Abiotic stress-inducible receptor-like kinases negatively control ABA signaling in Arabidopsis

Authors

  • Hidenori Tanaka,

    1. Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
    Search for more papers by this author
  • Yuriko Osakabe,

    1. Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
    Search for more papers by this author
  • Shogo Katsura,

    1. Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
    Search for more papers by this author
  • Shinji Mizuno,

    1. Graduate School of Science and Technology, Chiba University, Inage, Chiba, 263-8522, Japan
    2. Biological Resources and Post-Harvest Division, Japan International Research Center for Agricultural Sciences, Tsukuba, Ibaraki 305-8686, Japan
    Search for more papers by this author
  • Kyonoshin Maruyama,

    1. Biological Resources and Post-Harvest Division, Japan International Research Center for Agricultural Sciences, Tsukuba, Ibaraki 305-8686, Japan
    Search for more papers by this author
  • Kazuya Kusakabe,

    1. Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
    Search for more papers by this author
  • Junya Mizoi,

    1. Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
    Search for more papers by this author
  • Kazuo Shinozaki,

    1. RIKEN Plant Science Center, Yokohama, Kanagawa 230-0045, Japan
    Search for more papers by this author
  • Kazuko Yamaguchi-Shinozaki

    Corresponding author
    1. Laboratory of Plant Molecular Physiology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan
    2. Biological Resources and Post-Harvest Division, Japan International Research Center for Agricultural Sciences, Tsukuba, Ibaraki 305-8686, Japan
    Search for more papers by this author

(fax +81 3 5841 8009; e-mail akys@mail.ecc.u-tokyo.ac.jp).

Summary

Membrane-anchored receptor-like protein kinases (RLKs) recognize extracellular signals at the cell surface and activate the downstream signaling pathway by phosphorylating specific target proteins. We analyzed a receptor-like cytosolic kinase (RLCK) gene, ARCK1, whose expression was induced by abiotic stress. ARCK1 belongs to the cysteine-rich repeat (CRR) RLK sub-family and encodes a cytosolic protein kinase. The arck1 mutant showed higher sensitivity than the wild-type to ABA and osmotic stress during the post-germinative growth phase. CRK36, an abiotic stress-inducible RLK belonging to the CRR RLK sub-family, was screened as a potential interacting factor with ARCK1 by co-expression analyses and a yeast two-hybrid system. CRK36 physically interacted with ARCK1 in plant cells, and the kinase domain of CRK36 phosphorylated ARCK1 in vitro. We generated CRK36 RNAi transgenic plants, and found that transgenic plants with suppressed CRK36 expression showed higher sensitivity than arck1-2 to ABA and osmotic stress during the post-germinative growth phase. Microarray analysis using CRK36 RNAi plants revealed that suppression of CRK36 up-regulates several ABA-responsive genes, such as LEA genes, oleosin, ABI4 and ABI5. These results suggest that CRK36 and ARCK1 form a complex and negatively control ABA and osmotic stress signal transduction.

Introduction

Osmotic stresses such as drought and high salinity adversely affect plant growth and crop yield. Plants have evolved complex mechanisms to control the expression of various genes to adapt to such environmental stresses. Stress-inducible genes function not only in protecting cells by producing metabolic proteins, but also in regulating important genes for signal transduction in the stress response (Zhu, 2002; Bartels and Sunkar, 2005; Yamaguchi-Shinozaki and Shinozaki, 2006). The plant hormone abscisic acid (ABA) regulates a wide range of physiological events, including seed maturation and dormancy, control of vegetative growth and flowering, and tolerance to various abiotic stresses (Finkelstein et al., 2002; Himmelbach et al., 2003). High levels of ABA are produced under osmotic stress conditions, regulating the expression of various osmotic stress-responsive genes.

The cell surface is the site of sensors and receptors that interpret environmental conditions and transduce signals to other sites on the membrane, inside the cell and in distal parts of the plant to trigger direct and rapid responses to changing environmental conditions (Hwang et al., 2002; Kacperska, 2004; Barkla and Pantoja, 2011). Receptor-like protein kinases (RLKs) localized on the plasma membrane play important roles in plant growth, development and stress responses (Morris and Walker, 2003). In Arabidopsis, RLKs belong to a large gene family with at least 610 members encoding more than 400 receptor kinases and 200 receptor-like cytoplasmic kinases (RLCKs) (Shiu and Bleecker, 2001). Structurally, a typical receptor kinase consists of an extracellular domain, a transmembrane domain, and an intracellular Ser/Thr protein kinase domain. They may be classified into various sub-classes based on the structure of their extracellular domains (Shiu and Bleecker, 2001). RLKs are involved in diverse signaling events, such as brassinosteroid perception by BRASSINOSTEROID INSENSITIVE1 (BRI1) and BRI1-ASSOCIATED RECEPTOR KINASE1 (BAK1), perception of bacterial flagellin fragments by FLAGELLIN-SENSITIVE 2 (FLS2), which mediates a pathogen response, and recognition of a small peptide CLAVATA3 (CLV3) by CLV1, CLV2 and RECEPTOR-LIKE PROTEIN KINASE2 (RPK2) (Diévart and Clark, 2004; Mizuno et al., 2007; Kinoshita et al., 2010; Gish and Clark, 2011). BAK1 also functions in flagellin sensing via formation of the BAK1–FLS2 receptor complex (Chinchilla et al., 2007). These studies suggest that the cross-talk between hormone and pathogen responses is mediated by RLKs and their ligand peptides. An ABA-inducible leucine-rich repeat RLK, RPK1, acts as one of the key factors in ABA responses in Arabidopsis (Hong et al., 1997; Osakabe et al., 2005). Recently, it was shown that RPK1 is involved in reactive oxygen species (ROS) signaling during abiotic stresses (Osakabe et al., 2010) and in ABA-induced leaf senescence (Lee et al., 2011). Over-production of RPK1 enhanced both water and oxidative stress tolerance in Arabidopsis (Osakabe et al., 2010).

Cysteine-rich repeat RLKs (CRKs) make up a large sub-group of the RLK family with 46 members (Wrzaczek et al., 2010). The extracellular region of the protein contains two copies of the DUF26 domain containing four conserved cysteines, three of which form the motif C-8X-C-2X-C, which is presumed to be involved in formation of the 3D structure of the protein through disulfide bonds and has roles in protein–protein interactions. Several CRK genes are transcriptionally induced by ROS, biotic stress and salicylic acid (Wrzaczek et al., 2010). Over-expression of CRK5 resulted in increased resistance to the bacterial pathogen Pseudomonas syringae and enhanced growth of plant leaves (Chen et al., 2003). Enhanced defense responses upon over-expression of CRK13, with increased H2O2 levels, have also been reported (Acharya et al., 2007). To understand the molecular mechanisms of RLK-mediated signaling systems, the focus has been on identification of the receptor complex. Recent studies revealed that some RLCKs form a receptor complex with RLKs and play roles in transducing signals in this pathway. The RLCK BIK1 (BOTRYTIS-INDUCED KINASE1), associates with the FLS2–BAK1 receptor complex and mediates early flagellin signaling (Lu et al., 2010). BRASSINOSTEROID-SIGNALING KINASE (BSK) kinases are a sub-group of RLCKs that are substrates of the BRI1 kinase involving brassinosteroid signal transduction (Tang et al., 2008).

