Cooperative regulation of PBI1 and MAPKs precisely controls the master 1 transcription factor WRKY45 in rice immunity

The U-box type ubiquitin ligase PUB44 is targeted by the Xanthomonas oryzae XopP 3 effector and positively regulates pattern-triggered immunity in rice. Here we identified 4 PBI1, a protein that interacts with PUB44. Crystal structure analysis indicated that PBI1 5 forms a four-helix bundle structure. PBI1 also interacts with WRKY45, a master 6 transcriptional activator of rice immunity, and negatively regulates its activity. PBI1 is 7 degraded during the chitin response, and this is suppressed by silencing of PUB44 or 8 expression of XopP , indicating that PBI1 degradation depends on PUB44. These data 9 suggest that PBI1 suppresses WRKY45 activity when cells are in the unelicited state, 10 and during chitin signaling, PUB44-mediated degradation of PBI1 leads to activation of 11 WRKY45. In addition, phosphorylation of WRKY45 by MAP kinases releases WRKY45 12 from the PBI1-mediated inhibitory effect. These results demonstrate that chitin-induced 13 activation of WRKY45 is regulated by the cooperation between MAP kinase-mediated 14 phosphorylation and PUB44-mediated PBI1 degradation.

INTRODUCTION 1 espectively 9-11 . Upon ligand perception, CEBiP and LYP4/6 interact with the LysM 1 receptor-like kinase OsCERK1 at the plasma membrane, and then OsCERK1 transmits 2 the signal to intracellular components 11,12 . Thus, OsCERK1 is a common factor regulating 3 both fungal chitin-and bacterial peptidoglycan-triggered immunity. In response to the 4 chitin signal, the cytoplasmic kinase domain of OsCERK1 phosphorylates the rice 5 receptor-like cytoplasmic kinase OsRLCK185, which then phosphorylates MAP kinase 6 kinase kinases such as MAPKKK11, MAPKKK18, and MAPKKK24/MAPKKKε. These 7 kinases then trigger the intracellular activation of the MAP kinases MPK3 and MPK6 13-15 . 8 The activated MAP kinases (MAPKs) induce robust immune responses by 9 phosphorylating downstream immune factors including transcription factors. 10 Pb1 interacts with and stabilizes WRKY45 through its CC domain, probably by inhibiting 1 proteasome-mediated degradation of WRKY45, and this enhances the immune 2 responses mediated by WRKY45. 3 WRKY45 appears to be autoregulated, because artificial expression of 4 ( Fig. 2g), supporting the possibility that PBI1 degradation may be regulated by the 1 PUB44-mediated ubiquitination pathway. 2 3 PBI1 is composed of a four-helix bundle 4 Although PBI1 consists mainly of the DUF1110 domain, the molecular nature of 5 this domain was unknown. To elucidate the structure of PBI1, we used an E. coli protein 6 expression system 41 , purified the recombinant PBI1 protein, and determined its tertiary 7 structure. The crystal structure was solved at a resolution of 1.84 Å. PBI1 is composed 8 of a four-helix bundle (Fig. 3a, b) with a diameter of approximately 19 Å and a length of 9 about 70 Å. There are six molecules in the asymmetric unit. The root mean square 10 differences (r.m.s.d.) between each monomer are from 0.15 to 0.85 Å. We observed no 11 large conformational change induced by crystal packing. The calculated solvent content 12 was 63% (Matthews coefficient = 3.32 Å 3 Da -1 ). The four helices of PBI1 are arranged in 13 an up-down-up-down topology, and the bundle is leftward turning. The four helices are 14 part of a single polypeptide chain (Ala10-His39, Glu49-Gly90, and 15 Val110-Val191) and are connected to each other by three loops (Leu40-Asp48,  Tyr109, and His149-Cys154). The interfaces between the helices consist of hydrophobic 17 residues, whereas hydrophilic residues are exposed on the surfaces that interact with 18 the aqueous environment. The hydrophobic residues occur as repeats of 3 or 4 residues 19 per helical turn and form the core of the bundle structure. 20 We used the Dali server 42 to perform a database search for three-dimensional 21 structures that exhibit similarity to PBI1, and identified eight unique proteins with Z-22 scores higher than 10. Seven of these proteins are: Methyl-accepting chemotaxis 23 transducer 43 , SH2 domain 44 , Talin 1 45 , surface protein VSPA 46 , focal adhesion kinase 1 47 , 24 tyrosine kinase 2 beta 48 , and superoxide dismutase 49 . In addition, the four-helix bundle 1 structure occurs in the CC domains of plant CC-NB-LRRs including  In particular, the CC domain of Rx is very similar in structure to PBI1, with a high Z-score 3 of 4.5. A structural alignment of the Rx CC domain and PBI1 showed a high degree of 4 similarity (Extended Data Fig. 2). 