The OsGSK2 Kinase Integrates Brassinosteroid and Jasmonic Acid Signaling by Interacting with OsJAZ4

OsGSK2 directly interacts with and destabilizes OsJAZ4 to activate JA-mediated defense signaling.

Abstract

The crosstalk between brassinosteroid (BR) and jasmonic acid (JA) signaling is crucial for plant growth and defense responses. However, the detailed interplay between BRs and JA remains obscure. Here, we found that the rice (Oryza sativa) Glycogen synthase kinase3 (GSK3)–like kinase OsGSK2, a conserved kinase serving as a key suppressor of BR signaling, enhanced antiviral defense and the JA response. We identified a member of the JASMONATE ZIM-domain (JAZ) family, OsJAZ4, as a OsGSK2 substrate and confirmed that OsGSK2 interacted with and phosphorylated OsJAZ4. We demonstrated that OsGSK2 disrupted the OsJAZ4-OsNINJA complex and OsJAZ4-OsJAZ11 dimerization by competitively binding to the ZIM domain, perhaps helping to facilitate the degradation of OsJAZ4 via the 26S proteasome pathway. We also showed that OsJAZ4 negatively modulated JA signaling and antiviral defense and that the BR pathway was involved in modulating the stability of OsJAZ4 protein in an OsCORONATINE INSENSITIVE1–dependent manner. Collectively, these results suggest that OsGSK2 enhances plant antiviral defenses by activating JA signaling as it directly interacts with, phosphorylates, and destabilizes OsJAZ4. Thus, our findings provide a clear link between BR and JA signaling.

INTRODUCTION

Brassinosteroids (BRs), a class of plant-specific steroidal hormones, play vital roles in various developmental and physiological processes of plants (Clouse and Sasse, 1998; Bishop and Koncz, 2002; Kim and Wang, 2010; Clouse, 2011). When the receptor-like kinase BRASSINOSTEROID INSENSITIVE1 (BRI1) perceives BRs at the plasma membrane, it activates a signal transduction cascade leading to the transcriptional regulation of BR-responsive genes (Kim and Wang, 2010; Wang et al., 2012). Glycogen synthase kinase3 (GSK3)–like kinases serve as key negative regulators of BR signaling (Li and Nam, 2002). In Arabidopsis (Arabidopsis thaliana), BRASSINOSTEROID INSENSITIVE2 (BIN2), a GSK3-like kinase, phosphorylates and destabilizes the two homologous transcription factors BRASSINAZOLE RESISTANT1 (BZR1) and BZR2/BES1, thereby blocking BR signaling (He et al., 2002; Li and Nam, 2002; Yin et al., 2002; Kim et al., 2009). In rice (Oryza sativa), OsGSK2 (also named OsSK22, OsGSK7), the homolog of BIN2, also negatively mediates BR signaling through OsBZR1 (Tong et al., 2012a), together with other transcription factors such as DWARF AND LOW-TILLERING (OsDLT; Tong et al., 2012a), LEAF and TILLER ANGLE INCREASED CONTROLLER (OsLIC; Zhang et al., 2012), OVATE FAMILY PROTEIN8 (OsOFP8; Yang et al., 2016), and REDUCED LEAF ANGLE1/SMALL ORGAN SIZE1(RLA1/SMOS1; Qiao et al., 2017). GSK3-like kinases are involved in multiple developmental and physiological processes (reviewed by Saidi et al., 2012; Youn and Kim, 2015; Tong and Chu, 2018) and also mediate crosstalk between the auxin (Vert et al., 2008; Cho et al., 2014), abscisic acid (Cai et al., 2014), gibberellin (Guo et al., 2018), jasmonic acid (JA; Gan et al., 2015), and BR pathways.

JA and its derivatives, such as methyl jasmonate (MeJA) and jasmonyl-L-isoleucine (JA-IIe), not only regulate plant growth and development but also promote plant defense against insect attack and pathogen infection (Chini et al., 2016; Goossens et al., 2016; Howe et al., 2018; Wasternack and Feussner, 2018; Yan et al., 2018). JA and MeJA treatments induce JA responses through their conversion to JA-Ile (Wasternack and Feussner, 2018). The JA signaling pathway has been well elucidated in recent years (Gfeller et al., 2010; Howe et al., 2018). The key suppressors in JA signaling are JASMONATE ZIM-DOMAIN (JAZ) proteins (Chini et al., 2007; Thines et al., 2007; Pauwels and Goossens, 2011). In the absence of JA-IIe, JAZ proteins, together with NOVEL INTERACTOR OF JAZ (NINJA) and TOPLESS, bind to and inhibit transcription factors that promote the expression of JA-responsive genes (Pauwels et al., 2010). By contrast, when JA is present, high levels of JA-Ile lead to binding of its receptor CORONATINE INSENSITIVE1 (COI1) to JAZ proteins, resulting in SCF^COI1-dependent ubiquitination and degradation of JAZ proteins through the 26S proteasome that in turns activates JA signaling (Chini et al., 2007; Thines et al., 2007; Sheard et al., 2010). Thus, fast turnover of JAZ proteins holds the key to JA signal output (Gfeller et al., 2010; Mao et al., 2017; Wasternack and Feussner, 2018; Chen et al., 2019).

Rice black-streaked dwarf virus (RBSDV) is a double-stranded RNA virus belonging to the genus Fijivirus within the family Reoviridae (Shikata and Kitagawa, 1977; Mertens, 2004). RBSDV is propagatively and persistently transmitted to rice, maize (Zea mays), barley (Hordeum vulgare), and wheat (Triticum sp) via the small brown planthopper (SBPH; Laodelphax striatellus; Uyeda et al., 1995; Wei and Li, 2016). RBSDV infection causes acute growth abnormalities (particularly severe dwarfism) in plants and results in serious yield losses in cereal crops (Shikata and Kitagawa, 1977; Wu et al., 2013). We have previously shown that JA, BR, and abscisic acid signaling play different roles in the defense response against RBSDV infection (He et al., 2017; Xie et al., 2018; Zhang et al., 2019). While the BR pathway mediates susceptibility to RBSDV infection, the JA pathway plays a positive role in rice defense against viral infection (He et al., 2017). Recently, several studies have implicated GSK3-like kinases in the plant immune system (Karlova et al., 2006; Piroux et al., 2007; Mills-Lujan et al., 2015; Mei et al., 2018; Qiu et al., 2018), but the detailed mechanism of GSK3-like kinase-mediated antiviral defense remains unclear in rice.

In this study, RBSDV inoculation and primary root inhibition assays using transgenic plants overexpressing OsGSK2 (Go; Tong et al., 2012a) and OsGSK2 RNA interference (RNAi) transgenic plant (Gi) lines (Tong et al., 2012a) indicated that OsGSK2 acts as a positive regulator in JA signaling and antiviral defense. We demonstrate that OsGSK2 physically interacts with and phosphorylates OsJAZ4. We provide evidence that OsGSK2 destabilized OsJAZ4 via disrupting the OsJAZ4-OsNINJA complex and OsJAZ4-OsJAZ11 dimerization. In addition, our results reveal that OsJAZ4 negatively regulates JA signaling and antiviral defense. Therefore, we propose that OsGSK2 activates JA signaling by facilitating the degradation of OsJAZ4.

RESULTS

OsGSK2 Positively Regulates the JA Response and Antiviral Defenses

Our previous results showed that BR signaling promoted rice susceptibility to RBSDV infection (He et al., 2017; Zhang et al., 2019). Here, we found the levels of OsGSK2 protein in RBSDV-infected plants at 30 days postinfection (dpi) were higher than those in virus-free rice plants (Supplemental Figure 1A), suggesting a role for OsGSK2 in rice during RBSDV infection. When infested with RBSDV-carrying SBPH, infected Go  (overexpressing OsGSK2) plants displayed less severe symptoms and less stunting, whereas infected Gi (OsGSK2 RNAi) plants had more severe symptoms than the RBSDV-infected wild-type O. sativa subsp japonica cv Zhonghua 11 (Zh11) plants (Figure 1A). Approximately 75% of the control wild-type Zh11 plants became infected, while the Go plants had fewer infected plants (∼43%) and the Gi plants had more (∼93%) at 30 dpi (Figure 1B). Consistent with the symptoms, RBSDV coat protein (CP) RNA and protein levels were much lower in infected Go plants, but dramatically higher in Gi plants than in Zh11 plants (Figure 1C; Supplemental Figure 1B). Evaluation of SBPH resistance, as previously described (He et al., 2017), showed that Go, Gi, and Zh11 plants were similarly susceptible to SBPH (Supplemental Figure 1C). Thus, RBSDV resistance conferred by OsGSK2 is independent of SBPH resistance.

