Selective regulation of the chitin-induced defense response by the Arabidopsis receptor-like cytoplasmic kinase PBL27

Authors


Summary

Recognition of microbe-associated molecular patterns (MAMPs) initiates pattern-triggered immunity in host plants. Pattern recognition receptors (PRRs) and receptor-like cytoplasmic kinases (RLCKs) are the major components required for sensing and transduction of these molecular patterns. However, the regulation of RLCKs by PRRs and their specificity remain obscure. In this study we show that PBL27, an Arabidopsis ortholog of OsRLCK185, is an immediate downstream component of the chitin receptor CERK1 and contributes to the regulation of chitin-induced immunity in Arabidopsis. Knockout of PBL27 resulted in the suppression of several chitin-induced defense responses, including the activation of MPK3/6 and callose deposition as well as in disease resistance against fungal and bacterial infections. On the other hand, the contribution of PBL27 to flg22 signaling appears to be very limited, suggesting that PBL27 selectively regulates defense signaling downstream of specific PRR complexes. In vitro phosphorylation experiments showed that CERK1 preferentially phosphorylated PBL27 in comparison to BIK1, whereas phosphorylation of PBL27 by BAK1 was very low compared with that of BIK1. Thus, the substrate specificity of the signaling receptor-like kinases, CERK1 and BAK1, may determine the preference of downstream RLCKs.

Introduction

Plants have the ability to detect invading microbes through the perception of conserved microbial molecules (Boller and Felix, 2009; Silipo et al., 2010). Recognition of these microbe- (or pathogen-) associated molecular patterns (MAMPs or PAMPs, respectively) initiates pattern-triggered immunity (PTI) in host plants. Several MAMPs, including fungal chitin, bacterial flagellin, and peptidoglycan (PGN), have been reported to elicit PTI in a broad range of host plants. Cell surface receptor-like kinases (RLKs) and receptor-like proteins (RLPs) are the major components required for sensing these molecular patterns (Antolin-Llovera et al., 2012; Monaghan and Zipfel, 2012). Corresponding pattern recognition receptors (PRRs), such as CEBiP/CERK1 for chitin, LYM1/LYM3/CERK1 for PGN, and FLS2 for flagellin, are activated by ligand binding and trigger various immune responses (Gomez-Gomez and Boller, 2000; Kaku et al., 2006; Miya et al., 2007; Wan et al., 2008b; Shimizu et al., 2010; Willmann et al., 2011).

Two types of lysin motif (LysM)-containing membrane proteins, CEBiP and CERK1, a receptor-like protein and a receptor-like kinase, respectively, have been reported as the components of plant chitin receptors (Kaku et al., 2006; Miya et al., 2007; Wan et al., 2008b; Shimizu et al., 2010). The rice chitin receptor system requires both CEBiP and OsCERK1 for chitin perception and signaling, whereas the CERK1 receptor kinase is sufficient for both ligand perception and signaling in Arabidopsis (Shinya et al., 2012). Although Arabidopsis possesses a CEBiP homolog, LYM2/AtCEBiP, which specifically binds chitin oligosaccharides similar to rice CEBiP, LYM2 does not contribute to CERK1-mediated chitin signaling (Shinya et al., 2012; Wan et al., 2012). On the other hand, the lym2 mutant showed increased susceptibility to fungal pathogens, similar to the cerk1 mutant, which raises the question of the mode of action of LYM2 in disease resistance (Faulkner et al., 2013; Narusaka et al., 2013). Faulkner et al. (2013) recently reported that LYM2 contributes to disease resistance independent of CERK1 through the regulation of molecular flux via plasmodesmata, indicating the presence of a unique mechanism of disease resistance. In addition, rice homologs of Arabidopsis LYM1 and LYM3, namely LYP4 and LYP6, were reported to be involved in perception of both chitin and PGN (Liu et al., 2012a), although LYM1 and LYM3 were not involved in chitin perception in Arabidopsis (Willmann et al., 2011). In addition, another LysM receptor-like kinase, LYK4, was indicated to contribute to chitin signaling in Arabidopsis (Wan et al., 2012).

Ligand-induced formation of homo- and hetero-oligomeric receptor complexes is often an important step for ligand perception and activation of PRRs. In the rice chitin receptor system, CEBiP binds (GlcNAc)7/8 in an unusual sandwich-like manner, resulting in the dimerization of CEBiP and the formation of a receptor complex consisting of both CEBiP and OsCERK1 (Hayafune et al., 2014). In contrast to the rice chitin receptor complex, Arabidopsis does not require a CEBiP-like molecule for chitin perception (Shinya et al., 2012). The observation that CERK1 binds chitin directly indicates that the receptor kinase CERK1 serves for both the perception and the transduction of chitin oligosaccharides in Arabidopsis. Consistent with this notion, Liu et al. (2012b) demonstrated that CERK1 forms a homodimer in a ligand-dependent manner and that dimerization is critical for its activation and chitin-induced immune responses.

FLS2 is a leucine-rich repeat (LRR)-RLK that detects bacterial flagellin through the epitope flg22 (Chinchilla et al., 2006). The Arabidopsis FLS2–FLS2 complex is present before the perception of flg22 (Sun et al., 2012). In addition, FLS2 forms a ligand-induced complex with BAK1, which is another LRR-RLK belonging to the SERK family and a master regulator of immune responses mediated by several LRR-RLKs (Schulze et al., 2010; Roux et al., 2011; Schwessinger et al., 2011; Monaghan and Zipfel, 2012; Sun et al., 2013). After ligand-induced complex formation, both FLS2 and BAK1 are phosphorylated, a step that appears to be critical for the activation of downstream signaling.