5 6 PBI1 interacts with WRKY45 in the nucleus 7 To analyze the subcellular localization of PBI1 we made constructs encoding 8 green fluorescent protein (GFP) fused to the N-or C-terminal of PBI1. These constructs, 9 along with one encoding red fluorescent protein (RFP) containing a nuclear localization 10 signal (RFP-nls), were used to transfect rice protoplasts. Fluorescence from both the 11 GFP-PBI1 and PBI1-GFP hybrid proteins was detected predominantly in the nuclei (Fig. 12 3c), although some GFP fluorescence was also observed in the cytoplasm. 13 The presence of PBI1 in the nucleus suggests that it may be involved in 14 transcriptional regulation. We screened for rice factors that interact with PBI1 and 15 identified WRKY45 as a candidate. WRKY45 is a key regulator of rice immunity against 16 rice blast and bacterial blight diseases 24 . To analyze the interaction between PBI1 and 17 WRKY45, we performed a bimolecular fluorescence complementation (BiFC) assay 18 using rice protoplasts. WRKY45 was tagged with the N-terminal domain of the yellow 19 fluorescent protein Venus (WRKY45-Vn), and PBI1 was tagged with the C-terminal 20 domain of Venus (PBI1-Vc). Transient expression of these constructs together resulted 21 in fluorescence in the nucleus (Fig. 4a). We also examined the interaction between PBI1 22 and WRKY45 in a co-immunoprecipitation assay using rice protoplasts transiently 23 expressing GFP-PBI1 and Myc-tagged WRKY45. Myc-WRKY45 co-immunoprecipitated 24 with GFP-PBI1 (Fig. 4b), confirming the in vivo interaction between PBI1 and WRKY45. 1 The fact that WRKY45 interacts with PBI1 raises the possibility that WRKY45 2 may play a role in chitin-induced immunity. However, the involvement of WRKY45 in 3 pattern-triggered immunity has not been described thus far. Our quantitative real-time 4 PCR experiments demonstrated that expression of WRKY45 is activated after treatment 5 with chitin (Fig. 4c). We also found that chitin-induced expression of WRKY62, which 6 functions downstream of WRKY45 53 , was significantly suppressed in two WRKY45-7 knockdown lines ( Fig. 4d and Extended Data Fig. 3a) 53 . These data indicate that 8 WRKY45 participates in chitin-induced immunity. 9 The interaction between PBI1 and WRKY45 in nuclei suggests that PBI1 may 10 be involved in the regulation of transcription by WRKY45. We carried out transactivation 11 assays using effector constructs expressing Myc-tagged WRKY45 and PBI1, and a 12 reporter construct containing a promoter with four W-box sequences upstream of the 13 luciferase cDNA 24 . We transfected rice protoplasts with the Myc-tagged WRKY45 14 construct and the reporter, with or without the PBI1 construct, and examined the 15 luciferase activity. Luciferase activity was increased in the presence of WRKY45 (Fig 4e  16 and Extended Data Fig. 3b) but was significantly inhibited by co-expression of PBI1. This 17 result indicates that PBI1 negatively regulates the transcriptional activity of WRKY45. 18 19 pbi1 mutations cause increases in the protein levels of WRKY45. 20 To clarify the roles of PBI1 in rice immunity, we generated two PBI1 knock-out 21 mutants (pbi1-1 and pbi1-2) using the CRISPR/Cas9 system. pbi1-1 has a frameshift 22 mutation caused by a 2-bp deletion, and pbi1-2 has a 6-bp deletion causing the loss of 23 two aa residues and a 1-bp substitution within the PBI1-coding region (Extended Data Fig. 3c). Immunoblots with α-PBI1 showed that no PBI1 protein was detected in either 1 mutant (Fig. 5a). Since no truncated PBI1-2 protein was detected in pbi1-2, the deletion 2 of the two amino acid residues may have caused protein stability. 3 Both pbi1-1 and pbi1-2 exhibited a weak dwarf phenotype (Fig. 5b), which was 4 similar to the phenotype of WRKY45-oxerexpressing plants 24 . Therefore, we analyzed 5 the protein levels of WRKY45 by immunoblotting with α-WRKY45. The WRKY45 protein 6 levels were significantly increased in the pbi1 mutants compared with wild type (Fig. 5c). 