A fungal CFEM‐containing effector targets NPR1 regulator NIMIN2 to suppress plant immunity

Summary

Colletotrichum fructicola causes a broad range of plant diseases worldwide and secretes many candidate proteinous effectors during infection, but it remains largely unknown regarding their effects in conquering plant immunity. Here, we characterized a novel effector CfEC12 that is required for the virulence of C. fructicola. CfEC12 contains a CFEM domain and is highly expressed during the early stage of host infection. Overexpression of CfEC12 suppressed BAX‐triggered cell death, callose deposition and ROS burst in Nicotiana benthamiana. CfEC12 interacted with apple MdNIMIN2, a NIM1‐interacting (NIMIN) protein that putatively modulates NPR1 activity in response to SA signal. Transient expression and transgenic analyses showed that MdNIMIN2 was required for apple resistance to C. fructicola infection and rescued the defence reduction in NbNIMIN2‐silenced N. benthamiana, supporting a positive role in plant immunity. CfEC12 and MdNPR1 interacted with a common region of MdNIMIN2, indicating that CfEC12 suppresses the interaction between MdNIMIN2 and MdNPR1 by competitive target binding. In sum, we identified a fungal effector that targets the plant salicylic acid defence pathway to promote fungal infection.

Introduction

Natural plants face numerous microbial pathogens and have evolved complicated innate immune systems to protect themselves against infection (Chisholm et al., 2006; Jones and Dangl, 2006). When pathogens land on the host surface, a large number of plant immune receptors recognize conserved pathogen‐associated molecular patterns (PAMPs) and rapidly launch basal immune responses called pattern‐triggered immunity (PTI) (Boller and He, 2009; Dodds and Rathjen, 2010). Pathogens can deliver virulence proteins called effectors into host cells to interfere with plant immunity (Dangl et al., 2013). Plants also develop immune receptors to recognize specific effectors and activate effector‐triggered immunity (ETI) characterized by a rapid hypersensitive response (HR), which provides robust defence responses to pathogens (Cui et al., 2015; Wu et al., 2014). Plant innate immunity involves a series of signalling cascades and responses, such as Ca²⁺ flux across the plasma membrane, hormone dynamics, mitogen‐activated protein kinase activation, production of reactive‐oxygen species, callose deposition and expressions of pathogenesis‐related (PR) genes (Tsuda and Katagiri, 2010).

Salicylic acid (SA) as a key plant endogenous phytohormone is required for both local resistance and systemic acquired resistance (SAR) against biotrophic and hemibiotrophic pathogens (Fu and Dong, 2013). After pathogen challenge, the level of SA increases and activates the expression of PR genes (van Loon et al., 2006). In the signalling cascade, NPR1 was identified as a receptor for SA in different reports (Ding et al., 2018; Wang et al., 2020; Wu et al., 2012). NPR1 converts to a monomeric state by reduction of the redox‐sensitive disulphide bonds and then translocates into the nucleus (Mou et al., 2003). NPR1 interacts with a subset of TGA‐bZIP transcription factors and activates TGA‐bZIP transcription factors by enhancing their DNA‐binding affinity (Despres et al., 2000; Kinkema et al., 2000; Mou et al., 2003). TGA transcription factors can interact with the cis‐acting elements in the PR1 gene promoter, induce the expression of PR genes and the production of SAR (Despres et al., 2000; Zhang et al., 1999). It is also known that NPR1 interacts with TCP transcription factors in the nucleus to increase the expression of PR genes (Li et al., 2018). A recent study showed that NPR1 can recruit cyclin‐dependent kinase 8 (CDK8) and WRKY18 to the NPR1 promoter, facilitating its own expression (Chen et al., 2019). Two NPR1 paralogues, NPR3 and NPR4, are required for SA perception (Fu et al., 2012; Liu et al., 2020). Both NPR3 and NPR4 contain BTB domain and ankyrin repeats, which are typical adaptors for CUL3 substrate. Either NPR3 or NPR4 can directly bind to SA and modulate their interactions with NPR1 which results in NPR1 degradation through CUL3‐mediated ubiquitination (Fu et al., 2012; Moreau et al., 2012).

