RVE2, a new regulatory factor in jasmonic acid pathway, orchestrates resistance to Verticillium wilt

Summary

Verticillium dahliae, one of the most destructive fungal pathogens of several crops, challenges the sustainability of cotton productivity worldwide because very few widely‐cultivated Upland cotton varieties are resistant to Verticillium wilt (VW). Here, we report that REVEILLE2 (RVE2), the Myb‐like transcription factor, confers the novel function in resistance to VW by regulating the jasmonic acid (JA) pathway in cotton. RVE2 expression was essentially required for the activation of JA‐mediated disease‐resistance response. RVE2 physically interacted with TPL/TPRs and disturbed JAZ proteins to recruit TPL and TPR1 in NINJA‐dependent manner, which regulated JA response by relieving inhibited‐MYC2 activity. The MYC2 then bound to RVE2 promoter for the activation of its transcription, forming feedback loop. Interestingly, a unique truncated RVE2 widely existing in D‐subgenome (GhRVE2D) of natural Upland cotton represses the ability of the MYC2 to activate GhRVE2A promoter but not GausRVE2 or GbRVE2. The result could partially explain why Gossypium barbadense popularly shows higher resistance than Gossypium hirsutum. Furthermore, disturbing the JA‐signalling pathway resulted into the loss of RVE2‐mediated disease‐resistance in various plants (Arabidopsis, tobacco and cotton). RVE2 overexpression significantly enhanced the resistance to VW. Collectively, we conclude that RVE2, a new regulatory factor, plays a pivotal role in fine‐tuning JA‐signalling, which would improve our understanding the mechanisms underlying the resistance to VW.

Introduction

Verticillium wilt (VW), a vascular disease caused by Verticillium dahliae Kleb. (Vd), reduces yield as well as quality of several important crops including cotton, tomato, lettuce and potato (Fradin and Thomma, 2006; Klosterman et al., 2009). This disease is very difficult to control on infected plants owing to the long viability of resting structure, microsclerotia and invasion of pathogen into xylem vessels (Fradin and Thomma, 2006; Inderbitzin and Subbarao, 2014; Zhang et al., 2014). Varieties with improved resistance to the disease is the most economical, user and environment‐friendly strategy for combating VW (Fradin and Thomma, 2006; Klosterman et al., 2009). However, very few widely‐cultivated Upland cotton (2n = 4x = AADD = 52) varieties are resistant to VW. Gossypium australe F. Mueller (2n = 2x = GG = 26), a wild diploid species, is an important cotton species containing genes for multiple important traits including delayed gland morphogenesis and resistance to insect pests and diseases (Cai et al., 2020; Feng et al., 2019). Nevertheless, the molecular basis of VW resistance in G. australe remains enigmatic.

Jasmonate (JA), a lipid‐derived plant hormone, is a major immunity hormone which regulates plant defence response to mechanical wounding, herbivore attack and necrotrophic and hemibiotrophic pathogen infection (Browse, 2009; Chini et al., 2016; Wasternack and Hause, 2013). The JA signalling pathway consisting of multiple interconnected protein–protein interaction has been well studied in model plants, Arabidopsis and tomato. The basic helix–loop–helix (bHLH) transcription factor, MYC2, acts as a master regulator in diverse aspects of JA response (Boter et al., 2004; Dombrecht et al., 2007; Du et al., 2017; Kazan and Manners, 2013; Lorenzo et al., 2004; Zhai et al., 2013). In the resting stage, several JASMONATE‐ZIM‐DOMAIN (JAZ) proteins physically bind and inhibit MYC2 and related MYC transcriptional activators act as repressors by recruiting members of the TOPLESS (TPL) and TPL‐related (TPRs) proteins through Novel‐Interactor of JAZ (NINJA) (Chini et al., 2007; Fernández‐Calvo et al., 2011; Pauwels et al., 2010; Qi et al., 2015; Sheard et al., 2010; Thines et al., 2007; Zhang et al., 2015). It is believed that TPL can epigenetically repress MYC2‐regulated gene expression by recruiting repressive histone modification enzymes and chromatin remodelers (Zhang et al., 2017; Zhu et al., 2011). In response to stimuli, jasmonoyl‐isoleucine (JA‐Ile), bioactive JA ligand, is rapidly synthesized by JASMONATE‐RESISTANT1 (JAR1) (Fonseca et al., 2009; Staswick and Tiryaki, 2004). The JAZ proteins make a JA‐Ile‐dependent co‐receptor complex with CORONATINE‐INSENSITIVE1 (COI1), which is the F‐box subunit of the SCF^COI1 ubiquitin ligase, to perceive JA‐Ile (Chini et al., 2007; Sheard et al., 2010; Thines et al., 2007; Xie et al., 1998; Xu et al., 2002). The JAZ proteins are subsequently ubiquitinated and degraded through the 26S proteasome pathway, which liberates transcriptional activators involved in diverse JA‐mediated responses (Chini et al., 2007; Sheard et al., 2010; Thines et al., 2007; Xie et al., 1998; Xu et al., 2002; Yan et al., 2007, 2009). Previous studies have revealed that JA pathway plays an essential role in plant defence responses against necrotrophic and hemibiotrophic pathogens (Abuqamar et al., 2008; Kachroo and Kachroo, 2009; Stintzi et al., 2001; Thaler et al., 2004; Thomma et al., 1998; Vijayan et al., 1998). For instance, the Arabidopsis JA biosynthesis mutant fad3/fad7/fad8 exhibited increasing susceptibility to Pythium mastophorum (Stintzi et al., 2001), while a mutant jar1, deficient in catalysing biologically active form of JA, is highly susceptible to the V. dahliae. Additionally, mutation in COI1, jasmonate receptor, results in increasing susceptibility to necrotrophic fungi Botrytis cinerea and Fusarium oxysporum (Thaler et al., 2004; Thomma et al., 1998). In contrast, Arabidopsis mutant cev1 constitutively produces JA and constitutively expresses JA‐related downstream genes PLANT DEFENSIN1.2 (PDF1.2) and THIONIN2.1, and exhibits enhanced resistance to powdery mildew (Ellis and Turner, 2001). In transgenic Arabidopsis, overexpressing ETHYLENE RESPONSE FACTOR1 (ERF1) and ERF2 activated JA‐responsive genes PDF1.2, exhibited more resistance to F. oxysporum (Berrocal‐Lobo and Molina, 2004; Kidd et al., 2009; McGrath et al., 2005; Thaler et al., 2004; Thatcher et al., 2009) than that of the wild type. Whereas, esa1 mutant showed that delayed induction of PDF1.2 resulted in enhancing susceptibility to F. oxysporum (Van Hemelrijck et al., 2006). Although JA plays a pivotal role in protecting plants against various pathogens, the offensive‐defensive transformations are more complex than what we know (Yan and Xie, 2015). It is reported that some biotrophic and hemibiotrophic pathogens are able to produce and inject toxins and virulence effector proteins into host cells, thereby hijacking the JA‐signalling pathway to evade plant defence systems (Cole et al., 2014; Gimenez‐Ibanez et al., 2014; Jiang et al., 2013).

The REVEILLE (RVE) proteins belong to a subfamily of MYB‐like transcription factors that includes CIRCADIAN‐CLOCK‐ASSOCIATED1 (CCA1) and LATE–ELONGATED‐HYPOCOTYL (LHY) clock components, with a single Myb‐like domain containing a distinctive SHAQKYF motif (Schaffer et al., 1998; Wang and Tobin, 1998). The circadian central oscillator consists of CCA1, LHY and TIMING‐OF–CAB‐EXPRESSION1 (TOC1) in Arabidopsis (Carré and Kim, 2002; Harmer, 2009). In the family, the best‐studied were CCA1 and LHY, which regulate clock rhythms by inhibiting TOC1 expression and promoting expression of PSEUDO‐RESPONSE‐REGULATOR7 (PRR7) and PRR9 (Alabadí et al., 2001; Farré et al., 2005). The RVE1 involved in integrating the circadian clock and auxin pathways to regulate hypocotyl growth (Rawat et al., 2009), modulating chlorophyll biosynthesis and seedling de‐etiolation (Xu et al., 2015) and regulating cold acclimation (Meissner et al., 2013). The RVE2 or CIRCADIAN1 acts as a part of regulatory feedback loop of circadian outputs (Zhang et al., 2007), and along with RVE1 acts as activators of seed dormancy by restraining bioactive GA biosynthesis (Jiang et al., 2016; Yang et al., 2020). The RVE4 plays a Time‐of‐Day‐Specific role in regulating first wave of heat shock‐regulated gene expression (Li et al., 2019). The RVE7 or EARLY–PHYTOCHROME‐RESPONSIVE1 requires central oscillator for rhythmic expression and probably as a slave oscillator important for fine‐tuning the output pathways (Kuno et al., 2003). RVE8 with PRR5 forming a negative transcriptional feedback loop acts within circadian network and RVE8 represses anthocyanin biosynthesis (Farinas and Mas, 2011; Pérez‐García et al., 2015; Rawat et al., 2011). Nevertheless, whether RVE2 proteins are involved in regulating other hormone signalling pathways and plant defence responses is not fully understood.

