SlWRKY45 interacts with jasmonate-ZIM domain proteins to negatively regulate defense against the root-knot nematode Meloidogyne incognita in tomato

Abstract Parasitic root-knot nematodes (RKNs) cause a severe reduction in crop yield and seriously threaten agricultural production. The phytohormones jasmonates (JAs) are important signals regulating resistance to multiple biotic and abiotic stresses. However, the molecular mechanism for JAs-regulated defense against RKNs in tomato remains largely unclear. In this study, we found that the transcription factor SlWRKY45 interacted with most JA-ZIM domain family proteins (JAZs), key repressors of the JA signaling. After infection by the RKN Meloidogyne incognita, the slwrky45 mutants exhibited lower gall numbers and egg numbers per gram of roots than wild type, whereas overexpression of SlWRKY45 attenuated resistance to Meloidogyne incognita. Under M. incognita infection, the contents of jasmonic acid (JA) and JA-isoleucine (JA-Ile) in roots were repressed by SlWRKY45-overexpression. Furthermore, SlWRKY45 bound to and inhibited the promoter of the JA biosynthesis gene ALLENE OXIDE CYCLASE (AOC), and repressed its expression. Overall, our findings revealed that the SlJAZ-interaction protein SlWRKY45 attenuated RKN-regulated JA biosynthesis and repressed defense against the RKN M. incognita in tomato.


Introduction
The phytoparasitic root-knot nematodes (RKNs Meloidogyne spp.) are extensively distributed throughout the world. Among them, Meloidogyne incognita is known as one of the most harmful RKNs [1]. Infective juveniles (J2s) of RKNs infect roots of plants, parasitize vascular cylinders, and stimulate roots to develop giant cells and form root-knots (galls) [2][3][4]. These galls destroy the normal physiological activities of the roots, hinder transport of water and nutrients, reduce host growth and yield, and even lead to host plant death [5,6].
Previous studies demonstrated that exogenous application of JA enhances resistance to RKNs in tomato [20]. Consistently, spr2, a tomato JA-deficient mutant, displayed a RKN-susceptible phenotype compared with wild type [21,22]. Recent studies revealed that some factors affect JA-regulated defense against the RKN M. incognita in tomato. For instance, miR319 and the transcription factor TCP4 regulate resistance to M. incognita by affecting JA contents [23]. The bHLH-type JA signaling transcription factor SlMYC2 participates in crosstalk between JA, strigolactone (SL), and abscisic acid (ABA) to inhibit resistance to M. incognita [24]. SlCSN4 and SlCSN5 interact with SlJAZ2 to positively regulate defense against M. incognita [25]. Although JAs control tomato defense against RKNs, the underlying molecular mechanism has not been fully explored, and remains to be elucidated.
WRKY transcription factors modulate plant defense against abiotic and biotic stresses [26][27][28]. The WRKY family in tomato contains 83 members [29]. SlWRKY70 confers resistance to aphids and the RKN Meloidogyne javanica, and its transcript level is inducible by salicylic acid (SA) but suppressed by JA [30]. Overexpression of SlWRKY3 enhances resistance to the RKN M. javanica, whereas loss of function of SlWRKY3 causes susceptibility [31]. SlWRKY45 represses tomato resistance to M. javanica, regarding the larger numbers of galls and females in roots with Agrobacterium rhizogenes-mediated SlWRKY45-overexpression [32].
Although SlWRKYs control tomato defense against RKNs, the regulatory mechanism is still unclear.
Here, we provided deep insights into the molecular mechanism by which SlWRKY45 negatively regulated defense against the RKN M. incognita. SlWRKY45 physically interacted with most SlJAZ members (SlJAZ1, SlJAZ2, SlJAZ3, SlJAZ4, SlJAZ5, SlJAZ6, SlJAZ7, and SlJAZ11). Loss of function of SlWRKY45 enhanced resistance to the RKN M. incognita, whereas overexpression of SlWRKY45 decreased defense against M. incognita. Furthermore, SlWRKY45 overexpression reduced the contents of JA and JA-Ile under M. incognita infection, whereas SlWRKY45 bound to the promoter of the JA biosynthesis gene ALLENE OXIDE CYCLASE (AOC) and repressed its expression. Our results provide evidence that SlWRKY45 participates in both JA signaling and biosynthesis pathways, and inhibits resistance to M. incognita in tomato.

