figures 3A, B ), suggesting the role of NCRK in callose deposition and PD permeability regulation. The callose turnover of the plant showed defective growth and development or was even lethal except for some mutant genes such as gsl4. Interestingly, the ncrk-1 mutant plant displayed enhanced developmental growth compared with the wild-type, similar to the gsl4 mutant plant, showing a reduced PD callose level and increased PD permeability ( Figure 1F ). These data indicate that NCRK is a plasmodesmal RLK protein family member, which controls basal callose accumulation and PD permeability.

It has been shown that the Cys residue in the ED of RLK could perceive the eH₂O₂ and respond to the ROS stress (Hunter et al., 2019; Wu et al., 2020). Therefore, we investigated whether NCRK could be involved in ROS-dependent callose accumulation. As described earlier, the spraying of 10 mM H₂O₂ on the leaf surface of wild-type plants was reported to stimulate ROS-mediated callose accumulation after 2 hours of treatment (Cui and Lee, 2016). Therefore, we focused on the role of NCRK in ROS-induced callose accumulation by eH₂O₂ application. We first observed that callose deposition and PD permeability of the ncrk-1 mutant plant was impaired in 10 mM H₂O₂ treatment ( Figures 2A, B ). To confirm that the basal callose and ROS-induced callose impaired phenotype of ncrk-1 was indeed caused by the NCRK mutation, we showed that NCRK could rescue the ncrk-1 phenotype and callose accumulation was restored in the complementation lines upon eH₂O₂ treatment ( Figures 2E , 3A ). NCRK contains a highly conserved and unique Cys-rich repeat sequence in the ED ( Figure 1B ). To explore whether these extracellular Cys residues are essential for eH₂O₂ sensing, we performed ncrk-1 complementation using NCRK mutant, which included Cys mutations ( Figures 2C, D ). NCRK mutants with two (Cys30 and Cys32 to Ala; 2CA), three (Cys46, Cys50, and Cys52 to Ala; 3CA), or all five Cys residues to Ala (5CA) could complement the basal callose level of ncrk-1 but not ROS-dependent callose phenotype ( Figure 2F ), suggesting that NCRK-ED might not affect basal callose accumulation, but the Cys-rich motif seems specifically involved in ROS-induced callose regulation. The HPCA1 contains the Cys residues, which could sense the eH₂O₂ in Arabidopsis (Wu et al., 2020). To investigate whether the Cys residues of ED in NCRK could respond to eH₂O₂, we pursued a domain-swapping approach replacing the ED and transmembrane domain (TMD) of NCRK with the ED and TMD of HPCA1 that yielded the chimeric H/N (HPCA1/NCRK) protein. The predicted structure of the H/N receptor was comparable with the native NCRK protein except for the ED region ( Figure 3A ). However, the chimeric H/N protein was localized at PM but not in the PD, as observed in the transient N. benthamiana assay ( Figure 3B ).

Further, Arabidopsis transgenic plants in the ncrk-1 mutant background with a chimeric H/N protein driven by the endogenous NCRK promoter were generated, and eH₂O₂ treatment was applied ( Figure 3A ). We found that the chimeric H/N protein could not complement the callose in normal growth conditions but partially rescued the impaired callose deposition phenotype in the case of eH₂O₂ treatment ( Figures 3C, D ). These data suggest that NCRK-ED may serve a similar function as the HPCA1-HP domain regarding the H₂O₂-induced callose response. Overall, these results demonstrate the role of NCRK in eH₂O₂ perception at PD, and the Cys-rich motif of NCRK-ED might be involved in eH₂O₂-induced callose deposition.