Figure 1.

OsGSK2 Positively Regulates Antiviral Defense and JA Response.

(A) RBSDV symptoms on Zh11, Go, and Gi plants. Bar = 10 cm.

(B) Disease incidence in Zh11, Go, and Gi plants following RBSDV inoculation. The numbers of healthy and diseased plants in each treatment were determined by RT-PCR 30 d after inoculation, and the number of diseased plants was used to calculate the viral incidence (percentage of plants infected). Each treatment used at least 40 seedlings, and at least three biological replicates were performed. Different letters at the top of columns indicate significant difference between transgenic and control plants at P ≤ 0.05 by Fisher’s LSD test.

(C) Expression levels of the RBSDV CP gene as measured by RT-qPCR at 30 dpi. Data are relative expression levels of CP in Go and Gi plants compared with that in the wild-type Zh11 plants. OsUBQ5 was used as the internal reference gene. Error bars indicate the sd of three biological replicates. Asterisk (*) indicates significant difference between transgenic and control plants at P ≤ 0.05 by Fisher’s LSD test.

(D) Expression analysis of JA-responsive genes by RT-qPCR. Seven-day-old seedlings were collected for total RNA extraction. OsUBQ5 was used as the internal reference gene. Values are means ± se of three biological replicates. Asterisk (*) indicates significant difference between transgenic and the control plants at P ≤ 0.05 by Fisher’s LSD test.

(E) Levels of endogenous JA in 7-d-old Zh11, Go, and Gi plants. The limit of quantification to JA was 1 ng/mL. Values are means ± sd of three biological replicates. Different letters at the top of columns indicate significant difference between transgenic and control plants at P ≤ 0.05 by Fisher’s LSD test. FW, fresh weight.

(F) and (G) Images (F) and quantification of root length (G) of Zh11, Go, and Gi after MeJA treatment. The root lengths of 3-d-old seedlings grown in normal rice culture solutions supplemented with indicated concentrations of MeJA were measured. Data shown are the means from at least 10 seedlings for each indicated plant. Error bars represent sd. Different letters at the top of columns indicate significant difference between transgenic and control plants at P ≤ 0.05 by Fisher’s LSD test. Bar in (F) = 2 cm.

Specifically blocking GSK3-like kinase activity can attenuate JA signaling in rice, implying its potential role as a link between BR and JA signaling (Gan et al., 2015). Thus, we tested whether OsGSK2 enhanced plant resistance to RBSDV by activating JA signaling. The transcript levels of JA biosynthetic and signaling genes were greatly elevated in Go plants but lower in Gi plants than in the wild-type Zh11 plants (Figure 1D).

Quantification of hormone contents revealed that the production of JA was consistently induced in the Go plants but suppressed in the Gi plants. Moreover, significant upregulation of JA-IIe was observed in Go plants relative to that in Zh11 plants (Figure 1E; Supplemental Figure 2), suggesting activation of the JA pathway by OsGSK2. In addition, the inhibitory effect of MeJA on root growth was significantly enhanced in Go plants but suppressed in Gi plants compared with the wild-type Zh11 plants (Figures 1F and 1G), indicating that overexpression of OsGSK2 enhanced rice root sensitivity to JA signaling. These data together suggest a direct involvement of OsGSK2 in regulating both the JA pathway and RBSDV resistance.

OsGSK2 Interacts with and Phosphorylates OsJAZ4

To explore the role of OsGSK2 in JA signaling in detail, we performed a yeast two-hybrid (Y2H) screen assay using OsGSK2 bait and a normalized rice cDNA prey library. Interestingly, one of the interactors obtained was OsJAZ4, a JAZ family protein (Figure 2A). Further Y2H experiments showed that OsGSK2 interacted with another rice OsJAZ4 homolog, OsJAZ3 (Figure 2A; Supplemental Figure 3A). OsJAZ4 also interacted with the other OsGSK2 rice homologs OsGSK3, OsGSK5, and OsGSK7 (Figure 2B; Supplemental Figure 3B). Pull-down assays demonstrated that glutathione S-transferase (GST)-OsGSK2 could directly interact with His-OsJAZ4 in vitro (Figure 2C; Supplemental Figure 4).

Figure 2.

OsGSK2 Interacts with OsJAZ4 in Vitro and in Vivo.

(A) Y2H assay showing the interaction between OsGSK2 and OsJAZ1-15 proteins. Interactions were examined with SD base without Ade, His, Leu, and Trp.

(B) Y2H assay showing the interaction between OsJAZ4 and OsGSK1-8 proteins. Transformed yeast cells were grown on SD-Ade-His-Leu-Trp medium.

(C) Pull-down assay confirming that OsGSK2 interacts with OsJAZ4 in vitro. Immobilized GST and GST-OsGSK2 were used to pull down His-OsJAZ4, and immunoprecipitated fractions were detected using anti-His antibody. The bait proteins were probed with anti-GST antibody.

(D) BiFC assay showing the interaction between OsGSK2 and OsJAZ4 in N. benthamiana leaves. OsJAZ11 was used as negative control. Bar = 20 µm.

(E) Co-IP assays showing the interaction between OsGSK2 and OsJAZ4 in vivo. The proteins were extracted from N. benthamiana leaves and immunoprecipitated by anti-myc and anti-HA magnetic beads, respectively. The coimmunoprecipitated proteins were probed by either anti-Myc or anti-HA antibody. OsJAZ11-myc was used as negative control.

Subcellular localization analysis revealed that OsGSK2-GFP and OsJAZ4-mCherry colocalized in the cytoplasm and nucleus of Nicotiana benthamiana epidermal cells when these fusion proteins were coexpressed under strong cauliflower mosaic virus 35S promoter (Supplemental Figure 5). Bimolecular fluorescence complementation (BiFC) experiments showed that coexpression of either C-terminal part of yellow fluorescent protein (cYFP)-OsGSK2/OsJAZ4-N-terminal part of yellow fluorescent protein (nYFP) or cYFP-OsJAZ4/OsGSK2-nYFP, but not cYFP-OsGSK2/OsJAZ11-nYFP or cYFP-OsJAZ11/OsGSK2-nYFP, resulted in strong fluorescence signals (Figure 2D; Supplemental Figure 6). To further verify these results, we performed in vivo coimmunoprecipitation (Co-IP) assays using N. benthamiana leaves transformed with Human influenza hemagglutinin (HA)-OsGSK2 and OsJAZ4-myc/OsJAZ11-myc and found that HA-OsGSK2 and OsJAZ4-myc specifically coprecipitated with one another, but not with OsJAZ11-myc (Figure 2E). Together, these results demonstrated that OsGSK2 interacts with OsJAZ4 both in vitro and in vivo.

As a kinase, OsGSK2 can phosphorylate most of the proteins with which it interacts (Youn and Kim, 2015). To investigate whether OsJAZ4 is phosphorylated in vivo, we overexpressed OsJAZ4-MYC in the stably transformed wild-type Zh11 plants. We then used anti-myc beads to immunoprecipitate OsJAZ4-myc protein from OsJAZ4-MYC plants and treated the immunoprecipitation (IP) product with calf intestinal alkaline phosphatase (CIP). To avoid protein degradation, we added MG132 and a cocktail of proteinase inhibitors to the protein extracts. The phosphorylated OsJAZ4-myc (OsJAZ4-myc-P) was detected in phos-tag SDS-PAGE using an anti-myc antibody, and it disappeared after CIP treatment (Figure 3A), indicating that the band corresponded to phosphorylated OsJAZ4-myc.