SlJAZs interact with SlWRKY45
To explore the molecular basis of JA pathway in regulation of tomato defense against M. incognita, we sought to identify potential downstream transcription factors of SlJAZ repressors using the yeast two-hybrid (Y2H) system. We first analysed expression levels of SlJAZs at 1 d, 3 d, 7 d, and 14 d after M. incognita infection, and found that the SlJAZ11 expression was notably increased at these time points (Fig. S1, see online supplementary material). SlJAZ11 was selected as a bait and ligated with DNA binding domain (BD) in pLexA to screen a cDNA library of RKN-infected tomato. The WRKY transcription factor SlWRKY45 was identified as a candidate SlJAZ11-interaction protein (Fig. 1A). We further investigated whether other SlJAZs interacted with SlWRKY45 using Y2H assays. As shown in Fig. 1A, only BD-fused SlJAZ11 interacted with B42 activation domain (AD)-fused SlWRKY45 in yeast. We further fused SlJAZs with AD to produce AD-SlJAZs, and fused SlWRKY45 with BD to produce BD-SlWRKY45. As depicted in Fig. 1B, AD-SlJAZ1, SlJAZ2, SlJAZ3, SlJAZ4, SlJAZ5, SlJAZ6, SlJAZ7, and SlJAZ11 interacted with BD-SlWRKY45 in yeast, whereas AD-SlJAZ8, SlJAZ9, and SlJAZ10 did not.
Bimolecular fluorescence complementation (BiFC) assays showed that YFP signals were detected in the nucleus of epidermal N. benthamiana cells when SlJAZ7 or SlJAZ11 in fusion with N-terminal part of YFP (SlJAZ7-nYFP or SlJAZ11-nYFP) was coexpressed with cYFP (C-terminal part of YFP)-fused SlWRKY45 (cYFP-SlWRKY45), while YFP signals could not be detected when SlJAZ10-nYFP and cYFP-SlWRK45, or the negative control combinations were coexpressed (Fig. S2C, see online supplementary material). The results of BiFC assays suggest that SlJAZ7 and SlJAZ11 interact with SlWRKY45 in the nucleus.
SlJAZ7 and SlJAZ11 were truncated to generate N and C-terminal fragments (SlJAZ7NT, SlJAZ11NT, SlJAZ7CT, and SlJAZ11CT), containing the ZIM or Jas domain, respectively ( Fig. 2A), in order to investigate which domains of SlJAZs mediate interaction with SlJAZs and SlWRKY45. As illustrated in Fig. 2B, SlWRKY45 interacted with SlJAZ7NT and SlJAZ11NT, but not with SlJAZ7CT and SlJAZ11CT. LCI assays further showed that co-infiltration of SlJAZ7NT-nLUC or SlJAZ11NT-nLUC with cLUC-SlWRKY45 reconstituted strong LUC signals (Figs. 2C and D). These results consistently confirm that the N-terminal parts of SlJAZs participate in SlJAZ-SlWRKY45 interactions.

Nuclear localization and M. incognita infection-induced expression of SlWRKY45
We further ligated SlWRKY45 to the C-terminus of GFP (GFP-SlWRKY45) to analyse its subcellular localization. In contrast to the localization of GFP alone, GFP-SlWRKY45 protein signals were detected in the nucleus, suggesting that SlWRKY45 is localized to the nucleus (Fig. 3A). Consistent with the interaction of SlJAZ7/11 and SlWRKY45 in the nucleus, SlJAZ7-GFP and SlJAZ11-GFP were also localized to the nucleus (Fig. S3, see online supplementary material).
In addition, we explored the expression pattern of SlWRKY45, and discovered that the expression levels of SlWRKY45 in roots were strongly increased at 1 d, 3 d, 7 d, and 14 d after the RKN M. incognita infection, with a peak at 3 d (Fig. 3B), implying that SlWRKY45 may be involved in defense against M. incognita.