Changes in eH₂O₂ and intracellular H₂O₂ were found to be related to an increase in Ca²⁺ and consequently activate several CML protein family members (Magnan et al., 2008; Leba et al., 2012; Wu et al., 2017; Xu et al., 2017). Also, the CML members are localized at the nucleus, cytoplasm, and PM, and some are co-localized with NCRK ( Supplementary Figure 4 ). This subcellular co-localization might suggest the interaction of NCRK and the other partners. We hypothesized that NCRK might control ROS-induced callose deposition by directly interacting with CMLs members or GSL members and regulating its activity. To understand how NCRK controls the callose deposition during ROS stress and to identify the interaction partners of NCRK, we performed bimolecular fluorescent complementation (BiFC), co-immunoprecipitation (Co-IP) assay, and genetic experiments. As a positive control, RDK1-YFPⁿ and ABI4-YFP^c resulted in strong reconstituted YFP signals at PM (Kumar et al., 2017). However, a fluorescence signal was not observed in the leaves co-expressing NCRK-YFPⁿ and ABI4-YFP^c ( Figure 4A ). We detected a YFP signal on the PM when the NCRK-YFPⁿ was co-expressed with the YFP^c fused to CML19, CML20, CML41, GSL4, and GSL8. This data indicates that NCRK showed strong or weak interactions with several CML (CML19, CML20, CML41) and GSL (GSL4, GSL8) family members. We performed Co-IP assays to confirm the in planta interaction of NCRK with the CML and GSL family members. The result showed that NCRK coimmunoprecipitates with CML41, GSL4, and GSL8 but not with negative control (GFP) ( Figure 4B ). Consequently, BiFC and Co-IP observations revealed that NCRK strongly interacts with CML41, GSL4, and GSL8 in vivo. To test if the ROS can induce the interaction of NCRK and CML41, we performed the BiFC assay under the eH₂O₂ treatment. After 2 hours of post-spraying, the reconstituted YFP fluorescence resulting from NCRK-YFPⁿ and CML41-YFP^c was detected in the membrane and punctate spots. When treated with eH₂O_2, the accumulation of NCRK-CML41 interaction signal intensity was increased in leaves but not in RDK1-ABI4 (positive control) and NCRK-CML20 signal intensity ( Figure 4C ). These results suggest that eH₂O₂ may activate NCRK, stimulating the interaction between NCRK and CML41.

Since we observed the direct interaction of NCRK and GSL members, GSL4 and GSL8, we analyzed the role of GSL members in the ROS-dependent pathway of NCRK. To determine whether GSL4 is working in the NCRK-related pathway, we generated the ncrk gsl4 double mutant plant by crossing the ncrk-1 and gsl4 mutant. The callose intensity of the double mutant was similar to the ncrk-1 and gsl4 single mutants in normal growth and ROS-treated conditions ( Figure 4D ). It suggests that GSL4 acts in the NCRK-dependent-ROS pathway. Our data demonstrated that NCRK, CML41, and GSL4 are responsible for ROS-induced callose deposition pathways.

NCRK shares similar conserved catalytic subdomains with CRK2 and HPCA1, which require kinase activity for their functions (Stone and Walker, 1995; Kornev et al., 2006) ( Supplementary Figure 5 ). To explore whether NCRK (NCRK^CD) cytoplasmic domain has kinase activity or functions as a scaffold, we prepared recombinant proteins purified from E. coli. We could detect the autophosphorylation and phosphorylation of myelin basic protein (MBP) only with wild-type NCRK^CD but not with the kinase-inactive mutants NCRK(K238E)^CD (ATP binding) and NCRK(D339L)^CD (catalytic) ( Figure 5A ). Since NCRK physically interacts with CML41, we tested whether NCRK phosphorylates the CML41 in vitro. The ROP4 was used as a positive control, and NCRK could phosphorylate ROP4 and CML41. NCRK also could phosphorylate CML20, but it reduces the activity of NCRK ( Figure 5B ). The ncrk-1 transgenic plants with constructs of these two kinase inactive variants driven by the endogenous NCRK promoter were generated. The complementation lines showed partially rescued basal callose levels in mock condition but significantly reduced callose levels compared to wild-type plants under the ROS stress condition ( Figure 5C ). These data revealed that the NCRK phosphorylates CML20 and CML41 in vitro, and NCRK phosphorylation status influences ROS-induced callose accumulation at PD.

