It was previously reported that KIN10 degradation is strictly dependent on its kinase activity (Baena-González et al., 2007; Crozet et al., 2016). Since GRIK is the major kinase that phosphorylates and activates KIN10, we tested whether GRIK is involved in KIN10 degradation. GFP signal from GFP-tagged KIN10L was monitored upon transient co-expression of KIN10L with GRIK1 in Nicotiana benthamiana (N. benthamiana) leaves by fluorescence microscopy ( Figure 1A ) and western blotting with anti-GFP antibodies ( Figure 1B ). Previously, it was shown that Threonine-198 (T198) in the T-loop of KIN10L (equivalent to T175 in the T-loop of KIN10) is phosphorylated by GRIK and essential for KIN10 kinase activity (Shen et al., 2009). Consistent with previous reports, expression of the KIN10L T198A phosphorylation mutant resulted in increased protein accumulation relative to KIN10L ( Figures 1A–C ). Co-expression of GRIK1 with KIN10L greatly reduced KIN10L accumulation relative to the expression of KIN10L alone ( Figures 1A–C ). Mn²⁺-Phos-tag gel electrophoresis was used to separate phosphorylated from non-phosphorylated proteins, revealed that most of the residual KIN10L upon its co-transformation with GRIK1 was present in phosphorylated form ( Figure 1B ). There was only a single detected band for GFP-KIN10L (T198A) visible in the Mn²⁺-Phos-tag gel blot upon co-transformation with GRIK1-HA ( Figure 1B ), suggesting that T198 in the T-loop is the only phosphorylation site for GRIK1. That the T-loop phosphorylation mutant of KIN10L was strongly stabilized relative to its parental wild-type sequence also suggests posttranslational regulation. To confirm this, we compared the levels of GFP-KIN10L mRNA upon its expression alone versus upon its co-expression with GRIK1-HA. The levels of GFP-KIN10L transcripts were equivalent for both treatments ( Figure 1D ), confirming that the observed reduction in GFP-KIN10 protein occurs at the posttranslational level.

To further understand GRIK-mediated KIN10 degradation we engineered a GRIK1 mutant, K137A. K137 is a key residue in the ATP binding domain of GRIK1 reported to be essential for its kinase activity (Shen et al., 2009). Co-transformation of KIN10 with GRIK1(K137A) did not result in substantial KIN10 degradation, confirming that GRIK1-mediated KIN10L degradation is dependent on the kinase activity of GRIK1 ( Figures 1A, B , 2A, B ).

To substantiate GRIK1-mediated KIN10 degradation we observed in transient N. benthamiana leaf assays, KIN10 protein levels were quantified in two Arabidopsis grik double mutants: grik1-2 grik2-1 containing the weaker grik1-2 allele and grik1-1 grik2-1 containing the stronger grik1-1 allele of GRIK1. Consistent with published results (Glab et al., 2017), phosphorylated KIN10 was observed in the grik1-2 grik2-1 line, but not in grik1-1 grik2-1, as evidenced by probing with the phosphorylated KIN10-specific antibody. This confirms that GRIK1 and GRIK2 are major activating kinases of KIN10 in vivo ( Figure 3A ). Immunoblot assays probed with the KIN10-specific antibody showed that two distinct forms of KIN10 are detected in WT and the grik mutants but no KIN10 immunoreactive species are visible in the kin10 mutant. The higher molecular mass KIN10 form was significantly more abundant in grik1-1 grik2-1 than in either WT or grik1-2 grik2-1 ( Figure 3A ). Since KIN10 has two alternative splicing isoforms (i.e., KIN10L and KIN10) ( Figure 3B ), we hypothesize that the large and small KIN10s detected in grik1-1 grik2-1 are KIN10L and KIN10, respectively.

The individual coding sequence (CDS) corresponding to KIN10L or KIN10 was transiently expressed in N. benthamiana leaves. Three days after agroinfiltration, protein samples were extracted and separated along with protein samples from the grik1-1 grik2-1 double mutant and subjected to immunoblotting. Transiently expressed KIN10L and KIN10 polypeptides in N. benthamiana showed similar SDS-PAGE mobilities to those of the putative KIN10L and KIN10 products in grik1-1 grik2-1 respectively ( Supplementary Figure 1 ). To further confirm the identity of two sizes of KIN10 detected in the grik1-1 grik2-1 double mutant, KIN10s were immunoprecipitated with anti-KIN10 antibody from grik1-1 grik2-1 and separated by SDS-PAGE. The two protein bands were excised and analyzed with the use of tandem mass spectrometry. The faster migrating protein was confirmed to be KIN10. The slower migrating protein was identified as the KIN10L isoform ( Supplementary Figure 2 ). These data show that KIN10L, a minor KIN10 isoform in WT, is enriched in the grik1-1 grik2-1 double mutant background due to the increased protein stability of its unphosphorylated form in grik1-1 grik2-1.

