Interaction of the Arabidopsis Receptor Protein Kinase Wak1 with a Glycine Rich Protein AtGRP-3

The Wak1, is a member of family (Wak1-5) that links the plasma membrane to the extracellular matrix. By the yeast two hybrid screen, we found that a glycine rich extracellular protein, AtGRP-3, binds to the extracellular domain of Wak1. Further in vitro binding studies indicate that AtGRP-3 is the only isoform among the six tested AtGRPs that specifically interacts with Waks and the cysteine-rich carboxy-terminus of AtGRP-3 is essential for its binding to Wak1. We also show that Wak1 and AtGRP-3 form a complex with a molecular size of approximately 500 kDa in vivo in conjunction with the kinase associated protein phosphatase, KAPP, that has been shown to interact with a number of plant receptor-like kinases. Binding of AtGRP-3 to Wak1 is shown to be crucial for the integrity of the complex. Wak1 and AtGRP-3 are both induced by salicylic acid treatment. Moreover, exogenously added AtGRP-3 upregulates the expression of Wak1 , AtGRP-3 , and PR-1 (for pathogenesis-related) in protoplasts. Taken together, our data suggest that AtGRP-3 regulates Wak1 function through binding to the cell wall domain of Wak1 and that the interaction of Wak1 with AtGRP-3 occurs in a pathogenesis-related process in planta


INTRODUCTION
Extracellular matrix (ECM) is a dynamic zone harboring active components that regulate cell-cell interactions of developmental processes and responses to the environment (1). ECM molecules interact with plasma membrane proteins to initiate and modulate diverse signaling pathways that play key roles in essential cellular processes (2). While these events are well understood in animals (3), their study in land plants is in its infancy (4)(5)(6)(7)(8)(9). The interactions of cell surface receptors with ligands and extracellular molecules in ECM or cell wall may be involved in these events.
Waks, cell wall associated receptor kinases, physically link the plant ECM to the plasma membrane and can also serve as signaling molecules (10,11). They contain an extracellular domain that mediates tight association with the cell wall, a transmembrane domain, and a cytoplasmic protein kinase domain (12). While Waks are highly conserved in the kinase domains, they share only 40 to 64% sequence identities in the extracellular domains. Waks have in their extracellular domains epidermal growth factor (EGF)-like repeats and regions similar to the known ECM proteins that include tenascins, collagens, and neurexins (12). Considering the less conserved sequences and the sequence similarity with various ECM proteins in the extracellular domains, it is supposed that the Wak isoforms are functionally distinct and mediate diverse extracellular signals from ECM to the cells. Interaction of Wak1 with AtGRP-3 4 Wak1 was shown to be implicated in the defense mechanism against pathogen (13).
Wak1 is induced by pathogen or exogenously added salicylic acid (SA) or 2,6dichloroisonicotinic acid (INA), which requires a positive regulator NPR1/NIM1 (14). While Wak1 has thus been regarded as a pathogenesity-related (PR) protein, the physiological functions of Waks are largely unknown. Identification of the upstream and/or downstream molecules of Waks is crucial to elucidate the physiological roles of Waks and the molecular mechanisms underlying the Wak activities.
In this study, we report that a glycine rich secreted protein, AtGRP-3 (15), specifically binds to the extracellular domain of Wak1. The coimmunoprecipitation and size exclusion chromatography experiments combined with Western blotting show that Wak1, AtGRP-3, and kinase associated protein phosphatase (KAPP) are associated into a multimeric complex of 500 kDa in size in vivo and the binding of AtGRP-3 to Wak1 is essential for the structural integrity of the complex. We also demonstrate that AtGRP-3 induces PR-1, a pathogenrelated gene, and triggers a positive feedback loop that upregulates the expression of Wak1 and AtGRP-3. We propose that AtGRP-3 is an extracellular molecule that binds to and regulates Wak functions in Arabidopsis.

Yeast Two Hybrid Screen
The MATCHMAKER Two-Hybrid System (Clontech) was used for screening. The Arabidopsis cDNA library was constructed in GAL4 activation domain vector (pGAD424) as described (16). The bait construct was prepared by cloning the region for the extracellular domain of Wak1 (amino acids 178-334) into GAL4 binding domain vector (pGBT9). The HF7c yeast strain was sequentially transformed with the bait construct and then with the cDNA library. Transformants (9x10 6 ) were screened for activation of the HIS3 and lacZ reporter genes. The full-length AtGRP-3 and AtGRP-3S genes were isolated from the cDNA library by polymerase chain reaction (PCR). The interaction between the two proteins was quantitated by liquid β-galactosidase assay using o-nitrophenyl β-D-galactopyranoside (ONPG) as substrate.