Figure 3.

OsGSK2 Phosphorylates OsJAZ4.

(A) Immunoprecipitated OsJAZ4-myc protein from OsJAZ4-MYC plants was treated with CIP or water. OsJAZ4-myc protein was separated in a phos-tag SDS-PAGE gel and detected by anti-myc anti-body (top and middle panels). The slowly migrating band (red arrow) of OsJAZ4-myc in the phos-tag gel with short exposure time (Short exp.) or long exposure time (Long exp.) represents the phosphorylated form of OsJAZ4 (OsJAZ4-myc-P). As loading control (bottom), equal amounts of the immunoprecipitated OsJAZ4-myc protein were separated in a normal SDS-PAGE gel followed by immunoblot analysis.

(B) Potential phosphorylation sites in OsJAZ4.

(C) Immunoprecipitated OsJAZ4-myc and OsJAZ4Δ8-myc protein in N. benthamiana leaves transiently coexpressed with HA-empty vector, HA-OsGSK2, or HA-OsGSK2^K92R. The protein immunoprecipitated by anti-myc beads or CIP-treated OsJAZ4-myc protein from each combination were separated in a phos-tag SDS-PAGE gel and detected by anti-myc antibody (top). The slowly migrating band (red arrow) of OsJAZ4-myc in the phos-tag gel with short exposure time (Short exp.) or long exposure time (Long exp.) is the phosphorylated form of OsJAZ4 (OsJAZ4-myc-P). As loading controls, OsGSK2 (bottom) and different amounts of the immunoprecipitated OsJAZ4-myc protein (middle) were separated in a normal SDS-PAGE gel followed by immunoblot analysis.

To investigate the potential phosphorylation sites of OsJAZ4, the recombinant His-OsJAZ4 was incubated with the recombinant GST-OsGSK2 in an in vitro kinase assay buffer, separated by SDS-PAGE electrophoresis, and then subjected to liquid chromatography–tandem mass spectrometry analysis. As a result, eight potential phosphorylation sites were identified: Ser-100, Ser-133, Ser-134, Ser-147, Ser-148, Ser-252, Ser-267, and Ser-364 (Figure 3B; Supplemental Figures 7 and 8). We then constructed mutated forms of OsJAZ4 with alterations in these sites (S100A, S133A, S134A, S147A, S148A, S252A, S267A, and S364A) and named them OsJAZ4Δ8. Y2H assays showed that the physical interaction of OsJAZ4-OsGSK2 was not impaired among these site mutants (Supplemental Figure 9). When transiently coexpressed with HA-OsGSK2, but not with HA-empty vector, or HA-OsGSK2^K92R, kinase-dead mutant of OsGSK2 (Sun et al., 2018), in N. benthamiana leaves, two bands of OsJAZ4-myc were detected from myc bead–immunoprecipitated products, and the upper band could be eluted by CIP treatment (Figure 3C), showing the phosphorylation of OsJAZ4 by occurs OsGSK2 in vivo. Although OsJAZ4Δ8 was still phosphorylated by OsGSK2, the phosphorylation level of OsJAZ4Δ8 mutants was significantly reduced, indicating that these sites are likely the major sites being phosphorylated by OsGSK2 kinase. These results therefore demonstrate that OsGSK2 directly phosphorylates OsJAZ4.

OsGSK2 Promotes the Degradation OsJAZ4

GSK3-like kinases usually regulate both the activity and stability of their substrates by direct phosphorylation (Saidi et al., 2012; Youn and Kim, 2015; Tong and Chu, 2018). We therefore investigated whether OsGSK2 affected the stability of OsJAZ4. When transiently coexpressed with HA-OsGSK2 in N. benthamiana leaves, the accumulation of OsJAZ4-myc at 48 h was less than that when coexpressed with either HA-empty vector or HA-OsGSK2^K92R (Supplemental Figures 10A to 10D). Similarly, we found that OsJAZ4 degraded much faster in the extracts with GST-OsGSK2 than GST or GST-OsGSK2^K92R (Figures 4A and 4B; Supplemental Figure 11) using a cell-free protein degradation system (Qiao et al., 2017). However, accumulation of the mutant OsJAZ4Δ8-myc was not affected when it was transiently coexpressed with either HA-OsGSK2 or the HA-empty vector (Supplemental Figures 10E and 10F). Using the cell-free degradation system, we found that the degradation rate of the recombinant protein His-OsJAZ4Δ8 was much slower than that of His-OsJAZ4 and similar to the control, which contained 50 μM proteasome inhibitor MG132 (Figures 4C and 4D). Thus, we concluded that phosphorylation by OsGSK2 probably helped to accelerate OsJAZ4 degradation.

Figure 4.

OsGSK2 Promotes OsJAZ4 Degradation.

(A) Time course of OsJAZ4 degradation in the wild-type NIP protein extracts treated with GST, GST-OsGSK2, or GST-OsGSK2^K92R. Equal amounts of plant crude extracts were added to equal amounts of the recombinant proteins in the in vitro cell-free degradation assays. The Coomassie blue–stained ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) large subunit (Rbc L) was used as a loading control.

(B) Quantification analysis of (A). The relative levels of OsJAZ4 in the wild-type NIP plant protein extracts at 0 h were defined as 1. Data are means ± se (n = 3).

(C) Time course of degradation of His-OsJAZ4 and His-OsJAZ4Δ8 in the wild-type NIP protein extracts with or without MG132. Equal amounts of the recombinant proteins were incubated with equal amounts of plant crude extracts in the in vitro cell-free degradation assays.

(D) Quantification analysis of (C). The relative levels of His-OsJAZ4 or His-OsJAZ4Δ8 incubated with wild-type NIP plant protein extracts at 0 h were defined as 1. Data are means ± se (n = 3).

(E) Protein levels of OsJAZ4 in Go, Gi, and Zh11 leaves. The OsJAZ4 protein was detected with anti-OsJAZ4 antibody and Rbc L was used as a loading control. Two independent pools of leaves are shown.

(F) Quantification analysis of (E). The relative level of OsJAZ4 in the wild-type ZH11 was set as 1. Data are means ± se (n = 3).

To test this hypothesis, N. benthamiana leaves transiently coexpressing HA-OsGSK2 and OsJAZ4-myc were treated with Bikinin, a specific inhibitor of the kinase activity of GSK3-like kinases (Rozhon et al., 2014). As expected, Bikinin treatment inhibited OsGSK2-induced degradation of OsJAZ4, indicating that OsJAZ4 is stable in its unphosphorylated form (Supplemental Figures 10G and 10H). The degradation of OsJAZ4 induced by OsGSK2 was also suppressed by the proteasome inhibitor MG132, which suggested that the 26S proteasome pathway may be involved in OsGSK2-mediated OsJAZ4 degradation (Supplemental Figures 10I and 10J). Interestingly, the levels of OsJAZ4 protein were obviously lower in the Go plants but markedly more in the Gi plants than those in Zh11 plants (Figures 4E and 4F; Supplemental Figure 11), indicating a negative role for OsGSK2 in OsJAZ4 accumulation in vivo. These data together suggested that accumulation of OsGSK2 decreased levels of OsJAZ4 protein.

OsGSK2 Affects the Dimerization of OsJAZ4-OsJAZ11

We next used a domain deletion assay to define the domains involved in the OsGSK2-OsJAZ4 interaction. Based on the conserved N-terminal domain, C-terminal JA-associated (Jas) domain, and the ZIM domain found in all JAZ proteins (Pauwels and Goossens, 2011; Wasternack and Feussner, 2018), we constructed and examined the ability of various deletion mutants of OsJAZ4 (Figure 5A, top schematic diagrams) to interact with OsGSK2. The results showed that the ZIM domain of OsJZA4 was responsible for binding to OsGSK2 (Figure 5A). Previous reports showed that the conserved ZIM domain is responsible for forming homo- and heterodimers among JAZ proteins (Chini et al., 2007, 2009; Chung and Howe, 2009). We therefore conducted a Y2H assay using OsJAZ4 as bait and found that OsJAZ4 interacted with OsJAZ9, OsJAZ11, and OsJAZ12 (Figure 5B), consistent with previous reports of dimerization within a subset of OsJAZs (Yamada et al., 2012; Wu et al., 2015).