7 The WRKY45 transcript levels were also increased in the pbi1 mutants (Fig. 5d). Since 8 WRKY45 is known to be autoregulated 23 , these results suggest that loss of PBI1 may 9 result in the leaky autoactivation of WRKY45 transcription. 10 We examined the resistance of the pbi1 mutants to the compatible race Xoo 11 T7174 by inoculating the plants using the clipping method. The pbi1 mutants developed 12 disease lesions that were shorter than those of the wild type (Fig. 5e, f). Genomic 13 quantitative PCR using specific primers for the X. oryzae XopA gene indicated that 14 bacterial growth was also reduced in the pbi1 mutants (Fig. 5g). Thus, the pbi1 mutants 15 enhanced resistance to X. oryzae, possibly via accumulation of WRKY45. 16 We also tested whether the pbi1 mutations affect chitin-induced expression of 17 WRKY62 using cultured rice cells. Expression of WRKY62 was significantly enhanced in 18 the pbi1 mutants (Fig. 5h). This is consistent with results of the transient transcription we produced mapkkk11/mapkkk18 double mutant lines by using the CRISPR-CAS9 5 system and the mapkkk11-1 mutant background, which was generated by a Tos17 6 insertion in MAPKKK11 gene 15 . The mapkkk11-1/mapkkk18-1 and mapkkk11-7 1/mapkkk18-2 lines carry nonsense mutations caused by 1 bp insertions in MAPKKK18 8 gene (Extended Data Fig. 4a). Chitin-induced activation of MPK3 and MPK6 was 9 significantly reduced in the mutants (Fig. 6a). As shown in Fig. 6b, chitin-induced PBI1 10 degradation was suppressed in the mapkkk11/mapkkk18 mutants, indicating that MAPK 11 activity is required for the PBI1 degradation. 12 WRKY-type transcription factors are known to be activated by MAPK-mediated 13 phosphorylation 20 . The three amino acid residues Thr266, Ser294, and Ser299, located 14 in the C-terminal region of WRKY45, are phosphorylated by MPK6 19 . The 15 phosphorylation of Ser294 and Ser299 positively regulates the immune response, 16 whereas a phosphor-mimic mutation of Thr266 inhibits immunity. To examine whether 17 the phosphorylation of WRKY45 by MAPKs may affect the interaction between WRKY45 18 and PBI1, we analyzed the interaction using a split nano-luciferase assay. For this assay, 19 the full length, N-terminal or C-terminal regions of WRKY45 were fused to the Small BIT 20 (SmBiT) of NanoLuc, and PBI1 was fused to the Large BiT (LgBiT). The resultant 21 constructs were used to transfect rice protoplasts, and then total luciferase activities were 22 measured. The experiments indicated that PBI1 interacts more strongly with the C-23 terminal region of WRKY45 than with the N-terminal region (Fig. 6c). We produced a phosphor-mimic mutant (WRKY45 DD ) of WRKY45 in which Ser294 and Ser299 were 1 each substituted with Asp. The phosphor-mimic mutation suppressed the interaction 2 between WRKY45 and PBI1 (Fig. 6d). In addition, we tested the effect of the dominant-3 active phosphor-mimic mutant (MKK4 DD ) of MKK4, because expression of MKK4 DD 4 induces activation of MPK3 and MPK6 54 . The interaction between PBI1 and WRKY45 5 was also suppressed by co-expression with MKK4 DD (Fig. 6d). Our data indicated that 6 the phosphorylation of WRKY45 reduced the binding affinity between PBI1 and WRKY45. 7 It is possible that this reduced binding affinity may stimulate the PUB44-mediated 8 degradation of PBI1. In fact, chitin-induced expression of WRKY62 was strongly reduced 9 in the mapkkk11/mapkkk18 mutants (Fig. 6e), whereas expression of WRKY45 was less 10 affected by these mutations. These results suggest that the MAPK-mediated 11 phosphorylation of WRKY45 and the PUB44-mediated degradation of PBI1 function co-12 operatively in the activation of WRKY45. 13 During these experimental processes, we found a shifted band of PUB44 on the 14 immunoblots as shown by an arrow in Fig. 7a. The shifted band of PUB44 was detected 15 by treatment with chitin. However, the shifted band was undetectable in the Oscerk1 11 , 16 and it disappeared after treatment with λ phosphatase (Fig. 7b). These results indicate 17 that PUB44 is phosphorylated in an OsCERK1-dependent manner. The phosphorylation 18 of PUB44 was delayed and reduced in the mapkkk11/mapkkk18 mutant (Fig. 7c). In 19 addition, the transcript level of OsCERK1 in the mapkkk11/18 mutant was lower than in 20 wild type cells (Fig. 7d), suggesting that the steady-state levels of OsCERK1 transcript 21 are controlled through the MAPK pathway. The reduced levels of OsCERK1 transcript in 22 the mapkkk11/18 mutant may be the reason for the reduction and delay in PUB44 23 phosphorylation. Thus, it is possible that the defects in PBI1 degradation in the 24 mapkkk11/mapkkk18 mutants is partially associated with the reduction of PUB44 1 phosphorylation. 