Construction and Purification of Recombinant Proteins
For Waks, the coding region for the extracellular domain was cloned into pGEX vector

Antibody Generation and Protein Analysis
Polyclonal antibodies were raised in mouse to recombinant AtGRP-3 protein and in rabbit to Wak1-specific peptide (amino acids 98-112). The specificities of antibodies were confirmed with both recombinant proteins and plant extracts (Fig. 3 A and B). In Western analysis, the proteins were separated on 12.5% SDS-polyacrylamide gels, transferred onto nitrocellulose membranes, and incubated with these antibodies. The antibody-bound proteins were detected by incubation with secondary antibodies conjugated to horseradish peroxidase using the ECL system (Amersham).

Size Exclusion Chromatography
Protein extracts in the buffer B were applied to size exclusion fast protein liquid chromatography (FPLC). Proteins were fractionated over Superose 6 HR 10/30 column (Pharmacia) equilibrated with the running buffer (50 mM Tris, pH 7.6, 50 mM NaCl).
Fractions (0.5 ml) were collected at a flow rate of 0.3 ml/min.

RNA Analysis
RNA was extracted using 0.5 ml of Trizol reagent (GIBCO-BRL) per 10 6 cells of protoplasts. Contaminated genomic DNA from total RNA was removed as described (18).
Total RNA (1 µg) was treated with 6 units of RNase-free DNase I (GIBCO-BRL) for 5 min at

Interaction of Wak1 with Glycine Rich Proteins AtGRP-3 and AtGRP-3S
To identify proteins that bind to the extracellular domain of Wak1, we have performed the yeast two hybrid screen (19). A library of Arabidopsis seedling cDNAs fused to the GAL4 activation domain was screened against the construct encoding the extracellular domain of Wak1 fused to the GAL4 DNA binding domain. Among the positive clones, of interest was AtGRP-3 that is a member of a previously described family of glycine-rich secreted cell wall proteins (20)(21)(22). While we were trying to isolate full length AtGRP-3 cDNA by PCR, we identified a closely related sequence that is 81% identical to AtGRP-3 and lacks ca. 25 amino acids near the amino-terminus (Fig. 1A). This short form of AtGRP-3 is hereafter referred to as AtGRP-3S.
To confirm the interaction between Wak1 and AtGRP-3, the extracellular domain of

Isoform-Specific Interaction between Wak1 and AtGRP-3
Five Wak genes (Wak1-Wak5) (12) and multiple AtGRP genes (15,(23)(24)(25) have previously been identified from Arabidopsis. The complete Arabidopsis genome sequence also reveals that there exist additional Wak and GRP family members. Therefore, we tested whether the interaction between Wak1 and AtGRP-3 we have shown above is isoformspecific or it merely represents one of the general interactions between Waks and AtGRPs. Interaction of Wak1 with AtGRP-3 12 the complexes were analyzed (Fig. 2B). The results indicated that AtGRP-3 binds to Wak1, Wak3, and Wak5. We obtained the same results with AtGRP-3S (data not shown), implying that AtGRP-3 and AtGRP-3S may play a similar physiological role. Notably, AtGRP-2, 4, 6, 7, and 8 did not bind to any of the Wak isoforms (data not shown), indicating that AtGRP-3 and AtGRP-3S are unique among AtGRPs in their ability to interact with Waks.
We next localized the region in AtGRP-3 that is responsible for the interaction with Wak1. Constructs for the amino-and carboxy-terminus deleted forms (AtGRP-dN and AtGRP-dC, respectively) and carboxy-terminal region (AtGRP-CT) of AtGRP-3 were generated (Fig. 2C). The proteins were in vitro-translated and subjected to the in vitro binding assay with GST alone or GST-ED (Fig. 2D). Interactions were observed with full length AtGRP-3, AtGRP-dN and AtGRP-CT but not with AtGRP-dC, indicating that the cysteinerich carboxy-terminus is essential for the association of Wak1 and AtGRP-3.