Since the ncrk-1 mutant showed the deficiency in the callose deposition under the eH₂O₂ treatment, we tested whether ROS production triggered by another abiotic and biotic stress in the plant could affect the callose deposition in NCRK mutant background. Previously, herbicide methyl viologen (MV) treatment was shown to induce ROS production in chloroplast and mitochondria (Cui et al., 2019). The chloroplast and peroxisomes are the primary sources of intracellular ROS in plants (Noctor et al., 2002). To determine the effect of MV application on callose deposition, wild-type plants were air sprayed with different MV concentrations (0.1, 1, and 5 μM). Wild-type plants sprayed with 5 μM of MV showed considerably higher callose deposition at 2 hours post-treatment ( Supplementary Figure 6 ). Similarly, a higher level of callose accumulation was observed in the ncrk-1 mutant plants sprayed with 5 μM of MV ( Figure 5D ). This data revealed that NCRK is independent of MV-induced callose deposition. Next, we tested whether NCRK is required for apoplastic ROS-induced callose deposition. Since GSL4 is specifically responsible for apoplasmic ROS-mediated PD regulation during wounding stress and NCRK acts as a mediator of GSL4 in eH₂O₂ response, we quantified callose level after mechanical wounding to wild-type and ncrk-1 mutant plants. This experiment showed the ncrk-1 mutant is impaired to gain of callose intensity within 2 hours post-wounding ( Figure 5E ). These data suggest that NCRK plays a role in wounding-dependent callose deposition.

The previous study has shown that CML41 is a responsive component of callose deposition-pathogen defense, and an flg22 (P. syringae) also triggers apoplastic ROS in plant immunity (Xu et al., 2017). To examine the function of NCRK mediating ROS-induced plant signaling network during pathogenic stress, we infiltrated wild-type and ncrk-1 mutant plants with flg22 to assess the impact of the altered callose regulation. After 24 hours post-infiltration, PD callose of ncrk-1 mutant plants was slightly reduced compared to wild-type plants ( Figures 5F, G ). Additionally, we focused on assessing the significance of NCRK activity in bacterial resistance by surface inoculation with P. syringae. After three days post-infection (3 dpi), ncrk-1 mutant plants showed increased bacterial growth compared to wild-type plants, indicating that NCRK might be involved in an flg22-induced PD closure pathway.

Seedlings of Arabidopsis thaliana, ecotype Columbia-0 (Col-0) were grown in soil or in petri dishes in half-strength Murashige and Skoog (MS) salts (Duchefa Biochemie, P.O. Box 809 2003 RV Haarlem, The Netherlands), 1.0% (w/v) sucrose (BioShop, Canada Inc., 5480 Mainway, Burlington, Ontario L7L 6A4, Canada), and 0.6% (w/v) agar (Duchefa Biochemie) in controlled environmental rooms or plant growth chambers (20-22°C). The fluency rate of white light was approximately 100 μmol m^-2s^-1. The photoperiods were 16 hours/8 hours light/dark cycles. For seedling culture, seeds were sterilized with 20% bleach for 5 min, washed with sterilized water, and then sown on soil or half-strength MS media, placed at 4°C for 3 days in the dark, and then transferred to growth rooms or chambers. The hydroponic assay was performed in half-strength of MS media, and plants were grown at the culture room condition (22°C under 16 hours/8 hours light/dark cycles). T-DNA insertion mutants were obtained from the Arabidopsis Biological Resource Center (https://abrc.osu.edu). The T-DNA insertion mutants of NCRK (SALK_045757 and SALK_202953 designated as ncrk-1 and ncrk-2, respectively) identified by T-DNA insertion-based PCR using T-DNA left-border primer (LBb1.3) and two pairs of NCRK-specific primers (LP1-RP1 and LP2-RP2), respectively.