Figure 5.

OsGSK2 Affects OsJAZ4-OsJAZ11 Interaction.

(A) Y2H assay shows that the ZIM domain of OsJAZ4 is responsible for binding to OsGSK2. Schematic diagrams show the truncated versions of OsJAZ4. Interactions were examined with SD base without Ade, His, Leu, and Trp. aa, amino acids; NT, N-terminal domain.

(B) Y2H assay shows the interaction between OsJAZ4 and OsJAZ1-15 proteins. Transformed yeast cells were grown on SD-Ade-His-Trp-Leu medium.

(C) In vitro interaction between OsJAZ4-myc and His-JAZ11 is weakened by GST-OsGSK2. His-OsJAZ11 protein combined with GST-OsGSK2 was incubated with immobilized OsJAZ4-myc. The immunoprecipitated fractions were detected by anti-His antibody. The gradient indicates increasing amount of GST-OsGSK2. OsJAZ4-myc input was probed with anti-myc antibody, and the loading of His-OsJAZ11 and GST-OsGSK2 is shown in the lower panel by Coomassie blue (CBB) staining.

(D) HA-OsGSK2 affects accumulation of OsJAZ4-myc and OsJAZ11-myc. The respective vectors were cotransiently expressed in the N. benthamiana leaves. The infiltrated leaves were collected at 48 h after infiltration, and at least four plants were pooled. The Coomassie blue–stained ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) large subunit (Rbc L) was used as a loading control.

(E) The protein levels of OsJAZ11 in Go, Gi, and Zh11 plants. The OsJAZ11 protein was detected with anti-OsJAZ11 antibody and Rbc L was used as a loading control. Two independent pools of leaves are shown.

We then investigated whether OsGSK2 affected the interaction between OsJAZs and OsJAZ4. In a semi–pull-down assay, GST-OsGSK2 attenuated the interaction between OsJAZ4-myc and His-OsJZA11 (Figure 5C). When transiently coexpressed with OsJAZ11-myc in N. benthamiana leaves, the level of OsJAZ4-myc increased, suggesting that dimerization may contribute to the stability of JAZ proteins (Figure 5D). Although the amount of OsJAZ11-myc was not affected by HA-OsGSK2 (Supplemental Figures 12A and 12B), the amounts of both OsJAZ11-myc and OsJAZ4-myc were lower in the OsJAZ11-myc/OsJAZ4-myc/HA-OsGSK2 combination than in either OsJAZ11-myc/OsJAZ4-myc/HA or OsJAZ11-myc/OsJAZ4-myc/HA-OsGSK2^K92R (Figure 5D; Supplemental Figures 12C to 12F). Interestingly, the levels of OsJAZ11 protein were obviously less in the Go plants, but markedly more in the Gi plants, than in Zh11 plants (Figure 5E; Supplemental Figure 11). These results demonstrated that OsGSK2 could compete with OsJAZ11 for binding to OsJAZ4 to dissociate the OsJAZ4-OsJAZ11 complex, resulting in suppression of OsJAZs accumulation.

OsGSK2 and OsNINJA Compete for Binding to OsJAZ4

JAZ proteins are suppressors of JA signaling (Chini et al., 2007; Thines et al., 2007). In the domain deletion experiment, we demonstrated that the Jas domain of OsJAZ4 is necessary for binding to OsCOI1b in a coronatine (a JA-IIe analog)–dependent manner and that the ZIM domain of OsJAZ4 is responsible for interacting with OsNINJA (Figure 6A). These results are consistent with previous reports (Katsir et al., 2008; Chung and Howe, 2009; Fonseca et al., 2009; Sheard et al., 2010) and show that OsJAZ4 is involved in JA signaling. JAZs and NINJA are the corepressors in JA signing, and dissociation of JAZ-NINJA complexes results in degradation of JAZ proteins (Pauwels et al., 2010; Wasternack and Feussner, 2018). Since the ZIM domain was also responsible for the OsGSK2-OsJAZ4 interaction (Figure 5A), we investigated whether OsGSK2 affected the interaction between OsJAZ4 and OsNINJA. An in vitro pull-down assay showed that the interaction of OsJAZ4-myc and His-OsNINJA was impaired by an increased amount of GST-OsGSK2 (Figure 6B). This result demonstrated that OsGSK2 and OsNINJA compete for binding to OsJAZ4.

Figure 6.

Effect of OsGSK2 on the OsJAZ4-OsNINJA Interaction.

(A) OsNINJA and OsCOI1b interact with OsJAZ4. OsNINJA and OsCOI1b were fused to the GAL4 DNA BD, while OsJAZ4 and its mutants were fused to the GAL4 AD, respectively. Interactions were examined using SD base without Ade, His, Leu, and Trp. For the interactions between OsJAZ4 and OsCOI1b, 25 μM coronatine (COR) was added.

(B) OsGSK2 competes with OsNINJA for binding to OsJAZ4. In vitro interaction between OsJAZ4-myc and His-OsNINJA is weakened by GST-OsGSK2. His-OsNINJA protein combined with GST-OsGSK2 was incubated with immobilized OsJAZ4-myc. The immunoprecipitated fractions were detected by anti-His antibody. The gradient indicates increasing amounts of GST-OsGSK2. OsJAZ4-myc input was probed with anti-myc antibody, and the loading of His-OsNINJA and GST-OsGSK2 is shown bottom by Coomassie blue (CBB) staining.

OsJAZ4 Suppresses JA Signaling and Antiviral Defense

To further test the function of OsJAZ4 in JA signaling, we generated OsJAZ4 overexpression (OsJAZ4-OE) and OsJAZ4-RNAi transgenic rice plants in the O. s. japonica cv Nipponbare (NIP) background and confirmed the expected effects on the expression of OsJAZ4 by RT-qPCR (Figure 7A). Notably, the transcripts of JA-responsive genes, except HYDROPEROXIDE LYASE (OsHPL3, encoding a competitor of allene oxide synthase for the same substrate; Tong et al., 2012b), were significantly suppressed in OsJAZ4-OE lines (nos. 1 and 3) but elevated in OsJAZ4-RNAi lines (nos. 14 and 18), compared with the wild-type NIP (Figure 7A). When treated with MeJA, the root length reduced more slowly in OsJAZ4-OE lines (nos. 1 and 3) but faster in OsJAZ4-RNAi lines (nos. 14 and 18) than in NIP (Figures 7B and 7C), suggesting that OsJAZ4 serves as a suppressor of JA signaling in rice.

Figure 7.

OsJAZ4 Negatively Modulates JA Signaling and Rice Immunity.

(A) JA-responsive gene expression in indicated transgenic plants. RT-qPCR analysis of the mRNA levels of JA-responsive genes in the wild-type NIP, OsJAZ4-OE lines (nos. 1 and 3), and OsJAZ4-RNAi lines (nos. 14 and 18). OsUBQ5 was used as the internal reference gene. Values are means ± se of three biological replicates. Asterisk (*) indicates significant difference at P ≤ 0.05 (n = 3) by Fisher’s LSD test.

(B) and (C) Images (B) and quantification of root length (C) in indicated plants following MeJA treatment. The root lengths of 3-d-old seedlings grown in normal rice culture solutions supplemented with different concentrations of MeJA were measured. Data shown are the means from at least 15 seedlings for each plant type. Error bars represent sd. Different letters at the top of columns indicate significant difference at P ≤ 0.05 by Fisher’s LSD test. Bar = 2 cm.

(D) Disease incidence. The numbers of healthy and diseased plants in each treatment were determined by RT-PCR 30 d after inoculation, and the number of the diseased plants was used to calculate the viral incidence (percentage of plants infected). Each treatment used at least 40 seedlings, and at least three biological replicates were performed. Different letters at the top of columns indicate significant difference between transgenic and control plants at P ≤ 0.05 by Fisher’s LSD test.