With regard to the signaling components immediately downstream of PRRs, receptor-like cytoplasmic kinases (RLCKs) have been suggested to play a critical role in inducing immune responses (Monaghan and Zipfel, 2012; Yamaguchi et al., 2013b). Arabidopsis BIK1 and its closely related homolog PBL1 are RLCKs that directly interact with several PRRs, including FLS2 and CERK1 (Lu et al., 2010; Zhang et al., 2010; Liu et al., 2013). After the perception of flg22, BIK1 and PBL1 are phosphorylated and dissociate from the FLS2/BAK1 complex. It has been demonstrated that knockout mutants of BIK1 are compromised with respect to several flg22-induced defense responses. Zhang et al. (2010) reported that BIK1 and PBL1 are also involved in the regulation of CERK1-mediated chitin responses, including the accumulation of reactive oxygen species (ROS) and the induction of defense genes. On the other hand, Lu et al. (2010) reported that BIK1 phosphorylation was induced by flg22 but not by chitin. Interestingly, it has been reported that BIK1 does not regulate flg22-induced mitogen-activated protein kinase (MAPK) activation (Feng et al., 2012). These results suggest the presence of other signaling components downstream of these PRRs, particularly those involved in the regulation of the MAPK cascade.

A survey of the host targets of Xoo1488, an effector of the rice pathogen Xanthomonas oryzae, identified a receptor-like cytoplasmic kinase, OsRLCK185 (Yamaguchi et al., 2013a). OsRLCK185 directly interacts with OsCERK1 at the plasma membrane and regulates chitin- and PGN-induced defense responses. After chitin perception, OsRLCK185 is phosphorylated by and dissociates from OsCERK1 similar to flg22-mediated BIK1 activation in Arabidopsis. OsRLCK185 regulates MAMP-induced activation of MAPK, which is in contrast to the inability of BIK1 to activate MAPK in flg22 responses (Feng et al., 2012; Yamaguchi et al., 2013a). Suppression of OsRLCK185 by either Xoo1488 or RNA interference enhanced the susceptibility to a Xoo hrpX mutant, suggesting that OsRLCK185 plays an important role in OsCERK1-mediated plant immunity in rice. Although BIK1 and BAK1 have been reported as the components of the signaling complexes of multiple PRRs, it is unclear whether OsRLCK185 is sufficient for other PRR-mediated defense signaling.

Here we show that PBL27, an Arabidopsis ortholog of OsRLCK185, is an immediate downstream component of CERK1 and contributes to the regulation of chitin-induced immunity in Arabidopsis. On the other hand, we observed that the contribution of PBL27 to flg22 signaling seems very limited, suggesting that PBL27 selectively regulates defense signaling downstream of specific PRR complexes. In vitro phosphorylation experiments suggested that the substrate specificity of the signaling receptor-like kinases, CERK1 and BAK1, determines the preference of downstream RLCKs. It was also shown that PBL27 regulated chitin-induced activation of MAPKs but not accumulation of ROS, differently from BIK1.

Results

PBL27 selectively interacts with CERK1 at the plasma membrane

A database search for the Arabidopsis homolog of OsRLCK185 identified the closest homolog, PBL27 (At5g18610) – a receptor-like cytoplasmic kinase having the highest sequence similarity to OsRLCK185. PBL27 belongs to the RLCK VII subfamily, which includes several defense-related RLCKs such as the proteins BIK1, BSR1, OsRLCK185, PBS1, and PBS1-like (PBL; Shiu et al., 2004; Zhang et al., 2010). To evaluate the function of PBL27 in chitin signaling in Arabidopsis we first investigated the subcellular localization of PBL27 and its interaction with CERK1.

A fusion protein, PBL27–GFP, was transiently expressed in Nicotiana benthamiana to identify its subcellular localization. In addition, GFP itself was expressed in N. benthamiana as a control. Fluorescence of PBL27–GFP was localized on the plasma membrane (Figure 1a), whereas fluorescence of the control GFP was detected in the cytosol and nucleus. Plasmolysis experiments further clarified localization of PBL27–GFP at the plasma membrane. Transiently expressed PBL27-4Myc was detected in the microsomal fraction from N. benthamiana (Figure 1b, ‘Input α-Myc’). These results suggest that PBL27 is a plasma membrane-anchored protein.

Figure 1.

PBL27 interacts with CERK1 at the plasma membrane. (a) Subcellular localization of PBL27. PBL27-GFP and GFP were expressed under the control of the 35S promoter in Nicotiana benthamiana. Microscopic observation of PBL27-GFP after plasmolysis (plasmolyzed cell) is also shown (lower panel). The upper right corner shows a magnified image of the boxed area. Scale bar = 50 μm. (b) Analysis of the interaction of PBL27 and CERK1 by co-immunoprecipitation. Microsomal proteins from N. benthamiana expressing PBL27-4Myc and CERK1-3HA were prepared and co-immunoprecipitated with an anti-hemagglutinin (HA) antibody. The presence of PBL27-4Myc and CERK1-3HA in the complex was determined by immunoblot analysis with anti-Myc antibody or anti-HA antibody. (c) Visualization of the interaction between PBL27 and CERK1/FLS2 using bimolecular fluorescence complementation in N. benthamiana. CERK1D441V/FLS2-VenusN and PBL27-VenusC contained the N-terminal and C-terminal fragments of Venus, respectively. Venus fluorescence indicates the presence of the interaction between PBL27 and CERK1. Scale bar = 50 μm. (d) Interaction of PBL27 with the intracellular domain of CERK1 in yeast two-hybrid system.