In this study, a novel G. hirsutum‐G. australe alien chromosome translocation line resistant to Verticillium wilt were identified and employed to narrow the search for disease resistance genes in G. australe for the first time. Our results further indicate that RVE2 plays a new important role in regulating resistance to VW via JA‐signalling pathways, which elucidate the molecular mechanisms underlying the resistance to VW and provide new insights into the JA‐signalling pathway.

Results

TA01‐7G is a novel G. hirsutum‐G. australe translocation line resistant to Verticillium wilt

To narrow the search for disease resistance genes in G. australe. G. hirsutum‐G. australe translocation was developed (Wang et al., 2018). In V. dahliae pathogenicity assays, one line (TA01‐7G) exhibited enhanced resistance (Figure 1a). Its alien segment was from Chr. 7G flanking the markers between AGA7R2703 and AGA7R7454 in G. australe (Figure S1a‐l). The genomic in situ hybridization (GISH) assay and transcriptome sequencing confirmed that the translocated chromosome in G. hirsutum was Chr. A01 (Figures S1m,n and S2a).

Figure 1.

TA01‐7G cotton exhibits resistance to Verticillium wilt. (a) Verticillium wilt symptoms after inoculation with strain V991 in TA01‐7G, CK (genetic background control), Hai7124 and Junmian1 cotton plants as indicated time points. Sterile water treatment was used as mock. G. barbadense cv. Hai7124 and G. hirsutum cv. Junmian1 were used as resistant and susceptible control, respectively. The V. dahliae infection assays were repeated for three times with similar results. (b) Disease index at 10 dpi or 20 dpi (n = 3, at least 15 seedlings in each biological repeat). dpi, days post‐inoculation. (c) Vascular discoloration after inoculation with V991 in TA01‐7G, CK, Hai7124 and Junmian1 plants at indicated time points or Mock. Scale bar, 2 mm. (d) Fungal recovery experiments of stems in TA01‐7G, CK, Hai7124 and Junmian1 plants 4 days after recovery. (e) Relative fungal biomass in roots, stems and leaves at 20 dpi (n = 3). qRT‐PCR was performed to quantify V. dahliae DNA using a V. dahliae‐specific primer pair (ITS1‐F and STVe1‐R). Results are normalized to GhHiston3. All values are given in means ± s.d. The P‐values indicate the results from pairwise comparisons of one‐way ANOVA tests. **P < 0.01.

To evaluate the potential contribution of alien chromosome segment to disease resistance, the translocation line TA01‐7G and genetic background cotton (CK, isolated from TA01‐7G) were inoculated with V991. Compare with CK, TA01‐7G enhanced the resistance to Vd. The mild disease symptoms appeared in leaves of TA01‐7G and Hai7124 (Figure 1a). The disease index (DI) was significantly lower in TA01‐7G and Hai7124 plants than in CK and Junmian1 plants, respectively (Figure 1b). The resistance of TA01‐7G to V. dahliae was further corroborated by observing fungal accumulation in brown discoloration stems (Figure 1c), fungal recovery assays (Figure 1d) and quantification of the fungal biomass in roots, stems and leaves with qRT‐PCR (Figure 1e). These experiments confirmed enhanced resistance in TA01‐7G.

TA01‐7G alien chromosome‐mediated transcriptional reprogramming in roots during V. dahliae infection

To understand the molecular mechanism of resistance to Vd in TA01‐7G, transcriptome sequencing (RNA‐seq) was performed. It was observed that the number of uniquely mapped reads and distribution of extent of gene expression levels in TA01‐7G were indistinguishable from the profile of CK (Figure S2b,c and Table S1). The overlap analysis of differentially expressed genes (DEGs) indicated that TA01‐7G and CK were induced to express 1622 and 2122 genotype‐specific genes, respectively (Figure S2d‐h).

To investigate the molecular consequences of the alien segment in TA01‐7G, total of 4421 DEGs (|log₂FC| ≥ 1, FDR ≤ 0.05) were identified (Figure S3a and Data Set S1). Out of these, 450 DEGs from G. australe were potentially affecting the disease resistance phenotype (Data Set S2). Gene ontology (GO) enrichment analysis indicated that DEGs in TA01‐7G were largely associated with response to jasmonic acid (JA), JA stimulus, JA mediated signalling pathway and defence response (Figure S3b and Table S2), highlighting that regulation of resistance in TA01‐7G was closely associated with jasmonate signalling pathway. Co‐expression cluster analysis based on disease progression and genotype in DEGs revealed five unsupervised groups of transcripts (Figure S4a,b). Kyoto Encyclopedia of Genes and Genomes (KEGG) terms analysis revealed that the DEGs in TA01‐7G were enriched for annotations related to phenylpropanoid biosynthesis and alpha‐linolenic acid metabolism pathway (Figure S5). Gene set enrichment analysis (GSEA) among the DEGs revealed enrichment for annotated biological functions related to immune response, defence response to fungus and bacterium (Figure S4c,d). Taken together, these results indicated that alien genetic components of TA01‐7G played a dominant role in Vd‐induced root transcriptome programming.

The weighted‐gene co‐expression network analysis (WGCNA) divided the DEGs into 12 gene modules (Figure S6a–f). Subsequently, association analysis of gene modules with pathogen, disease progression and genotype indicated that ‘magenta’ (Pearson correlation r = 0.8, P = 0.0002) and ‘red’ (r = 0.71, P = 0.002) modules were with the highest positive correlation in TA01‐7G at 72 hpi (Figure S3c,d, Figure S6g‐j). Combining the results of co‐expression clustering analysis, ‘magenta’ and ‘red’ modules were selected as targets for selecting the candidate genes (Figures S4a and S6k,l and Data Set S3). Target modules were mainly involved in response to JA, JA‐mediated signal pathway, MAPK cascade and JA biosynthetic process (Nominal P value <0.01, FDR q‐value <0.005) (Table S3). Totally, 12 hub genes, Gaus01 to Gaus12, derived from G. australe, were significantly up‐regulated in TA01‐7G by pathogen, which were as resistant candidate genes for undertaking further functional analysis (Figure S3e,f and Table S4).

GausRVE2 is essentially required for conferring JA‐mediated disease resistance in TA01‐7G

RNA‐seq analysis demonstrated that TA01‐7G‐mediated disease resistance was closely associated with the jasmonate signalling pathway. Transcript levels of JA biosynthesis‐related genes (LOX1, LOX2, LOX3, AOS and AOC1) were significantly higher in TA01‐7G than in CK (Figure S7a). The expression of JA signalling‐related genes including MYC2, ERF1 and ERF2 was also significantly high in TA01‐7G compared to CK but no significant differences were observed for genes, COI1, JAZ6 and JAZ10 (Figure S7b). It is well known that COI1 acts as the JA receptor (Xie et al., 1998), JAZ proteins act as repressors in regulating JA signalling (Pauwels et al., 2010), MYC2 is a master regulator of JA signalling (Kazan and Manners, 2013), ERF1 and ERF2 are transcription factors that regulate JA‐mediated defence response branches for pathogens (Lorenzo et al., 2003). To investigate whether JA‐mediated disease resistance response was activated in TA01‐7G, the expression level of PDF1.2 (Spoel et al., 2003), a marker of JA‐regulated pathogen response and pathogenesis‐related (PR) genes, were monitored. The transcript levels of PDF1.2 were significantly higher in roots after inoculation with V991 in TA01‐7G than in CK at the corresponding time points (Figure S7c). The transcript levels of the PR genes PR1, PR4, PR5, PR6 and PR10 were higher in TA01‐7G than in CK plants. And no significant differences were found for PR2 and PR3 (Figure S7d).

To determine whether the candidate genes discovered in RNA‐seq analysis were involved in regulating disease resistance, the tobacco‐rattle‐virus (TRV)‐based virus‐induced gene silencing (VIGS) strategy was deployed to knockdown the candidate genes in TA01‐7G. Two weeks after VIGS treatment, the decreased expression levels shown by all candidate genes were confirmed through qRT‐PCR analysis in VIGS treated plants (Figure S7e). Subsequently, the VIGS‐silenced plants were exposed to V. dahliae. Ten days later, Gaus10‐ and Gaus12‐silenced plants exhibited reduced disease resistance with symptoms of wilting and etiolated leaves as compared to TRV:00 plants (Figure S8a). The DI and the fungal biomass in roots were remarkably high in Gaus10‐ and Gaus12‐silenced plants than in TRV:00 plants (Figure S8b,c). These experiments suggested a possible role of Gaus10 and Gaus12 in TA01‐7G defence against Vd. The transcript levels of JA biosynthesis‐related genes were impaired in the Gaus10‐silenced plants as compared to TRV:00 plants, but not in Gaus12‐silenced plants (Figures S7f and S8d). Moreover, knockdown of Gaus10 significantly altered the transcript levels of JA signalling‐related genes, PR genes and JA‐responsive genes (Figure S8e–g), indicating a critical role of Gaus10 in the JA‐mediated resistance to VW. Gaus10‐silenced plants further showed the most sensitive than others. Subsequently, Gaus10 was determined as the excellent candidate resistant gene and named as GausRVE2 (REVILLE2, EPI10_012925) by protein sequence alignment with the G. australe genome (Cai et al., 2020). Taken together, these results indicated that GausRVE2 may be required for conferring JA‐mediated disease resistance in TA01‐7G.