(E) Relative expression levels of RBSDV CP gene measured by RT-qPCR at 30 dpi. Data represent relative expression levels of the CP gene in the mutant compared with that in the wild-type NIP plants. OsUBQ5 was used as the internal reference gene. Error bars represent sd. Asterisk (*) indicates significant difference at P ≤ 0.05 (n ≥ 3) by Fisher’s LSD test.

(F) Viral symptoms in OsJAZ4-OE lines (nos. 1 and 3) and OsJAZ4-RNAi lines (nos. 14 and 18). Bar = 10 cm.

The expression of OsJAZ4 in RBSDV-infected leaves at 30 dpi was decreased relative to that in uninfected controls (Supplemental Figures 13A and 13B), indicating that OsJAZ4 was involved in regulating rice immunity. Therefore, the sensitivity of OsJAZ4 transgenic plants to RBSDV infection was assessed. OsJAZ4-OE lines were more susceptible to RBSDV, while OsJAZ4-RNAi lines were more resistant than the NIP control (Figures 7D to 7F; Supplemental Figure 13C). The susceptibility to RBSDV infection promoted by OsJAZ4 was independent of any effect on SBPH resistance (Supplemental Figure 14). These results together suggest that OsJAZ4 acts as an important repressor of JA signaling and plays a negative role in rice antiviral defense.

Effect of the OsGSK2-OsJAZ4 Interaction on JA and BR Signaling Crosstalk

To further investigate the role of the OsGSK2-OsJAZ4 interaction in JA and BR signaling, we first treated coi1-13 (OsCOI1 knockdown mutant; Yang et al., 2012) and NIP control seedlings with brassinolide (BL), which inactivates and degrades OsGSK2 (Kim et al., 2009), and Bikinin, respectively. Treatment of rice seedlings with either BL or Bikinin significantly increased the accumulation of OsJAZ4 protein (Figures 8A and 8B). However, the BR- and Bikinin-mediated stabilization of OsJAZ4 was inhibited in coi1-13 mutants (Figures 8C and 8D), indicating involvement of OsCOI1 in BR signaling-mediated OsJAZ4 stability.

Figure 8.

Effect of OsGSK2-OsJAZ4 Interaction on JA and BR Pathway Crosstalk.

(A) and (B) OsJAZ4 accumulation increases in response to BL (A) or Bikinin (B) treatment. The leaves of the 7-d-old wild-type NIP seedlings were treated with 1 μM BL or 20 μM Bikinin, and protein was extracted from the treated leaves 0, 3, 6, or 12 h later. The Coomassie blue (CBB)–stained ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) large subunit (Rbc L) was used as a loading control. Two independent pools of leaves are shown.

(C) and (D) Accumulation of OsJAZ4 induced by BL and Bikinin was inhibited in coi1-13 mutants. Leaves of the 7-d-old wild-type NIP and coi1-13 seedlings were treated with 1 μM BL (C) or 20 μM Bikinin (D), and the treated leaves were used for protein extraction at different times after treatment, and detected by anti-OsJAZ4 antibody. Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) large subunit (Rbc L) was used as a loading control. Two independent pools of leaves are shown.

(E) Effect of Bikinin on MeJA hypersensitivity in OsJAZ4-RNAi plants. Germinated seeds were grown in normal rice culture solutions containing 0 or 1 μM MeJA, with or without 200 μM Bikinin for 5 d, and the root length was then measured. Relative root elongation is expressed as a percentage of root elongation in solutions with (right section) or without (left section) 200 μM Bikinin. Error bars represent se (n ≥ 20). Different letters at the top of columns indicate significant difference at P ≤ 0.05 by Fisher’s LSD test.

(F) BL sensitivity test of the wild-type NIP, OsJAZ4-OE lines (nos. 1 and 3), and OsJAZ4-RNAi lines (nos. 14 and 18) by lamina joint assay. The plus and minus symbols indicate with and without BL (100 ng), respectively.

(G) Quantification of the data shown in (E). Data shown are the means from at least 15 seedlings for each indicated plant. Error bars represent se. Different letters at the top of columns indicates significant difference at P ≤ 0.05 by Fisher’s LSD test.

The chemical Bikinin can specifically inhibit the activity of GSK3-like kinases including BIN2 and OsGSK2, resulting in a BR signaling defect (De Rybel et al., 2009). Thus, we used Bikinin treatment to analyze the function of OsGSK2 in OsJAZ4-RNAi plants. The root sensitivity of MeJA-treated OsJAZ4-RNAi lines was similar in the presence or absence of Bikinin (Figure 8E). These results indicated that suppression of OsGSK2 did not affect JA sensitivity in OsJAZ4-RNAi plants. Together, these data suggest that OsJAZ4 is required for OsGSK2-mediated JA signaling and that it acts downstream of OsGSK2.

In addition, the BR sensitivity, as shown by the lamina joint assay, was enhanced in OsJAZ4-OE plants but suppressed in OsJAZ4-RNAi lines compared with that in the wild-type NIP controls (Figures 8F and 8G). Consistent with previous reports that the JA pathway was antagonized by the BR pathway (Ren et al., 2009; Kim et al., 2013; Nahar et al., 2013; Gan et al., 2015; He et al., 2017), our results confirmed that OsJAZ4 could enhance BR responses by inhibiting JA signaling.

DISCUSSION

Plant growth and defense responses are coordinated by the interplay among various phytohormones, such as BRs and JA (Kim and Wang, 2010; Vidhyasekaran, 2015). BRs and JA are growth-promoting and defense-related hormones, respectively (Vidhyasekaran, 2015), and there is extensive crosstalk between the BR and JA signaling pathways during both plant growth and plant defense responses (Ren et al., 2009; Gan et al., 2015). In Arabidopsis, BR suppresses the inhibition of root elongation by JA, whereas a defect in BR biosynthesis increases sensitivity to the JA response and reduces the negative effect of BR signaling on JA-inhibitory root growth (Ren et al., 2009). In rice, MeJA inhibits the BR-induced increase in lamina joint inclination (Gan et al., 2015). Interactions between BR and JA are also involved in modulating plant immunity (Kim et al., 2013; Nahar et al., 2013; He et al., 2017). The mutant gulliver3-D overproduces BR and has reduced sensitivity to MeJA, and JA interrupts BR signaling by repressing DWF4 expression upon Pseudomonas syringae infection (Kim et al., 2013). In rice roots, BR and JA are antagonistic, and physiological BR levels suppress the JA-induced resistance to the root-knot nematode Meloidogyne graminicola (Nahar et al., 2013). Here, we found that the Go plants overexpressing OsGSK2 with its upregulated JA pathway were more resistant to RBSDV infection, while the OsGSK2 RNAi Gi plants with suppressed JA pathway were more susceptible compared with the wild-type Zh11 plants. This indicates that the JA pathway is involved in OsGSK2-mediated rice defense (Figure 1; Supplemental Figure 1). We then demonstrated that OsGSK2 could interact with and phosphorylate OsJAZ4 protein (Figures 2 and 3). We also showed that OsGSK2 disrupted the OsJAZ4-OsNINJA corepressor and OsJAZ4-OsJAZ11 complex via the ZIM domain and that these dual effects of OsGSK2 on OsJAZ4 helped to promote degradation of OsJAZ4 by the 26S proteasome (Figures 4 to 6). It cannot be ruled out that altered JA content in OsGSK2 overexpressing and RNAi plants may affect the levels of OsJAZ proteins. We further confirmed that OsJAZ4 suppressed JA signaling and antiviral defense (Figure 7). The BR- and Bikinin-treatment assays demonstrated that OsCOI1 was involved in BR signaling-mediated OsJAZ4 stability (Figures 8). In addition, the MeJA hypersensitivity in OsJAZ4-RNAi lines were not enhanced by exogenous Bikinin (Figure 8), suggesting that OsGSK2-mediated JA signaling may depend on OsJAZ4. Although the levels of OsGSK2 protein were not altered in OsJAZ4 mutants (Supplemental Figure 15), the BR sensitivity analysis in OsJAZ4 mutants demonstrated that suppression of JA signaling by OsJAZ4 may enhance the BR-induced lamina joint inclination (Figure 8). Hence, these results suggested that OsJAZ4-mediated BR sensitivity might be via the downstream signaling pathway of OsGSK2. Our previous study showed that JA mediated resistance and that BR mediated susceptibility to RBSDV infection (He et al., 2017). OsGSK2 also suppresses the BR pathway by inactivating the transcription function of OsBZR1 in rice (Tong et al., 2012a). Thus, we suggest that OsGSK2 positively modulates rice antiviral defense by coordinating JA and BR signaling. The physical interaction between OsGSK2 and OsJAZ4 reveals a direct crosstalk between JA and BR signaling at the molecular level.