To evaluate whether there are differences between KIN10L and KIN10, GFP-KIN10L or GFP-KIN10 were transiently co-expressed for 3 days in N. benthamiana leaves with either GRIK1 or an empty vector. As shown in Figure 4 , compared with GFP-KIN10, more intense GFP fluorescence corresponding to GFP-KIN10L was observed as the puncta in the cytosol, consistent with reports by Williams (Williams et al., 2014) ( Figure 4A ). Immunoblot assays showed the levels of GFP-KIN10L were significantly higher than GFP-KIN10 ( Figure 4B ) upon co-expression with EV. Co-expression with GRIK1 dramatically reduced both KIN10L and KIN10 protein levels ( Figure 4B ). The accumulation of KIN10L seems related to its subcellular localization because a putative nuclear localization signal mutant of KIN10L (K250A, K251A, K253A) in the sequence LFKKIKG which is a match to monopartite nuclear localization signal K·(K/R) ·X·(K/R) (Chelsky et al., 1989), accumulated to higher levels than native KIN10L. Conversely, fusing KIN10L with the SV40 NLS resulted in almost complete retention of KIN10L within the nucleus and promoted its degradation ( Supplementary Figure 3 ).

Next, we tested whether the kinase activity of KIN10L is equivalent to that of KIN10 i.e., whether the extra 23AA at the N-terminus of KIN10L has any effect on its in vitro kinase assay. A His-trigger factor (TF) followed by a factor Xa protease cleavage site domain was fused to the N-termini of KIN10L or KIN10. The constructs were expressed in E. coli and the resulting protein products were purified with the use of Ni-NTA chromatography. KIN10L or KIN10 were recovered after factor Xa protease digestion to remove the affinity tag, yielding proteins with equivalent, i.e., approximately 95% purity as assessed by SDS-PAGE and Coomassie Brilliant blue staining ( Supplementary Figure 4 ). For the kinase assays, a recombinant KIN10 kinase domain (KIN10KD) and GRIK1 were used as a positive kinase control and activator, respectively. The expected low i.e., basal levels of phosphorylation activity were observed in for KIN10L or KIN10 in the absence of GRIK1 activation, although KIN10KD showed higher activity than either isoform. Upon GRIK1 activation, the kinase activities of both KIN10L and KIN10 were elevated dramatically during a 10-minute time course during which no significant differences in kinase activity were detected between KIN10L and KIN10 ( Figure 5 ).

grik mutants: grik1-2 grik2-1 and grik1-1 (+/−) grik2-1 (−/−) were obtained from Nathalie Glab (Institute of Plant Sciences Paris-Saclay, France) (Glab et al., 2017). Other mutants are used in this research including: grik1-1 (CS2103211 or GK-713C09), grik1-2 (SALK_142938), grik2-1 (SALK_015230), kin10-2 (SALK_093965). For experiments with the grik1-1 grik2-1 double mutants, double homozygous null individuals grik1-1(−/−) grik2-1(−/−) were selected from the progeny of the grik1-1(+/−) grik2-1(−/−) sesqui parental line by genotyping with primers listed in Supplementary Table S1 . For Arabidopsis, seeds were surface sterilized with 70% ethanol, then with 30% bleach (Clorox^®) containing 0.01% Tween20 for 15min. Seeds were rinsed five times with sterile water before planted on the half strength of Murashige and Skoog (MS) medium supplemented with 1% of sucrose and 0.7% of agar. After stratified at 4°C in dark for 3 days, seeds were germinated and cultured in Percival^® plant tissue culture chamber. 10-day-old seedlings then were transplanted onto moist BM2 potting soil (Berger) and grown on the shelf of a walk-in growth chamber. Plants were grown with a light/dark cycle of 18h/6h at 23°C, photosynthetic photon flux density of 250 μmol m⁻² s⁻¹, and 75% relative humidity. Nicotiana benthamiana (N. benthamiana seeds) were directly germinated on BM2 potting soil (Berger). 4-week-old N. benthamiana plants were used for transient gene expression by agroinfiltration.