2 3 4 DISCUSSION 5 PUB44 was originally identified as the target for X. oryzae type III effector XopP. 6 Previous study indicated that PUB44 plays an important role in immune activation in 7 response to bacterial peptidoglycan as well as fungal chitin. In rice, upon perception of 8 peptidoglycan and chitin, the corresponding PRRs transmit the immune signals into 9 intracellular components through OsCERK1 9-11 . Therefore, PUB44 is most likely 10 activated downstream of OsCERK1. However, the molecular mechanisms of how PUB44 11 is activated and how it regulates the downstream immune responses had been unknown 12 so far. In this study, we found that upon perception of chitin, PUB44 is phosphorylated in 13 an OsCERK1-dependent manner. We also identified PBI1 as an interactor with PUB44. 14 During the chitin response, PBI1 is degraded in a PUB44-dependent manner, suggesting 15 that PUB44 may control immunity through degradation of PBI1. In addition, PBI1 16 interacts with and inhibits WRKY45, a key regulator of rice immunity. PBI1 degradation 17 is also regulated by MAPKs. The data presented here demonstrate that the chitin-18 induced activation of WRKY45 is regulated by both MAPK-mediated phosphorylation 19 and PUB44-mediated PBI1 degradation. 20 PBI1 is a novel protein carrying the DUF1110 domain, and it forms a small 21 protein family with PBI2, PBI3, and PBI4. The biological function of this family has not 22 been elucidated so far. In this study, we determined the crystal structure of PBI1 and 23 found that it forms a four-helix bundle. Many other proteins have four-helix bundle structures, including the CC domains of the CC-NB-LRR-type immune receptors 50,52 . In 1 fact, the tertiary structure of PBI1 is very similar to that of the CC domain of Rx, which is 2 a CC-NB-LRR receptor. Interestingly, it has been reported that WRKY45 interacts with 3 the CC domain of Pb1, a CC-NB-LRR protein involved in rice blast resistance 22 . Pb1 is 4 predicted to positively regulate the abundance of WRKY45 protein by protecting it from 5 degradation by the ubiquitin proteasome system, however, the molecular mechanisms 6 have not been elucidated in detail. In contrast to Pb1, PBI1 appears to negatively 7 regulate the abundance of WRKY45 protein, because WRKY45 protein levels are higher 8 in the pbi1 mutants than in the wild type. 9 The plant PUB family regulates a variety of biological responses, but the 10 mechanisms of PUB activation remain largely unknown. In Arabidopsis, the activation of 11 PUB22 is regulated by MAPK-mediated phosphorylation 55 . In this study, we found that 12 PUB44 is phosphorylated in an OsCERK1-dependent manner upon chitin perception. 13 The phosphorylation of PUB44 was also observed in the mapkkk11/mapkkk18 mutants, 14 but it was delayed and reduced. Therefore, it is unlikely that MAPKs phosphorylate 15 PUB44. The reduced level of phosphorylation may be explained by the fact that 16 OsCERK1 expression was reduced in the mapkkk11/mapkkk18 mutants. The 17 identification of protein kinases that phosphorylate PUB44 will be required for a further 18 understanding of PUB44 activation. 19 The co-expression of PBI1 and WRKY45 in rice protoplasts indicated that PBI1 20 inhibits the transcriptional activity of WRKY45. Therefore, it is likely that PBI1 functions 21 as a negative regulator of WRKY45 by direct interaction. In fact, the pbi1 plants contained 22 increased levels of WRKY45 protein, possibly because of the leaky auto-activation of 23 WRKY45 transcription. These increased levels of WRKY45 resulted in enhanced 24 resistance to Xoo, which is consistent with previous observations that overexpression of 1 WRKY45 enhanced rice immunity 17 . On the other hand, the enhanced levels of WRKY45 2 mRNAs negatively affect plant growth 24,25 . Therefore, it is possible that negative 3 regulation of the WRKY45 transcriptional activity via PBI1 under unelicited condition is 4 important for growth and reproduction. 