Interactions between PBL27 and CERK1 were assessed by co-immunoprecipitation and bimolecular fluorescence complementation (BiFC) assays using transient expression in N. benthamiana. PBL27-4Myc was observed to co-immunoprecipitate with both CERK1-3HA and a kinase-inactive CERK1D441V mutant that carried an amino acid substitution at the active site of the kinase domain (Figure 1b). However, the recovery of PBL27 was much higher in the immunoprecipitate with CERK1D441V than in that with wild-type CERK1. Because a major part of the wild-type CERK1 that was expressed in N. benthamiana was already autophosphorylated (Figure S1), these observations indicated that PBL27 preferentially interacts with the unphosphorylated CERK1. As a phosphorylation-dependent dissociation from corresponding receptor-like kinases has been reported for OsRLCK185 and other cytoplasmic kinases (Lu et al., 2010; Zhang et al., 2010; Yamaguchi et al., 2013a), the above observation might indicate a similar case for PBL27, though further experiments would be required to confirm such a possibility. In addition, we observed PBL27–CERK1D441V interaction at the plasma membrane by BiFC assay using the transient expression of PBL27-VenusC and CERK1D441V-VenusN in N. benthamiana (Figure 1c). PBL27–CERK1D441V interaction was also detected by a yeast two-hybrid system (Figure 1d). These results indicate that PBL27 interacts with CERK1 at the plasma membrane similar to OsRLCK185, suggesting a function in CERK1-mediated chitin signaling in Arabidopsis.

Selective regulation of MAMP-induced responses by PBL27

To identify the biological function of PBL27, we investigated MAMP-triggered responses of two lines of pbl27 mutants (Figure S2). In addition to chitin responses, we investigated flg22-induced defense responses to determine whether PBL27 functions in different MAMP signaling pathways. Chitin-induced MAPK activation and callose deposition in the pbl27 mutants were clearly reduced compared with Col-0 (Figure 2a,b). The time course of MAPK activation (Figure S3) indicated that pbl27 mutation caused a decrease of MAPK activation rather than delaying activation. In addition, chitin-induced expression of two defense-related genes significantly decreased in the pbl27 mutants (Figure 2c). On the other hand, chitin-induced generation of ROS evaluated in both Arabidopsis seedlings and leaf disks did not show a decrease in the pbl27 mutants (Figures 2d and S4). Interestingly, flg22-induced callose deposition and ROS generation were not impaired in the pbl27 mutants (Figure 2a,d). The activation of MAPK by flg22 treatment was even slightly increased in the pbl27 mutants (Figure 2b). Flg22-induced defense gene expression was not significantly affected in the pbl27 mutants (Figure 2c). These results indicate that PBL27 selectively regulates chitin-induced defense responses. The fact that hemagglutinin (HA)-tagged PBL27 successfully complemented the pbl27 mutant for chitin-triggered MPK3/6 activation further supported the importance of PBL27 in chitin responses (Figure 2e).

Figure 2.

PBL27-mediated chitin responses in Arabidopsis. (a) (GlcNAc)7 and flg22-induced callose deposition in the pbl27 mutants. Seedlings were treated with 10 μm (GlcNAc)7 or 100 nm flg22 for 18 h. Data represent the mean of seven or eight cotyledons ± SD. The asterisks indicate statistical significance from the wild-type (WT) controls by Student's t-test (< 0.01). (b) Chitin and flg22-induced activation of mitogen-activated protein (MAP) kinase (MAPK) in pbl27 mutants. Total proteins were extracted from seedlings treated with 10 μm (GlcNAc)7 or 1 μm flg22. Activation of each MAPK was analyzed by immunoblot with anti-pMAPK antibody. The protein loading control was shown by Ponceau staining. Experiments were repeated at least three times with similar results. (c) Expression of defense-related genes induced by (GlcNAc)7 in the pbl27 mutants. Gene expression was evaluated 180 min after treatment with 40 μm (GlcNAc)7 or 100 nm flg22. Data are shown as the average of three independent biological replicates ± SD. The asterisks indicate statistically significant differences from the WT controls by Student's t-test (< 0.05). (d) Generation of reactive oxygen species (ROS) of pbl27 measured by leaf disk assay. The ROS were evaluated 30 min after treatment with 20 μg ml−1 (GlcNAc)7 or 100 nm flg22. Data are shown as the mean of 36 leaf disks ± SD. Statistical treatment of the data by anova followed by the Tukey HSD test (< 0.01) did not show significant differences between pbl27-2 and wild type Col-0. (e) Complementation of chitin-induced MAPK activation in pbl27 mutant by PBL27-3HA. Total proteins were extracted from seedlings treated with 10 μm (GlcNAc)7. The expression of PBL27-3HA was detected by immunoblot with anti-HA antibody. Activation of each MAPK was analyzed by immunoblot with anti-pMAPK antibody. The protein loading control was shown by Ponceau staining. Experiments were repeated at least three times with similar results.