In this study, we analyzed the function of the ABA- AND OSMOTIC-STRESS-INDUCIBLE RECEPTOR-LIKE CYTOSOLIC KINASE1 (ARCK1) gene, which encodes an RLCK that belongs to the CRK sub-family. We found that ARCK1 interacts with CRK36 in the plasma membrane, and together they exert negative effects in ABA signaling during the post-germinative growth phase. Our studies on ARCK1 and CRK36 suggest a mode of RLK behavior which an RLCK and its RLK partner at the plasma membrane transduce signals in response to ABA and osmotic stress.

Results

ARCK1 mRNA levels increase in response to abiotic stress and ABA

To elucidate important factors for mediating stress signaling, we identified genes whose expression was induced by ABA and abiotic stresses from our previous microarray data (Lehti-Shiu et al., 2009; Maruyama et al., 2009) and public databases (Zimmermann et al., 2004; https://www.genevestigator.com/). Among these genes, we selected a stress-responsive RLCK gene named ABA- AND OSMOTIC STRESS-INDUCIBLE RECEPTOR-LIKE CYTOSOLIC KINASE1 (ARCK1), which belongs to the CRK sub-family. To examine the physiological function of ARCK1 under abiotic stress conditions, we used quantitative RT-PCR to analyze the expression pattern of ARCK1 in leaves and roots of 3-week-old plants and 4-day-old seedlings under abiotic stresses and ABA treatment (Figure 1a). Expression of ARCK1 was strongly induced within 1 h under high-salinity and drought stress, and slightly induced after 10 h under cold stress. The accumulation of ARCK1 mRNA in leaves was increased after 3 h of ABA treatment. Similar results were obtained by histochemical analysis of ARCK1 promoter:GUS transgenic plants (Figure 1b). The GUS activity was increased, especially in the leaves, by osmotic stress and ABA treatment. These data indicate that ARCK1 is induced, mainly in the leaves, by abiotic stress and ABA treatment. We tested the subcellular localization pattern of the ARCK1 protein by using transgenic Arabidopsis plants expressing the ARCK1 protein fused to the N-terminus of GFP (GFP–ARCK1) under the control of the CaMV 35S promoter (35S:GFP–ARCK1). Confocal microscopic analyses showed that, under normal growth conditions, the green fluorescent signals of the GFP–ARCK1 proteins were mainly localized in the cytosol and at the cell surface (Figure 1c). ARCK1 has a protein kinase catalytic domain homologous to those of the CRK sub-family. Figure S1 shows a phylogenic tree based on the amino acid sequences of 44 Arabidopsis CRK proteins and the well known leucine-rich repeat RLKs (Figure S1).

Figure 1.

ARCK1 expression patterns and phenotypes of ARCK1 knockout plants.
(a) Expression of ARCK1 in response to ABA treatment and abiotic stress conditions. Total RNAs were extracted from 4-day-old seedlings and 3-week-old plants (shoots or roots) of wild-type Arabidopsis (ecotype Columbia) treated with ABA or various abiotic stresses. Plants grown on MS agar plates were dehydrated (drought) on parafilm, transferred to distilled water containing 200 mm NaCl or 50 μM ABA, or transferred to and maintained at 4°C (cold) for the indicated times. The value for the control (0 h, 3-week-old shoot) was set to 1.0. To normalize ARCK1 expression, 18S rRNA was amplified as an internal control. Error bars indicate SD (= 3).
(b) Tissue-specific expression of the ARCK1 promoter–GUS fusion. Transgenic plants were treated with 50 μM ABA, 300 mm mannitol or 200 mm NaCl for 5 h before histochemical GUS staining. Scale bar = 2 mm.
(c) GFP–ARCK1 fluorescence in the root of a transgenic plant expressing 35S:GFP-ARCK1 was visualized by confocal microscopy. The GFP fluorescence image (GFP–ARCK1) and the image merged with the bright field image (merged) are shown. Scale bar = 10 μm.
(d) Nine-day-old arck1-1, arck1-2 and wild-type (WT) seedlings grown on MS agar plates with or without 0.5 μM ABA, and 6-day-old arck1-1, arck1-2 and WT seedlings grown on MS agar plates with 150 mm NaCl. Control plants were grown on MS agar plates). Scale bar = 4 mm.
(e–g) Cotyledon greening rate of ARCK1 knockouts. arck1-1, arck1-2 and WT plants were grown on control MS medium (e) or MS medium containing 0.5 μM ABA (f) or 150 mm NaCl (g). For these experiments, seedlings with expanded green cotyledons were counted. SD values were calculated from three individual experiments (= 40 seeds per experiment).

arck1 mutants are more sensitive to ABA during the post-germinative growth phase

To investigate the function of ARCK1 in plant cells, we obtained two ARCK1 T-DNA insertional alleles (SALK_057538 and SALK_037588) from the Arabidopsis Biological Resource Center (http://www.arabidopsis.org/), and named them arck1-1 and arck1-2, respectively (Figure S2a). The genomic sequences of arck1-1 and arck1-2 contain a T-DNA insertion at 68 bp upstream and 211 bp downstream of the translational start site of ARCK1, respectively. To assess the expression level of ARCK1 in homozygous T-DNA insertion lines, we performed RT-PCR using cDNAs synthesized from the 3-week-old wild-type (Col), arck1-1 and arck1-2 plants under high-salinity stress. The PCR products for ARCK1 were not detected in arck1-1 and arck1-2 homozygous lines (Figure S2b, suggesting that these T-DNA insertion lines are ARCK1 knockout mutants. Under normal growth conditions, these plants did not show any morphological alterations (Figure 1d). Because the ARCK1 gene was induced by ABA and osmotic stress (Figure 1a), we examined the ABA sensitivity of the arck1 knockout mutants (Figure 1). The germination rates for the mutants were the same as for the wild-type on both ABA-containing and ABA-free media (Figure S2c,d). However, the cotyledon greening rates of the knockout mutants during the post-germinative growth phase were significantly lower than in wild-type plants on ABA-containing medium, but not on ABA-free medium (Figure 1d–f). We then analyzed post-germinative growth under high-salinity conditions, and observed that the arck1 mutants were more NaCl-sensitive than wild-type plants (Figure 1d,g). We performed root growth inhibition experiments and root bending assays on ABA- or NaCl-containing media using the arck1 mutants. The arck1 mutants exhibited similar sensitivities to those of the wild-type plants (Figure S2e,f). These results suggest that ARCK1 has a negative effect on the response to ABA and high-salinity stress in terms of cotyledon greening during the post-germinative growth phase.