GausRVE2 confers JA‐mediated disease resistance in Arabidopsis

To further investigate whether GausRVE2 contributes towards resistance to V. dahliae, a total of 90 overexpression lines of Arabidopsis (OE1 to OE90) were produced. Out of these, five overexpression lines (OE97, OE23, OE105, OE10 and OE32) with varying expression levels were selected (Figure S9a,g). In infection assays, three independent GausRVE2‐OE transgenic lines (OE97, OE23 and OE105) exhibited enhanced resistance to V. dahliae than that of the wild‐type (WT) (Figure S9b), including more healthy‐rosette leaves, less fungal biomass and lower DI (Figure S9c–e), which were corroborated by quantification of fungal biomass in root, stem and leaf by qRT‐PCR (Figure S9f). Two GausRVE2‐OE lines (OE97 and OE105) showing very strong resistance were chosen for further analysis. As was the case for VIGS, GausRVE2‐OE lines evidently activated JA response as compared to WT (Figure S9h–j). The transcript levels of PR1, PR4 and PR5 genes were greatly increased in GausRVE2‐OE lines (Figure S9k), and also PDF1.2 expression level was elevated during the V. dahliae inoculation (Figure S9l). These results indicated that overexpression of GausRVE2 might confer JA‐mediated disease resistance in Arabidopsis.

To further confirm the JA signalling was crucial for GausRVE2‐mediated resistance to Vd, the disease resistances of GausRVE2 overexpression lines in JA biosynthesis (aos) and signalling deficient mutant background (coi1 and myc2) was investigated (Figure 2a,b). The qRT‐PCR assays then confirmed that GausRVE2 expressed in these transgenic Arabidopsis (Figure 2c). Inoculation experiments showed that the GausRVE2‐overexpressed lines in aos, coi1 and myc2 background were more susceptible to Vd than that in Col‐0 background (Figure 2a), which was further corroborated by the fungal biomass analysis and the DI (Figure 2d). Taken together, these results suggest that GausRVE2‐regulating Vd resistance was closely related to JA signalling pathway.

Figure 2.

Jasmonic acid (JA) signalling was crucial for GausRVE2‐mediated resistance to Vd. (a) Disease symptom of transgenic GausRVE2 Arabidopsis inoculation with Vd for 14 d in JA biosynthesis (aos) and signalling deficient mutant background (coi1 and myc2). (b) The expression levels of AOS, COI1 and MYC2 in mutants and Col‐0 Arabidopsis. The roots of Arabidopsis were harvested for RNA isolation. Results were normalized by the internal control gene actin (n = 3). (c) Relative expression levels of GausRVE2 in transgenic Arabidopsis and Col‐0. Results were normalized by the internal control gene actin (n = 3). (d) Quantification of Vd biomass in roots of transgenic Arabidopsis at 20 dpi (n = 3). qRT‐PCR was performed to quantify Vd DNA using a Vd‐specific primer pair (ITS1‐F and STVe1‐R). The disease index of VIGS plants after V991 infection at 20 dpi. (e) GausRVE2‐OE tobacco enhanced resistance to Vd infection. Four‐week‐old tobacco seedlings were inoculated Vd using the root‐dipping method. The images were taken at 14 dpi. (f) Relative expression levels of GausRVE2 in transgenic tobacco and wild‐type (WT). Results were normalized by the internal control gene actin (n = 3). (g) Quantification of Vd biomass in roots of transgenic tobacco and WT plants at 20 dpi (n = 3). The disease index of transgenic tobacco and WT plants at 20 dpi. (h) Expression of AOS and MYC2 in VIGS and control plants. (i) Disease symptoms of TRV:00, AOS‐ and MYC2‐silenced plants in GausRVE2‐OE6 transgenic tobacco. After 2 weeks of VIGS, the indicated VIGS‐silenced plants were treated with 200 μM MeJA or sterile water. (j) Quantification of Vd biomass on roots of VIGS tobacco at 20 dpi (n = 3). The disease index of VIGS tobacco at 20 dpi. The data are shown as the means ± s.e. from three independent repeats. The P‐values indicate the results from pairwise comparisons of one‐way ANOVA tests. *P < 0.05, **P < 0.01, n.s., not significant. The above experiments were repeated at least three times with similar results.

JA signalling has a vital role for GausRVE2‐mediated disease resistance in plants

To determine whether jasmonate signalling was required for disease resistance to Vd in TA01‐7G, the VIGS strategy was employed to verify this. The lipoxygenase gene (LOX) (Wasternack and Song, 2017) and the master transcription factor MYC2 (Kazan and Manners, 2013) were silenced for attenuating JA biosynthesis and JA signal transduction in TA01‐7G, respectively (Figure S10a,b and File S1). The CK plants was also treated by VIGS as a genotype control. And, the VIGS (TRV:00) treatment did not change the disease resistance of TA01‐7G and CK (Figure 11Sa). LOX‐silenced and MYC2‐silenced plants of TA01‐7G exhibited severe wilting with etiolated leaves as compared to TRV:00 plants (Figure S11a,b). Moreover, exogenous application of MeJA alleviated the disease symptoms shown in the LOX‐silenced plants, but not in the MYC2‐silenced plants (Figures S10d and S11a,b), indicating a vital role of JA signalling in resistance to Vd in TA01‐7G. The response of resistance in CK knockdown‐plants remained consistent after the application of MeJA. The JA‐mediated disease resistance was required for TA01‐7G, which was corroborated by the observing fungal accumulation in brown discoloration stems (Figure S11c), fungal recovery assays (Figure S11d), the DI (Figure S11e) and quantification of fungal biomass in roots, stems and leaves using qRT‐PCR (Figure S11f). When the transcript levels of LOX and MYC2 were significantly reduced in VIGS‐plants (Figure S10e), the contents of JA, JA‐Ile and OPDA were also concomitantly decreased in LOX‐silenced plants but not in MYC2‐silenced plants as compared in TRV:00 plants (Figure S11g), suggesting that the VIGS assay was applied successfully in TA01‐7G and CK plants.

To investigate whether GausRVE2‐mediated resistance is dependent on jasmonate signalling, N. benthamiana containing knockdown of JA biosynthesis and signalling transduction pathways was used to carryout transient overexpression assays. When the transcript levels of AOS and MYC2 were substantially reduced in VIGS‐tobacco, GausRVE2 were transiently expressed in the newly developed leaves of AOS‐silenced or MYC2‐silenced (Figure S11h,i), which exhibited higher disease sensitivity than TRV:00 tobacco leaves even though they had the similar transcript of GausRVE2 in AOS‐silenced and MYC2‐silenced leaves (Figure S11h,j). The lesion area was substantially reduced after the exogenous application of MeJA in LOX‐silenced leaves but not in MYC2‐silenced leaves (Figure S11k). Subsequently, the potential role of JA pathway in GausRVE2‐mediated resistance to Vd was examined in transgenic tobacco lines (Figure 2e). Inoculation experiments showed that overexpression of GausRVE2 enhanced resistance to Vd in tobacco (Figure 2e–g). Moreover, compared with control, AOS‐silenced and MYC2‐silenced plants exhibited severe disease symptoms in transgenic GausRVE2 tobacco background (Figure 2h,i). These results indicated that JA pathway was involved in GausRVE2‐mediated disease resistance, which was further corroborated by the fungal biomass analysis and DI (Figure 2j). And, exogenous MeJA treatment only slightly increased disease resistance in the GausRVE2‐silenced plants (Figure 3a–c), indicating that disease resistance in TA01‐7G is partially due to GausRVE2‐mediated activation of JA defence response. To further explore whether JA pathway contributed to GausRVE2‐mediated Vd resistance in cotton, we first generated transgenic cotton plants to examine the potential role of GausRVE2 in Vd resistance (Figure 3d,f). Inoculation experiments showed that overexpression of GausRVE2 enhanced resistance to Vd in cotton (Figure 3d,h). And, LOX‐silenced and MYC2‐silenced cotton exhibited increased susceptibility to Vd in transgenic GausRVE2 cotton background with higher fungal biomass and DI than the VIGS control (Figure 3e,g,i). Taken together, these results suggest that JA signalling plays a vital role in GausRVE2‐mediated resistance to Vd.

Figure 3.