To evaluate the contribution of PBL27 to disease resistance, the development of lesions after the inoculation of a fungal pathogen, Alternaria brassicicola, was evaluated for both the pbl27 mutants and wild type Col-0. Lesion development significantly increased in the pbl27 mutants, similar to the cerk1 mutant (Figure 3a). Resistance to Pseudomonas syringae pv. tomato DC3000 hrcC, a non-pathogenic strain in which secreting type-III effectors are absent, was also impaired in the pbl27 mutants, as evidenced by the increased growth of the bacteria in these mutants (Figure 3b). These results indicated the contribution of PBL27 to disease resistance against both fungal and bacterial infections.

Figure 3.

Disease resistance of pbl27 mutants. (a) pbl27 mutants are compromised with respect to resistance to Alternaria brassicicola. Lesion sizes were measured 6 days after inoculation. Data are shown as mean ± SE of 98–113 lesions. Asterisks indicate significant a difference from the wild-type (WT) controls by Student's t-test (< 0.01). (b) pbl27 mutants are compromised with respect to resistance to Pseudomonas syringae pv. tomato DC3000 hrcC. Bacterial titers were measured at 3 days after inoculation. Data are shown as means ± SD of six independent samples. Six independent experiments were performed with similar results. Asterisk indicates significant difference from the WT controls by Student's t-test (< 0.05).

We note that the MAPK activation and disease resistance phenotypes of the pbl27-2 mutant allele were sometimes unstable. We speculate that this instability was because of the insertion of the T-DNA into the C-terminal region of the PBL27 gene.

Substrate specificity of signaling receptor-like kinases correlates with the choice of downstream RLCKs

Differently tagged CERK1 kinase domains and PBL27 were recombinantly expressed in Escherichia coli and tested for the ability of CERK1 to phosphorylate PBL27. Recombinant kinase-inactive PBL27K112E carrying a site-specific substitution for the ATP-binding site was used as a substrate for transphosphorylation assay. The recombinant CERK1 showed autophosphorylation activity, whereas PBL27K112E did not. The transphosphorylation assay showed that the PBL27K112E was directly phosphorylated by CERK1 in vitro (Figure 4a).

Figure 4.

Selective phosphorylation of cytoplasmic kinases by CERK1 and BAK1. (a–c) Substrate specificity of CERK1, BAK1, and FLS2 for the transphosphorylation of PBL27 and BIK1. Specificity of transphosphorylation was analyzed by in vitro kinase assay with recombinant proteins of intercellular domains of PBL27K112E, CERK1, FLS2, BAK1, and BIK1K105A/K106A. The protein loading control was shown by staining with Coomassie Brilliant Blue. Asterisks indicate artificial bands. Although the autoradiograms and Coomassie Brilliant Blue-stained gel images were obtained from different gels, we confirmed the correspondences of these bands from replicated experiments. As the results on the different gels were obtained on different days, the band intensities are not comparable between the gels. This is mainly because of the differences in the specific activity of the radioisotope as well as the adjustment of detector system. (d) Chitin oligosaccharide but not flg22 induces PBL27 phosphorylation in Arabidopsis. PBL27 phosphorylation was evaluated using Arabidopsis Col-0 or a cerk1 mutant expressing PBL27-HA under the control of its own promoter. Seedlings were treated with 10 μm (GlcNAc)7 or 100 nm flg22 for 10 min. The results are the representatives of two to three experiments.

Mutant analysis showed that PBL27 selectively regulates defense responses downstream of CERK1. To understand the molecular basis of such selectivity, we examined whether the substrate specificity of the corresponding receptor kinases could explain the observation. Tagged BAK1, FLS2 kinase domain, and BIK1 were again expressed in E. coli and used for these experiments. Recombinant kinase-inactive BIK1K105A/K106A carrying a site-specific substitution for the ATP-binding site was used as a substrate. Although the recombinant BAK1 phosphorylated BIK1K105A/K106A, phosphorylation of PBL27K112E by BAK1 was barely detectable (Figure 4b). In contrast, CERK1 preferentially phosphorylated PBL27K112E and the phosphorylation of BIK1K105A/K106A was very weak (Figure 4a), indicating differences in the substrate specificity between these two receptor kinases involved in MAMP signaling. In addition, the kinase activity of FLS2 was very low, as previously reported (Schwessinger et al., 2011), and in vitro phosphorylation of PBL27 by FLS2 was not detected (Figure 4c).

In vivo phosphorylation of PBL27 was examined in a seedling assay with the Arabidopsis plants expressing PBL27-3HA (Figure 4d). The band of PBL27-3HA detected by the anti-HA antibody was shifted upward by chitin treatment, but this chitin-induced band shift was not observed in the cerk1 mutant, showing that the chitin-induced phosphorylation of PBL27 depends on the presence of CERK1. We could not confirm by phosphatase treatment that the band shift was caused by phosphorylation, because the extraction of the samples with SDS-PAGE sample buffer did not allow us to try such experiments. However, we observed the reproducibility of such a band shift only in the transformants based on the wild type Col-0. In addition, PBL27 phosphorylation upon flg22 treatment was assayed, and the results showed that flg22 did not induce PBL27 phosphorylation (Figure 4d). Taken together, these results were in accord with the preferred usage of corresponding cytoplasmic kinases downstream of CERK1 and FLS2/BAK1.