CRK36, an abiotic stress-inducible CRR RLK, interacts with ARCK1

The RLCKs have important roles in the RLK signaling pathway as a result of their ability to form a receptor complex (Tang et al., 2008; Lu et al., 2010). Thus, it is possible that ARCK1 functions coordinately with other RLKs in ABA and high-salinity stress signal transduction pathways. We searched for candidate RLKs that could form a complex with ARCK1 using the web-based co-expression analysis tools ATTED-II (http://atted.jp/) (Obayashi et al., 2006) and Cluster Cutting (http://prime.psc.riken.jp/?action=agetree_index)(Akiyama et al., 2008). ATTED-II provides co-regulated gene relationships based on co-expressed genes deduced from microarray data and predicted cis elements (Obayashi et al., 2006). Cluster Cutting shows the results of analyses based on similarities of gene expression patterns derived from AtGenExpress data (Akiyama et al., 2008). Analysis using ATTED-II indicated that 15 candidate genes, including six RLK genes, showed similar expression patterns to that of ARCK1 (Table S1). Cluster Cutting analysis indicated that eight genes were present in the same cluster as ARCK1. Of these eight genes, one RLK gene, At4g04490 (CRK36), matched a gene identified by ATTED-II (Figure S3). Thus, two co-expression analyses indicated that the RLK gene CRK36 shows a similar expression pattern to that of ARCK1.

Next, we investigated the interaction between ARCK1 and the six RLKs identified by ATTED-II using the yeast two-hybrid system. Only the kinase domain of CRK36 interacted with ARCK1 in yeast (Figure 2a). CRK36 belongs to the CRK sub-family, which has 46 members in Arabidopsis (Wrzaczek et al., 2010), and the CRK36 kinase domain is similar to those of four other CRKs: CRK37, CRK38, CRK39 and CRK40 (Figures S1 and S4). Next, we examined the interaction between ARCK1 and these CRKs, except CRK38, which cannot be expressed in yeast cells. Determination of the interaction of ARCK1 with the kinase domains of these CRKs in the yeast two-hybrid system showed that CRK36 was the only ARCK1-interacting RLK among the members of this sub-family expressed in yeast cells (Figure 2b). Moreover, the yeast two-hybrid assay indicated that the kinase domain of CRK36 interacted with itself, and that the kinase domains of CRK39 and CRK40 interacted with each other (Figure 2b).

Figure 2.

 Identification of an abiotic stress-induced CRK that interacts with ARCK1.
(a) Yeast two-hybrid analysis of ARCK1 and the kinase domain (KD) of various RLKs. ARCK1 (bait) and kinase domain of RLKs (prey) were introduced into yeast cells as indicated, and grown on SD medium minus leucine and tryptophan (-LW) or leucine, tryptophan, histidine, and adenine (-LWHA) + 6 mM 3-aminotriazole. Empty vectors pGBKT7 and pGADT7 were used as controls.
(b) Yeast two-hybrid analysis of ARCK1 and the kinase domain (KD) of various CRKs. ARCK1 and CRKs kinase domains were cloned into the bait vectors and CRK kinase domains were cloned into the prey vectors. They were introduced into yeast cells as indicated.
(c, d) Expression pattern of CRK36 in 3-week-old plants (c) and 4-day-old seedlings (d) subjected to abiotic stress and ABA treatment. Analyses were performed as described in Figure 1(a). Error bars indicate SD (= 3).
(e, f) CRK36–GFP fluorescence in roots of transgenic plants expressing 35S:CRK36-GFP. For plasmolysis analysis of CRK36–GFP, we examined transgenic plants treated with (f) or without (e) 0.8 M mannitol. Scale bar = 20 μm.

We used quantitative RT-PCR to analyze the expression pattern of CRK36 in leaves and roots of 3-week-old plants and 4-day-old seedlings under abiotic stress and ABA treatment (Figure 2c,d). CRK36 was mainly induced in leaves in response to drought and high-salinity stress. The public microarray database showed that the expression of other CRKs is not strongly induced by abiotic stress (Figure S5). These data indicated that CRK36 showed a similar expression pattern to ARCK1 under drought and high-salinity stress conditions. We tested the subcellular localization pattern of the CRK36 protein using transgenic Arabidopsis plants expressing the CRK36 protein fused to the C-terminus of GFP (CRK36–GFP) under the control of the CaMV 35S promoter (35S:CRK36–GFP). Green fluorescent signals of the CRK36–GFP proteins were observed in the cell periphery of root tips (Figure 2e). These signals retracted from the cell wall due to plasmolysis after treatment with 0.8 m mannitol (Figure 2f), indicating that CRK36 is localized to the plasma membrane. These results suggest that CRK36 is a plasma membrane-localized CRK that interacts with an RLCK (ARCK1).

CRK36 interacts with ARCK1 in vivo

To confirm the interaction between ARCK1 and CRK36 in plant cells, we performed bimolecular fluorescence complementation (BiFC) and co-immunoprecipitation assays. We constructed fusions of ARCK1, CRK36 and CRK39 with the N-terminal region of the Venus variant of yellow fluorescent protein (VYFP) (VYFPN1–173; VYNE) and the C-terminal region of cyan fluorescent protein (CFP) (CFPC156–239; SCYCE), respectively, under the control of the CaMV 35S promoter (Figure 3a). These constructs were introduced into Nicotiana benthamiana and transiently expressed (Figure 3b,c), and green fluorescence signals were observed by confocal microscopy. Co-expression of VYNE–ARCK1 and CRK36–SCYCE produced green fluorescence at the plasma membrane in N. benthamiana (Figure 3b), whereas co-expression of VYNE–ARCK1 and CRK39–SCYCE did not (Figure 3c). We also generated transgenic Arabidopsis plants expressing both VYNE–ARCK1 and CRK36–SCYCE. Green fluorescence was observed at the plasma membrane in the transgenic Arabidopsis cells (Figure 3d). Moreover, transient expression of 3xFLAG–ARCK1 and CRK36–GFP proteins in N. benthamiana, and constitutive expression of VYNE–ARCK1 and CRK36–SCYCE in transgenic Arabidopsis plants, were detected after co-immunoprecipitation experiments with CRK36 using anti-GFP (α-GFP) antibody microbeads and anti-hemagglutinin (α-HA) antibody agarose beads, respectively (Figure 3e,f). These results indicate that CRK36 interacts with ARCK1 at the plasma membrane of plant cells.

Figure 3.

 Physical interaction between ARCK1 and CRKs.
(a) Constructs used for BiFC analysis. ARCK1, CRK36 and CRK39 full-length cDNAs with no stop codon were sub-cloned into 35S:VYNE and 35S:SCYCE vectors. c-myc, c-myc epitope tag; HA, hemagglutinin epitope tag.
(b, c) BiFC analysis in Nicotiana benthamiana. The 35S:VYNE-ARCK1 construct was co-expressed with 35S:CRK36-SCYCE (b) or 35S:CRK39-SCYCE (c). The GFP fluorescence image (left) and the image merged with the bright field image (right) were shown. Scale bar = 50 μm.
(d) BiFC analysis in Arabidopsis thaliana. Arabidopsis was transformed with 35S:VYNE-ARCK1 and 35S:CRK36-SCYCE, and leaves of transgenic plants were analyzed. Scale bars = 50 μm (left, epidermal cells) and 5 μm (right, guard cell).
(e) Co-immunoprecipitatioin of 3xFLAG–ARCK1 and CRK36–GFP using Nicotiana benthamiana. Total protein extracts (upper panel) or proteins immunoprecipitated (IP) using anti-GFP (α-GFP) antibody from the solubilized microsomal and cytosolic fractions (middle and lower panels) were analyzed by Western blotting (WB) using α-GFP or α-FLAG antibodies.
(f) Co-immunoprecipitation of VYNE-myc-ARCK1 and CRK36-HA-SCYCE using transgenic plants. Total protein extracts (upper panel) or proteins immunoprecipitated (IP) using anti-hemagglutinin (α-HA) antibody from the microsomal fractions (middle and lower panels) were analyzed by Western blotting (WB) using α-myc or α-HA antibodies. The asterisk indicates non-specific binding.