Loss of SlWRKY45 function enhances tomato defense against M. incognita
To investigate the function of SlWRKY45, we generated slwrky45 mutants in tomato (Solanum lycopersicum) cv Castlemart (CM) using CRISPR/Cas9 technology. Two targets specific for the first and second exons of SlWRKY45, respectively, were selected and ligated with the guide RNA scaffold in the CRISPR/Cas9 vector pCBSG012 (Fig. S4 , see online supplementary material). Thirteen independent T0 transgenic lines (slwrky45-cr-1 to slwrky45-cr- 13) were obtained through Agrobacterium-mediated transformation (Fig. S5A, see online supplementary material). Sequencing-based genotyping showed that the slwrky45 mutations were homozygous in slwrky45-cr-1 and slwrky45-cr-2, heterozygous in slwrky45cr-10 and slwrky45-cr-13, bi-allelic in slwrky45-cr-3, slwrky45-cr-9 and slwrky45-cr-11, and chimeric in slwrky45-cr-5 and slwrky45-cr-7, and it was wild type (WT) in the remaining four lines ( The two representative T0 homozygotes, slwrky45-cr-1 (16-bp and 1-bp deletions, respectively, in the first and second target) and slwrky45-cr-2 (3-bp and 1-bp deletions, respectively, in the first and second target) were chosen for further study (Fig. S5A, see online supplementary material). We first analysed the presence and absence of the slwrky45 mutations and Cas9 in 15 T1 generation plants of slwrky45-cr-1 and slwrky45-cr-2, respectively. As shown in Fig. S6 (see online supplementary material), all the detected T1 generation contained the same slwrky45 mutations as their corresponding parental T0 lines, suggesting that these mutations were stably inherited by their T1 progeny. Meanwhile, Cas9-free T1 homozygotes were identified (three plants from slwrky45-cr-1 and four plants from slwrky45-cr-2). Furthermore, we investigated the three most probable off-target sites of the two single guide RNAs in the slwrky45-cr-1 and slwrky45-cr-2 T1 plants, and found that no mutations occurred at these off-target sites ( Fig. S7 and Table  S1, see online supplementary material). Therefore, Cas9-free T2 Figure 1. SlJAZ proteins interact with SlWRKY45. A-B Yeast two-hybrid (Y2H) assays to assess interactions of SlJAZs with SlWRKY45. SlJAZs and SlWRKY45 were fused to BD domain in pLexA (marked as BD-SlJAZs or BD-SlWRKY45) or B42AD domain in pB42AD (marked as AD-SlJAZs or AD-SlWRKY45). C Interactions of SlJAZs and SlWRKY45 were detected by firefly luciferase (LUC) complementation imaging (LCI) assays. SlJAZ1, SlJAZ2, SlJAZ3, SlJAZ4, SlJAZ7, SlJAZ11, and SlWRKY45 were fused with nLUC or cLUC (N or C-terminal fragments of LUC) to produce SlJAZ1-nLUC, SlJAZ2-nLUC, SlJAZ3-nLUC, SlJAZ4-nLUC, SlJAZ7-nLUC, SlJAZ11-nLUC, and cLUC-SlWRKY45, respectively. Luciferase activities were evaluated at 50 h after the infiltration of corresponding Agrobacterium strains in leaves of N. benthamiana. D Pull-down assays show that SlJAZ7 and SlJAZ11 associate with SlWRKY45. MBP and MBP-SlWRKY45 were immobilized on amylose resin, and incubated with transiently expressed myc-SlJAZ7 and myc-SlJAZ11 proteins. The samples were detected with an anti-myc antibody. E Bimolecular fluorescence complementation (BiFC) analyses show interaction of SlJAZ7/11 and SlWRKY45. SlJAZ7, SlJAZ11, and SlWRKY45 were fused with nYFP or cYFP (N or C-terminal parts of YFP), respectively. YFP signals were observed at 50 h after the expression of corresponding combinations of Agrobacterium strains in leaves of N. benthamiana. compared with those in WT. Furthermore, we found that at 35 d post inoculation, the gall numbers per plant, and gall numbers and egg numbers per gram of roots in slwrky45-cr mutants were also significantly lower compared with those in WT ( Fig. 4A-C), consistently indicating that SlWRKY45 plays a negative role in defending against M. incognita. PLANT DEFENSE FACTOR (PDF) and PROTEINASE INHIBITOR 2 (PI-2) are two defensive genes against pathogens, including M. incognita [24]. SlPDF and SlPI-2 expression was responsive to M. incognita infection as shown in Fig. 4D and E. Consistent with the increased resistance of slwrky45-cr plants, the expression levels of SlPDF and SlPI-2 were obviously higher in slwrky45-cr roots with the RKN M. incognita infection compared with those in the M. incognita-infected WT roots ( Fig. 4D and E).