NIMIN is a family of proteins first found in Arabidopsis thaliana, whose members are NIMIN1, NIMIN2 and NIMIN3, differently interacting with NPR1 (Hermann et al., 2013). NIMIN3 interacts with the N‐terminal of NPR1, whereas NIMIN1 and NIMIN2 have similar motifs that bind to the C‐terminal of NPR1. Based on the finding by Hermann et al. (2013), different NIMIN‐NPR1 complexes are formed at different stages of systemic acquired resistance (SAR) to promote the activation of defence genes. When NIMIN3 inhibits PR1 gene expression in unattacked plants, NIMIN2 is induced by low SA concentration to relieve NIMIN3 inhibition of PR1 expression by binding to C‐terminus of NPR1. This process activates the early SA‐dependent PR1 gene. However, the interaction between NIMIN2 and the C‐terminal of NPR1 is not sufficient to activate PR1 gene of late SAR. The effect of NIMIN2 on NPR1 was temporary, followed by NIMIN1 replacing NIMIN2. NIMIN1 has a more transient effect on NPR1 than NIMIN2, in which case the late SAR gene will be activated through the direct effect of SA on NPR1, resulting in the removal of NIMIN1 from the NPR1 complex (Maier et al., 2011). NIMIN proteins regulate plant homeostasis by acting differently on the regulatory factor NPR1 to ensure swift up‐regulated expressions of defence genes, thus successfully fighting invading pathogens.

Colletotrichum fructicola is a worldwide phytopathogen and can infect over 50 plant species, causing serious yield and economic losses (Farr and Rossman, 2021). On apple, C. fructicola causes Glomerella leaf spot (GLS) on susceptible cultivars such as Gala and Golden Delicious, and also causes fruit bitter rot in a cultivar non‐discriminative manner. In a period of high temperature and humidity, GLS can cause more than 90% defoliation, seriously reducing fruit quality and yield (Velho et al., 2019). C. fructicola is the predominant pathogen among Colletotrichum species causing GLS in China (Wang et al., 2015). During infection, C. fructicola first goes through a biotrophic stage with infectious vesicles and primary hyphae and switches to necrotrophic stage with secondary hyphae (Shang et al., 2020a). In the establishment of hemibiotrophic stage, fungal phytopathogens release a set of effectors into plant cells, which interfere with the host's normal physiological metabolism related to immune responses (Irieda et al., 2014; Kleemann et al., 2012; Rivas and Genin, 2011). In planta specifically expressed candidate effectors have been identified from C. fructicola and certain members have been demonstrated to play vital roles in the pathogenicity (Liang et al., 2018; Shang et al., 2020b). However, the virulence functions of most candidate effectors of C. fructicola are still largely unknown.

CFEM (common in several fungal extracellular membrane proteins) domain encompasses about 60 amino acids (aa) including eight conserved cysteine residues (Kulkarni et al., 2003). The domain is specific to the fungal kingdom and commonly exists in membrane proteins (Kulkarni et al., 2003). CFEM domain‐containing proteins were reported to be related to fungal pathogenicity. Pth11, a Magnaporthe oryzae protein containing seven‐transmembrane regions, functions as a G‐protein‐coupled receptor and is an essential regulator for appressorial development and full pathogenicity (Kou et al., 2017). In Botrytis cinerea, BcCFEM1 is associated with conidial production, stress tolerance and pathogenicity (Zhu et al., 2017). CFEM domain‐containing proteins with host‐manipulative activity have also been reported. For example, MoCDIP2 from Magnaporthe oryzae induces rice cell death (Chen et al., 2013), and CgCFEM effectors from Colletotrichum graminicola suppress HR response in non‐host and may function inside different cellular compartments (Gong et al., 2020). In general, however, the functional mechanism of this category of secretory proteins is not clear.