Suppression of CRK36 expression enhances ABA sensitivity during post-germinative growth

To investigate the function of CRK36 in plant cells, we generated CRK36 knockdown plants using double-stranded RNA interference (RNAi). A 400 bp fragment of CRK36 was used for trigger silencing to generate transgenic CRK36 knockdown plants in the wild-type (CRK36 RNAi) or arck1-2 (arck1-2/CRK36 RNAi) backgrounds (Figure 4a,b). The levels of CRK36 gene suppression in the transgenic plants were confirmed by quantitative RT-PCR (Figure 4b). We then examined the responsiveness to ABA during the post-germinative growth phase by measuring the cotyledon greening rates and chlorophyll content. The cotyledon greening rates of CRK36 RNAi were lower than those of control plants and arck1-2 on ABA-containing medium, but similar on ABA-free medium (Figure 4c–f).

Figure 4.

 Phenotypes of CRK36 RNAi transgenic plants.
(a) Generation of CRK36 RNAi lines. The first 400 bp downstream of the translational start site of CRK36 was cloned.
(b) Expression levels of CRK36 in CRK36 RNAi lines. Total RNAs were extracted from wild-type plants, CRK36 RNAi lines a and b, and arck1-2/CRK36 RNAi lines a and b transferred into distilled water for 24 h.
(c–e) Cotyledon greening rate of CRK36 RNAi lines. Plants of CRK36 RNAi lines, arck1-2/CRK36 RNAi lines, arck1-2/vec (arck1-2 transformed with the empty vector) and Col transformed with empty vector (col/vec) were grown on MS medium (c) or MS medium containing ABA at a concentration of 0.3 (d) or 0.5 μM (e), and the cotyledon greening rate was determined. SD values were calculated from three individual experiments (= 40 seeds per experiment).
(f) Young seedlings 6 days after the end of stratification. Seeds were germinated and allowed to grow on MS medium containing 0 or 0.3 μM ABA. Scale bar = 3 mm.
(g) Chlorophyll content of CRK36 RNAi lines. Six-day-old seedlings grown on MS agar medium or MS medium containing 0.3 μM ABA were examined. SD values were calculated from three individual experiments (= 30 seedlings per experiment). Asterisks show significance level compared to seedlings grown in the absence of ABA (*< 0.05, **< 0.005; Student’s t test).
(h,i) Cotyledon greening rate of CRK36 RNAi lines under high-salinity conditions (165 mm NaCl (h) or osmotic stress conditions (300 mm mannitol) (i). The cotyledon greening rate was determined. SD values were calculated from three individual experiments (= 40 seeds per experiment).

To analyze genetic interaction between CRK36 and ARCK1, we tested the ABA sensitivity of the arck1-2/CRK36 RNAi double mutant. The cotyledon greening rates of arck1-2/CRK36 RNAi were similar to those of CRK36 RNAi and arck1-2 on ABA-containing medium (Figure 4d–f). In the presence of ABA, the chlorophyll contents of arck1-2/CRK36 RNAi and CRK36 RNAi were lower than those of wild-type and arck1-2 (Figure 4g). We then investigated the sensitivity to high-salinity stress (165 mm NaCl) and osmotic stress (300 mm mannitol) (Figure 4h,i). The CRK36 RNAi lines were more sensitive to high-salinity and osmotic stresses than the wild-type and arck1-2 mutant plants, as determined by cotyledon greening rates (Figure 4h,i and S6). We also examined the sensitivity of the CRK36 RNAi lines to various plant hormones other than ABA. The CRK36 RNAi lines did not show obvious differences from the wild-type plants (Figure S7). These results suggest that CRK36 negatively controls ABA and osmotic stress signaling in a specific manner during the post-germinative growth phase in Arabidopsis.

CRK36 knockdown enhances transcription levels of ABA-responsive genes

To identify the molecular events in the CRK36 signaling pathway, we compared the mRNA profiles of CRK36 RNAi and wild-type plants in the presence of ABA. We performed a microarray analysis using an Agilent microarray using RNA from young seedlings of CRK36 RNAi and wild-type plants incubated with 0.5 μm ABA. Genes with an expression ratio (CRK36 RNAi/wild-type) <0.5 (for down-regulated genes) or >2 (for up-regulated genes) in two independent transgenic lines were selected. We identified 191 up-regulated genes and 140 down-regulated genes in the CRK36 RNAi plants (Tables S2 and S3). Using Genevestigator (https://www.genevestigator.com/gv/index.jsp) (Zimmermann et al., 2004), we obtained expression profiles of the genes that were down- and up-regulated in the CRK36 RNAi plants (Tables S4 and S5). We used the descriptions in the TAIR10 database (http://www.arabidopsis.org/) to functionally categorize the down- and up-regulated genes. The majority of down-regulated genes encoded seed storage proteins and proteins with roles in lipid biosynthesis, redox processes and secondary metabolism (Figure 5a and Table S4). The majority of up-regulated genes encoded late embryogenesis abundant (LEA) proteins and proteins with roles in transcription and proteolysis (Figure 5b and Table S5). We analyzed the responsiveness of down- and up-regulated genes to various stimuli, and found that they appeared to be ABA-responsive in germinating seeds (Figure 5c,d). Approximately 50% of the down-regulated genes showed negative responses to ABA in germinated seeds (Figure 5c), and approximately 80% of the up-regulated genes showed positive responses to ABA (Figure 5d). The down-regulated genes were expressed throughout various plant developmental stages, whereas the up-regulated genes were mainly expressed in germinating seeds and mature siliques (Figure S8).

Figure 5.

 Ontology and responsiveness of genes showing down- or up-regulated expression in young CRK36 RNAi seedlings treated with ABA.
(a, b) Functional classification of down- (a) or up-regulated (b) genes. Results were obtained from the PageMan database (https://http://mapman.mpimp-golm.mpg.de/index.shtml) and the TAIR10 database (http://www.arabidopsis.org/).
(c, d) Expression patterns of down- (c) or up-regulated (d) genes under various conditions. Results are based on data from Genevestigator (https://www.genevestigator.com/gv/html.jsp). Data were not available for analysis for some genes. Genes on the left side have high fold changes in CRK36 RNAi lines.