Overexpression of SlWRKY45 reduces resistance to M. incognita in tomato
We further generated SlWRKY45-overexpressing lines of tomato. Flag-SlWRKY45-OE-5 and flag-SlWRKY45-OE-13 with approximately 39-fold and 28-fold of the WT level regarding SlWRKY45 expression, respectively, were used as representatives (  Fig. 5D and E, the expression of SlPDF and SlPI-2 in the roots of flag-SlWRKY45-OE lines at 1 d, 3 d, 7 d, and 14 d after infection was significantly lower than those in the infected WT roots. Altogether, the results in Figs. 4, 5, and S9 (see online supplementary material) demonstrate that SlWRKY45 acts as a repressor to regulate resistance to M. incognita in tomato.

SlWRKY45 overexpression attenuates RKN-affected JAs biosynthesis in tomato
To further explore the mechanism of SlWRKY45 in regulating the resistance to M. incognita, we examined the concentrations of JA and JA-Ile in the roots of WT and SlWRKY45-overexpressing plants (flag-SlWRKY45-OE-5 as a representative) without or with M. incognita inoculation. In contrast to those in WT roots without M. incognita infection, the contents of JA and JA-Ile in M. incognitainfected roots of WT were increased at 1 d after infection, and were restored to normal levels or even decreased at 3 d and 7 d ( Fig. 6A and B), which were consistent with previous observations [33]. The contents of JA and JA-Ile in the roots of flag-SlWRKY45-OE-5 at 1 d, 3 d, and 7 d after M. incognita infection were lower compared with those in the roots of M. incognita-infected WT ( Fig. 6A and B).
In addition, the CM wild type and flag-SlWRKY45-OE-5 plants were treated with MeJA, and 24 h later, these plants were infected with M. incognita for 7 d and 35 d. The results in Figs. S10 (see online supplementary material) and 6C-E showed that MeJA

SlWRKY45 binds to and represses the JA biosynthesis gene SlAOC
A previous study showed that two JA biosynthesis genes LIPOXYGENASE D (LOXD) and ALLENE OXIDE CYCLASE (AOC) were induced by the RKN M. incognita infection in tomato [25]. Thus, we explored whether the decreased JAs contents in M. incognita-infected SlWRKY45-overexpressing plants were due to reduced expression of SlLOXD or SlAOC under M. incognita infection. To demonstrate this, we analysed the transcript levels of SlLOXD and SlAOC, and discovered that, in response to M. incognita infection, the expression levels of SlAOC were lower in SlWRKY45-overexpression lines compared with those in WT (Fig. 7A), while SlLOXD expression was not affected (Fig. S11 , see online supplementary material).
We next investigated whether SlWRKY45 was able to bind to and regulate SlAOC. We carried out chromatin immunoprecipitationquantitative PCR (ChIP-qPCR) analysis using M. incognita-infected flag-SlWRKY45-OE-5 transgenic plants, and found that SlWRKY45 bound to the first and second typical WRKY factor target sequences (W-box, TTGACT) in the promoter of SlAOC (Fig. 7B and C). We further examined the regulatory function of SlWRKY45 on the SlAOC promoter using Dual-LUC assays in which the SlAOC promoter drove the luciferase (LUC) reporter gene (SlAOC pro -LUC), and SlWRKY45 controlled by the CaMV35S promoter (35S-SlWRKY45) served as an effector (Fig. 7D). As shown in Fig. 7E, coexpression of SlAOC pro -LUC and 35S-SlWRKY45 in N. benthamiana leaves resulted in a lower LUC/REN ratio than the coexpression of SlAOC pro -LUC and the control 35S-GFP, consistently suggesting that SlWRKY45 obviously represses the promoter and expression of SlAOC. Moreover, we found that the repression activity of SlWRKY45 was attenuated by SlJAZ11 (Fig. 7E), suggesting that SlJAZ11 inhibits the function of SlWRKY45.