Golden Gate cloning system (Werner et al., 2012) was used to generate pNCRK::NCRK-mCherry, pNCRK::NCRK(C30A/C32A)-mCherry, pNCRK::NCRK(C46A/C50A/C52A)-mCherry, pNCRK::NCRK(C30A/C32A/C46A/C50A/C52A)-mCherry, pNCRK::NCRK^K238E-mCherry, pNCRK::NCRK^D339L-mCherry, pNCRK::HPCA1^ED-NCRK^CD-mCherry. The full-length NCRK genomic region (2 kb upstream and 1 kb downstream of the NCRK coding region), HPCA1 ED region (2 kb), NCRK promoter region (2 kb), and NCRK1 full-length genomic were amplified by PCR from genomic DNA. These fragments were cloned into the pAGM4723 expression vector (Addgene #48015). Transgenic Arabidopsis lines were generated by Agrobacterium-mediated transformation (Zhang et al., 2006), and homozygous transgenic T3 lines carrying a single insertion were used. The ncrk-1 mutant was complemented by a genomic fragment and NCRK-mCherry driven by the native NCRK promoter. The site-directed mutagenesis was used to introduce Cys-to-Ala mutations into the pNCRK::NCRK-mCherry plasmid. For the pNCRK::GFP-GUS, p35S::CML19-GFP, p35S::CML19-GFP, p35S::CML20-GFP, was cloned by Gateway cloning. The NCRK promoter fragment, CML9, CML19, CML20, and CML41 cDNA were amplified from genomic DNA. These fragments were moved into pDONR207 plasmid (Invitrogen, 5791 Van Allen Way PO Box 6482 Carlsbad, California 92008, USA) and then cloned into the Gateway binary vector pKGWFS7, pMDC43, and pMDC83, respectively.

A phylogenetic tree was created by multiple sequence alignment method using MEGA (v11.0.9) (Tamura et al., 2021), in which bootstrap was performed with 1,000 replicates, and gaps were treated as pairwise deletions. The neighbor-joining method performed a bootstrap with 1,000 replicates for the phylogenetic trees. Multiple sequence alignments of the full-length amino acid sequence of Figure 1B (OsNCRK, GRMZM2G063897, Glyma.11G144800, Glyma.12G077300, and AtNCRK) and Supplementary Figure S5 (CRK2, HPCA, and NCRK) were aligned by ClustalW (https://www.ebi.ac.uk/Tools/msa/clustalo) (Sievers et al., 2011). The protein structure of NCRK, H/N fusion protein, and NCRK-ED were predicted by AlphaFold2 using MMseq2 (Mirdita et al., 2022). The 3D structure was visualized by the PyMOL program.

Histochemical GUS staining of the NCRK promoter-driven GUS (pNCRK::GFP-GUS) transgenic lines was detected by infiltration with 5-bromo-4-chloro-3indolyl-b-D-glucuronic acid (X-Gluc), as described in an earlier report (Kumar et al., 2017). Briefly, plant materials were incubated at 37°C overnight in a GUS staining buffer containing 2 mM X-Gluc, 2mM potassium ferricyanide, 2 mM potassium ferrocyanide, 0.2% Triton X-100, and 50 mM potassium phosphate buffer, pH 7.0 and followed by chlorophyll clearing in 75% ethanol. GUS images were taken using a stereomicroscope.