A recent study showed that three kinases belonging to sub-class III of the SNF1-related protein kinase 2 (SnRK2) family, i.e. SRK2D/SnRK2.2, SRK2E/OST1/SnRK2.6 and SRK2I/SnRK2.3, act as global positive regulators in ABA signaling (Umezawa et al., 2010). To determine how CRK36 is linked to the known ABA signaling pathway, we compared the downstream genes of CRK36 with those of SRK2D/E/I in ABA signaling (Fujita et al., 2009; Nakashima et al., 2009). Of the genes up-regulated in CRK36 RNAi lines, 29% (56/191) and 38% (72/191) showed expression levels that were reduced more than twofold in seeds and 12-day-old plants, respectively, of the srk2d/e/i triple mutant (Figure 6a and Table S5). In total, 51% (97/191) of the genes up-regulated in CRK36 RNAi lines were expected to be downstream genes of SRK2D/E/I (Figure 6a and Table S5). Of the genes up-regulated in CRK36 RNAi lines that were not downstream genes of SRK2D/E/I, 57% (54/94) were ABA-responsive (Table S5). These data indicate that CRK36 target genes partially overlap with SRK2D/E/I target genes in the response to ABA.

Figure 6.

 Expression levels of CRK36-related genes in arck1 and CRK36 RNAi lines.
(a) Number of down-regulated genes in seeds and seedlings of the srk2d/e/i mutant, and genes up-regulated by at least twofold in CRK36 RNAi lines.
(b) Quantitative RT-PCR was used to determine relative mRNA levels in arck1-2 and CRK36 RNAi plants grown on MS medium with or without 0.5 μM ABA for 10 days (white bars, without ABA; black bars, with ABA). Similar results were obtained from three independent experiments, and typical results are shown. SD values were calculated from three independent PCRs from the same biological sample. The At3g02480 (LEA), ABI4, ABI5, OLEO4, At1g72070 (chaperone DnaJ-domain superfamily protein), At5g58390 (peroxidase) genes were analyzed. The value for wild-type without ABA was set to 1.0. Error bars indicate SD (= 3).

To verify the microarray results, we used quantitative RT-PCR analyses to investigate the expression of several down-regulated and up-regulated genes in wild-type, arck1-2 and CRK36 RNAi plants (Figure 6b). The ABA-responsive gene At3g02480 (an LEA gene) and the gene encoding the transcription factor ABA-INSENSITIVE 5 (ABI5) were expected as the SRK2D/E/I target genes (Table S5). The expression level of At3g02480 (LEA gene) was modestly increased in arck1-2 and greatly increased in CRK36 RNAi lines in the presence of ABA. This trend was also observed for other ABA-responsive genes, including the transcription factors ABI4 and ABI5, an oleosin gene (OLEO4) and a chaperone DnaJ-containing superfamily gene (At1g72070) (Figure 6b). The expression level of the ABA-repressible gene At5g58390 (peroxidase) was modestly decreased in the arck1 mutant and greatly decreased in the CRK36 RNAi lines (Figure 6b). These results indicate that CRK36 plays an inhibitory role in ABA signaling in germinating seeds. CRK36 appears to suppress expression of ABA-inducible genes such as LEA genes, oleosin, ABI4 and ABI5, and enhance expression of ABA-repressible genes, such as the peroxidase gene At5g58390.

Suppression of ARCK1 and CRK36 enhances norflurazon sensitivity

We analyzed the responsiveness of down-regulated genes to various stimuli using Genevestigator, and found that some of the down-regulated genes were repressed by norflurazon (NF) treatment (Figure 5c). NF inhibits carotenoid biosynthesis at the phytoene desaturase step (Kleudgen, 1979; Chamovitz et al., 1991). To understand the relationship between NF and the signaling in which ARCK1 and CRK36 were involved, we investigated the sensitivity of arck1-2 and CRK36 RNAi plants grown on NF-containing medium. The CRK36 RNAi lines and arck1-2 mutant showed higher NF sensitivity and lower chlorophyll contents than wild-type plants (Figure S9a,b). We also measured transcript levels of LHCB1.2 (At1g29910), which was previously reported to be down-regulated in NF-treated wild-type plants (Batschauer et al., 1986). Expression of LHCB1.2 was down-regulated more strongly in both the arck1 mutant and CRK36 RNAi lines compared with wild-type (Figure S9c), indicating that suppression of ARCK1 and CRK36 increased sensitivity to NF. However, the relationship between NF and the signaling involving ARCK1 and CRK36 is not well understood, and further analyses are required to understand this relationship.

CRK36 is an active serine/threonine kinase and phosphorylates ARCK1 in vitro

To test whether ARCK1 encodes a functional protein kinase, we analyzed the autophosphorylation activity of ARCK1 expressed as a recombinant fusion protein with glutathione S-transferase (GST–ARCK1) in Escherichia coli. The GST–ARCK1 protein was capable of autophosphorylation when incubated with [γ-32P]ATP (Figure S6a). Site-directed mutagenesis of a conserved active-site residue (K65E) (GST–mARCK1) abolished this activity, indicating that incorporation of 32P requires an active kinase domain (Figure S10a). This amino acid substitution has been reported to abolish protein kinase activity of other RLKs (Horn and Walker, 1994; Braun et al., 1997). ARCK1 showed low autophosphorylation activity compared with that of other RLKs, such as BRI1 (Figure S10a). Compared with the amino acid sequences of other RLKs, ARCK1 has several variations within the amino acids of the conserved kinase domain, and the DFG motif of domain VII is not conserved in ARCK1 (Figure S10a). The maize (Zea mays) kinase MARK (Llompart et al., 2003) and Arabidopsis TMKL1 (Valon et al., 1993), which have no autophosphorylation activity, also show amino acid variations in the same domain (Figure S10b). Substitution of these amino acids to create the DFG motif increased the autophosphorylation activity of ARCK1, resulting in a constitutively active form of ARCK1 (ARCK1-CA) (Figure S10a).

We next examined the autophosphorylation activity of the kinase domain (KD) of CRK36, and showed that CRK36 is an active protein kinase (Figure 7a). To determine whether CRK36 could phosphorylate or be phosphorylated by ARCK1, we created enzymatically inactive forms (mutated kinase domain, mKD) of CRK36 by substituting a conserved lysine in sub-domain II by glutamic acid. We assayed phosphorylation of the mutant proteins and found that they had no protein kinase activity (Figure 7a, lane 4). Combinations of the mutant and wild-type ARCK1 and CRK36 proteins were tested for protein kinase activity. The results showed that CRK36 could phosphorylate ARCK1 in vitro. In addition, combination of the two kinases noticeably increased the phosphorylation level of ARCK1 (Figure 7a, lane 5). These data suggest that ARCK1 could be a substrate for CRK36 in vitro.

Figure 7.