5 PBI1 is degraded upon chitin perception, and this is suppressed by silencing of 6 the PUB44 gene or expression of XopP. Thus, it is possible that PBI1 degradation occurs 7 via PUB44-mediated ubiquitination of PBI1. However, the ligase activity of full-length 8 PUB44 is very weak 40 , and we failed to detect ubiquitination of PBI1 by PUB44 in multiple 9 in vitro ubiquitination assays. On the other hand, treatment of rice cells with the 10 proteasome inhibitor MG132 resulted in the accumulation of PBI1, suggesting that PBI1 11 protein levels are likely regulated by the ubiquitin-proteasome pathway. 12 PBI1 inhibits the activity of WRKY45. Therefore, it is possible that the chitin-13 induced degradation of PBI1 releases WRKY45 and activates WRKY45-mediated 14 transcription. If the activation of WRKY45 is regulated only via PBI1 degradation, then 15 WRKY45-mediated transcription would not occur in the absence of PBI1. However, 16 chitin-induced expression of WRKY62 was still observed in the pbi1 mutants, indicating 17 the existence of a positive regulatory mechanism for WRKY45 activation. Previous 18 studies indicated that WRKY45 activity is regulated through the phosphorylation of its C-19 terminal region by MPK6 19 . In fact, ectopic activation of MPK6 increases the 20 transcriptional activity of WRKY45 19 . Consistent with this, we found that the 21 mapkkk11/mapkkk18 mutations greatly reducing the activation of MPK3 and MPK6 22 strongly suppressed the chitin-induced expression of WRKY62. These results indicate 23 that the MAPKs regulate the chitin-induced activation of WRKY45. It has been shown that the DNA-binding activity of WRKYs is regulated via MAPK-mediated 1 phosphorylation 20 , however, it hasn't yet been shown that MPK3 and MPK6 control 2 WRKY45 activity in a similar manner. 3 This study and previous reports indicate that WRKY45 is regulated by both 4 MAPK-mediated phosphorylation and PBI1 degradation. In addition, we also revealed a 5 connection between MAPK-mediated phosphorylation and PBI1 degradation. The 6 phosphorylation of WRKY45 by MAPKs reduces the binding affinity between PBI1 and 7 WRKY45, suggesting that phosphorylation may stimulate the release of WRKY45 from 8 PBI1. Furthermore, PBI1 degradation was suppressed in the mapkkk11/mapkkk18 9 mutants. Thus, it is possible that the PUB44-mediated degradation of PBI1 may require 10 the disassociation between PBI1 and WRKY45. 11 Our study has revealed two regulatory mechanisms for WRKY45 activation, with 12 both positive and negative regulation (Extended Data Fig. 4b). Under unelicited 13 conditions, PBI1 inhibits WRKY45 activation in order to maintain its basal activity. Upon 14 chitin perception, the MAPK cascade is activated, and the MAPKs phosphorylate 15 WRKY45. This stimulates the release of WRKY45 from PBI1. At the same time, PUB44 16 is phosphorylated and then PBI1 is degraded, possibly following the disassociation from 17

WRKY45. 18
The perception of microbe-associated molecular patterns induces the rapid 19 transcription of immune-related genes, which is important for effective inhibition of 20 pathogen growth. The protein phosphorylation-and ubiquitination-based mechanisms 21 that control the activities of transcription factors are likely able to induce expression of 22 downstream genes much more rapidly than mechanisms involving the transcriptional 23 control of genes encoding transcription factors. Therefore, it seems that the cooperative regulation of WRKY45 via both the PUB44-PBI1 and MAPK-pathways contributes to the 1 rapid activation of immunity in rice. Although WRKY45 is a key factor for the activation 2 of rice immunity, its enhanced activation negatively affects plant growth 24,25 . Therefore, 3 the strict regulation of WRKY45 may be required for balancing immunity and growth. 4 5 6 Methods 7

Rice transformation 16
Calli generated from rice embryos were transformed using Agrobacterium tumefaciens 17 EHA101 lines carrying each construct, as described previously 58 . The transformed calli 18 were selected by resistance to hygromycin and used for the generation of suspension-19 cultured cells. 20 21 Yeast two-hybrid assays 22 The yeast two-hybrid screening and interaction assays were based on the requirement 23 for histidine for yeast growth, as described previously 40 .