Discussion
Previous studies have indicated that WRKY transcription factors associate with JA pathway to exert their biological functions. For example, AtWRKY57 integrates both auxin and JA signaling by interacting with AtIAA29 and AtJAZs, and mediates leaf senescence [34]. AtWRKY51 associates with AtJAZ8/AtJAV1 to comprise a JAV1-JAZ8-WRKY51 complex for controlling defense against insects [35]. SlWRKY31 cooperates with SlVQ15, and participates in JA signaling and resistance to Botrytis cinerea [36]. Nevertheless, the relationship between JA and WRKYs in tomato remains poorly understood. Here, we reveal that SlWRKY45 is involved in both JA biosynthesis and signaling pathways to attenuate resistance to the RKN M. incognita (Figs 1-7).
WRKYs play regulatory roles via interacting with various proteins. AtWRKY50 interacts with AtTGA2 or AtTGA5 to form a protein complex and synergistically activates the expression of the resistance-related gene PATHOGENESIS-RELATED 1 (PR1) [37]. AtWRKY8 recruits AtVQ10 to enhance its binding to target DNA, and promote resistance to B. cinerea [38]. AtVQ9 associates with and decreases the transcriptional activity of AtWRKY8, and modulates tolerance to salt stress [39]. AtWRKY12 and AtWRKY13 interact with SQUAMOSA PROMOTER BINDING-LIKE 10 to antagonistically regulate their transcriptional functions and age-mediated flowering [40]. AtWRKY38 and AtWRKY62 associate with HISTONE DEACETYLASE 19 to participate in defense responses [41]. Here, we demonstrate that SlJAZs act through their N-terminal regions to physically interact with SlWRKY45 using Y2H, LCI, pull-down, and BiFC assays (Figs. 1 and 2).
WRKYs bind to target genes and their own promoters to activate or repress expression through the combination of the WRKY domain and W-box. For instance, AtWRKY57 binds to the W-box region of the AtJAZ1 and AtJAZ5 promoters and activates their expression to repress JA-mediated defense against B. cinerea [42]. AtWRKY51 binds to the promoter regions of the Arabidopsis JA biosynthesis gene ALLENE OXIDE SYNTHASE (AOS) via the W-box sequence [35]. It also associates with AtJAV1 and AtJAZ8 to repress AtAOS expression, which inhibits JAs biosynthesis and controls resistance to insect attack [35]. PcWRKY1 binds to the W-boxes in its own promoter, as well as the PcWRKY3 and PcPR1-1 promoters [43]. Here, our ChIP-qPCR results showed that SlWRKY45 binds to fragments spanning the W-boxes in the promoter of the JA biosynthesis gene SlAOC (Fig. 7C). Dual-LUC assays suggested that SlWRKY45 attenuates the expression of SlAOC (Fig. 7E). It would be interesting to further investigate whether SlWRKY45 could bind to its own promoter and regulate its own expression.
One previous study reported that SlWRKY45 expression is enhanced within 5 d after M. javanica infection, and is maintained through feeding-site development and gall formation [32]. The phytohormones cytokinin (CK), auxin, and SA induce SlWRKY45 expression, whereas JA inhibits its expression. In response to the RKN M. javanica infection, the roots with transient SlWRKY45 overexpression displayed higher numbers of developed females, gall formation, and overall feeding site area than WT roots. In this study, we found that SlWRKY45 was localized to the nucleus, and was induced at 1 d, 3 d, 7 d, and 14 d after M. incognita infection (Fig. 3). We generated stable SlWRKY45-overexpressing transgenic tomato and slwrky45 mutants, and demonstrated that SlWRKY45 is a negative regulator of tomato defense against M. incognita (Figs. 4, 5, and S9, see online supplementary material). Moreover, we deeply explored and revealed the mechanism that SlWRKY45 interacts with SlJAZs, directly binds to the SlAOC promoter, and inhibits SlAOC expression and JA biosynthesis to reduce defense against the RKN M. incognita (Figs. 1, 2, 6, and 7).
SlMYC2, a master transcription factor, interacts with 11 SlJAZs [44], and plays positive or negative roles to modulate diverse physiological responses in tomato, including positively regulating fruit chilling tolerance [45], resistance to the necrotrophic pathogen B. cinerea [44,46], growth and developmental processes such as flower formation, fruit set, and fruit shape [47], and negatively controlling resistance to M. incognita by mediating the interplay of SL, ABA, and JA [24]. In our results, we discovered that SlWRKY45 interacts with most SlJAZ proteins, and negatively controls defense against M. incognita (Figs. 1, 4, 5, and S9, see online supplementary material). It remains to explore whether SlWRKY45 also controls other JAregulated responses in tomato. As shown by the finding that the transcription factors SlMYC2 and SlWRKY45 both repress resistance to M. incognita in tomato [24] (Figs. 4, 5, and S9, see online supplementary material), it remains to investigate whether SlMYC2 and SlWRKY45 target some mutual downstream genes to synergistically or antagonistically regulate their expression, and control defense responses. Additionally, it will be interesting to isolate the master factors positively modulating JA-mediated tomato defense against RKNs, which will contribute to further understanding the molecular basis of JA-controlled defense responses.
A summary of our findings is shown in Fig. 8. SlJAZ repressors interact with and repress SlWRKY45 to attenuate its function. JAs induce SlJAZs degradation to release SlWRKY45. The released SlWRKY45 binds to and inhibits SlAOC expression to reduce RKNregulated JAs biosynthesis, and represses defense against the RKN M. incognita.