 ARCK1 phosphorylation in vitro and in vivo.
(a) Auto- and trans-phosphorylation activity of CRK36. Proteins were separated by 10% SDS–PAGE after incubation in protein kinase assay buffer and [γ-32P]ATP. Each lane represents an independent reaction: lane 1, GST–ARCK1KD; lane 2, GST–ARCK1mKD (K65E); lane 3, GST–CRK36KD; lane 4, GST–CRK36mKD (K368E); lane 5, GST–ARCK1KD + GST–CRK36KD; lane 6, GST–ARCK1KD + GST–CRK36mKD; lane 7, GST–ARCK1mKD + GST–CRK36KD; lane 8, GST–ARCK1mKD + GST–CRK36mKD. KD, kinase domain; mKD, mutated kinase domain. Autoradiographic results are from the same gel after exposure to an imaging plate overnight. Protein abundance was visualized by Coomassie brilliant blue (CBB) staining.
(b) ARCK1 phosphorylation on threonine residues in vivo. Total membrane proteins of transgenic lines expressing VYNEARCK1 and SCYCE or VYNEARCK1 and CRK36–SCYCE were analyzed by Western blotting (WB) using α-HA antibodies (upper panel). Proteins immunoprecipitated (IP) from the total proteins using anti-myc (α-myc) antibody were analyzed by Western blotting (WB) using α-myc (middle panel) or anti-phosphothreonine (α-pThr, lower panel) antibodies. Three-week-old transgenic plants were incubated in distilled water (−) or 150 mm NaCl solution (+) for 5 h. The asterisk indicates non-specific binding.

Extent of VYNE–ARCK1 phosphorylation is increased by over-expression of CRK36 under salt stress

To determine the extent of ARCK1 phosphorylation in vivo, we next analyzed the phosphorylation status of ARCK1 using transgenic Arabidopsis plants constitutively expressing the proteins VYNE–ARCK1 and SCYCE, or VYNE–ARCK1 and SCYCE–CRK36. In this experiment, we used 3-week-old transgenic plants to obtain sufficient protein for immunoprecipitation studies. Moreover, we examined the phosphorylation status of ARCK1 under salt-stress conditions, as the expression analysis of ARCK1 and CRK36 indicated that these genes were highly induced under salt-stress conditions in 3-week-old plants (Figures 1a and 2c). Immunoprecipitates of VYNE–ARCK1 were hardly detected by anti-phosphothreonine antibody in the transgenic lines expressing VYNE–ARCK1 and SCYCE, regardless of whether plants were exposed to salt stress or not (Figure 7b). However, immunoprecipitates of VYNE–ARCK1 were detected by anti-phosphothreonine antibody in the transgenic lines expressing VYNE–ARCK1 and SCYCE–CRK36 under the salt stress condition (Figure 7b). These data indicate that the extent of VYNE–ARCK1 phosphorylation is affected by over-expression of CRK36 under salt-stress conditions. To investigate the function of phosphorylated ARCK1 in planta, we constructed transgenic plants over-expressing GFP–ARCK1 and GFP–ARCK1-CA. We performed post-germinative growth inhibition experiments on ABA- or NaCl-containing media using these transgenic lines, and found that they showed similar sensitivities to those of the control plants (Figure S11). These data indicate that, in addition to phosphorylation status, other factors may also be involved in the regulation of ARCK1 function.

Discussion

To analyze cellular responses to abiotic stress in plants, we focused on osmotic stress-inducible receptor-like kinases, specifically an Arabidopsis receptor-like cytosolic kinase (RLCK) ARCK1, which belongs to the CRK sub-family. The ARCK1 gene is strongly induced by osmotic stress and weakly induced by ABA, and knockout mutants showed hypersensitivity to ABA and high-salinity stress during the post-germinative growth phase (Figure 1). These observations suggest that ARCK1 functions negatively in ABA and high-salinity stress signaling in plant cells in the post-germinative growth phase.

Using a GFP–ARCK1 fusion protein, we confirmed that ARCK1 is mainly localized to the cytosol and the cell surface (Figure 1). Recently, several RLCKs have been reported to form a complex with an RLK. For example, one RLCK, BIK1, associates with two RLKs (FLS2 and BAK1), and plays an important role in mediating early flagellin signaling (Lu et al., 2010). Moreover, other RLCKs and brassinosteroid-signaling kinases interact with the RLK BRI1, and are phosphorylated by BRI1 kinase, which activates downstream brassinosteroid signal transduction (Tang et al., 2008). Based on these reports, we assumed that there must be a protein that interacts and forms a complex with ARCK1. Using co-expression analyses, we isolated a candidate RLK, CRK36, belonging to the CRK sub-family. Our results indicated that ARCK1 and CRK36 interact with each other in vivo in both yeast and plant cells (Figures 2 and 3). The results suggest that ARCK1 and CRK36 form a complex in the plasma membrane.

Both ARCK1 knockout and CRK36 knockdown mutants showed enhanced sensitivity to ABA and osmotic stress during the post-germinative growth phase (Figures 1 and 4). These data suggest that CRK36 and ARCK1 form a complex and function as mediators that negatively control ABA and osmotic stress signaling during the post-germinative growth of Arabidopsis. As expression of CRK36 and ARCK1 is induced by ABA and abiotic stress (Figures 1 and 2), they appear to function in negative feedback control of ABA and osmotic stress signaling. These negative effects might be important for plants in order to modulate the stress signal transduction flexibly in response to environmental conditions. Furthermore, the ABA and osmotic stress sensitivities of the arck1-2 CRK36 RNAi double mutant were similar to those of CRK36 RNAi transgenic lines and higher than those of arck1-2 (Figure 4). This implies that CRK36 RNAi lines are epistatic to arck1 mutant, and additional factors other than ARCK1 interact with CRK36 and mediate ABA and osmotic signaling. Further investigations on CRK36-interacting proteins are required to elucidate the mechanisms controlling the CRK36 and ARCK1 complex.

In contrast to the ABA hypersensitivity of arck1-2 and CRK36 RNAi seedlings during the post-germinative growth (Figure 4), germination of the mutants was not altered (Figure S2). Our quantitative RT-PCR analysis indicated that expression of ARCK1 and CRK36 was elevated during seed germination and post-germinative growth (data not shown). These data suggest that their differing expression in seeds and seedlings affects their phenotype. To analyze the ABA hypersensitivity of the CRK36 RNAi seedlings during post-germinative growth, we measured the cotyledon greening rate (Figure 4d) and chlorophyll content of seedlings (Figure 4g). The results of these analyses indicated that CRK36 RNAi plants were more hypersensitive to ABA than wild-type plants, but the levels of ABA hypersensitivity differed between the analyses. The cotyledon greening rate differed by more than threefold between wild-type and CRK36 RNAi plants (Figure 4d), whereas differences in chlorophyll content were less than twofold (Figure 4g). Seedlings with expanded green cotyledons were counted to obtain the results shown in Figure 4(d), while total chlorophyll content of seedlings was measured to obtain the results shown in Figure 4(g). We speculate that chlorophyll was also extracted from non-expanded cotyledons, thus weakening the difference in Figure 4(g).

The microarray data showed that 20 LEA genes were up-regulated in the CRK36 RNAi plants (Table S5). LEA class proteins are widely assumed to play crucial roles in cellular tolerance, protecting macromolecules such as proteins and membranes from dehydration in both vegetative tissues and seeds (Hundertmark and Hincha, 2008). The Arabidopsis genome contains 51 LEA genes, 34 of which are expressed in seeds (Hundertmark and Hincha, 2008). Among the 34 seed-expressed genes, 19 were up-regulated in CRK36 RNAi transgenic plants (Table S5). Thirteen genes for seed storage proteins were also up-regulated in CRK36 knockdown plants, seven of which encoded oleosin family proteins (Table S5). Oleosins are abundant proteins in oil bodies and play a role in stabilizing oil bodies in seeds (Siloto et al., 2006). Nineteen genes involved in transcription were up-regulated in the CRK36 knockdown plants, including ABI4, encoding an AP2 domain transcription factor, and ABI5, encoding a bZIP transcription factor (Table S5). These two proteins are key regulators of ABA-responsive transcription in seeds (Finkelstein et al., 1998; Finkelstein and Lynch, 2000; Reeves et al., 2011). These results suggest that the ARCK1 and CRK36 have negative effects on the expression of various ABA-inducible seed-expressed genes, including those encoding important regulatory proteins for ABA-responsive gene expression in seeds.