Chitin treatments 2
Rice suspension-cultured cells were subcultured for 3 days in fresh medium, divided into 3 12-well plates (150 mg cells, 2 ml fresh medium per well), and treated with 2 μg/ml 4 (GlcNAc)7. 5 6 RNA isolation and quantitative real time PCR 7 Total RNA was isolated from rice suspension-cultured cells and leaves using TRIzol 8 reagent (Invitrogen) and then treated with RNase-free DNase I (Roche). First-strand 9 cDNA was synthesized from 1 μg total RNA with an oligo-dT primer and ReverTra Ace 10 reverse transcriptase (Toyobo). Expression levels were quantified by quantitative real 11 time PCR using the SYBR Green master mix (Applied Biosystems) in a Step-One Plus 12 Real-Time PCR system (Applied Biosystems). The expression levels were normalized 13 against a ubiquitin reference gene. Three biological replicates were used for each 14 experiment, and two quantitative replicates were performed for each biological replicate. 15 16

Protoplasts were isolated from cultured rice cells by digestion of the cell walls with 13
Cellulase RS (Yakult) and Macerozyme R-10 (Yakult) as described previously 58 . Aliquots 14 (100 μl) of protoplasts (2.5 × 10 6 cells/ml) were transformed with plasmid DNA using the 15 polyethylene glycol (PEG) method 59 . For the localization analysis and the BiFC assays, 16 transfected protoplasts were observed using a fluorescence microscope, the Axio Imager 17 M2 (Carl Zeiss) with the ApoTome2 system (Carl Zeiss). The transactivation assay of 18 WRKY45 was carried out as described previously 24  The over-expression, purification, crystallization, and preliminary X-ray analysis of native 3 and selenomethionine-labeled PBI1 were performed as described previously 41 . The 4 experimental phase and density modification were calculated from SAD data using 5 SHELXC/D/E 60 . Thirty-four of 36 selenium sites were identified. After density modification, 6 the figure of merit improved from 0.38 to 0.66. Automatic model building was performed 7 using Buccaneer 61 . Further structure refinement was performed with Coot 62 and 8 REFMAC5 63 . The coordinates and structure factors have been deposited with the PDB 9 (http://pdbj.org) with accession code 7CJC. Data collection and refinement statistics are 10 given in Table S1. The structural model was evaluated using Rampage 64 . 11 12 Split Nano Luciferase assay 13 DNA fragments of PBI1, WRKY45 1-326 , WRKY45 1-174 , WRKY45 175-326 , and WRKY45 DD 14 were transferred using the Gateway system with LR clonase reactions into p35S-  T7-GW or p35S-SmBiT-T7-GW (K. Taoka, paper in preparation). The plasmid containing 16 MKK4 DD was described previously 13 . The Firefly Luciferase gene under the control of 17 CaMV 35S promoter was used as an internal control. The indicated combinations of 18 plasmids were used to transfect rice protoplasts. After 18 h incubation at 30°C, the 19 activities of the Firefly and NanoLuc luciferases were measured on a TriStar2 LB942 20 luminometer (Berthold) using the ONE-Glo Luciferase Assay System (Promega) and the 21

Nano-Glo Live Cell Assay System (Promega). 22
Fully expanded rice leaves were inoculated with a compatible race of bacterial blight 1 pathogen Xanthomonas oryzae pv. oryzae T7174 by clipping the leaf tips with scissors 2 that had been immersed in bacterial suspension (OD600 = 0.2). Symptoms were scored 3 by measuring lesion length 14 days after infection. The bacterial population of Xoo T7174 4 was also analyzed by quantitative real-time PCR. The DNA levels of the Xoo XopA gene 5 relative to those of the rice ubiquitin gene were measured using genomic DNAs purified 6 from the infected leaves. replicate consisted of two technical replicates. The asterisks indicate statistically 2 significant differences from the WT controls by Student's t-test (P < 0.05). g, PBI1 protein 3 levels in XopP-ox cells after treatment with 2 μg/ml (GluNAc)7, determined by 4 immunoblot analysis with α-PBI1. 5 6 Fig. 3. PBI1, with a four-helix bundle structure, localizes mainly to the nucleus. 