Discussion

PBL27 regulates chitin signaling downstream of CERK1 and contributes to disease resistance

The perception of MAMPs by PRRs leads to the activation of intracellular PTI signaling in which RLCKs regulate downstream responses as cytosolic components (Monaghan and Zipfel, 2012; Yamaguchi et al., 2013b). In this study, we observed that an Arabidopsis receptor-like cytoplasmic kinase, PBL27, plays an important role in CERK1-mediated chitin signaling. Analysis of knockout mutants indicated that PBL27 regulates multiple defense responses triggered by chitin, such as the activation of MPK3/6, callose deposition, and the expression of several defense-related genes (Figure 2). These observations are consistent with the finding that the pbl27 mutants showed a significant decrease in disease resistance against the fungal pathogen A. brassicicola.

In addition, the impaired resistance to P. syringae pv. tomato DC3000 hrcC, shown by the pbl27 mutants, suggests that PBL27 is involved in the perception of bacterial PGN, similar to OsRLCK185 (Yamaguchi et al., 2013a), and contributes to resistance against bacterial infections.

CERK1 directly interacts and phosphorylates PBL27

After the perception of MAMPs on the cell surface, several steps would be required to activate the intracellular PBL27. In Arabidopsis, the binding of chitin oligosaccharides to CERK1 leads to dimerization and phosphorylation of CERK1, which are essential steps in chitin signaling (Petutschnig et al., 2010; Liu et al., 2012b). The following observations suggest that PBL27 is phosphorylated by CERK1 after chitin perception: (i) direct interaction of PBL27 and CERK1 as shown by in vitro and in vivo experiments; (ii) phosphorylation of PBL27 by the kinase domain of CERK1 expressed in E. coli: and (iii) the loss of chitin-induced phosphorylation of PBL27 in the cerk1 mutant. Preferential interaction of PBL27 with the kinase-inactive CERK1 (Figure 1) might indicate the dissociation of the phosphorylated PBL27 from CERK1. However, considering the observation that the overexpression of the wild-type CERK1 in N. benthamiana led to cell death in the host plant (Pietraszewska-Bogiel et al., 2013), comparison of the results obtained by co-expression of PBL27 with the kinase-active and kinase-inactive CERK1 should be made with care.

PBL27 selectively regulates defense responses downstream of CERK1: indication of different uses of RLCKs by different PRRs

To understand whether PBL27 functions in different MAMP signaling pathways, we evaluated flg22-induced defense responses in pbl27 mutants. Flg22-induced defense responses, including activation of MPK3/6, generation of ROS, callose deposition, and expression of defense-related genes were not significantly affected by pbl27 mutation, indicating that PBL27 selectively regulates defense signaling downstream of CERK1 (Figure 2). Furthermore, we investigated the molecular basis of such selectivity for the downstream cytoplasmic kinases between CERK1 and FLS2/BAK1. In vitro phosphorylation experiments showed that PBL27 is phosphorylated by CERK1 (Figure 4a). In addition, the in vitro phosphorylation of PBL27 by BAK1 was very low compared with that of BIK1 (Figure 4b). These results indicated that the substrate specificity of these RLKs essentially determines the use of specific RLCKs for downstream signaling. The observation that treatment with chitin oligosaccharide induced phosphorylation of PBL27 in wild-type Col-0 but not in the cerk1 mutant, and that the phosphorylation of PBL27 was not induced by flg22 treatment are in good agreement with this notion (Figure 4d). Shi et al. (2013) also reported that BSK1 regulates a subset of flg22-induced responses but not elf18-induced responses, providing another example of selective recruitment of RLCK by distinct PRRs.

To further extend this model, we evaluated the phosphorylation of BIK1 by the CERK1 kinase domain expressed in E. coli. As described in the Introduction, two papers reported different observations concerning the involvement of BIK1 in chitin-induced defense responses (Lu et al., 2010; Zhang et al., 2010). Our in vitro phosphorylation experiments showed that CERK1 phosphorylates BIK1 much more weakly than PBL27 (Figure 4a). These results appear to be consistent with the results of Zhang et al. (2010), indicating the involvement of BIK1 in chitin signaling in Arabidopsis. However, the extent to which BIK1 contributes to the regulation of chitin-induced responses remains unclear.

PBL27 and BIK1 regulate partially overlapping but also different MAMP signaling pathways

Significant differences were found when the defense responses downstream of BIK1 and PBL27 were compared. It appears that both RLCKs are involved in the regulation of MAMP-induced callose deposition, but only BIK1 is involved in MAMP-induced accumulation of ROS in Arabidopsis. It was shown that PBL27 is involved in the activation of MAPK, in contrast to BIK1, which has been reported not to activate the MAPK cascade (Feng et al., 2012). These results suggest that PBL27 and BIK1 regulate overlapping, but partially different, MAMP signaling pathways in Arabidopsis (Figure 5).

Figure 5.

A model of chitin and flg22 signaling mediated by PBL27 and BIK1.

Although the perception of different MAMPs results in mostly similar biological responses (Zipfel et al., 2006; Gust et al., 2007; Wan et al., 2008a), the present study suggests the possibility that each PRR uses specific RLCK(s) for the regulation of downstream responses, reflecting the substrate specificity of signaling receptor-like kinases. Although the MAPK cascade is a common and important signaling system in PTI and could be a crossing-point of the signaling pathways activated by distinct MAMPs, the molecule that triggers the activation of the MAPK cascade remains unknown. How PBL27 triggers the activation of the MAPK cascade is an intriguing question.