Jasmonic acid (JA) pathway was involved to GausRVE2‐mediated Vd resistance in cotton. (a) Disease symptoms of TRV:00 and GausRVE2‐silenced plants in TA01‐7G. After 2 weeks of VIGS, the indicated VIGS‐silenced plants were treated with 200 μM MeJA or sterile water. (b) Expression of GausRVE2 in VIGS and control plants. (c) Quantification of Vd biomass on roots of VIGS and control cotton at 20 dpi (n = 3). The disease index of VIGS and control cotton at 20 dpi. (d) GausRVE2‐OE transgenic cotton enhanced resistance to Vd infection. Two‐week‐old cotton seedlings were inoculated with Vd using the root‐dipping method. The images were taken at 15 dpi. (e) Disease symptoms of TRV:00, LOX‐ and MYC2‐silenced plants in GausRVE2‐OE21 transgenic cotton. (f) Relative expression levels of GausRVE2 in transgenic cotton and WT. Results were normalized by the internal control gene GhHiston3 (n = 3). (g) Expression of LOX and MYC2 in VIGS and control plants. (h) Quantification of Vd biomass in roots of transgenic cotton and WT plants at 20 dpi (n = 3). The disease index of transgenic cotton and WT plants at 20 dpi. (i) Quantification of Vd biomass in roots of VIGS cotton at 20 dpi (n = 3). The disease index of VIGS cotton at 20 dpi. The data are shown as the means ± s.e. from three independent repeats. The P‐values indicate the results from pairwise comparisons of one‐way ANOVA tests. *P < 0.05, **P < 0.01, n.s., not significant. The above experiments were repeated at least three times with similar results.

GausRVE2 participates in JA‐mediated disease resistance by attenuating the binding between GhNINJA and GhTPL or GhTPR1

It is well known that JAZ proteins act as repressor in JA signal pathway (Chini et al., 2007). The JAZ proteins recruit the corepressors TPL and TPRs through interacting with NINJA for preventing MYC2 from regulating the downstream JA‐responsive genes (Kazan and Manners, 2013; Pauwels et al., 2010). Recalling our finding that GausRVE2‐mediated disease resistance is closely associated with the JA signalling (Figures 2 and 3), compelled us to unravel the role of GausRVE2 in JA signalling. The physical interaction of RVE2 proteins with TPL or TPRs in Arabidopsis (Causier et al., 2012) provoked that GausRVE2 could be directly interacted with GhTPL or GhTPRs in cotton. As expected, the interactions of GausRVE2‐GhTPL and GausRVE2‐GhTPRs (GhTPR1, 2 and 3) were confirmed by yeast two‐hybrid (Y2H) assays (Figure S12a,c), bimolecular fluorescence complementation (BiFC) assays (Figures S13a and S14a) and split‐LUC assay Figure S13b,c). And, the GST‐pull down assays further confirmed that GausRVE2 directly interacted with TPL and TPR1 in vitro (Figure 4a). To validate the interaction of GausRVE2 with TPL and TPR1 in vivo, cotton protoplasts were co‐transfected with the constructs encoding GFP‐GausRVE2 and Flag‐TPL, ‐TPR1 or Flag‐GUS (control) in co‐immunoprecipitation (Co‐IP) assays. The Co‐IP results showed that Flag‐TPL and Flag‐TPR1 interacted with GausRVE2 in planta (Figure 4b). Subsequently, the pairwise protein interactions between GhNINJA, GhTPL and GhTPRs were verified by Y2H assays (Figure S12b,d), Split‐LUC (Figures S13e,f and S14b,c), Pull‐down and Co‐IP assays (Figure 4c–f). BiFC assays demonstrated that the interaction also occurred in the nucleus (Figure S13d), which drive us to examine whether GausRVE2 affects the protein interactions of GhNINJA‐GhTPL and GhNINJA‐GhTPRs. In split‐LUC assays, the LUC activity of GhNINJA‐GhTPL and GhNINJA‐GhTPR1, but no of GhNINJA‐GhTPR2 and GhNINJA‐GhTPR3, was attenuated when GausRVE2‐Myc was added in the split‐LUC system (Figure S13g). The reconstituted LUC activity was detected approximately half greater than empty vector controls (cLUC‐Myc) (Figure S13h), indicating that the protein interactions of GhNINJA‐GhTPL and GhNINJA‐GhTPR1 were attenuated by GausRVE2 in plant. To further confirm this phenomenon of protein interaction interference, we used Co‐IP assays in vivo and Pull‐down assays in vitro for verification. The results showed that the amount of NINJA enriched by GST‐TPL/TPR1 was significantly reduced when GausRVE2 protein was added, indicating that GausRVE2 could interfere with the formation of NINJA and TPL/TPR1 complex in vitro. In Co‐IP assays, the amount of TPL/TPR1 protein deposited by NINJA was significantly decreased, indicating that this interference phenomenon still existed in vivo (Figure 4c–f). GausRVE2 attenuates physical interaction of GhNINJA‐GhTPL and GhNINJA‐GhTPR1 probably liberates a partially suppressed JA response.

Figure 4.

TPL/TPR1 interacts with GausRVE2 and NINJA, and GausRVE2 attenuates the interaction between NINJA and TPL/TPR1. (a) Pull‐down assays for analysis of the interactions between GausRVE2 and TPL or TPR1. Recombinant GST‐GausRVE2 protein immobilized on anti‐GST beads was incubated with an equal amount of MBP (control), MBP‐TPL or MBP‐TPR1 protein separately, and probed with anti‐GST and anti‐MBP antibodies for immunoblot analysis. (b) TPL and TPR1 both interacted with GausRVE2 in co‐immunoprecipitation (Co‐IP) assay. Flag‐TPL or Flag‐TPR1 were co‐transfected with GFP‐GausRVE2 into cotton protoplasts. Total proteins were extracted with anti‐GFP beads for Co‐IP assays, and the proteins were analysed by immunoblotting with anti‐GFP, or anti‐Flag antibodies. Protein loading is indicated by Ponceau S staining of RBC. (c) The GausRVE2 reduces NINJA association with TPL in vitro GST pull‐down assays. Recombinant GST‐TPL protein immobilized on anti‐GST beads was incubated with an equal amount of MBP (control) or protein combination of MBPGausRVE2 and His‐NINJA, and probed with anti‐GST and anti‐MBP antibodies for immunoblot analysis. (d) The GausRVE2 reduces NINJA association with TPR1 in vitro GST pull‐down assays. The pull‐down assays were performed as described in (c). (e) The GausRVE2 interferes NINJA association with TPL in vivo by co‐immunoprecipitation (Co‐IP) assays. Flag‐GUS or Flag‐GausRVE2 were co‐transfected with the construct combination of GFP‐NINJA and Flag‐TPL into cotton protoplasts. Total proteins were extracted with anti‐GFP beads for Co‐IP assays, and the proteins were analysed by immunoblotting with anti‐GFP, or anti‐Flag antibodies. Protein loading is indicated by Ponceau S staining of RBC. (f) The GausRVE2 interferes NINJA association with TPR1 in vivo by Co‐IP assays. The Co‐IP assays were performed as described in (e). The above experiments were performed three times with similar results.

GhMYC2 enhances the transcription of cotton RVE2 by binding to their promoters

In Arabidopsis, MYC2 is a master regulator of JA signalling pathway (Kazan and Manners, 2013) especially in triggering the defence response. Here, overexpressing GausRVE2 increased the resistance in plants and altered the transcription of genes involved in JA‐mediated defence pathway (Figure 2 and Figure S9). Interestingly, three G‐box were found in GausRVE2 promoter, which is the binding sites for MYC2 transcription factor. Thus, does the MYC2 regulate the transcription of GausRVE2? This hypothesis was tested by the cotton RVE2 promoter activity when inoculating with pathogens (V991 or Vd8, two highly aggressive defoliating V. dahliae strains) and or chemicals (MeJA or H₂O₂) in tobacco. When tobacco leaf cells harbouring the cotton RVE2pro‐LUC (GausRVE2pro‐, GhRVE2proA/D‐ or GbRVE2proA/D‐LUC) reporter construct were treated with V991, Vd8 or MeJA, the promoter activity was significantly increased compared with the Mock treatment but except the treated with H₂O₂ (Figure S15a,b). The β‐glucuronidase (GUS) staining signals were detected and GUS enzyme activity was significantly increased in tobacco treated with V991, Vd8 or MeJA compared with the Mock treatment (Figure S15c,d). Together, these results suggested that the cotton RVE2 promoter activity was activated by V991, Vd8 or MeJA.

Next, to explore whether the GhMYC2 directly activates the cotton RVE2 promoter, a well‐established dual‐LUC system assay was performed. Co‐expression of cotton RVE2pro‐LUC reporter with the effectors GhMYC2 in tobacco led to substantially increase the firefly fluorescence signal (Figure 5a,b) and quantification of LUC enzyme activity assay substantiated that MYC2 activated the cotton RVE2 promoter activity (Figure 5c). Consistent with the dual‐LUC assay, the GhMYC2 activation of cotton RVE2 promoter activity was also confirmed by native promoter driven‐NLS‐GFP assay (Figure 5d–f). To further investigate whether MYC2 plays a critical role in V. dahliae‐induced activation of GausRVE2 promoter activity, the MYC2‐silenced tobacco leaves were used in dual‐LUC assay (Figure S15g,h). When MYC2‐silenced tobacco leaves harbouring GausRVE2pro‐LUC reporter were induced with V. dahliae, the promoter activities of GausRVE2 markedly attenuated compared with TRV:00 leaves (Figure S15g,i). These results elucidated that V. dahliae‐induced activation of cotton RVE2 promoter was positively regulated by MYC2.