7 a, Side view of PBI1, which forms a four-helix bundle. Coloring is from blue at the N-8 terminus to red at the C-terminus. b, End view, with N-and C-termini at the front. c, 9 Detection of GFP-PBI1 and PBI1-GFP after transient expression in rice protoplasts. 10 mCherry with a nuclear localization signal was used as a nuclear localization marker. to a co-immunoprecipitation assay. Proteins were precipitated using an antibody against 20 GFP (α-GFP), and the input proteins and precipitated proteins were probed with α-Myc 21 and α-GFP. c, WRKY45 transcript levels in rice suspension-cultured cells treated with 22 (GluNAc)7, analyzed using quantitative real-time PCR. Data are means ±SD from three 1 independent biological replicates. The asterisks indicate statistically significant 2 differences between the wild-type and WRKY45-kd leaves by the Student's t-test (P < 3 0.05). e, Transactivation assay using a dual were prepared from rice suspension-cultured cells after treatment with 2 μg/ml (GluNAc)7 6 and subjected to immunoblots with α-pMAPK. b, Chitin-induced PBI1 degradation was 7 inhibited in the mapkkk11/mapkkk18 mutants. Total proteins were prepared as for (a) and 8 probed with α-PBI1. c, The interactions between PBI1 and full length WRKY45 or 9 WRKY45 fragments were analyzed using split NanoLuc assays. Constructs were made 10 to produce PBI1 fused to LgBiT and the WRKY45 fragments fused to SmBiT. Fluc was 11 used as an internal control. Rice protoplasts were transfected with the constructs and 12 the interactions were indicated by the Nluc to Fluc ratios. Different letters above the data 13 points indicate significant differences (p < 0.01, Welch's t test). d, Phosphorylation of 14 WRKY45 inhibits the interaction between PBI1 and WRKY45. Split NanoLuc assays 15 were carried out by transient expression of PBI1-LgBiT and WRKY45-SmBiT or 16 WRKY45 DD -SmBiT with or without MKK4 DD in rice protoplasts. Values are means ±S.E. 17 Different letters above the data points indicate significant differences (p < 0.01, Welch's 18 t test). e, The expression levels of WRKY45 and WRKY62 in mapkkk11/mapkkk18 19 suspension-cultured cells treated with 2 μg/ml (GluNAc)7 were analyzed using 20 quantitative real-time PCR. Data are means ±SD from three independent biological 21 replicates, where each biological replicate consisted of two technical replicates. The 22 asterisks indicate statistically significant differences from the WT controls by Student's t- a, Chitin-induced MAPK activation in two mapkkk11/mapkkk18 mutants. Total proteins were prepared from rice suspension-cultured cells after treatment with 2 μg (GluNAc) 7 and subjected to immunoblots with α-pMAPK. b, Chitin-induced PBI1 degradation was inhibited in the mapkkk11/mapkkk18 mutants. Total proteins were prepared as for (a) and probed with α-PBI1. c, The interactions between PBI1 and full length WRKY45 or WRKY45 fragments were analyzed using split NanoLuc assays. Constructs were made to produce PBI1 fused to LgBiT and the WRKY45 fragments fused to SmBiT. Fluc was used as an internal control. Rice protoplasts were transfected with the constructs and the interactions were indicated by the Nluc to Fluc ratios. Different letters above the data points indicate significant differences (p < 0.01, Welch's t test). d, Phosphorylation of WRKY45 inhibits the interaction between PBI1 and WRKY45. Split NanoLuc assays were carried out by transient expression of PBI1-LgBiT and WRKY45-SmBiT or WRKY45 DD -SmBiT with or without MKK4 DD in rice protoplasts. Values are means ±S.E. Different letters above the data points indicate significant differences (p < 0.01, Welch's t test). e, The expression levels of WRKY45 and WRKY62 in mapkkk11/mapkkk18 suspension-cultured cells treated with 2 μg (GluNAc) 7 were analyzed using quantitative real-time PCR. Data are means ±SD from three independent biological replicates, where each biological replicate consisted of two technical replicates. The asterisks indicate statistically significant differences from the WT controls by Student's t-test (P < 0.05).