Plant materials and growth conditions
Seeds of CM wild type, SlWRKY45-overexpressing plants and slwrky45 mutants were germinated at 28 • C on moistened filter paper for 2-3 days, and then grown in a greenhouse (

Generation of SlWRKY45-overexpression plants
The CDS of SlWRKY45 was amplified with specific primers (Table  S2, see online supplementary material), and ligated into a reformative pCAMBIA1300 vector through the Sal I and Spe I sites to generate the SlWRKY45-overexpression plasmid (flag-SlWRKY45-OE), in which SlWRKY45 was fused with three flag tags and was driven by the CaMV35S promoter. Through Agrobacterium (GV3101)-mediated cotyledon explant transformation, this construct was transformed into CM wild type. Hygromycin B was used to select the transgenic lines. T3 homozygous SlWRKY45overexpressing plants were used for further experiments.

Generation of SlWRKY45 gene-edited plants
CRISPR/Cas9 technology was used to generate slwrky45 mutants. We used CRISPR-P to choose two sgRNAs that targeted the first and second exons of SlWRKY45 (Fig. S4A, see online supplementary material). The PCR primers included the two target sites and the Bsa I site, which are shown in Table S2, see online supplementary material. The PCR fragment was amplified using the plasmid pSG-SlU6 (Biogle GeneTech, Jiangsu, China) as a template to generate the PCR product containing the Bsa I site, Data represent means (±SD) of three independent biological replicates. Significant differences were analysed by ANOVA with Duncan's multiple range test (P < 0.05) and indicated with letters. B Diagram of SlAOC promoter. Red lines represent the regions detected in ChIP-qPCR assays. C ChIP-qPCR assays to detect the binding of flag-SlWRKY45 to SlAOC promoter. Chromatin from RKN-infected CM wild type and the flag-SlWRKY45 transgenic plants was immunoprecipitated without (−) or with (+) anti-flag antibody. A promoter of SlACTIN2 was used as a negative control. Data represent means (±SD) of three independent biological replicates. Significant differences were analysed by ANOVA with Duncan's multiple range test (P < 0.05) and indicated with different letters. D Diagram displaying the constructs used in the Dual-LUC assays in (E). E Dual-LUC assays showing that SlWRKY45 attenuates SlAOC promoter activity, and that SlJAZ11 inhibits this effect. Error bars represent SD (n = 6). Significant differences were analysed by ANOVA with Duncan's multiple range test (P < 0.05) and indicated with letters. two targets, gRNA scaffold, and the tomato U6 promoter (Fig. S4B, see online supplementary material). The product was purified and inserted into the pCBSG012 vector (Biogle GeneTech, Jiangsu, China) through the Bsa I site (Fig. S4C-D, see online supplementary material). Through Agrobacterium (GV3101)-mediated cotyledon explant transformation, this construct was transformed into CM wild type. Hygromycin B was used to select the transformants. slwrky45 mutants of Cas9-free T2 homozygotes were analysed by PCR and sequencing, and used for further experiments.

Analysis of mutation types and off-target mutations
Tomato DNA was extracted using a DNA extraction kit (GeneBette, Beijing, China), and used as a template to amplify the target sites using PrimeSTAR Max DNA polymerase (TaKaRa, Ohtsu, Japan). To analyse mutation types of each T0 line, we cloned the PCR fragments into the pMD20-T vector (TaKaRa, Ohtsu, Japan), sent 15 individual clones for sequencing, and analysed the mutations. To analyse mutation types of T1 and T2 lines, the PCR products were sequenced. The primers used for amplification are shown in Table S3, see online supplementary material. To identify Cas9, the PCR products were amplified using primers specific for Cas9 (Table S4, see online supplementary  material).
To analyse off-target mutations, we used CRISPR-P to predict the potential off-target sites (Table S1, see online supplementary material). The corresponding primers (Table S5, see online supplementary material) for each site were used for PCR amplification using PrimeSTAR Max DNA polymerase (TaKaRa, Ohtsu, Japan), and the PCR products were sequenced.