Association of Wak1, AtGRP-3, and KAPP
We have generated anti-Wak1 and anti-AtGRP-3 antibodies as described in the "Experimental Procedures". The anti-Wak1 antibody specifically recognized only the ED of Wak1 but not those of other tested Wak proteins (Fig. 3A). Western blotting of Arabidopsis protein extracts with anti-Wak1 antibody detected two bands that migrate as approximately 78 kDa and 100 kDa proteins on a SDS-gel. Since the expected size of Wak1 is 78 kDa, the Interaction of Wak1 with AtGRP-3 13 slowly migrating band is likely to be a posttranslationally modified form (e.g. glycosylated) of Wak1 (Fig. 3B, lane 1). Anti-AtGRP-3 antibody recognized an 18 kDa band that is larger than the expected 12 kDa of AtGRP-3 (lane 2). This discrepancy is unlikely to be due to a posttranslational modification, because the in vitro translated (Fig. 1C) or recombinant AtGRP-3 protein purified from E. coli (Fig. 3A) also migrated as an 18 kDa protein. Protein extracts from Arabidopsis seedlings were subjected to immunoprecipitation with preimmune or anti-AtGRP-3 antibody (Fig. 3C). The precipitated proteins were separated on a SDS-gel and probed with anti-Wak1 antibody. The results showed that both 78 and 100 kDa Wak1 proteins were co-precipitated with AtGRP-3, indicating that Wak1 and AtGRP-3 form a complex in vivo.
As KAPP has been shown to interact with a number of plant receptor-like kinases (RLKs) (26), we tested whether KAPP is also associated with the Wak1-AtGRP-3 complex.
The proteins immunoprecipitated by anti-AtGRP-3 antibody were probed with anti-KAPP antibody that recognized a 65 kDa KAPP protein on a Western blot (Fig. 3B, lane 3). The presence of KAPP in the immunoprecipitates (Fig. 3C) suggests that KAPP is a component of the complex that includes Wak1 and AtGRP-3.

Association of Wak1 and AtGRP-3 in the Multimeric Complex in Vivo
To further analyze the protein complex of Wak1, AtGRP-3, and KAPP, protein extracts Interaction of Wak1 with AtGRP-3 14 of Arabidopsis seedlings were subjected to size exclusion chromatography. The fractions were then probed with anti-Wak1, anti-AtGRP-3, and anti-KAPP antibodies (Fig. 4A). The molecular sizes were extrapolated from a standard curve prepared with a set of marker proteins. Wak1 proteins were present in two complexes with the apparent molecular sizes of 500 and 200 kDa, respectively. It is noticeable that 100 kDa and 78 kDa Wak1 proteins were separated into the 500 and 200 kDa complexes. AtGRP-3 was fractionated along with the 500 kDa complex, but not with the 200 kDa complex. This suggests that AtGRP-3 is stably associated with the 100 kDa Wak1 in the 500 kDa complex. AtGRP-3 that runs as a 48 kDa is likely the unbound form of AtGRP-3. KAPP was detected mostly in the range of 500-600 kDa, implying the association of KAPP with a number of RLKs as reported (26).
To elucidate the role of AtGRP-3 in the formation of the 500 kDa complex, we analyzed protein extracts of protoplasts by size exclusion chromatography (Fig. 4B). AtGRP-3 was completely absent in the protoplasts as determined by Western blotting. In the absence of AtGRP-3, Wak1 was detected mostly in the 200 kDa complex (−AtGRP-3). Upon the addition of the recombinant AtGRP-3 protein to protoplasts, Wak1 reappeared associated with the 500 kDa complex and the added AtGRP-3 was also eluted along with the 500 kDa complex (+AtGRP-3). These data suggest that AtGRP-3 is essential for the structural integrity and/or probably the formation of 500 kDa complex that includes Wak1 and AtGRP-  15 ineffective (+AtGRP-dC), suggesting that the direct interaction between Wak1 and AtGRP-3 is critical for the formation and/or maintenance of the integrity of the 500 kDa complex.

Co-expression of Wak1 and AtGRP-3 in Tissues and Co-induction by Salicylic Acid Treatment
In an effort to evaluate the biological relevance of the interaction between Wak1 and AtGRP-3, we examined the expression pattern of Wak1 and AtGRP-3 in Arabidopsis tissues by northern analysis. Wak1 and AtGRP-3 were co-expressed predominantly in leaves and stems (Fig. 5A), which is consistent with previous reports (12,13,15). It was also shown that Wak1 expression is induced by SA and this induction requires the positive regulator NPR1/NIM1 (12). We thus analyzed the expression of Wak1 and AtGRP-3 in response to SA.
Northern analysis showed that both AtGRP-3 and Wak1 were induced upon SA treatment (Fig.   5B). This induction was abolished by mutation in npr1 and by overexpression of nahG which degrades SA (27,28) (Fig. 5C). These data suggest that the functions of Wak1 and AtGRP-3 are closely related and implicated in the SA signaling.