For analysis of NCRK-mCherry, HPCA1/NCRK-mCherry in stable transgenic Arabidopsis lines, seedlings grown in half-strength MS media in Petri dishes for 7 days were subjected to mCherry confocal imaging with the FV1000 or FV1000MPE confocal microscope (Olympus, Japan). Data represent more than 3 independent lines examined, which displayed similar mCherry subcellular localization. To analyze wild-type NCRK and HPCA1/NCRK protein subcellular location, the transient expression of fluorescent fusion proteins in N. benthamiana leaf tissue was performed as previously described (Iswanto et al., 2020). The full-length coding sequences of NCRK, HPCA1/NCRK were amplified from Col-0 and fused with p35S and mCherry into pAGM4723 vector by Golden Gate cloning. For BiFC assays, LR recombination of appropriated coding sequence regions (NCRK, CML9, CML19, CML20, CML41, GSL4, and GSL8) in pDONR207 were performed with the split-YFP destination vectors pDEST-^GWVYNE/pDEST-VYNE^®GW, which allow C-terminal fusions with C-terminal YFP moieties (Gehl et al., 2009). Agrobacterium tumefaciens strain GV3101 containing the binary plasmid was grown in LB medium with the appropriate antibiotics at 28°C for 24 hours for the primary culture and then subculture at 28°C for 12 hours. The cells were centrifuged for 15 min at 3500 rpm, resuspended to a final OD₆₀₀ of 0.8, and infiltrated into tobacco leaves by a needle-less syringe. Leaf pieces were viewed under a confocal microscope after 2-3 days.

For Co-IP in N. benthamiana, Golden Gate was used to clone PDRLK1 full-length genomic to generate HA-tagged fusions. pDONR207-CML19, CML20, CML41 and CML9, GSL8 and ROP4 were recombined into pMDC43 and pMCD83, respectively, to generate GFP-tagged fusions by Gateway cloning. To clone the p35S::GSL4-GFP construct, GFP with Nos terminator was amplified with additionally flanked with restriction sites for SmaI and SpeI. The amplicon was excised and cloned into the construct pDEST-^GWVYNE-GSL4 via SmaI/SpeI. These constructs were electroporated into A. tumefaciens strain GV3101 (pMP90). Agrobacterium suspensions carrying p35S::PDRLK1-HA, p35S::GFP-CML19, p35S::GFP-CML20, and p35S::GFP-CML41, p35S::GSL8-GFP, pDEST::GSL4-GFP were infiltrated in various combinations into N. benthamiana leaves (OD₆₀₀ = 0.3). One gram of leaf tissue was harvested at 48-72 hpi, and total protein was resuspended in 1 ml IP buffer (100 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 10% glycerol, 3 mM DTT, 1 mM Na₂MoO₄, 1.5 mM sodium orthovanadate [Na₃VO₄], 2 mM sodium fluoride [NaF], 1 mM phenylmethylsulfonyl fluoride [PMSF], 50 µM MG132, a complete protease inhibitor [Roche] and 0.5% IGEPAL). The samples were clarified by centrifugation at 13,000 rpm, at 4°C for 10 min twice. The supernatant was incubated with 10 μl of protein A agarose incubated with anti-GFP antibody for 2 hours at 4°C with 360° rotating shaker. Beads were washed five times with IP buffer. 40 μl of 2× SDS sample buffer was added to the beads, and the beads were heated at 70°C for 15 min and subjected to further SDS-PAGE and immunoblotting analysis (Kadota et al., 2016; Kumar et al., 2017).

For hypocotyl experiments, five-day-old, etiolated Arabidopsis seedlings were first cut at the base of the hook using a razor blade (Iswanto et al., 2020). A coverslip was placed between each hypocotyl cut surface and the MS agar plate. For dye loading, individual agar blocks containing 8-hydroxypyrene-1,3,6-trisulfonic acid trisodium salt (5 mg/ml) were placed on the cut hypocotyl surface. After a 5 min loading period, seedlings were washed in water for 15 min, and then fluorescent probe movements were evaluated by confocal microscopy. For leave experiments, intact 2.5 to 3.5-week-old Arabidopsis plants were air sprayed once with H₂O₂, MV, or wounding. Following a designated incubation time, the fifth and sixth entire leaves were subjected to DANS assays. DANS dye-loading assay using 1 mM 5(6)-carboxy-2’-7’-dichlorofluorescein diacetate for cell-to-cell movement assay as described (Cui and Lee, 2016). Briefly, the dye was loaded as a 1 µl droplet on the central regions of the adaxial side of each half-leaf blade, followed by confocal imaging of the abaxial side after 5 min loading and rinsing with a distilled water. Plasmodesmal callose staining was performed by staining the mature leaves or hypocotyl, for 1-1.5 hours in the dark before confocal imaging, with 0.02% aniline blue solution containing 0.6 M glycine-NaOH, pH 9.5, and 0.001% silwet L-77 (Han et al., 2014; Cui and Lee, 2016). Quantifying aniline blue stained plasmodesmal callose levels was performed using ImageJ as described before (Zavaliev and Epel, 2015) by automatically extracting plasmodesmata-associated fluorescent spots along the cell boundaries of epidermal leaf pavement cells and plotting fluorescence intensity per unit area.