Yeast two-hybrid screening and yeast two-hybrid assays
SlJAZ11 was ligated to the pLexA vector with BD, and the cDNA library generated with RKN-infected tomato was used for yeast two-hybrid (Y2H) screening. Y2H screening was carried out based on the manufacturer's instructions (Clontech, CA, USA). For Y2H assays, full-length CDSs or fragments of SlJAZ1, SlJAZ2, SlJAZ3, SlJAZ4, SlJAZ5, SlJAZ6, SlJAZ7, SlJAZ8, SlJAZ9, SlJAZ10, SlJAZ11, and SlWRKY45 were fused with the pLexA or pB42AD vector. Yeast transformation and analysis of protein interactions were carried out in line with the manufacturer's instructions (Clontech, CA, USA). All Y2H experiments were repeated three biological times.

BiFC assays
The CDSs of SlWRKY45, SlJAZ7, SlJAZ10, and SlJAZ11 were fused into the cYFP or nYFP vector [49]. Agrobacterium GV3101 carrying the indicated construct pairs was injected into leaves of N. benthamiana. At 50 h after injection, YFP signals were captured using a confocal microscope (TCS-SP5, Leica, Wetzlar, Germany). BiFC assays were repeated three biological times. The primers are shown in Table S2, see online supplementary material.

Pull-down assays
The CDS region of SlWRKY45 was cloned to the pMAL-c5X vector (NEB, MA, USA) via the Sal I and EcoR I sites for MBP fusion. Escherichia coli strains Transetta (DE3) containing the MBP-fused SlWRKY45 vector or empty vector were cultured at 16 • C overnight in LB liquid medium with 0.3 mM IPTG to induce the expression of the corresponding proteins. Amylose resin (NEB, MA, USA) was used to purify these proteins.
Pull-down assays were adopted using previously described methods [49] with slight modification. MBP-fused SlWRKY45 and MBP proteins were respectively added to the amylose resin for 4 h at 4 • C, after which 200 μL myc-SlJAZ7, myc-SlJAZ10, or myc-SlJAZ11 protein was added and incubated for 2 h at 4 • C. After washing four to six times, the samples were boiled with 100 μL SDS loading buffer at 95 • C for 10 min. The proteins were separated via 10% SDS-PAGE for 2 h, and then transferred to PVDF membranes (Millipore, MA, USA). The PVDF membranes were blocked using PBS buffer with 5% nonfat milk, and subsequently incubated with an anti-myc antibody (Abmart, Shanghai, China; 1:5000 dilutions) for 1 h and a goat anti-mouse secondary antibody (Abmart, Shanghai, China; 1:3000 dilutions) for 1 h. The proteins were observed using enhanced chemiluminescence (ECL) (Solarbio, Beijing, China) by a Tanon 5200Multi instrument (Tanon, Shanghai, China). The pull-down assays were repeated three biological times.

Quantitative real-time PCR
RNA isolation and cDNA synthesis were respectively conducted using kits (DP432, Tiangen, Beijing, China; and AT311-02, Transgen, Beijing, China) in line with the manufacturer's instructions. qRT-PCR using SYBR Green Mix (TaKaRa, Ohtsu, Japan) with specific primers (Table S6, see online supplementary material) on the Bio-Rad CFX96 qPCR instrument was conducted to detect the expression levels of genes. The qRT-PCR reaction conditions were as below: 3 min at 95 • C, 39 cycles of 15 s at 95 • C, 10 s at 56 • C, and 72 • C for 15 s. The internal control gene was tomato ACTIN2. Values of relative gene expression were analysed using the 2 − Ct method [50]. The qRT-PCR experiments were repeated three biological times.