Many genes encoding proteins putatively involved in proteolysis were also up-regulated in CRK36 knockdown plants (Table S5). These gene products include hypothetical proteins with RING finger motifs, F-boxes or U-boxes, which may interact with target proteins in the ubiquitin–proteasome pathway (del Pozo and Estelle, 2000; Azevedo et al., 2001). The proteolysis mediated by the ubiquitin–proteasome system is a key regulatory component of many cellular processes, including cell-cycle control, transcription and receptor desensitization (Smalle and Vierstra, 2004). In ABA signal transduction, the RING-finger ubiquitin E3 ligase KEG (KEEP ON GOING) functions as a negative regulator of ABA signaling, acting via degradation of ABI5 (Liu and Stone, 2010). DWD (DDB1-BINDING WD40 PROTEIN) HYPERSENSITIVE TO ABA 1 (DWA1) and DWA2, which are components of CUK4-based E3 ligases, also function as negative regulators (Lee et al., 2010). Although both KEG and DWA1/2 function negatively in ABA signaling, products of the up-regulated proteolysis-related genes in the knockdown plants may function positively in ABA signaling, indicating that ARCK1 and CRK36 could negatively control expression of these genes to reduce the effects of ABA and osmotic stress on germinating seeds.

Comparison of the downstream genes of CRK36 with those of SRK2D/E/I in ABA signaling indicated that CRK36 target genes partially overlap with SRK2D/E/I targets in responses to ABA (Figure 6 and Table S5). Moreover, our quantitative RT-PCR analysis indicated that various non-SRK2D/E/I target genes that are also ABA-responsive were up-regulated in arck1-2 and CRK36 RNAi lines compared with wild-type (Figure 6 and Table S5). These data suggest that CRK36 and ARCK1 indirectly influence ABA signaling via SRK2D/E/I during post-germinative growth of Arabidopsis. Furthermore, these data also raise the possibility that CRK36 and ARCK1 mediate unknown ABA signaling pathways other than the SnRK2-related pathway.

We also found that CRK36 phosphorylates ARCK1 in vitro, and over-expression of CRK36 leads to an increase in ARCK1 phosphorylation under salt stress in planta (Figure 7). On the other hand, in the in vitro phosphorylation assay, we detected only weak autophosphorylation activity of ARCK1 due to variations in the conserved amino acids of the kinase domain (Figures 7 and S6). RLKs containing alternative amino acids in place of the conserved amino acids within the kinase domain have been reported as atypical RLKs (Castells and Casacuberta, 2007). In fact, up to 20% of plant RLKs are atypical kinases (Castells and Casacuberta, 2007). Based on our findings regarding phosphorylation of ARCK1 by CRK36, the signal transduction mediated by atypical RLKs may be controlled by interaction with active kinases in protein complexes. Furthermore, perception of the unknown ligand(s) by the extracellular cysteine-rich repeat domain of CRK36 may strengthen ARCK1 phosphorylation under osmotic stress conditions.

The results of this study indicated that the abiotic stress-inducible protein ARCK1 physically interacts with CRK36 and forms a complex in the plasma membrane. The CRK36–ARCK1 complex plays a negative role in ABA and osmotic stress signaling during post-germinative growth. The CRK36–ARCK1 complex may be required for appropriate adjustment of plant growth in response to environmental conditions.

Experimental procedures

Plant materials and growth conditions

Arabidopsis thaliana plants of the Columbia (Col) ecotype were grown in Petri dishes on germination medium (MS medium containing 1% sucrose and 0.8% agar) for 3 weeks under a 16 h light/8 h dark regime. Two ARCK1 T-DNA knockout mutants of Arabidopsis [arck1-1 (SALK_057538) and arck1-2 (SALK_037588)] were obtained from the Arabidopsis Biological Resource Center (http://abrc.osu.edu/). The T-DNA insertion site was confirmed by PCR using a T-DNA left border primer (LBh1) and a pair of ARCK1-specific primers (ARCK1-RP1 and 2). All primer sequences are shown in Table S6.

RNA analyses and abiotic stress treatments

Total RNAs were extracted from the leaves or roots of 3-week-old Arabidopsis plants grown on GM (germination medium) agar plates using RNAiso reagent (TaKaRa, http://www.takara-bio.com). For real-time quantitative PCR analysis, cDNA was synthesized from total RNA using a high-capacity cDNA reverse transcription kit (Applied Biosystems, http://www.appliedbiosystems.com/) with random primers according to the manufacturer’s instructions, and PCR was performed using the ABI 7300 real-time PCR system (Applied Biosystems) and Power SYBR Green PCR Master Mix (Applied Biosystems). The primer pairs used for real-time PCR are listed in Table S6. For NaCl and ABA treatments, the plants were transferred into distilled water for 24 h prior to treatments, and subsequently cultured with 200 mm NaCl and 50 μm ABA for the indicated times. For drought and cold treatments, plants grown on MS agar plates were dehydrated on Parafilm (Sigma, http://www.sigmaaldrich.com/) or transferred to and maintained at 4°C for the indicated times.

Construction of transgenic plants

The ARCK1 promoter:GUS reporter plasmid was constructed by cloning PCR-amplified fragments containing a 2000 bp sequence of the ARCK1 promoter region. For promoter:GUS analysis, the primers ARCK1pro-BamHI-F and ARCK1pro-XhoI-R were used to amplify the ARCK1 promoter fragment. These fragments were ligated into the pGK-GUS vector (Qin et al., 2008), and GUS activity was determined as previously described (Mizuno et al., 2007). A construct containing the GFP–ARCK1 or CRK36–GFP fusion was generated as previously described (Mizuno et al., 2007). All primer sequences are shown in Table S6.

Germination assay

For germination assays to score ABA sensitivity, 120 seeds per treatment were surface-sterilized in 5% hypochlorite solution. Seeds were then rinsed four times with sterile water before plating on MS medium containing 1% sucrose and mixed isomers of ABA as indicated (Sigma). The dishes were incubated for 3 days at 4°C to break any residual dormancy, and then transferred to 22°C under a 16 h light/8 h dark regime. The number of seedlings with green cotyledons was scored at the indicated times.

Yeast two-hybrid analysis

Yeast two-hybrid analysis was performed using MatchMaker GAL4 Two-Hybrid System 3 (Clontech, http://www.clontech.com/) according to the manufacturer’s instructions. A yeast strain (AH109) was transformed with pairs of pGBKT7 vectors (Clontech, http://www.clontech.com/) harboring ARCK1 and CRK36 and pGADT7 vectors (Clontech) harboring genes of interest. The transformants were tested on SD (Synthetic Dextrose) screening medium.