All chemical treatments were performed by spraying them on intact and mature plants at various concentrations before DANS assays or plasmodesmal callose quantification. Mechanical wounding was done by snipping the tip of the fifth or sixth leaf with a pair of sharp forceps (Cui and Lee, 2016). Infected plants were grown in the plant chamber and subjected to DANS assays or plasmodesmal callose quantification 48 hours after infection. For flg22 treatment, either distilled water or distilled water containing 1 µM of flg22 was infiltrated into rosette leaves. After 24 hours, the infiltrated leaves were incubated in callose staining buffer (aniline blue) for 1-1.5 hours in the dark (Xu et al., 2017). For bacterial growth curve assay, we performed infections by surface inoculation with the less virulent, coronatine-deficient strain DC3000 (cor^-). Briefly, Pst cultured at 28°C and resuspended in MgCl₂ to final A₆₀₀ nm between 0.02 and 0.04 were generously infiltrated onto leaf abaxial surfaces of 5-6-week‐old plants. Plants were covered, and leaf discs were taken 3 dpi from three leaves per plant, with three plants per genotype per independent trial. Bacterial growth was assessed by colony counting (Kadota et al., 2014).

For detection of aniline blue-stained plasmodesmal callose, confocal images were taken under a 100X UplanXApo objective with oil, using 405-nm laser excitation with a 420 to 480 nm emission filter. Free or tagged mCherry was imaged with 543 nm laser excitation and a 587 to 683 nm bandpass emission filter. GFP was imaged with 488 nm laser excitation and a 505 to 550 nm bass pass emission filter. DANS and HPTS assay images were taken under a 10X UplanXApo objective using 488 nm Argon laser excitation, with a 505 to 550 nm band pass emission filter. All confocal images were acquired and processed using Olympus FV1000MPE or FV3000 confocal microscope or ImageJ software (Iswanto et al., 2020).

Recombinant protein purification and kinase assays were conducted as previously described (Harper and Speicher, 2011). NCRK^CD (wild type, (K238E)^CD and (D339L)^CD) and other substrates (CML20, CML41, and ROP4) were amplified and ligated into the pGEX4T-1 vector containing GST tag. Overexpression of those proteins in E. coli strain BL21(DE3) and purification on glutathione sepharose 4B R10 (GE Healthcare, USA) were conducted as described. For the kinase assay, these GST-fusion cytoplasmic domain kinase proteins (3 μg), GST (1 μg; negative control) MBP (0.5 μg; positive control) were incubated in kinase reaction buffer (25 mM Tris-HCl, pH 7.5, 1 mM DTT, 20 mM MgCl₂, 2 mM MnCl₂, and 50 μM [γ-³²P] ATP) and initiated using 1 μCi [γ-³²P] ATP. The final reaction volume was 20 μl. After 30 min at 30°C, the reaction was stopped by adding 6 μL of 4 × SDS sample buffer and boiling for 5 min. Samples containing 1 μg of protein from each autophosphorylation reaction were separated by SDS-PAGE, and the gel was stained, dried, and autoradiographed (Kim et al., 2017).

Column plots were created using GraphPad Prism version 9.0.0 for windows, GraphPad Software, San Diego, California, USA (www.graphpad.com). Student’s t-test was performed to test the statistical significance of differences.