KIN10 refers to AT3G01090. Coding sequences (CDS) of KIN10L, KIN10, GRIK1 were amplified by PCR from cDNAs using primers listed in Supplementary Table S1 . CDS of KIN10L (T198A), KIN10L (K250A, K251A, K253A), GRIK (S261A), GRIK1 (K137A) and SV40-KIN10L-SV40 were generated by overlapping PCR with primers listed in Supplementary Table S1 .

For expression in plants, the PCR products were cloned into the Invitrogen GATEWAY^M pDONR/Zeo vector (Thermo Fisher Scientific, Waltham, MA; www.thermofisher.com) using the BP reaction and sub-cloned (LR reaction) into the plant GATEWAY™ binary vector: pGWB414 (HA-tag at C terminal) or pMDC43 (GFP-tag at N terminal) (Nakagawa et al., 2007).

Transient gene expression in N. benthamiana by agroinfiltration was carried out according to a previous described procedure (Schütze et al., 2009). Leaves were harvested 3 days after agroinfiltration for imaging with a Leica SP5 confocal laser scanning microscope or protein content analysis.

RNA isolation and RT-qPCR were conducted according to the method described in (Zhai et al., 2021)with primers listed in Supplementary Table S1 .

Recombinant KIN10L, KIN10, KIN10 kinase domain and GRIK1 proteins fused with N-terminal His-tag were expressed in E. coli BL21 (DE3). His-tagged protein purification was performed as previously reported (Nallamsetty and Waugh, 2007). For production and purification of KIN10L and KIN10, CDS of KIN10L and KIN10 were cloned into pCold-TF (Takara Bio) for fusion with trigger factor (TF) between the NdeI and XbaI restriction sites. Subsequently, the TF was removed from fusion proteins in digestion with Xa factor protease.

A total of 50 mg of freshly harvested leaves tissues were ground in liquid nitrogen and then mixed with 200 µL of preheated protein extraction buffer (8 M urea, 2% SDS, 0.1 M DTT, 20% glycerol, 0.1 M Tris-HCl, pH (6.8), and 0.004% Bromophenol Blue). After incubated at 80°C for 5min, Samples were centrifuged at 17,000 g before loading supernatants into SDS-PAGE (SurePAGE™, Bis-Tris, 4-20%, precast gel, Genscript). Primary antibodies: anti-KIN10 1:1000 (Catalog No. AS10919, Agrisera), anti-phosphorylated-KIN10/11 (Phospho-AMPK alpha-1,2 (Thr183, Thr172) Polyclonal Antibody) 1:1000 (Catalog No. PA5-17831, Invitrogen), anti-GFP 1:2000 (Catalog No. A6455, Invitrogen), and anti-HA 1:2000 (Catalog No. 71-5500, Invitrogen). Immunoblots of targeted proteins were visualized using HRP-conjugated secondary antibodies with 1:10000 dilution (Catalog No. AP187P, Millipore) with SuperSignal™ West Femto Maximum Sensitivity Substrate (Catalog No. 34095, ThermoFisher). Immunoblot signals were detected and digitalized with Image Quant LAS4000. Phos-tag SDS-PAGE was performed using Acrylamide-pendant phos-tag™ according to the manual of the Acrylamide-pendant Phos-tag™ kit obtained from Wako Chemicals (Richmond, VA). 50μM of Mn²⁺-phos-tag™ SDS-PAGE (10%) is to separate phosphorylated protein from its non-phosphorylated form.

For the KIN10 activity assay, 50 nM of purified KIN10 kinase domain (KD) or KIN10 or KIN10L or GRIK1 was diluted into 25-μL of Kinase Reaction Buffer containing 50 mM HEPES-NaOH, pH7.5, 5 mM MgCl2, 200 μM SPS peptide (RDHMPRIRSEMQIWSED), 4 mM DTT, 0.5 μM okadaic acid, 0.2 mM ATP, 12.2 kBq [γ-32P] ATP and incubated at 30°C for 5 min. The assay was stopped by transferring 10 μL of the assay mixture to 4-cm2 squares of Whatman P81 Phosphocellulose paper (Whatman, Maidstone, UK; www.gelifesciences.com/whatman), immersing it in 1% (v/v) phosphoric acid, then washing with four 800-mL volumes of 1% phosphoric acid. The paper squares were immersed in acetone, dried, and transferred to liquid scintillation vials to which 2 ml of Ultima Gold XR (PerkinElmer) was added before liquid scintillation counting using a Tri-carb (PerkinElmer) was used to determine radioactivity associated with phosphorylated SPS peptide (Zhai et al., 2018).