It has been reported that Arabidopsis and rice JAZ proteins form homo- and heterodimers and that the ZIM domain is responsible for this dimerization (Chini et al., 2007, 2009; Chung and Howe, 2009; Ren et al., 2009; Yamada et al., 2012; Chen et al., 2019). Our Y2H data also showed that OsJAZ4 formed heterodimeric complexes with OsJAZ9, OsJAZ11, and OsJAZ12 (Figure 5). In Arabidopsis, strongly JA-insensitive phenotypes conferred by overexpression of JAZ10.4 (an alternatively spliced form of JAZ10 that lacks the Jas domain) were suppressed by mutations in the ZIM domain that block JAZ10.4-JAZ interactions (Chung and Howe, 2009). AtJAZ3ΔJas does not interact with AtMYC2, but overexpression of AtJAZ3ΔJas confers jasmonate insensitivity to Arabidopsis transgenic plants because its ZIM is able to interact with other JAZ proteins (Chini et al., 2007, 2009). These findings suggest that the dimerization mediated by the ZIM domain suppresses JA signaling and that disruption of JAZ dimerization therefore increases JA signaling output. It has been reported that the ZIM domain is essential for regulating JAZs stabilization (Chini et al., 2007; Mao et al., 2017; Chen et al., 2019). Although the ZIM domain is not itself required for the JAZ-COI1 interaction, the heterodimerization of JAZs via the ZIM domain would affect the spatial structure of the C terminus of JAZs and would likely therefore affect the JAZs-COI1 interaction. For example, AtJAZ3ΔJas protein could interact with AtJAZ1 and AtJAZ9 and inhibit degradation of AtJAZ1 and AtJAZ9 by interfering with COI1 activity (Chini et al., 2007). AtSPL9 and AtHARP1 interact with JAZ proteins via the ZIM domain to inhibit degradation of JAZs by interfering with the COI1-JAZs interaction (Mao et al., 2017; Chen et al., 2019). Here, we demonstrated that OsGSK2 interacts with OsJAZ4 via the ZIM domain and competed with OsJAZ11 for binding to OsJAZ4 (Figure 5), which may disrupt the OsJAZ4-OsJAZ11 heterodimer complex. We also found that the accumulation of OsJAZ11 was suppressed when OsGSK2 promoted degradation of OsJAZ4 in vivo (Figure 4 and 5), indicating that OsGSK2 disrupts the stability of the OsJAZ4-OsJAZ11 heterodimer. Thus, the reduced accumulation of OsJAZs resulting from the dissociation of OsJAZ4-OsJAZs by OsGSK2 may, at least partially, contribute to the JA sensitivity and antiviral defense induced by OsGSK2.

The protein phosphorylation pathway has been reported to be involved in the JA signaling system (Kazan and Manners, 2013). MYC2 phosphorylation is functionally coupled with its action to regulate JA-responsive gene transcription and JA-mediated immunity (Zhai et al., 2013). Phosphorylation of jasmonate-associated VQ domain protein1 (JAV1) disintegrates the JAV1-JAZ8-WRKY51 complex to derepress JA biosynthesis and defend against herbivory (Yan et al., 2018).

JAZ proteins serve as a key negative regulator in JA-mediated transcriptional responses (Pauwels et al., 2010; Pauwels and Goossens, 2011). However, there has so far been little research on the phosphorylation of JAZ proteins, and no reports of direct physical interactions between JAZ and any of the kinases involved in JA signaling (Gfeller et al., 2010; Yan et al., 2018; Liu et al., 2019). Here, we first demonstrated that OsJAZ4 physically interacted with and was phosphorylated by OsGSK2 (Figures 2 and 3). We then showed that OsGSK2 can disrupt the OsJAZ4-OsNINJA corepressor complex to destabilize OsJAZ4 via the 26S proteasome in an OsCOI1-dependent manner (Figures 4 and 8). In addition, we confirmed that OsJAZ4 suppressed JA signaling (Figure 6). Furthermore, overexpression of OsGSK2 resulted in decreased levels of OsJAZ proteins, thus increasing JA signaling by alleviating the repression mediated by JAZ (Figures 1, 4, and 5). This is consistent with previous reports that fast turnover of JAZ proteins holds the key to JA signal output (Gfeller et al., 2010; Mao et al., 2017; Wasternack and Feussner, 2018; Chen et al., 2019). Therefore, our results provide insight into the physical interaction between a JAZ family member, OsJAZ4, and a GSK3-like kinase, OsGSK2, involved in JA signaling.

In conclusion, we propose a model illustrating how OsGSK2 integrates the JA and BR signaling pathways and triggers rice antiviral resistance (Figure 9). As the levels of OsGSK2 increase, the repression of BR signaling and BR-induced susceptibility is enhanced. On the other hand, the increased amount of OsGSK2 binds to OsJAZ4 to dissociate the OsJAZ4-OsNINJA corepressor and the OsJAZ4-OsJAZ11 heterodimer complex. The phosphorylated OsJAZ4 and free OsJAZ11 are degraded by the 26S proteasome in an OsCOI1-dependent manner that in turn enhances the JA response and JA-mediated antiviral resistance. Our findings reveal an important mechanism for the positive role of OsGSK2 in rice antiviral immunity and provide insight into the crosstalk mechanism between JA and BR signaling.

Figure 9.

Model of OsGSK2-Mediated Plant Defense Signaling in Rice.

OsGSK2 binds to OsJAZ4 to disrupt OsJAZ4-OsNINJA corepressor and OsJAZ4-OsJAZ11 dimerization, which promotes the degradation of phosphorylated OsJAZ4 and free OsJAZ11 by the 26S proteasome in an OsCOI1-dependent manner. The increased amount of OsGSK2 elevates the JA response but suppresses the BR response, thereby enhancing rice antiviral defense. Lines ending with arrows show activation; a solid line ending with a perpendicular line indicates suppression or an antagonistic interaction.

METHODS

Plant Materials and Insect Vectors

Rice (Oryza sativa subsp japonica) cv Wuyujing No. 3 and cv NIP were used in this study. Go and Gi plants with the corresponding wild-type Zh11 were previously described by Tong et al. (2012a). The JA-insensitive mutant coi1-13 with its wild-type NIP was described by Yang et al. (2012). OsJAZ4-OE (line nos. 1 and 3), OsJAZ4-RNAi (line nos. 14 and 18), and OsJAZ4-MYC transgenic plants were constructed in this study (constructs are described below). RBSDV-infected rice plants were collected from fields in Shandong Province, China. Virus-free SBPHs were kept and reared on healthy Wuyujing No. 3 seedlings in glass beakers in a glasshouse at 25°C under artificial light. Rice plants were grown in a greenhouse at 28 to 30°C with a 14-h-light/10-h-dark cycle and light intensity of 600 µmol m⁻² s⁻¹. Nicotiana benthamiana plants were grown in a growth chamber at 25°C and with a 16-h-light/8-h-dark cycle.