Figure 5.

MYC2 directly binds to the promoters of cotton RVE2 to enhance their transcription. (a) Schematic diagrams of the effector and reporter were used in Dual‐LUC assays of (b). GhMYC2A/D, GhMYC2A and GhMYC2D; REN, Renilla luciferase; LUC, firefly luciferase. (b) Dual‐LUC assays in N. benthamiana leaves showing the activation of cotton RVE2 promoters by MYC2. The empty vector (35S:GUS) was used as negative control. Scale bar, 2 cm. (c) Activation ratio of cotton RVE2 promoters by MYC2 in N. benthamiana cells (n = 9). REN was used as an internal control. REN, Renilla luciferase; LUC, firefly luciferase. These experiments were repeated independently three times with similar results. (d) Schematic diagrams of the effector and reporter were used in native promoter‐driven NLS‐GFP assays of (e). SV40NLS, nuclear localization signal of simian virus 40. (e) Native promoter‐driven NLS‐GFP assays to detect the activation of cotton RVE2 promoters by MYC2 in N. benthamiana leaves. Colocalization of GFP and nuclei was determined by DAPI staining. The empty vector (35S:GUS) was used as negative control. Scale bars, 50 μm. (f) Relative optical density was detected in native promoter‐driven NLS‐GFP assays, and quantified by ImageJ software (n = 10). DAPI was used as an internal control. These experiments were repeated independently three times with similar results. All values are given in means ± s.d. The P‐values indicate the results from pairwise comparisons of one‐way ANOVA tests. **P < 0.01.

To explore whether MYC2 binds to the promoter of RVE2 in vivo, we performed chromatin immunoprecipitation followed by qPCR (ChIP‐qPCR) assays with three pairs of primers amplifying different regions (P1, P2 and P3) of RVE2 promoters in cotton protoplasts expressing MYC2‐GFP (Figure 6a). The ChIP‐qPCR assays revealed that MYC2 strongly bound to the P3 region in promoters of RVE2 (Figure 6b). Next, we further tested whether MYC2 directly binds to the promoter of RVE2 in vitro by electrophoretic mobility shift assay (EMSA) with the MBP‐MYC2 recombination protein to assay the binding affinity of MYC2 to the promoter of RVE2 (Figure 6c,d and Figure S24a). Notably, EMSA results showed that the retardation of DNA migration was observed for RVE2pro ^G‐box probes, and the binding could be inhibited by the competitive probes but not mutant probes (Figure 6c,d). Taken together, these results suggested that MYC2 could bind to the G‐box in the promoter of RVE2 genes in vitro and in vivo to regulate its expression.

Figure 6.

MYC2 directly binding to G‐box of cotton RVE2 promoters in vivo and in vitro. (a) Immunoblot analysis of transfected MYC2‐GFP cotton protoplasts (TA01‐7G and Hai7124). Proteins were analysed by immunoblotting with anti‐GFP antibodies. Protein loading is indicated by Ponceau S staining of RBC. (b) Binding of MYC2‐GFP to the G‐box motif of cotton RVE2 promoters was confirmed by ChIP‐qPCR. IgG was used as the negative control. The ChIP signals were normalized to input. (c) Schematic diagram of cotton RVE2 promoter G‐box sequence. The consensus of the motif was analysed by WebLogo3. (d) MYC2 directly binds to the G‐box motif of cotton RVE2 promoters in the EMSA assays. MBP‐MYC2 were expressed in E. coli, purified using amylose resin. Biotin‐labelled DNA probes were incubated alone or with purified recombinant MBP‐MYC2. The 125‐fold excess of the corresponding unlabelled probes was used as competitor. (e) Schematic diagram of G‐box motif in RVE2 promoters and their mutants. The DNA probes are shown in promoter and are labelled as GausRVE2pro‐P1, GausRVE2pro‐P2 and GausRVE2pro‐P3. The G‐box motifs of DNA probes mutated to adenine in blue (a) are labelled as GausRVE2pro‐P1mut, GausRVE2pro‐P2mut and GausRVE2pro‐P3mut. (f, g) GhRVE2D does not affect the binding affinity of GhMYC2 with G‐box motifs in GausRVE2 and GhRVE2A promoters. The experiments were performed as in (d). The data are shown as the means ± s.d. from three independent repeats. The P‐values indicate the results from pairwise comparisons of one‐way ANOVA tests. *P < 0.05, **P < 0.01, n.s., not significant. The above experiments were repeated at least three times with similar results.

The integrity of RVE2 is vital in mediating disease resistance in N. benthamiana

Totally, 54 SNPs were identified in RVE2 gene sequences from G. australe, G. hirsutum (TM‐1) and G. barbadense (Hai7124) (Data Set S4). Among them, 13 nonsynonymous SNPs were identified in G. australe, suggesting the extent of diversity present in RVE2 across the cotton species (Figure S16a). The GhRVE2D genes carried 1‐bp insertion at 1970‐bp downstream of the start codon (ATG), which caused frameshifts and early termination of translation after Amino acid 326 (Figure S16b). In addition, subcellular localization assay confirmed that GhRVE2D, the truncated variation, causing short tails of RVE2, fails to localize in cell nucleus (Figure S16c). All RVE2 genes were transcriptional upregulated following V. dahliae inoculation, but the truncated‐GhRVE2D exhibited delayed response to pathogen induction (Figure S16d). Furthermore, transcriptional upregulation of truncated‐GhRVE2D attenuated the corresponding At‐subgenome alleles to respond to pathogen stimulation. These studies have shown the involvement of GhRVE2D in regulating the transcription of GhRVE2A (Figure S16d). But no such interference phenomenon was found between GhRVE2D and GausRVE2 in TA01‐7G during V. dahliae inoculation (Figure S16d).

The effect of truncated GausRVE2 protein upon its mediation in disease resistance was explored (Figure S17a). GausRVE2‐N was unable to localize in nucleus, being the same as GhRVE2D. GausRVE2‐C showed the same as the empty vector (EV) (Figure S17b). Subsequently, tobacco leaves transiently overexpressing truncated‐GausRVE2 were inoculated with V. dahliae. The GausRVE2‐N and GausRVE2‐C tobacco leaves exhibited higher sensitivity to Vd than the GausRVE2‐overexpressed tobacco leaves (Figure S17c) even though they showed similar transcript levels. The loss of disease resistance in truncated‐GausRVE2, was corroborated by lesion area and the relative luminescence of V991‐GFP (Figure S17d). To examine whether the deletion mutant of RVE2 protein is responsible for the attenuated disease resistance, the disease phenotypes of GhRVE2D‐GausRVE2‐C, a gain‐of‐function variant of RVE2, was examined (Figure S17a). As expected, GhRVE2D‐GausRVE2‐C re‐localized to the nucleus and also enhanced the disease resistance in tobacco (Figure S17b,e–f). Together, these results confirmed that RVE2‐mediated disease resistance is dependent on the protein integrity.

Association of cotton RVE2 natural variants with disease resistance

The complexity of tetraploid cotton species and loss of resistance by the mutated REV2 gene, led us to think if sequence polymorphism in RVE2 genes across the species may alter the resistance to VW. To test this possibility, native cotton RVE2 protein was alone expressed in tobacco. It was observed that the only GhRVE2D‐expressed tobacco exhibited susceptibility to disease. These results were further confirmed through deploying pathogen growth assays (Figure S18a,b), indicated that only GhRVE2D lost its mediated disease resistance.

The transiently co‐expressed native cotton RVE2 proteins in tobacco showed fluctuated response towards disease resistance. For example, necrotic lesion symptoms clearly appeared when GhRVE2A was co‐expressed with GhRVE2D (Figure S18c,d), illustrating that GhRVE2D specifically suppressed the GhRVE2A‐mediated disease resistance in tobacco. Whereas, no symptoms appeared when GbRVE2A was co‐expressed with GbRVE2D (Figure S19a,b). The pathogenicity assay also revealed that susceptibility of GhRVE2A/D (GhRVE2A with GhRVE2D) to VW was suppressed when co‐expressed with either GausRVE2, GbRVE2A or GbRVE2D in tobacco (Figures S18e,f and S19c,d). It can be concluded that the RVE2‐mediated disease resistance in tobacco was affected by RVE2 interactions.