Here, we report that C. fructicola employs a CFEM domain‐containing effector CfCE12 as a virulent factor to suppress plant defence responses including BAX‐triggered cell death and callose and ROS accumulations. CfEC12 targets apple NIMIN2 protein that acts as a positive regulator against pathogen infection. Our analyses revealed that CfEC12 competes with MdNIMIN2 for MdNPR1 binding and then reduces plant resistance regulated by MdNIMIN2‐MdNPR1 module. This study not only provides insights into the molecular pathogenic mechanisms of C. fructicola but also reveals a novel immunity‐regulating mechanism of the fungal CFEM domain‐containing effector.

Results

CfEC12 is required for C. fructicola virulence

To identify the important secretory CFEM protein, we first searched C. fructicola 1104‐7 genome and acquired 28 CFEM‐containing proteins (Figure S1, Appendix S1). Of them, 26 proteins have predicted signal peptides and 18 proteins have at least 1 transmembrane domain. After calculated using EffectorP program, only MDCF17236_17, MDCF17848_51, MDCF17683_125, MDCF17499_01 and MDCF17836_0141 were predicted to be effectors. In our previous transcriptomic analysis, MDCF17236_17 was significantly up‐regulated in planta (Liang et al., 2018; Figure S1), which may be a key C. fructicola effector candidate to be further studied, named as CfEC12. CfEC12 contains 128 aa with conserved 8 cysteines and has a putative signal peptide at the N terminus (Figure S2). A BlastP search against the non‐redundant (nr) database of the National Center for Biotechnology Information (NCBI) revealed CfEC12 homologues were in a number of Colletotrichum species but not in genera other than Colletotrichum (E‐value cut‐off = 1 × 10⁻⁵) (Figure S3). To characterize the expression pattern of CfEC12 in more detail, we quantified CfEC12 transcript accumulation levels in different fungal tissues including mycelia, conidia, appressoria on cellophane and apple leaves following inoculation with conidial suspension. Results showed that the transcript accumulation levels of CfEC12 were highly up‐regulated from 8 to 24 hpi in comparison with mycelia, conidia, appressoria and other infection stages (Figure S4). Peaked gene expression was at 24 hpi, which corresponds to appressorium‐mediated penetration based on a histological study (Shang et al., 2020a). The results revealed that the high transcript level of CfEC12 was induced by the interaction with host, and suggest that CfEC12 likely plays an important role at the early infection stage.

In order to explore the potential biological role of CfEC12 in C. fructicola, we constructed deletion mutants in which CfEC12 gene was replaced with hph gene by a split marker approach (Figure S5a). Gene deletion was confirmed by PCR, RT‐PCR and Southern blot (Figure S5b–d). The mutant strain ΔCfEC12‐5 was complemented by the complementation vector PHZ100‐CfEC12 (Figure S5b–d).

CfEC12 gene deletion did not affect colony morphology or growth rate on PDA medium (Figure S6a), nor conidial sporulation in PDB shake culture (Figure S6b). Additionally, CfEC12 deletion did not affect appressorium development or appressorium‐mediated penetration on artificial cellophane membranes (Figure S6c).

On leaves inoculated with deletion mutant strains, visible lesions incurred by ΔCfEC12‐5 and ΔCfEC12‐7 were obviously decreased compared to WT and ΔCfEC12‐5‐C strains at 60 hpi (Figure 1a). The relative fungal biomass in leaves inoculated with gene deletion mutants was reduced by approximately 30% relative to WT or ΔCfEC12‐5‐C strains as determined by qPCR (Figure 1b). On apple fruit, visible lesions elicited by mutant inoculations were also significantly reduced (Figure 1c,d). Overall, these results indicated that CfEC12 was required for the full virulence of C. fructicola.

Figure 1.