RBSDV Transmission Experiment

RBSDV transmission via SBPHs was performed as previously described by Tong et al. (2011) and He et al. (2017), with some modifications. Briefly, SBPHs carrying RBSDV were transferred to rice seedlings at the 1.5- to 2.0-leaf stage (approximately three viruliferous insects per seedling) and allowed to feed for 3 d, after which they were removed from the seedlings. The inoculated plants grown in the greenhouse to develop symptoms. Plants infected with RBSDV exhibited symptoms such as stunting and darkening of leaves at 30 dpi, and the presence of RBSDV in each plant was confirmed by RT-PCR using virus-specific primers S10-F and S10-R (Supplemental Table). The number of diseased plants was used to calculate the viral incidence (percentage of plants infected). For each independent experiment, at least three biological replicates were used, and at least 40 seedlings were used for each replicate.

Total RNA Extraction and RT-qPCR

Total RNA from leaves was extracted using the TRIzol protocol (Invitrogen) in accordance with the manufacturer’s instructions. First-strand cDNA was synthesized from 1 μg of total RNA using a HiScript II Q RT for qPCR (+gDNA viper) kit (Vazyme). RT-qPCR was performed on a QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) using a CHamQ SYBR qPCR Master Mix kit (Vazyme), following the supplier’s protocol. The RT-qPCR conditions were as follows: 95°C for 3 min; 40 cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 20 s. The mRNA expression levels were normalized against the expression of housekeeping gene OsUBQ5, and the fold change was calculated by the comparative Ct method (2^–∆∆Ct method; Livak and Schmittgen, 2001). At least three biological replicate samples were used. Differences were considered significant at P ≤ 0.05. The primers used in this study are listed in the Supplemental Table.

Primary Root Inhibition Assay

Germinated seeds were cultured in normal rice culture solutions (Yoshida et al., 1976) supplemented with different concentration of meJA (TCI) and incubated in a growth chamber at 30°C with 8 h of light followed by 25°C with 16 h of darkness. Three days later, root lengths of seedlings were measured. For each treatment, at least 10 seedlings for each plant were treated and measured. Two independent experiments were performed.

Y2H Screening and Y2H Assays

For Y2H screening assay, the coding sequences of OsGSK2 were cloned into the pGBKT7 vector and used as the bait to screen a normalized rice cDNA prey library according to the manufacturer’s instructions (630498, Clotech).

For Y2H assays, the open reading frames (ORFs) of OsGSKs, OsCOI1b, OsNINJA, and OsJAZs with its mutants were cloned into the pGBKT7 (binding domain [BD]) or pGADT7 (activation domain [AD]) vectors. These constructs or the corresponding empty vectors were cotransformed into the yeast strain AH109 and incubated at 30°C on synthetic defined medium lacking Leu and Trp and then spotted on selective media lacking Ade, His, Leu, and Trp. Primers used are provided in the Supplemental Table.

Colocalization Experiments and BiFC Analysis

All binary vectors used in these studies were derived from the pCV1300 plasmid (Sun et al., 2013). For colocalization experiments, the ORFs of OsGSK2 or OsJAZ4 were cloned into pCV-mCherry-N1 or pCV-GFP-N1, respectively, to obtain pCV:OsJAZ4-mCherry and pCV:OsGSK2-GFP constructs as previously described by Sun et al. (2013). These constructs were cotransiently expressed in N. benthamiana leaves by Agrobacterium tumefaciens infiltration.

For BIFC assay, the full-length cDNA sequences of OsGSK2 and OsJAZ4 were cloned into the cYFP and nYFP vectors to obtain the OsGSK2-nYFP, OsGSK2-cYFP, OsJAZ4-nYFP, and OsJAZ4-cYFP constructs, respectively. The constructs were transformed into A. tumefaciens strain GV3101 and then transiently coexpressed in N. benthamiana leaves. The fluorescence signal for each combination was visualized using a TCS SP5 confocal laser scanning microscope system (Leica Microsystems) 40 to 44 h after infiltration.

Pull-Down Assay

The full-length coding sequence of OsGSK2 was cloned into the GST fusion vector (pGEX-6p-1), that of OsJAZ4 into the His fusion vector (pCOLD-TF), and those of OsJAZ11 and OsNINJA into the His fusion vector (pET-32a). The fusions were then transformed into Escherichia coli BL21 (DE3). To induce protein expression, a final concentration of 1 mM isopropyl β-d-thiogalactopyranoside was added when the OD₆₀₀ of the cultured cells was 0.6 to 0.8. For induction of recombinant protein, the cultures were incubated at 28°C for 8 h for GST-OsGSK2, OsNINJA, and His-OsJAZ11 and at 16°C for 16 h for His-OsJAZ4. For the pull-down assay, GST or GST-OsGSK2 was incubated with GST beads (Beaver) at 4°C for 1 h and then His-OsJAZ4 was added. The incubation continued for 2 h and then the beads were washed thoroughly, resolved by SDS-PAGE, and detected using anti-His antibody (1:4000 dilution, ab18184; Abcam). The bait proteins were probed with anti-GST antibody (1:3000 dilution, ab92; Abcam). The primers used are listed in the Supplemental Table.

For competitive pull-down assays, 3 μg of His-OsJAZ11 or His-NINJA with 2, 6, or 15 μg of GST-OsGSK2 or GST was incubated with immobilized OsJAZ4-myc at 4°C for 1 h. Proteins retained on the beads were resolved by SDS-PAGE and detected with anti-His antibody. The loading of OsJAZ4-myc was probed with anti-myc antibody (1:4000 dilution, ab9132; Abcam). His-OsJAZ11 and GST-OsGSK2, His-OsNINJA and GST-OsGSK2 were stained with Coomassie Brilliant Blue R 250. Supplemental Figure 16.

Co-IP Assay

For the Co-IP assay in N. benthamiana leaves, the coding sequences of OsGSK2, OsJAZ11, and OsJAZ4 were cloned into pCV-4HA-N1 and pCV-3myc-N1 vectors, respectively, to obtain pCV-HA-OsGSK2, pCV-OsJAZ11-myc, and pCV-OsJAZ4-myc constructs as previously described by Sun et al. (2013). Next, pCV-HA-OsGSK2 and pCV-OsJAZ11-myc or pCV-HA-OsGSK2 and pCV-OsJAZ4-myc were transiently coexpressed in N. benthamiana leaves. The leaves were collected and ground in liquid nitrogen and then extracted by Pierce^T IP lysis buffer (87,788; Thermo Fisher Scientific) with 1 mM DTT and 1× complete protease inhibitor cocktail. Next, 30 μL of anti-HA or anti-myc magnetic beads was added to the protein extraction and incubated at 4°C for 4 h. The precipitated samples were washed thoroughly, resolved by SDS-PAGE, and detected with the corresponding antibodies (anti-HA antibody, 1:3000 dilution, 2999S; Cell Signaling). Full scans of immunoblots are shown in Supplemental Figure 16. The primers used are listed in the Supplemental Table.

Kinase Assay

For the in vivo kinase assay, OsJAZ4-myc from pooled T1 OsJAZ4-MYC transgenic plants or N. benthamiana leaves coexpressed with different vector combinations were immunoprecipitated with anti-myc beads, and the IP product was treated with CIP. To avoid protein degradation, MG132 and a cocktail of proteinase inhibitors were added. The IP products were separated by 7.5% phos-tag (50 μM) SDS-PAGE (Wako) and analyzed with anti-myc antibody. Full scans of immunoblots are shown in Supplemental Figure 16.

Determination of Phosphorylation Sites of OsJAZ4 by OsGSK2 Kinase

His-OsJAZ4 was phosphorylated by GST-OsGSK2 as described by Wang et al. (2013). The fusion proteins (kinase:substrate, 1:5) were added in 25 μL of reaction buffer (25 mM Tris, pH 7.5, 12 mM MgCl₂, and 1 mM DTT) with 50 mM ATP in a 37°C water bath for 1 h. The phosphorylated His-OsJAZ4 was separated from the SDS-PAGE gel and subjected to in-solution alkylation/tryptic digestion followed by liquid chromatography–tandem mass spectrometry as described by Wang et al. (2013).