The comparison of RVE2 gene sequences was performed in 130 and 70 resequenced accessions of G. hirsutum and G. barbadense, respectively. The unique GhRVE2D sequence was identical to TM‐1, widely existed in all resequenced accessions of G. hirsutum (Table S5). However, three vital malfunctional mutants were identified among 70 accessions of G. barbadense (Table S6), indicating that RVE2 gene in G. hirsutum was highly conserved than that in G. barbadense. To evaluate the effect of natural variation in cotton RVE2 genes on disease resistance, three G. barbadense cultivars (Xinhai29, Xinhai12 and Xinhai21, a stop gain mutation in the third exon of GbRVE2D) were inoculated with V991. Pathogenicity assays revealed that three G. barbadense cultivar exhibited higher susceptibility than Hai7124 (Figure S20). GbRVE2D^Xinhai29, GbRVE2D^Xinhai12 and GbRVE2D^Xinhai21 were found to unable to localize into nucleus (Figure S21a,b). The three truncated‐GbRVE2D exhibited EV‐comparable necrotic lesion areas (Figure S21c,d). However, the GbRVE2A‐mediated disease resistance was not affected when co‐expressed with natural truncated GbRVE2D (Figure S21e,f). Taken together, it is concluded that natural variation in RVE2 genes of G. barbadense is associated with disease resistance to VW.

GhRVE2D attenuates the MYC2‐activated GhRVE2A promoter activity but not the GausRVE2 promoter

The transcript level of GhRVE2A was gradually decreased with the corresponding increase in accumulation of truncated‐GhRVE2D (Figure S16d), thus suppressing the resistance mediated by GhRVE2A (Figure S18c). The subcellular localization of GhRVE2D in nucleus was recorded just after 3 days of inoculation with V. dahliae but not in the nucleus of non‐inoculated cells (Figure S22). In dual‐LUC assays, GhRVE2A significantly activated itself promoter activity. Similar results were observed for GhRVE2Dpro‐LUC and GausRVE2pro‐LUC reporter vector (Figure S23a,b). The GhRVE2A/D and GausRVE2 activations of their own promoter activity was also confirmed by GUS reporter system assays (Figure S23c,d). Taken together, these results indicated that GhRVE2A, GhRVE2D and GausRVE2 were activated by their own promoters. Afterwards, we investigated whether RVE2 was functionally involved in JA pathway through its transcriptional activity. To this end, we cloned promoters of JA pathway genes (LOX2, JAZ6, PDF1.2 and PR4) to perform Dual‐LUC assays in cotton protoplasts. The Dual‐LUC assay results indicated that RVE2 could not activate the promoter activity of JA pathway genes (Figure S23e).

Next, a dual‐LUC system assay was performed to study the impact of GhRVE2D on MYC2‐mediated transactivation of GhRVE2A and GausRVE2. MYC2 alone was able to significantly activate the GhRVE2A and GausRVE2 promoters, and this induction was not obviously altered by MYC2 co‐expressing with GhRVE2A or GausRVE2 effector (Figure 7a–c). In GhRVE2Apro‐LUC reporter vector, co‐expression with effectors GhRVE2D significantly decreased the promoter activity of GhRVE2A in the presence MYC2 effector, demonstrating that GhRVE2D repressed the transcriptional activation activity of MYC2 (Figure 7a,b,d,e). However, when co‐expression of GausRVE2pro‐LUC reporter with the effectors GhRVE2D and MYC2, the promoter activity of GausRVE2 remained unchanged (Figure 7a,b). Consistent with the dual‐LUC assays result, GhRVE2D repressed the MYC2‐activated promoter activity of GhRVE2A was further confirmed by deploying GUS reporter system assay (Figure 7f,g). Subsequently, we further confirmed whether GhRVE2D suppressed the binding affinity of MYC2 to the promoters of GausRVE2 and GhRVE2A by the EMSA assays. Biotin‐labelled DNA probes were designed at P1‐P3 sites in the GausRVE2 and GhRVE2A promoters to ensure the integrity of the detection (Figure 6e). These results indicated that His‐GhRVE2D does not bind to these DNA probes, nor did it affect the binding affinity of MYC2 to these DNA probes (Figure 6f,g). The interference of GhRVE2D on the transcriptional activation of GhRVE2A may be due to other levels of regulation, such as directly affecting the activation ability of MYC2, and inhibiting the activation activity of MYC2 by recruiting transcriptional suppressor.

Figure 7.

GhRVE2D interferes with MYC2 activated the GhRVE2A promoter and lowered the GhRVE2A transcriptional activation but not the GausRVE2 promoter. (a) Dual‐LUC assays showing that GhRVE2D attenuates the activation of GhRVE2A promoter by MYC2 but not GausRVE2 promoter in N. benthamiana leaves. Schematic diagrams of the effector and reporter were used in the Dual‐LUC assays as indicated top. The empty vector (35S:GUS) was used as negative control. Scale bar, 2 cm. (b) Activation ratio of GausRVE2 and GhRVE2A promoters by MYC2 in N. benthamiana cells (n = 9). REN was used as an internal control. REN, Renilla luciferase; LUC, firefly luciferase. These experiments were repeated independently three times with similar results. (c) The protein expression level of MYC2 in the Dual‐LUC assays. Proteins were analysed by immunoblotting with anti‐GUS antibodies. Protein loading is indicated by Ponceau S staining of RBC. (d) The negative control of Dual‐LUC assays. (e) The LUC activity of co‐expression EV and RVE2pro‐LUC in tobacco leaves. (f) β‐glucuronidase (GUS) staining assays validated that GhRVE2D inhibits the activation of GhRVE2A promoters by MYC2 but not GausRVE2 promoter. Schematic diagrams of the effector and reporter were used in the GUS staining assays as indicated top. The empty vector (35S:GFP) was used as negative control. Scale bar, 2 cm. (g) GUS enzyme activity of GausRVE2 and GhRVE2A promoters by MYC2 (n = 9). These experiments were repeated independently three times with similar results. In (b, d) the box limits indicate the 25th and 75th percentiles, whiskers indicate the full range of the data, and the centre line indicates the median. Individual data points are plotted. n represents sample size. All values are given in means ± s.d. The P‐values indicate the results from pairwise comparisons of one‐way ANOVA tests. **P < 0.01, n.s., not significant.

Taken together, these results demonstrated that disease resistance suppressed by GhRVE2D was possibly due to its attenuating MYC2‐activated promoter activity of GhRVE2A. This could be one of the reasons why most accessions of G. barbadense exhibit higher resistance to V. dahliae than G. hirsutum.

Discussion

Wild relatives of several crop species were used as genetic resource for introgressing genes in cultivated crop species. Thus, studying the genetic mechanisms of underlying important traits in wild relatives is important for using in crop improvement programs. For unravelling the genetics of important traits, a complete set of alien chromosome addition lines, chromosome introgression and translocation lines was produced (Chen et al., 2014; Wang et al., 2018). These preliminary studies were supplemented with the cracking of the genome sequence of G. australe that can prove a landmark in utilizing this information in cotton breeding program (Cai et al., 2020). However, the molecular genetic insight of resistance to VW remains unknown in G. australe. In this study, we demonstrated that RVE2 participates in the JA pathway by fine‐tuning the JAZ‐MYC module in cotton. Our study revealed that RVE2 mediates the VW resistance in G. australe involved in the JA‐signalling pathway (Figure 8).

Figure 8.

Proposed working model of RVE2 in fine‐tuning jasmonic acid (JA) signalling to enhanced resistance to Vd. RVE2 disturbs JAZ recruiting TPL and TPR1 in NINJA‐dependent manner and relieves inhibited‐MYC2 activity. The liberated MYC2 binds to the promoters of RVE2 and activates its transcription, forming a feedback loop. This process results in the activation of JA signalling, leading to cotton resistance to Vd. The tRVE2 (naturally truncated‐GhRVE2D) represses the ability of the MYC2 to activate the GhRVE2A promoter. CERK1, Chitin Elicitor Receptor Kinase 1; LYK, LysM‐containing receptor‐like kinase.

RVE2 belongs to the Myb‐like family transcription factor, whose members are associated with the circadian clock in Arabidopsis (Alabadí et al., 2001; Carré and Kim, 2002; Farinas and Mas, 2011; Kuno et al., 2003; Rawat et al., 2009, 2011; Zhang et al., 2007). It is reported that RVE2 acts as part of a regulatory feedback loop of circadian outputs (Zhang et al., 2007) and restraining bioactive GA biosynthesis (Jiang et al., 2016). Here, we found that RVE2 is JA responsive and the direct target of MYC2 (Figures 6 and S16). The RVE2 disturbs JAZ recruiting TPL and TPR1 in NINJA‐dependent manner, relieves inhibited‐MYC2 activity, and thus activates the JA signalling pathway. Furthermore, the overexpression and down‐regulation of GausRVE2 activated and or repressed the JA‐mediated disease resistance (Figures S8 and S9). Thus, it is possible that RVE2 may have functions of participating in plant innate immunity via fine‐tuning of JA signalling pathway. Although we uncovered the molecular mechanisms underlying RVE2‐mediated JA signalling, it would also be worth investigating whether RVE2 integrates the circadian clock and JA signalling in the future.