CfEC12 is required for Colletotrichum fructicola virulence and can suppress plant defence responses. (a) The lesion symptom on apple leaves inoculated with different strains at 60 hpi. (b) The virulence was evaluated using fungal biomass versus foliar biomass by quantitative PCR analysis of C. fructicola β‐tubulin gene versus apple UBQ gene. All the DNA extracted from leaves is inoculated with conidia after 60 h. Significant differences are indicated by different letters. (One‐way analysis of variance (ANOVA) test with Tukey's comparisons, P < 0.01). (c) ΔCfEC12‐deletion strains showed fewer spot lesions on apple fruit relative to WT and complementation strains. (d) The virulence was estimated by lesion diameter. Values are presented as the average ± SD (n = 9). Different letters indicate significant differences (ANOVA test with Tukey's comparisons, P < 0.01). (e) CfEC12 inhibited plant cell death (PCD) triggered by BAX and INF1 respectively. (f) Callose deposition in tobacco leaves expressing GFP or CfEC12‐GFP after treatment by 10 μm Flg22 or 1 μm Chitin. (g) Statistical analysis of the callose deposition values. Bars represent means ± SD, n = 3. Asterisks indicate statistically significant differences (unpaired one‐tailed t‐test. P < 0.01). (h) ROS accumulations elicited by Flg22 or 1 μm Chitin were inhibited by CfEC12 expression in tobacco leaves. Bars represent means ± SD, n = 3.

CfEC12 suppresses plant basal immunity

We transiently overexpressed CfEC12‐GFP in the N. benthamiana leaves using Agrobacterium infiltration method to explore whether CfEC12‐GFP was able to suppress BAX‐induced PCD. At 7 days post‐infiltration, pGR107‐CfEC12‐GFP treatment did not cause PCD, but suppressed PCD triggered by BAX (Figure 1e). In contrast, the control sites infiltrated with pGR107 neither showed PCD nor diminished challenged infiltration with BAX (Figure 1e). We also tested the effects of CfEC12 on PCD triggered by INF1 (Kamoun et al., 1997) and found that INF1‐induced cell death was also suppressed by CfEC12 (Figure 1e). We then examined effects of CfEC12 on callose and ROS accumulation by Flg22 and Chitin treatments. Compared with GFP, transient expression of CfEC12 suppresses Flg22 and Chitin elicited callose and ROS accumulation in N. benthamiana (Figure 1f–h). Accumulation of GFP, CfEC12‐GFP, BAX and INF1 proteins in infiltrated leaves was confirmed by Western blot analysis (Figure S7). Together, these results revealed that ectopic expression of CfEC12 in N. benthamiana significantly suppressed plant immunity.

CfEC12 is secreted inside plant cells

Yeast signal peptide screen trap assay confirmed its secretory function (Figure 2a). To further test the SP sequence importance for the virulence function of CfEC12, a vector‐expressing SP‐truncated protein (CfEC12^ΔSP) was constructed and introduced into the ΔCfEC12 mutant. The obtained ΔCfEC12/CfEC12 ^ΔSP transformant was still defective in virulence, similar to that of the ΔCfEC12 (Figure 2b,c). Thus, SP‐truncated CfEC12 (CfEC12^ΔSP) failed to complement the virulence defect of ΔCfEC12, supporting the functional importance of SP.

Figure 2.