Experimental Procedures

Plant materials and growth conditions

Arabidopsis plants were grown in soil in a growth chamber at 22°C under constant light. Arabidopsis plants for leaf disk assay were grown under an 8-h light–16-h dark cycle at 22°C. Arabidopsis seedlings for defense response analysis were grown in MS medium containing 1% sucrose under a 16-h light (22°C)–8-h dark (15°C) cycle. The KO mutants of PBL27, pbl27-1 (GABI_001C07), and pbl27-2 (GABI_088H03) were obtained from the GABI-Kat. The AtCERK1-KO mutant cerk1-2 (096F09) was as previously described (Miya et al., 2007). To confirm the knockout of the target genes by reverse transcription PCR (RT-PCR), total RNA was extracted from the Arabidopsis seedlings using an RNeasy Plant Mini kit (Qiagen, http://www.qiagen.com/). Approximately 10–12 seedlings were combined to obtain an appropriate amount of total RNA per treatment. For two-step RT-PCR, first-strand cDNA synthesis was performed using 1 μg of total RNA and a QuantiTect Reverse Transcription Kit (Qiagen). Polymerase chain reaction was performed using Takara Ex Taq (TAKARA, http://www.takara-bio.com/) with the gene-specific primers described in Table S1.

Generation of transgenic Arabidopsis plants

Using PBL27 full-length cDNA as a template, the open reading frame (ORF) of PBL27 was amplified using PCR and the specific primer set 1 (Table S1). The amplified fragments were subcloned into pENTR/D-TOPO (Invitrogen, https://www.lifetechnologies.com/). To express PBL27-3HA driven by the native PBL27 promoter in Arabidopsis, a DNA fragment including both the promoter region and the ORF of PBL27 was amplified using PCR with primer set 2. The amplified fragment was subcloned into pENTR/D-TOPO and inserted into binary vector pGWB13 (Nakagawa et al., 2007). These plasmids were transformed into Agrobacterium tumefaciens C58C1 by electroporation. To generate transgenic Arabidopsis, flowering Arabidopsis plants were infected by A. tumefaciens carrying the expression vector using the floral-dip method.

Analysis of defense responses

Leaves from 5-week-old Arabidopsis plants were used to prepare leaf disks. The leaf disks were pre-incubated overnight in fresh MGRL medium containing 1% sucrose in a 24-well microtiter plate. The medium was replaced with fresh medium 2 h before (GlcNAc)7 or flg22 treatment. After the elicitor treatments, the ROS released by Arabidopsis leaves were quantified by L-012 chemiluminescence using 100 μm L-012 (Wako Pure Chemical Industries Ltd, http://www.wako-chem.co.jp/english/) (Albert et al., 2006).

For the detection of the defense gene expression, seedlings were collected 180 min after the elicitor treatment and total RNAs were extracted from the seedlings using a FavorPrep Plant total RNA Purification Minikit (FAVORGEN, http://www.favorgen.com/). Real-time quantitative RT-PCR with Fast SYBR Green Master Mix (Applied Biosystems, http://www.appliedbiosystems.com/) was performed using the synthesized cDNA as a template according to the manufacturer's protocol (Life Technologies, http://www.lifetechnologies.com/). The gene-specific primer sets used for the experiments are described in Table S1. Actin was used as an internal control.

Arabidopsis seeds were sterilized and allowed to germinate in MGRL medium containing 0.1% agarose and 1% sucrose. Seedlings were grown for 8 days in a 16-h light–8-h dark cycle at 23°C (light and dark, respectively). The seedlings were pre-incubated in MGRL liquid medium for 3 h, and 10 μm (GlcNAc)7 or 1 μm flg22 was added. Proteins were extracted from five seedlings in an extraction buffer [50 mm HEPES pH 7.4, 50 mm β-glycerophosphate, 5 mm EGTA, 5 mm EDTA, 10 mm NaF, 10 mm Na3VO4, 2 mm DTT, and protease inhibitor cocktail (Roche, http://www.roche.com/)]. Phosphorylation of MPK proteins was detected by immunoblotting with anti-pMPK antibody (1:2000 dilution; Cell Signaling Technology, http://www.cellsignal.com/).

The amount of callose deposition was determined by aniline-blue staining (Millet et al., 2010), with some modifications. The seedlings used in this assay were grown under a 16-h light–8-h dark cycle. Nine-day-old seedlings grown in a 24-well microtiter plate containing MS or MGRL medium with 1% sucrose were treated with elicitors. After the elicitor treatment, the seedlings were fixed in a 3:1 ethanol:acetic acid solution for several hours. Seedlings were subsequently rehydrated in 70% ethanol for 2 h, 50% ethanol for an additional 2 h, and water overnight. After one or two water washes, seedlings were suspended in 10% NaOH and shaken for 1.5–2 h at 37°C. After one or two water washes, seedlings were incubated in 150 mm K2HPO4, pH 9.5, and 0.01% aniline blue for several hours on a shaker. The fluorescence of stained callose was observed using BIOREVO BZ-9000 fluorescence microscope (KEYENCE, http://www.keyence.com/) at 390 nm excitation and 460 nm emission. The amount of callose deposition on the cotyledon was quantified using BZ-H2A/BZ-H1M/BZ-H1C (KEYENCE).