Accumulating evidence revealed that JA regulates the host resistance against a diverse strain of fungi (Abuqamar et al., 2008; Kachroo and Kachroo, 2009; Stintzi et al., 2001; Thaler et al., 2004; Thomma et al., 1998; Vijayan et al., 1998). In JA signalling pathway, JAZ proteins recruiting the TPL/TPRs through NINJA play vital roles for negative regulation of jasmonate responses (Pauwels et al., 2010; Zhang et al., 2015). In the present study, it was found that GausRVE2 disturbs the NINJA recruiting TPL and TPR1, relieves a partially inhibited‐MYC2 activity (Figure 4). Moreover, the liberated MYC2 is able to activate the GausRVE2 promoter activity (Figure 5) and form a feedback loop. Furthermore, GausRVE2 is able to interfere with NINJA recruiting TPL/TPR1, but does not affect the NINJA interactions with TPR2 and TPR3, indicating that GausRVE2 may precisely regulate JA signalling and release the partially inhibited‐MYC2 activity. In TA01‐7G and GausRVE2‐OE lines, the expression level of PDF1.2 and PRs genes are pre‐activated before stimulation by pathogen, which partly explains the GausRVE2‐mediated disease resistance. Therefore, our study demonstrates that GausRVE2‐mediates VW resistance in cotton by fine‐tuning the JAZ‐MYC2 module in JA‐signalling pathway.

To increase disease resistance plants must reallocate more energy from growth to defence, which is often accompanied a fitness penalty (Ning et al., 2017). In this study, we investigated the lint percent and fibre length of TA01‐7G. The result indicated that the lint percent and fibre length of TA01‐7G was decreased compared to TM‐1 (Figure S24b). And, the natural truncated‐GhRVE2D attenuated the GhRVE2A to respond the pathogen infection (Figure S16d). In the GausRVE2‐OE Arabidopsis, two lines with low expression level exhibited sensitivity to V. dahliae (Figure S9). Furthermore, truncated‐GhRVE2D repressed the ability of MYC2 to activate the GhRVE2A promoter activity (Figure 7). It would also be worth investigating whether the expression level of RVE2 is related to trade‐off between growth and defence in cotton.

It is generally believed that the hyphopodia of V. dahliae forms penetration peg infect the xylem vessels by root tips in cotton plants. However, under laboratory conditions, V. dahliae was able to infect plant leaves (Duan et al., 2016; Gao et al., 2016; Munis et al., 2010; Tong et al., 2021; Yang et al., 2015). It is worth noting that leaf‐inoculation method circumvents the first line of defence mechanism against the infection by V. dahliae in root system. To evaluate whether the interactions of cotton RVE2 affect disease resistance, we used the leaf‐inoculation method as it is hard to obtain a variety of cotton RVE2 overexpressing plants. The overexpression of RVE2 in leaves exhibited enhanced resistance to V. dahliae, indicating that RVE2 could confer resistance to VW in plant leaves (Figure S18). Although we performed the qRT‐PCR analysis to study the transcript levels of overexpressed genes in leaves, but the stable multiple RVE2 gene overexpressing plants are still required for further study. Whether RVE2 involved in constructing the first line of defence against infection in plant root system and RVE2‐mediated disease resistance should be evaluated under natural field condition.

RVE2 gene in G. hirsutum was highly conserved than the genes in G. barbadense (Tables S5 and S6). The natural truncated‐GhRVE2D delayed the response to pathogen and attenuated the ability of MYC2 to activate the GhRVE2A promoter (Figure 7). These findings highlighted that GhRVE2D interferes with GhRVE2A‐mediated disease resistance in cotton. However, more short‐tailed GbRVE2D^{Xinhai29,12,21} could not interfere with GbRVE2A‐mediated disease resistance. It seems that cotton has evolved a precise mechanism to balance RVE2‐meidiated resistance to V. dahliae and limits the energy consumption in defence response.

In summary, GausRVE2 plays a pivotal role in fine‐tuning JA signalling and enhancing resistance to VW in cotton. This can be used as a candidate gene for engineering disease resistance in cotton and other crop species. Moreover, our study provides new insights into the roles of RVE2 in regulating JA signalling and biotic stress resistance in plants.

Experimental procedures

Plant materials and growth condition

Cotton varieties used in this study included G. hirsutum cv. TM‐1 and Junmian1 (susceptible cultivar), G. barbadense cv. Hai7124 (resistant cultivar), Xinhai29, Xinhai12 and Xinhai21 (commercial cultivars). The resistant translocation line G. australe, named ‘TA01‐7G’, was used in this study. A genetic background cotton, named ‘CK’, without alien chromosomal segment was used as a control for TA01‐7G. A. thaliana ecotype Columbia (Col‐0) and N. benthamiana were also used. The mutants myc2 (SALK_040500) (Hong et al., 2012), aos and coi1 (SALK_017756C, SALK_035548) (Thatcher et al., 2009) were obtained from the Arabidopsis Biological Resource Center.

Cotton plants were grown in a growth chamber under normal conditions. Procedures are provided in Supplemental Detailed Experimental Procedures.

Genomic in situ hybridization

The genomic in situ hybridization was conducted as described (Guan et al., 2008; Wang et al., 2018).

RNA‐seq library preparation and data analysis

Roots from inoculated cotton seedlings were collected at indicated time points, and the RNA (2 μg) was used for library construction with a NEBNext® Ultra™ Directional RNA Library Prep Kit for Illumina® (NEB, USA). Procedures are provided in Supplemental Detailed Experimental Procedures.

Constructs for genetic transformation

To overexpress GausRVE2 in Arabidopsis, a full‐length coding sequence was cloned into pBI121 vector digested by XbaI and SmaI using the ClonExpress Entry One Step Cloning Kit (C114‐01, Vazyme). The recombinant vector was transformed into the Agrobacterium tumefaciens strain GV3101, and then the transformed GV3101 cells were used to generate transgenic Arabidopsis plants by the floral dip method (Zhang et al., 2006). And the same construct was used for generating transgenic cotton and tobacco. The plants hypocotyls were used to generate explants and Agrobacterium‐mediated transformation was carried out as previously described (Qin et al., 2020). Homozygous T₃ and T₄ transgenic Arabidopsis plants were used in experiments. Primers used for generating these constructs are listed in Data Set S5.

RNA isolation and qRT‐PCR

The RNA samples had three biological replicates. A total of three genes representing cotton (GhHistone3, AF024716), Arabidopsis (AtUbq5, At3g62250) and tobacco (NbActin) were used as internal controls for normalization purpose. The qRT‐PCR analysis was performed by adopting a standard published protocol (Nolan et al., 2006). The primers used to determine the transcript levels are listed in Data Set S5. Procedures are provided in Supplemental Detailed Experimental Procedures.

VIGS assays

Fragments of candidate genes, Gaus01 to Gaus12, GhLOX1‐3 and GhMYC2 were amplified using the corresponding primers and then cloned into the pTRV2 vector (Clontech). The newly emerged leaves were collected for extracting total RNA followed by conducting the qRT‐PCR assay to validate the silencing efficiency. The plants successfully silenced were selected for subsequent V. dahliae infection assay and hormone content determination assays.

The VIGS assays were performed in tobacco by adopting a published protocol (Senthil‐Kumar and Mysore, 2014) with minor modifications. Primers used for generating these constructs are listed in Data Set S5. All VIGS assays were performed three times independently. Procedures are provided in Supplemental Detailed Experimental Procedures.

Verticillium dahliae inoculation assays

The highly aggressive defoliating V. dahliae strains V991 and Vd8 were incubated on potato dextrose agar (PDA) at 25 °C for 7 days, and then inoculated into potato dextrose broth (200 g/L potato, 12.5 g/L glucose, pH 7.0) on a shaker at 150 rpm at 25 °C for 3–4 days. To inoculate the cotton, seedlings were infected with V. dahliae spore suspension liquid (1 × 10⁷ spore mL⁻¹) by impaired‐root dip inoculation for 1 min, and then returned to their original pots. To inoculate the Arabidopsis, 4‐week‐old seedling roots were infected with V. dahliae spore suspension liquid (1 × 10⁶ spore mL⁻¹) by root‐dipping inoculation method. Control plants in every treatment were inoculated with an equal volume of sterile distilled water. Fungal accumulation was observed under stereoscope (Olympus MVX10). These V. dahliae infection assays were repeated at least three times (Gong et al., 2018). Procedures are provided in Supplemental Detailed Experimental Procedures.

Measurement and treatment of plant hormones and exposure to H₂O₂

For cotton, hormone content was determined as previously described with minor modifications (Cheng et al., 2021; Stitz et al., 2014). MeJA (200 μM, Sigma‐Aldrich, USA) and H₂O₂ (1 mM, Sigma‐Aldrich, USA) were used for seedlings treatment. Procedures are provided in Supplemental Detailed Experimental Procedures.