Identification of signal peptide (SP) secretory activity and subcellular location of CfEC12. (a) Functional validation of the SP of CfEC12 using the yeast invertase secretion assay. The invertase‐defective yeast YTK12 carrying pSUC2‐SP (Avr1b) and pSUC2‐SP (CfEC12) could grow on both CMD‐W and YPRAA medium. YTK12 and YTK12 carrying empty vector pSUC2 or pSUC2‐SP (Mg87) were used as negative controls. Invertase activity was detected by 2,3,5‐triphenyltetrazolium chloride (TTC). The pink colour represents the invertase activity. (b) Representative images of apple leaves infected with Colletotrichum fructicola 1104‐7, ΔCfEC12 mutant and the ΔCfEC12/CfEC12 and ΔCfEC12/CfEC12 ^ΔSP transformants. (c) The relative biomass of fungal in apple leaves was quantified by quantitative PCR analysis. Different letters indicate significant differences (ANOVA test with Tukey's comparisons, P < 0.01). (d) Overexpression of CfEC12 (with or without SP) can suppress cell death induced by BAX in N. benthamiana leaves. Leaves were infiltrated with A. tumefaciens carrying PGR107‐GFP (negative control), PGR107‐CfEC12 or PGR107‐CfEC12^ΔSP. After 24 h, the infiltrated sites were injected with A. tumefaciens containing PGR107‐BAX again. (e) The accumulations of CfEC12‐mCherry were detected by Western blot in the total proteins of fungal tissue and culture filtrates. WT strain was used as a negative control. GAPDH served as internal control of total fungal proteins. (f) Fluorescence microscopy showed hypha of C. fructicola expressing mCherry or CfEC12‐mCherry (Bars = 5 μm) and the localization of CfEC12‐mCherry inside the nucleus of onion epidermal cells (Bars = 50 μm). Conidial suspensions of the C. fructicola strains expressing free mCherry and CfEC12‐mCherry were inoculated on the surface of inner layer of onion epidermal cells. At 2 dpi, the onion epidermal cells were observed and photographed.

On N. benthamiana, CfEC12^ΔSP still suppressed PCD triggered by BAX (Figure 2d). Furthermore, we overexpressed free mCherry and CfEC12‐mCherry in C. fructicola respectively. The accumulations of CfEC12‐mCherry were detected by Western blot in both the total proteins of fungal tissue and culture filtrates while free mCherry was only detected in fungal tissue (Figure 2e). Furthermore, conidial suspensions of the C. fructicola strain expressing free mCherry or CfEC12‐mCherry were inoculated on the inner layer of onion epidermal cells. At 2 dpi, the observation of onion epidermal cells showed that CfEC12‐mCherry localized in the cell nucleus (Figure 2f). These results suggested that the secreted CfEC12 protein was translocated inside host cell to perform its virulence function.

CfEC12 interacts with apple MdNIMIN2

Construct pGBKT7‐CfCE12^ΔSP was used as the bait to screen a prey yeast two‐hybrid (Y2H) library generated with a combination of RNA from apple leaves at 8, 12, 24, 36 and 48 h after C. fructicola infection. After screening for three rounds, 11 putative interaction clones were identified (Table S1). Further point‐to‐point identification showed that only the full‐length MdNIMIN2, a regulator of NPR1 in nucleus, strongly interacted with CfEC12 (Figure 3a). In Arabidopsis, AtNIMIN1, AtNIMIN2 and AtNIMIN3 can interact with NPR1 sequentially in SA‐triggered immunity (Weigel et al., 2001). We further determined the interaction relationship of MdNIMIN1 and MdNIMIN3 with CfEC12 using Y2H assays. Neither MdNIMIN1 nor MdNIMIN3 was found to interact with CfEC12 (Figure 3a).

Figure 3.

CfEC12 targets an apple MdNIMIN2 protein. (a) Interaction between CfEC12 and MdNIMIN2 in yeast. Yeast containing CfEC12 and MdNIMIN2 grew on DDO and QDO/X/A medium while both CfEC12 and MdNIMIN1 or MdNIMIN3 did not grow on QDO/X/A plates. The yeast carrying empty vector pGBKT7 and MdNIMIN2 were used as negative control. (b) BIFC assay confirmed interaction of CfEC12 with MdNIMIN2 in the plant cell nucleus. Agrobacterium tumefaciens containing CfCE12‐NYFP and MdNIMIN1‐CYFP/MdNIMIN2‐CYFP/MdNIMIN3‐CYFP was co‐infiltrated into Nicotiana benthamiana leaves. The co‐infiltration of CfEC12‐NYFP and CYFP was used as negative control. Confocal images were taken after 3 days. Bars, 20 μm. (c) Co‐immunoprecipitation (Co‐IP) confirmed that CfEC12 interacts with MdNIMIN2. Co‐overexpression of CfEC12‐GFP and MdNIMIN1‐Flag/MdNIMIN2‐Flag/MdNIMIN3‐Flag in N. benthamiana leaves. Total proteins were extracted and the immune complexes were pulled down by GFP‐Trap beads. The coprecipitation was detected by Western blot.