Upregulation of Gene Expression by Exogenously Added AtGRP-3 in Protoplasts
We tested by RT-PCR whether the treatment of protoplasts with AtGRP-3 resulted in the regulation of gene expression (Fig. 6). The expression of Wak1 and AtGRP-3 was upregulated by the addition of AtGRP-3 in a dose-dependent manner (Fig. 6 A and C), Interaction of Wak1 with AtGRP-3 16 indicating that Wak1 and AtGRP-3 are in a positive feedback loop that enhances the AtGRP-3-Wak1 signaling. The addition of AtGRP-3 to protoplasts also induced PR-1, a molecular marker of SA-dependent defense response, but not PDF1.2 (for plant defensin), a molecular marker of JA-dependent defense response (29,30). It is worth noting that the expression of Wak1 and AtGRP-3 was induced by SA (Fig. 5B). In contrast to AtGRP-3, the exogenously added AtGRP-dC and bovine serum albumin (BSA) barely affected the expression level of the tested genes ( Fig. 6 B and D). Taken together with the fact that AtGRP-3 binds to Wak1 in a specific manner, these data suggest that AtGRP-3 is a physiological ligand of Wak1, and the AtGRP-3-Wak1 signaling pathway is involved in the SA-dependent defense response. Waks are cell wall-associated kinases and the different isoforms vary significantly in the extracellular domains (12). This raises the possibility that the extracellular domains of Wak proteins may be involved in the interaction with diverse wall components in cell type-or stage-dependent manner. Waks may perceive differential signals from the changing environment through the interaction of extracellular domain with cell wall molecules and transduce the signals to the cell.
AtGRP-3 is a member of GRP family and likely a cell wall protein with a predicted signal peptide at the amino-teminus (15). Its amino acid sequence has high content of glycine (31%) and its mid-region is marked by a glycine-rich domain of 57 amino acids with the sequence Gly 4 -Asn/Arg-Tyr-Gln repeated six times. The signal peptide and carboxy-terminal cysteine-rich motif show sequence similarity to several nodule-specific plant-encoded proteins (nodulins) (31). In a recent report, the cysteine-rich carboxy-terminal region of by guest on March 21, 2020 http://www.jbc.org/

Downloaded from
Interaction of Wak1 with AtGRP-3 18 NtTLRP, a cell wall protein in tobacco that is highly homologous to that of AtGRP-3 was sufficient for cross-linking of proteins to cell walls (32). However, the physiological functions of GRPs are unknown. The sequences of the GRP family members are divergent with little conserved regions, which suggests that GRPs may play diverse biological roles (21).
In this study, we have shown that AtGRP-3 specifically binds to Wak1 both in vitro and in vivo. We also provide several lines of evidence suggesting that Wak1-AtGRP-3 interaction is physiologically relevant and the binding of AtGRP-3 regulates the signaling activity of Wak1. First, Wak1, AtGRP-3 and KAPP are associated into a 500 kDa complex in vivo that may represent the activated signalosome (33)(34)(35) and the direct binding of AtGRP-3 to Wak1 is critical for the integrity of the 500 kDa complex. These interactions share many features with the regulation of CLAVATA1 (CLV1) receptor kinase (36,37). Second, exogenously added recombinant AtGRP-3 to protoplasts upregulated the expression of PR-1 as well as The noticeable feature of the carboxy-terminal region of AtGRP-3 is the presence of six conserved cysteine residues. The sequence alignment of Wak isoforms also exhibits the conserved cysteine residues in twenty-four places (12). It would be interesting to see whether the conserved cysteine residues in AtGRP-3 and Wak1 participate in the interaction between two molecules and with others. Disulfide bonding is reversibly dependent on the redox state in the cell wall that would be changing in different developmental stages. The reactive oxygen species generated under various stress conditions such as wounding, pathogen attack, cold or ozone stress would also affect the redox condition of the cell wall (38)(39)(40)(41). There have been reports that the oxidative cross-linking of the cell wall proteins plays an important role by guest on March 21, 2020 http://www.jbc.org/

Downloaded from
Interaction of Wak1 with AtGRP-3 20 as a defense mechanism in response to pathogen attack (42,43). It is conceivable that the oxidative modifications of cysteine and other residues may regulate the interaction of AtGRP-3, Wak1, and other proteins and therefore the signaling of Wak1. In view of the finding that both Wak1 and AtGRP-3 are induced in response to SA and modulate PR-1 in a positive way as shown in this study and others (13,15), it appears likely that Wak1 and AtGRP-3 are involved in pathogen response. AtGRP-3 and its homologues in tobacco and petunia have expression patterns that suggest a role in the defense mechanism (44,45). It is also conceivable that oxidative change in AtGRP-3 and Wak1 proteins may be the underlying mechanism of pathogenesis-related response of Wak1 and AtGRP-3.
To date only a few plant plasma membrane receptors have been characterized in association with their ligands. A good example is the meristem regulator CLV1, the Arabidopsis receptor kinase that regulates cell proliferation and organ formation at the meristems (46). Both genetic and biochemical studies provided evidence that CLV3 acts as the ligand for CLV1 as part of a multimeric comlex (34,35,47). Another illustrative example would be BRI1 and SRK that are implicated in brassinosteroid signaling and selfincompatibility, respectively (48)(49)(50). Here the results add to the observation from other RLKs, that plant receptor kinases can be regulated by soluble secreted ligands.