Protein Degradation Assay

For the degradation assay in N. benthamiana, individual cultures were adjusted to OD₆₀₀ = 1, and equal volumes were mixed before leaf infiltration. The infiltrated leaves of at least four plants were collected and pooled at 36 h and again at 48 h after infiltration. Next, 50 μM MG132 or 20 μM Bikinin was infiltrated at 36 h; 12 h later, proteins were collected.

For the cell-free protein degradation assay, the 7-d-old wild-type NIP seedlings were harvested and ground to a fine power in liquid nitrogen. Total protein was extracted in degradation buffer (25 mM Tris-HCl, pH 7.5, 10 mM NaCl, 10 mM MgCl₂, 5 mM DTT, and 10 mM ATP; Qiao et al., 2017). Extracts containing equal amounts of recombinant proteins were added to the tubes and incubated at 37°C for the times indicated.

Plant Hormone Treatment

For BL or Bikinin treatments, 7-d-old rice seedlings were sprayed with 1 μM BL (Sigma-Aldrich) or 20 μM Bikinin (Sigma-Aldrich) dissolved in 0.1% Triton X-100. Leaves were collected for protein extraction at the indicated time points. Three independent experiments were performed.

For in vivo lamina joint assays, the micro-drop method was performed as described previously by Hong et al. (2003). The lamina joints of the second leaf of 4-d-old seedlings were spotted with 1000 ng of BL in 1 μL ethanol by micro-syringe. The angles between the leaf lamina of the second leaf blade and sheath were measured 3 d after treatment by analyzing digital images using Motic Images Plus 2.0 software (China Group). At least 20 plants were used for each treatment. Three independent experiments were performed.

Vector Construction and Plant Transformation

To generate OsJAZ4-OE plants, the ORF of OsJAZ4 was amplified using CV-OsJAZ4-F/R primers and cloned into the pCAMBIA1300 vector driven by the 35S promoter. To produce the RNAi lines, two fragments of OsJAZ4 (nucleotides 960 to 1218) were amplified using the primer pairs RNAi-OsJAZ4-F1/R1 and RNAi-OsJAZ4-F2/R2 and then inversely inserted into pTCK303 vector driven by the ubiquitin1 promoter. The constructs described above were introduced into A. tumefaciens strain EHA105 and transformed into the NIP background. pCV-OsJAZ4-myc vector was used to generate OsJAZ4-MYC transgenic plants in a Zh11 background. The T4 generation of OsJAZ4-OE and OsJAZ4-RNAi lines and T1 hemizygous OsJAZ4-myc plants were used. The primers used are listed in the Supplemental Table.

JA Measurement

Seven-day-old total leaves from Go, Gi, and ZH11 plants were collected, ground in liquid nitrogen, and then used for hormone extraction and analysis as previously described by Fu et al. (2012) and He et al. (2017). Three biological replicates were used, each of which consisted of at least 15 pooled plants.

Phylogenetic Analysis

The amino acid sequences of OsJAZs and OsJGSK2s were download from the Rice Genome Annotation Project (http://rice.plantbiology.msu.edu/index.shtml) and aligned in ClustalW (https://myhits.sib.swiss/cgi-bin/clustalw; Supplemental File). Phylogenetic analyses were conducted using MEGA version 6, and the tree was generated using the neighbor-joining method (complete deletion and 1000 bootstrap replications; Tamura et al., 2013).

Antibody Generation and Validation

The purified recombinant His-OsJAZ11, GST-OsGSK2 protein and peptide of OsJAZ4 (CSSNRDESLSLGQPR) were used as antigens to immunize New Zealand rabbits to produce antiserum. The polyclonal antibodies (anti-OsJAZ4, anti-OsJAZ11, and anti-OsGSK2) from the generated antisera were purified by protein G chromatography (Bio-Rad) according to the manufacturer’s protocol. Immunoblotting was performed to detect the purified antibody. Anti-OsJAZ4, anti-OsJAZ11, and anti-OsGSK2 were used as primary antibody at 3, 4.5, and 1.5 μg/mL, respectively, and one specific band for each antibody was detected within the total protein fraction of plants tested (Supplemental Figure 11). The protein levels of OsJAZ4 and OsJAZ11 decreased upon MeJA treatment (Supplemental Figure 11).

Statistical Analysis

Differences were analyzed using ANOVA with Fisher’s least significant difference (LSD) tests. A P-value ≤ 0.05 was considered statistically significant. All analyses were performed using ORIGIN 8 software. Statistical data are provided in the Supplemental Data Set.

Accession Numbers

Sequence data from this article can be found in the rice genome annotation project databases under the following accession numbers: OsJAZ1 (Os04g55920); OsJAZ2 (Os07g05830); OsJAZ3 (Os08g33160); OsJAZ4 (Os09g23660); OsJAZ5 (Os04g32480); OsJAZ6 (Os03g28940); OsJAZ7 (Os07g42370); OsJAZ8 (Os09g26780); OsJAZ9 (Os03g08310); OsJAZ10 (Os03g08330); OsJAZ11 (Os03g08320); OsJAZ12 (Os10g25290); OsJAZ13 (Os10g25230); OsJAZ14 (Os10g25250); OsJAZ15 (Os03g27900); OsGSK1 (Os01g14860); OsGSK2 (Os05g11730); OsGSK3 (Os02g14130); OsGSK4 (Os01g19150); OsGSK5 (Os03g62500); OsGSK6 (Os05g04340); OsGSK7 (Os01g10840); OsGSK8 (Os06g35530).

Supplemental Data

• Supplemental Figure 1. Effect of OsGSK2 on Rice black-streaked dwarf virus (RBSDV) infection.

• Supplemental Figure 2. Levels of endogenous JA-IIe in seven-day-old Zh11, Go and Gi plants.

• Supplemental Figure 3. Phylogenetic analysis of OsGSK and OsJAZ amino acid sequences in rice using the neighbor-joining method.

• Supplemental Figure 4. SDS-PAGE analysis of recombinant GST-OsGSK2 and His-OsJAZ4 proteins.

• Supplemental Figure 5. Co-localization of OsGSK2-GFP and OsJAZ4-mCherry in N. benthamiana leaf epidermal cells.

• Supplemental Figure 6. RT-qPCR analysis of OsJAZ4, OsJAZ11 and OsGSK2 transcript levels for Figure 2D.

• Supplemental Figure 7. Identification of OsJAZ4 phosphorylation sites by OsGSK2 Kinase using LC-MS/MS.

• Supplemental Figure 8. The 8 potential phosphorylation motifs of OsGSK2 in OsJAZ4.

• Supplemental Figure 9. Effects of potential phosphorylation site mutants of OsJAZ4 on the OsJAZ4-OsGSK2 interaction.

• Supplemental Figure 10. Effects of OsGSK2 on OsJAZ4 accumulation in N. benthamiana leaves.

• Supplemental Figure 11. Antibody validation.

• Supplemental Figure 12. Accumulation of OsJAZs-myc in N. benthamiana leaves.

• Supplemental Figure 13. Effect of OsJAZ4 on RBSDV infection.

• Supplemental Figure 14. Survival rates of SBPH on OsJAZ4-OE lines (nos. 1 and 3), OsJAZ4-RNAi lines (nos. 14 and 18) and NIP.

• Supplemental Figure 15. OsGSK2 protein levels in wild type Nipponbare (NIP) and OsJAZ4 mutant plants.

• Supplemental Figure 16. Full scan data of the immunoblots in this work.

• Supplemental Table. Primers used in this work.

• Supplemental File. Multiple sequence alignment for Supplemental Figure 3.

• Supplemental Data Set. Data for all statistical analyses performed in this study.

DIVE Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• DWF4 Gramene: At3g50660

• DWF4 Araport: At3g50660

• brassinosteroid CHEBI: CHEBI:22921

• ATP CHEBI: CHEBI:15422

• COI1 Gramene: AT2G39940

• COI1 Araport: AT2G39940

• JAZ10 Gramene: AT5G13220

• JAZ10 Araport: AT5G13220

• MYC2 Gramene: Medtr8g067280

• MYC2 Araport: Medtr8g067280

• glycogen CHEBI: CHEBI:28087