A BIFC assay was performed to determine the subcellular localization of interaction between the CfEC12 and MdNIMIN2 in plant cells. CfCE12‐NYFP and MdNIMIN2‐CYFP fusion constructs were generated and co‐infiltrated into N. benthamiana leaves by A. infiltration. In epidermal cells, GFP signals were observed in the nuclei, indicating that the interaction of CfCE12 and MdNIMIN2 occurred in the plant nucleus (Figure 3b). In contrast, GFP signal was not observed in N. benthamiana cells co‐expressing CfCE12‐NYFP and MdNIMIN1‐CYFP or co‐expressing CfCE12‐NYFP and MdNIMIN3‐CYFP, supporting that the interaction between CfCE12 and MdNIMIN2 was specific.

We further verified the CfCE12‐MdNIMIN2 protein interaction by Co‐IP assay. Recombinant CfCE12‐GFP was co‐expressed with MdNIMIN1‐Flag, MdNIMIN2‐Flag and MdNIMIN3‐Flag in N. benthamiana leaves respectively. Total proteins of infiltrated leaves were extracted, and immunoprecipitation (IP) experiment was performed using antibody magnetic beads that bind GFP. When detected with anti‐Flag and anti‐GFP monoclonal antibodies, MdNIMIN2 and CfCE12 could be detected in immunoprecipitates respectively. However, MdNIMIN1 and MdNIMIN3 could not be detected after immunoprecipitation (Figure 3c).

MdNIMIN2 positive regulates plant immunity

We characterized the expression pattern of MdNIMIN2 during C. fructicola infection at different timepoints post‐inoculation. It was found that MdNIMIN2 was up‐regulated in apple leaves upon C. fructicola infection, and reached the highest expression level at 24 hpi, which was similar to the expression pattern of CfEC12 (Figure 4a).

Figure 4.

MdNIMIN2 positively regulates apple defence against Colletotrichum fructicola. (a) Expression pattern of MdNIMIN2 during C. fructicola infection. Asterisks indicate statistically significant differences (an unpaired one‐tailed t‐test was performed. P < 0.01). (b) Symptoms in Nicotiana benthamiana leaves silencing NbNIMIN2 or overexpression of MdNIMIN2. (c) Relative biomass of P. nicotianae was quantified based on qPCR test. Bars represent means ± SD, n = 3. (d) Rot symptoms on apple fruit silencing MdNIMIN2, transiently overexpressing mCherry or MdNIMIN2 after C. fructicola infection. (e) Diameters of rot lesions on inoculated apple fruit. Values represent means ± SD of five independent samples. (f) Spot symptom on transgenic apple leaves inoculated with WT C. fructicola strain at 60 dpi. (g) The relative fungal biomass in transgenic apple leaves was detected by quantitative PCR. Different letters indicate significant differences (ANOVA test with Tukey's comparisons, P < 0.01).

We use a tobacco system to test the defence function of MdNIMIN2. First, we searched for a homologue of MdNIMIN2 in N. benthamiana, which was clustered together with MdNIMIN2 and AtNIMIN2 (Figure S8a). All of them contained a nuclear localization signal, an NPR1 interaction motif and an LXL motif and were annotated as NIMIN2 proteins in the InterPro database (Figure S8b,c). In the silenced plants, the silencing efficiency of NbNIMIN2 reached around 65%–75% (Figure S9a). The plants were more susceptible to P. nicotianae infection, with the relative biomass increasing by fivefolds compared to that in control leaves (Figure 4b,c). We further overexpressed MdNIMIN2 in the silenced N. benthamiana, which markedly improved its resistance against P. nicotianae (Figure 4b,c). The proper proteins in N. benthamiana leaves were detected by Western blot (Figure S9b). Additionally, both MdNIMIN2 and NbNIMIN2 could interact with NbNPR1 (Figure S9c). These results suggested that MdNIMIN2 and NbNIMN2 have the conserved function in the term of regulation of disease resistance.