Subcellular localization
SlWRKY45 was in fusion with GFP in the pEGAD vector through the EcoR I and BamH I sites to produce the GFP-SlWRKY45 construct. SlJAZ7 and SlJAZ11 (without a stop codon) were cloned into the Super1300-GFP vector via the Hind III and Spe I sites, which generated the SlJAZ7-GFP and SlJAZ11-GFP vectors, respectively. These constructs and the corresponding empty vectors were expressed in N. benthamiana leaves through Agrobacterium strain GV3101-mediated expression. GFP signals were captured after 50 h of infiltration with a confocal microscope (TCS-SP5, Leica, Wetzlar, Germany). The primers used in this experiment are shown in Table S2, see online supplementary material. These experiments were repeated three biological times.

Meloidogyne incognita inoculation assays
T3 homozygous SlWRKY45-overexpressing plants, Cas9-free T2 homozygous slwrky45 mutants, and their CM wild type plants were used for the experiments. Eggs of M. incognita were obtained from the infected tomato roots according to previously described methods [51], and incubated at 28 • C to hatch J2s. When tomato seedlings grew to four true leaves, each plant was inoculated with approximately 400 J2s. At 7 d and 35 d after infection, the roots were soaked with 1.5% sodium hypochlorite for 5 min, washed with water twice, and stained by 3.5% acid fuchsin with 25% acetic acid. After staining, the roots were washed with water and decolored with the solution of glycerol, acetic acid, and H 2 O (1:1:1). Then, we weighed the root, and counted the number of galls on each. For counting the egg numbers, nematode eggs of each root were collected according to previously described methods [51], and the egg numbers were counted under a microscope (SMZ-140, Motic, Guangdong, China) in 20 aliquots of 10 μL. For Figs. 6C and S10 (see online supplementary material), the indicated plants were treated without or with 100 μM MeJA, and 24 h later, the plants were infected with M. incognita. These experiments were repeated three biological times.

Quantification of JAs contents
H 2 JA (OlChemIm Ltd, Olomouc, Czech Republic) was used as the internal standard. Extraction and quantitative analysis of JA and JA-Ile contents was performed according to previously described methods [52]. Briefly, frozen roots were ground into powder. Then extraction buffer was added (2-propanol/H 2 O/concentrated HCl) and the internal standard, oscillated at 4 • C for 30 min. After adding dichloromethane and centrifuging, the supernatants were evaporated to dryness under N 2 gas, redissolved with methanol (0.1% formic acid), centrifuged, and filtrated using a 0.22 μM membrane. Finally, 2 μL of the solution was analysed using the ESI-HPLC-MS/MS system (HPLC, Agilent 1290, Agilent Technologies, CA, USA; MS, Applied Biosystems 6500 Quadrupole Trap, Applied Biosystem, CA, USA). These experiments were repeated three biological times.

ChIP assays
ChIP assays were carried out in accordance with previously described protocols [53]. Briefly, 21-day-old CM and flag-SlWRKY45-OE-5 plants were infected with the RKN M. incognita for 24 h. The harvested samples were soaked with 1% formaldehyde for cross-linking, neutralized with 0.125 M Glycine, and then ground into powder. The chromatin-protein was isolated, sonicated to cut the DNA to 300-500 bp, and then incubated with or without anti-flag-tag antibody (Agarose Conjugated) (Abmart, Shanghai, China) for 4 h at 4 • C. The buffer with 0.1 M NaHCO 3 and 0.5% SDS was used to elute the immunoprecipitated chromatin. Then the samples were added 5 M NaCl and incubated at 65 • C overnight. Finally, DNA was extracted using a DNA extraction kit (GeneBette, Beijing, China). qPCR analysis was used to measure enrichment of promoter fragments by the % input method [54]. A fragment of the SlACTIN2 promoter served as a negative control. The primers used for the experiment are listed in Table S6, see online supplementary material. These experiments were repeated three biological times.

Dual-LUC assays
SlWRKY45, SlJAZ11, and GFP were ligated to the pGreenII 62-SK vector [55] via the BamH I and EcoR I sites. The ∼1300 bp promoter of SlAOC was fused to the pGREENII 0800 LUC vector [55] via the BamH I and Nco I sites. Agrobacterium strains GV3101 (pSoup) carrying the corresponding constructs were cultured overnight, combined with the indicated recombinant pairs, incubated for 3-5 h, and co-infiltrated into N. benthamiana leaves. At 50 h after injection, LUC and REN activities were assessed with a dualluciferase assay kit (E1910, Promega, WI, USA). The primers used for Dual-LUC assays are shown in Table S2, see online supplementary material. These experiments were repeated three biological times.