The ROS accumulation assay with Arabidopsis seedlings was performed as previously described (Albert et al., 2006). Seedlings were pre-incubated in fresh MGRL medium containing 1% sucrose in a 48-well microtiter plate for 2 h and (GlcNAc)7 was added. The ROS released by the Arabidopsis seedlings was quantified by L-012 chemiluminescence.

Localization analysis and BiFC assay

Transient expression system with N. benthamiana was used for localization analysis and BiFC assays. Subcloned PBL27 in pENTR/D-TOPO was inserted into the binary vector pGWB5 (Nakagawa et al., 2007) for localization analysis and pDEST-GWVYCE (Gehl et al., 2009) for BiFC assays. CERK1 and FLS2 in pENTR/D-TOPO were inserted into pDEST-GWVYNE (Gehl et al., 2009) for BiFC assays. The expression vectors were transformed into A. tumefaciens LBA4404 by electroporation. The transformed A. tumefaciens was pressure-infiltrated into N. benthamiana leaves as reported (Sainsbury et al., 2009). The plasmolysis experiment was performed by treating the leaves with 1 m NaCl for 10 min. The expressed proteins were observed with a FV1000-D confocal microscope (Olympus, http://www.olympus.com/). Green fluorescence protein was excited with a 488-nm laser and the fluorescence was detected using a 500–600 nm bandpass filter. Venus was excited with a 515 nm laser and the fluorescence was detected at 530–630 nm.

Immunoprecipitation assay

To construct plasmids for immunoprecipitation, pENTR/D-TOPO-CERK1 was first inserted into a binary vector pGWB14 (Nakagawa et al., 2007) to fuse the C-terminal 3HA tag. A fragment containing CERK1-3HA was amplified from pGWB14-3HA using PrimeSTAR GXL DNA polymerase (Takara Bio Inc., http://www.takara-bio.com/) with primer set 3 (Table S1) and subsequently subcloned into pENTR/D-TOPO again. pENTR/D-TOPO-CERK1D441V was amplified from pENTR/D-TOPO-CERK1 using PrimeSTAR Max DNA polymerase with primer set 4 (Table S1). To construct expression plasmids for N. benthamiana, subcloned CERK1-3HA and CERK1D441V-3HA in pENTR/D-TOPO were inserted into the binary vector pEAQ-DEST1 (Sainsbury et al., 2009) and PBL27 was inserted into pGWB17 (Nakagawa et al., 2007).

The expression plasmids were transformed into A. tumefaciens LBA4404 by electroporation. The transformed A. tumefaciens was pressure-infiltrated into N. benthamiana leaves. The leaves were collected 24 h after infiltration. A microsomal membrane fraction was prepared from the leaves and solubilized with 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)-buffered saline (TBS) containing 1% IGEPAL as previously described (Shinya et al., 2012). Co-immunoprecipitation using a solubilized protein fraction was performed with Monoclonal Anti-HA-Agarose beads (Sigma-Aldrich, http://www.sigmaaldrich.com/) according to the manufacturer's protocol. The solubilized protein fractions were incubated overnight with Anti-HA-Agarose beads at 4°C and the beads were washed five times with 0.05% IGEPAL in TBS. Immunoprecipitates were eluted by boiling with the SDS-PAGE sample buffer for 5 min and used for SDS-PAGE. Western blotting was performed using an Immun-Blot PVDF Membrane (Bio-Rad, http://www.bio-rad.com/). Detection of PBL27-4Myc was performed using rabbit polyclonal antibodies against Myc (Cell Signaling Technology) as a primary antibody and horseradish peroxidase-conjugated goat anti-rabbit IgG (Chemicon, https://www.millipore.com/) as a secondary antibody. Detection of CERK1-3HA was performed using Anti-HA High Affinity Rat monoclonal antibody (Roche) as a primary antibody and horseradish peroxidase-conjugated goat anti-rat IgG (Santa Cruz Biotechnology, http://www.scbt.com/) as a secondary antibody. Signals were detected by chemiluminescence with Immobilon Western Detection reagents (Millipore, http://www.millipore.com/).

Detection of PBL27 phosphorylation with Phos-tag SDS-PAGE

SDS-PAGE with Phos-tag Acrylamide (Wako Pure Chemical Industries) was used to evaluate the phosphorylation of CERK1-3HA and PBL27-3HA. Microsomal proteins from N. benthamiana leaves expressing CERK1-3HA were separated by SDS-PAGE on a 7.5% acrylamide gel containing 25 μm Phos-tag Acrylamide. Arabidopsis seedlings expressing PBL27-3HA were treated with (GlcNAc)7 or flg22 for 10 min. After the elicitor treatments, the seedlings were mixed with SDS sample buffer and boiled for 5 min and the extracts were separated by SDS-PAGE on a 7.5% acrylamide gel containing 25 μm Phos-tag Acrylamide. The band shift kinetics of CERK1-3HA and PBL27-3HA were detected by Western blotting with anti-HA antibody as described above. The lambda phosphatase (Sigma) treatment was performed according to the manufacturer's protocol.

Yeast two-hybrid assays

The coding region of PBL27 and the intracellular domain of kinase-inactive mutant of CERK1 were transferred into pBTM116-GW and pVP16-GW, respectively, by the Gateway system using an LR clonase reaction. The resultant vectors were introduced into cells of Saccharomyces cerevisiae L40 (MATatrp1 leu2 his3 LYS2::lexA-HIS3 URA3::lexA-lacZ). Transformants were selected on minimal medium lacking tryptophan and leucine. The yeast two-hybrid interaction was analyzed based on the requirement of histidine for yeast growth, as previously described (Yamaguchi et al., 2012).