Bimolecular fluorescence complementation (BiFC) assay

To generate BiFC constructs, ARCK1, CRK36 and other CRK full-length cDNAs with no stop codons were sub-cloned via SpeI/BamHI restriction sites into pre-digested 35S:VYNE and 35S:SCYCE vectors (Waadt et al., 2008). For transient expression, Agrobacterium tumefaciens strain GV3101 carrying each construct was used together with the p19 strain to infiltrate leaves of 5-week-old N. benthamiana plants (Voinnet et al., 2003). For microscopic analysis, lower epidermis cells were observed using a confocal laser scanning microscope (Zeiss M205C-SP) with LSM Image Browser software (http://www.zeiss.com/).

Protein extraction and protein gel-blot analysis

To prepare total protein extracts, leaves of 5-week-old N. benthamiana or 3-week-old transgenic Arabidopsis plants were ground into fine powder in liquid nitrogen and thawed in cold extraction buffer [50 mm Tris/HCl, pH 8.0, 5 mm MgCl2, 0.25 m sucrose, 5 mm EDTA, 50 mm NaCl and a protease inhibitor mixture (Roche Applied Science, http://www.roche.com)] using a mortar and pestle. The extracts were then centrifuged at 10 000 g for 10 min at 4°C, and the microsomal fraction was isolated by centrifugation of the supernatant at 100 000 g for 40 min at 4°C. The resulting microsomal membrane fraction was suspended in extraction buffer containing 0.5%n-octyl-β-glucoside (Wako Pure Chemical, http://wako-chem.co.jp). The concentration of extracted proteins was determined using the Bio-Rad protein assay (http://www.bio-rad.com/). The microsomal fractions were then separated by SDS–PAGE and blotted onto a polyvinylidine fluoride membrane (Immobilon-P; Millipore, http://www.millipore.com). The CRK36 protein was detected using polyclonal anti-GFP antibody (see Appendix S1 for description of antibody production), or monoclonal anti-hemagglutinin (HA) antibody (Sigma). Signals were developed using SuperSignal West Dura extended duration substrate (Pierce, http://www.piercenet.com).

Immunoprecipitation experiments

Plants were ground in liquid nitrogen and homogenized in an extraction buffer containing 50 mm Tris/HCl, pH 8.0, 5 mm MgCl2, 0.25 m sucrose, 5 mm EDTA, 50 mm NaCl and a protease inhibitor mixture (Roche Applied Science). The homogenate was filtered through Miracloth (Merck, http://www.merck.be), and centrifuged at 10 000 g for 10 min, and the supernatant was then ultracentrifuged at 100 000 g for 30 min to precipitate the microsomal membranes. This pellet was then solubilized using 0.5%n-octyl-β-glucoside after incubation for 2 h at 4°C with gentle shaking. CRK36–GFP and CRK36–SCYCE proteins were immunoprecipitated from the solubilized microsomal membranes using anti-GFP antibody microbeads (Miltenyi Biotech, http://www.miltenyibiotec.com) and anti-HA agarose beads (Sigma), respectively. The resulting immunoprecipitates were subjected to Western blot analysis using antibodies against GFP or HA (Sigma). The co-immunoprecipitated 3 x FLAG-tag–ARCK1 and VYNE–ARCK1 fusions were analysed by Western blotting using anti-FLAG M2 (Sigma) and anti-myc (Millipore) antibodies, respectively. For immunoprecipitation of VYNE–ARCK1, single and double transgenic plants were ground in liquid nitrogen and homogenized in extraction buffer with 0.5%n-octyl-β-glucoside, and the homogenates were incubated for 2 h at 4°C with gentle shaking. This preparation was filtered through Miracloth, and centrifuged at 10 000 g for 10 min. The supernatant was incubated with anti-c-myc antibody microbeads (Miltenyi Biotech), and the resulting immunoprecipitates were analyzed by Western blot using anti-myc and polyclonal anti-phosphothreonine (Cell Signaling Technologies, http://www.cellsignal.com) antibodies.

In vitro kinase assay

Protein kinase activity was demonstrated by incubating 1 μg of the protein to be tested in a total volume of 10 μl kinase buffer (50 mm Hepes, pH 7.4, 50 mm NaCl, 0.1% Triton X-100, 10 mm MgCl2, 10 mm MnCl2, 1 mm DTT and 10 μCi [γ-32P]ATP) for 30 min at 28°C. The reaction was terminated by adding an equal volume of SDS–PAGE sample buffer, and the mixture was then incubated for 15 min at 55°C. The phosphorylated proteins were separated by 10% SDS–PAGE, and the gel was subsequently dried and exposed to a Fujifilm imaging plate (http://www.fujifilm.com/). The phosphorylated proteins were visualized by autoradiography using a BAS (Bioimaging Analysis System) imaging plate scanner (GE Healthcare, http://www.gelifesciences.com/).

Chlorophyll content measurement

Chlorophyll content measurements were performed as described by Arnon (1949) by measuring the absorbance of 80% acetone extracts at 664 and 647 nm.

Microarray analysis

Genome-wide expression studies were performed using an Agilent Arabidopsis 3 Oligo Microarray (44K-feature format, Agilent Technologies, http://www.agilent.com) for wild-type and CRK36 RNAi transgenic plants grown on ABA-containing plates. In all experiments, we compared gene expression between wild-type and CRK36 RNAi plants. We grew young seedlings on plates containing 0.5 μm ABA for 10 days. Total RNAs were extracted from young seedlings using RNAiso reagent (TaKaRa), and used for microarray analysis. Microarray experiments with CRK36 RNAi plants were performed using two sets of wild-type plants and two sets of independent transgenic plants grown under the same conditions. For each experiment, two slides were analyzed by cy3 and cy5 dye swapping (Qin et al., 2008). Statistical analysis of microarray data was performed as previously described (Qin et al., 2008). All microarray data are available through ArrayExpress (http://www.ebi.ac.uk/arrayexpress/) (accession number E-MEXP-3354).

Accession numbers

Sequence data are available at the Arabidopsis Genome Initiative database under the following accession numbers: ARCK1 (At4g11890), CRK36 (At4g04490), CRK37 (At4g04500), CRK38 (At4g04510), CRK39 (At4g04540), CRK40 (At4g04570), RLK191 (At1g66880), RLK403 (At4g23140), RLK405 (At4g23160), BRI1 (At4g39400), RPK1 (At1g69270), RPK2 (At3g02130), FLS2 (At5g46330), ABI4 (At2g40220), ABI5 (At2g36270), OLEO4 (At3g27660) and LHCB1.2 (At1g29910).

Acknowledgments

We are grateful for the excellent technical assistance provided by K. Yoshiwara and M. Toyoshima (Japan International Research Center for Agricultural Sciences, Tsukuba, Japan). This work was supported by the Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (BRAIN), Grants-in-Aid for Scientific Research in Priority Areas from the Ministry of Education, Culture, Sports, Science, and Technology (number 17078003 to K.Y.-S.), and a Grant-in-Aid for Scientific Research C from the Japan Society for the Promotion of Science (number 21580125 to Y.O.).

Ancillary