Transient expression and inoculation assays in N. benthamiana

For the transient expression assay, the coding region of cotton RVE2 (GausRVE2, GhRVE2A/D and GbRVE2A/D) were amplified from the cDNA, and then cloned into the pBInGFP4 vector (Clontech). The coding region of the GausRVE2 N‐terminal (GausRVE2‐N) and the GausRVE2 C‐terminal (GausRVE2‐C) were amplified and cloned into the pBInGFP4 vector by following the same procedure.

The inoculation assays in detached tobacco leaf with V. dahliae were performed by adopting a protocol with a few modifications (Duan et al., 2016; Gao et al., 2016; Munis et al., 2010; Tong et al., 2021; Yang et al., 2015). Procedures are provided in Supplemental Detailed Experimental Procedures.

Subcellular localization of cotton RVE2

The entire cotton RVE2 coding region without the stop codon was amplified and cloned into the KpnI and BamHI sites of binary vector pBInGFP4 vector (Clontech). The fusion cotton RVE2 and GFP were driven by the cauliflower mosaic virus 35S promoter. The GV3101 strains harbouring recombinant plasmids were transiently expressed in N. benthamiana leaf epidermal cells via A. tumefaciens infiltration. The green fluorescent protein signals were monitored with confocal laser microscopy (Leica TCS SP8, Germany) at 48 h after infiltration. Nuclei were stained with DAPI (Sigma‐Aldrich, USA). Primers used for generating these constructs are listed in Data Set S5. Transient expression and microscopy observation for protein localization signals were repeated for three times.

Native promoter driven‐NLS‐GFP assays

The native promoter driven‐NLS‐GFP assays was performed with little modifications in a previously described protocol (Tian et al., 2020). The pBInNLS‐GFP4 vector was constructed by fusing the nuclear localization signal of simian virus 40 (SV40NLS) with GFP. Primers used for generating these constructs are listed in Data Set S5. The native promoter driven‐NLS‐GFP assays were repeated for three times. Procedures are provided in Supplemental Detailed Experimental Procedures.

Protein–protein interaction assays

Yeast‐two‐hybrid (Y2H), bimolecular fluorescence complementation (BiFC), Pull‐down, Co‐IP and Split‐LUC assays were used to measure the interaction between special proteins. Procedures are provided in Supplemental Detailed Experimental Procedures.

Dual‐LUC assays

The Dual‐LUC assays were performed by adopting a protocol (Ma et al., 2018) with minor changes. The coding regions of GhMYC2A and GhRVE2D were amplified and cloned into the pBI121 vector (Clontech), which were used as effectors. The empty vector was used as a negative control for the effector. The promoter sequences of indicated cotton RVE2 were amplified from the genomic DNA and cloned into the pGreenII‐0800‐LUC vector (Clontech), which was used as a reporter. Primers used for generating these constructs are listed in Data Set S5. All experiments were repeated at least three times for each plasmid combination. Methods for EMSA assays and ChIP‐qPCR assays are provided in Supplemental Detailed Experimental Procedures.

Statistical analysis

GraphPad Prism 8.0.1 (GraphPad Software) or Excel (Microsoft, version 16) was used for all statistical analyses. The fluorescence intensity, DAPI and GFP were calculated using the image processing program ImageJ (version 1.53q). All source data for figures were listed in Data Set S6. ANOVA and t test results were in Data Set S7. Procedures are provided in Supplemental Detailed Experimental Procedures.

Accession numbers

The sequences of the genes cloned in this study are deposited in GenBank under the following accession numbers: GausRVE2 (OP208764), GhRVE2A (OP208765), GhRVE2D (OP208766), GbRVE2A (OP208767), GbRVE2D (OP208768), GausRVE2‐N (OP208770), GausRVE2‐C (OP208771), GhRVE2D‐GausRVE2‐C (OP208772), GbRVE2^{xinhai12,21,29} (OP208769), GhNINJA (OP208773), GhTPL (OP208774), GhTPR1 (OP208775), GhTPR2 (OP208776), GhTPR3 (OP208777), GhMPK3‐1 (OP208778), GhMPK3‐2 (OP208779), GhMPK3‐3 (OP208780), GhMPK3‐4 (OP208781), GhMPK6‐1 (OP208782), GhMPK6‐2 (OP208783), GhMPK6‐3 (OP208784), GhMPK6‐4 (OP208785), GhMYC2A (OP208786), GhMYC2D (OP208787).

Author contributions

B.L.Z and F.J.L conceived and designed the project. F.J.L and S.C performed the experiments. S.C, Z.F.M, H.R.Y, L.W, L.S.X, H.W, J.B.G, L.J.D, S.L.F, Y.Y.W and M.F.C assisted in completing the part of experiment. MUR criticized and edited the manuscript. B.L.Z supervised and revised the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Supporting information

Data Set S1 FPKM values for DEGs in CK and TA01‐7G after V. dahliae infection.

Data Set S2 RNA‐seq expression profiling of DEGs in TA01‐7G and CK.

Data Set S3 FPKM value of the target gene modules.

Data Set S4 Single nucleotide polymorphisms in the RVE2 gene among G. australe, TM‐1 and Hai7124.

Data Set S5 Names and sequences of primers used in all experiments of this study.

Data Set S6 Source data.

Data Set S7 ANOVA and t test results.

File S1 Alignment data used to generate the phylogenetic tree in Figure S9a.

Figure S1 Identification of TA01‐7G as a new G. hirsutum‐G. australe translocation line.

Figure S2 Comparative differential transcriptome analysis in TA01‐7G and CK.

Figure S3 TA01‐7G‐mediated transcriptional reprogramming in roots during V. dahliae infection.

Figure S4 Functional analysis of DEGs in comparative transcriptomes.

Figure S5 Kyoto Encyclopedia of Genes and Genomes analysis of DEGs detected in TA01‐7G.

Figure S6 Weighted correlation network analysis of DEGs detected in TA01‐7G and CK.

Figure S7 Relative expression level of genes involved in jasmonic acid (JA) signal pathway, pathogenesis‐related (PR) genes and virus‐induced gene silencing (VIGS) candidate genes.

Figure S8 Knockdown of RVE2 reduced JA‐mediated disease resistance in TA01‐7G.

Figure S9 GausRVE2 overexpression enhance JA‐mediated disease resistance in Arabidopsis.

Figure S10 VIGS strategy to interfere with cotton JA signalling pathway.

Figure S11 Disruption of JA signalling in TA01‐7G results in loss of GausRVE2‐mediated disease resistance.

Figure S12 GausRVE2 and GhNINJA interact with TPL and TPRs by yeast two‐hybrid (Y2H) assay.

Figure S13 The interaction between GausRVE2 and TPL/TPRs, and GausRVE2 interfered with the interaction between NINJA and TPL/TPR1.

Figure S14 The negative controls of GausRVE2 and TPL/TPRs interaction in BiFC and split‐LUC assays.

Figure S15 MYC2 is important for V991‐induced transcriptional activation of cotton RVE2 promoters.

Figure S16 Naturally truncated RVE2 protein loses nuclear localization and expression of cotton RVE2 genes in response to V. dahliae infection.

Figure S17. A non‐conserved amino acid fragment in the C terminus of RVE2 is required to mediate disease resistant in N. benthamiana.

Figure S18 Divergent disease resistance in transiently expressing various combinations of native cotton RVE2 proteins in N. benthamiana.

Figure S19 Divergent disease resistance in transiently expressing various combinations of RVE2 proteins.

Figure S20 Polymorphisms of RVE2 gene led to divergent disease resistance in cotton natural population.

Figure S21 Divergent disease resistance of naturally truncated RVE2 proteins in G. barbadense.

Figure S22 V991‐induced nuclear relocalization of GhRVE2D in tobacco epidermal cells.

Figure S23 RVE2 self‐activated transcription by binding to its own promoter.

Figure S24 EMSA assessment of MYC2 directly binding to G‐box of cotton RVE2 promoters, and lint percent and fibre length of TA01‐7G, CK and TM‐1.

Figure S25 Unprocessed western blot in Figure 4.

Figure S26 Unprocessed western blot in Figure 4.

Figure S27 Unprocessed western blot in Figure 4.

Figure S28 Unprocessed western blot in Figure 4.

Figure S29 Unprocessed western blot in Figure 6.

Figure S30 Unprocessed western blot in Figure 7.

Table S1 RNA sequencing reads at 24‐, 48‐ and 72‐h post inoculation (hpi).

Table S2 Top biological processes by Gene ontology (GO) enrichment analysis.

Table S3 Biological processes by Gene ontology (GO) enrichment analysis of target modules.

Table S4 List of candidate genes by WGCNA.

Table S5 Polymorphism of RVE2 gene in natural populations of G. hirsutum.

Table S6 Polymorphism of RVE2 gene in natural populations of G. barbadense.

Data availability statement

The RNA‐seq data from this study are openly available in NCBI Sequence Read Archive with the BioProject ID PRJNA713405 and PRJNA212007, BioSamples ID SRR13950210–SRR13950233. Source data are provided with this paper. The unprocessed western blots are provided in Figures S25–S30.