In vitro kinase activity assay

For the construction of GST-tagged CERK1, FLS2 and BAK1 plasmid, the kinase domains of these receptor-like kinases were amplified by PCR with primer sets 5, 6 and 7 (Table S1). Full-length cDNAs of CERK1, FLS2 and BAK1 were used as templates. The amplified fragments were subcloned into pENTR/D-TOPO and inserted into pDEST15 vector (Life Technologies). Kinase-dead PBL27 mutant plasmid was generated by site-directed mutagenesis. PBL27K112E was amplified from pENTR/D-TOPO-PBL27 using Prime STAR Max DNA polymerase (Takara Bio Inc.) with primer set 8. pENTR/D-TOPO-BIK1K105A/K106A, which lacks kinase activity, was kindly provided by Dr Hidenori Matsui of RIKEN. The fragments of Strep-PBL27K112E and Strep-BIK1K105A/K106A were amplified from the coding regions of PBL27K112E and BIK1 K105A/K106A using primer sets 9 and 10 and then subcloned into pENTR/SD/D-TOPO, respectively. Strep-PBL27K112E and Strep-BIK1K105A/K106A in pENTR/SD/D-TOPO were transferred into pDEST14 vector by the LR reaction (Life Technologies). GST-CERK1, GST-FLS2, GST-BAK1, Strep-PBL27K112E and Strep-BIK1K105A/K106A proteins were expressed in E. coli BL21-AI and purified with GSTrap HP (GE Healthcare, http://www3.gehealthcare.com/) columns for GST-tagged proteins or StrepTrap FF (GE Healthcare) columns for Strep-tagged proteins.

The kinase activity assay was performed in 50 mm HEPES (pH 7.6), 10 mm MgCl2, 1 mm ATP, 1 mm DTT, 1 μg of the kinase domain of CERK1, FLS2 or BAK1, and 1 μg PBL27 or BIK1 as substrates, in a total volume of 40 μl. The assay was initiated by addition of 0.4 μl (4 μCi) γ32P-ATP and incubated for 1 h at 25°C. The reaction was terminated by the addition of Laemmli loading buffer and subsequent incubation at room temperature (25°C) for 10 min. Samples were separated by a SDS-PAGE. The gel was subsequently dried and exposed to an imaging plate. The 32P-labeled band was analyzed using an FLA-7000 imaging analyzer (Fuji Film, http://www.fujifilm.com/) and a Multi Gauge version 3.0 software (Fuji Film).

Pathology assays

Arabidopsis seeds were planted in pots covered with nylon mesh. Plants were grown and incubated in a growth chamber in an 8-h light–16-h dark cycle at 23°C (light or dark, respectively). Completely expanded rosette leaves of 5-week-old plants were inoculated with P. syringae pv. tomato DC3000 hrcC grown on King's B (KB) medium with 100 μg ml−1 rifampicin (Rif) at 28°C. Bacteria were diluted to 108 colony-forming units (c.f.u.) ml−1 in water containing 0.03% Silwet L-77. The inverted plant pots were entirely submerged in the bacterial suspension, and immediately placed under a plastic dome to maintain high humidity for 3 days. To measure bacterial growth, three leaf disks were ground in water and serial dilutions were spotted onto KB medium with Rif. The c.f.u. were counted 2 days after incubation at 28°C.

Alternaria brassicicola (isolate O-264) has been described by Narusaka et al. (2005). Arabidopsis plants were grown in soil for 28–30 days in a growth chamber at 22°C under a 12-h light–dark cycle. Plants were inoculated by spraying the leaves with a spore suspension [(6–7) × 105 spores ml−1 in distilled water] of A. brassicicola. On the other hand, two 5-μl drops of the spore suspension were placed on each leaf. The inoculated plants were placed in a growth chamber at 22°C with a 12-h light–dark cycle and maintained at 100% relative humidity. Lesion sizes were measured at 6 days after inoculation. Control plants were treated with distilled water.

Acknowledgements

We thank Dr Kenji Hashimoto of University of Münster for useful advice. We are indebted to Yaizu Suisankagaku Industrial Co. Ltd. for the supply of chitosan oligosaccharides. We thank Dr Cyril Zipfel of Sainsbury Laboratory for the supply of FLS2 cDNA clones. We thank Dr Hidenori Matsui of RIKEN for the supply of BIK1 cDNA clones. We are grateful to Dr Jörg Kudla of University of Münster, Dr Tsuyoshi Nakagawa of Shimane University, and Dr George Lomonossoff of John Innes Center for BiFC vectors, pGWB vectors, and pEAQ-DEST1 vectors, respectively. We thank Mr Kota Kamiya, Kenkichi Suto, Shohei Takahashi, and Takuya Maruyama of Meiji University for assistance in the experiments. This work was supported in part by Grants-in-Aid for Scientific Research (no. 22248041 to NS, no. 23380028 to TK, no. 22570052 to HK, no. 24780334 to TS, and no. 24580071 to YN), Grants-in-Aid for Scientific Research on Innovative Areas (24113519 and 25114517 to TK and 25114516 to NS), and by Strategic Project to Support the Formation of Research Bases at Private Universities: Matching Fund Subsidy 2011–2015 (S1101035) to TK from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

Ancillary

Advertisement