In order to verify the role of MdNIMIN2 during infection, MdNIMIN2 and control mCherry were transiently expressed on apple fruit (Figure S9d). When inoculated with C. fructicola, the diameter of lesion in apple expressing MdNIMIN2‐mCherry was smaller than that expressing the mCherry control (Figure 4d,e). We also silenced the MdNIMIN2 in apple fruit. The transcripts of MdNIMIN2 transcripts were significantly reduced in the silencing fruit based on RT‐qPCR analysis (Figure S9e). The lesion of MdNIMIN2‐silenced apples was larger than that in fruit expressing the mCherry control (Figure 4d,e). For vilification of the results, we generated transgenic apple plants overexpressing or silencing MdNIMIN2 (Figure S10a,b). RT‐qPCR showed that MdNIMIN2 was significantly overexpressed or silenced in transgenic plants (Figure S10c,d). After inoculation, the spot lesion symptom on MdNIMIN2‐overexpressing apple leaves reduced significantly in comparison with control leaves, and the lesion on MdNIMIN2‐silencing leaves increased more than on control leaves (Figure 4f,g). These results demonstrated that MdNIMIN2 enhances the resistance to C. fructicola infection.

CfCE12 requires CFEM domain for suppressing BAX‐induced PCD and MdNIMIN2 interaction

Through protein 3D modelling by Swissmodel, the eight cysteines in CFEM domain of CfEC12 produce four disulphide bridges to stabilize the protein structure (Figure S11a). We separately constructed a CfEC12 mutant lacking CFEM domain (termed CfEC12^ΔCFEM) and a variant CfEC12 with two cysteines mutated to alanines (termed CfEC12^C26A/C30A) that disrupted two disulphide bridges (Figure S11b), and then assessed the effects of mutations on the suppression of BAX‐induced PCD. Unlike wild CfEC12, CfEC12^ΔCFEM and CfEC12^C26A/C30A failed to suppress BAX‐induced PCD in N. benthamiana leaves (Figure 5a). It indicates the functional CFEM domain is critical for CfEC12 to suppress BAX‐induced cell death. Western blot assays showed that co‐infiltration did not interfere with gene expressions of CfEC12, CfEC12^ΔCFEM, CfEC12^C26A/C30A, and BAX (Figure S12). We further test whether CFEM domain is critical for MdNIMIN2 interaction. We constructed pGBKT7‐CfCE12^ΔCFEM and pGBKT7‐CfCE12^C26A/C30A vectors for Y2H assay with pGADT7‐MdNIMIN2. Transformants could only grow on defective medium SD‐Trp‐Leu (DDO) but not on SD‐Trp‐Leu‐His‐Ade (QDO)‐containing X‐α‐Gal and ABA (Figure 5b). Furthermore, luciferase complementation imaging (LCI) assays revealed strong luciferase activity in N. benthamiana leaves when CfEC12‐cLuc and MdNIMIN2‐nLuc were co‐expressed, whereas no activity was observed when CfCE12^ΔCFEM‐cLuc or CfCE12^C26A/C30A‐cLuc was co‐expressed with MdNIMIN2‐nLuc (Figure 5c). Finally, we performed a Co‐IP assay to indicate that the full‐length CfEC12 interacted with MdNIMIN2, but CfCE12^ΔCFEM and CfCE12^C26A/C30A did not (Figure S13). These results indicate that the functional CFEM domain of CfEC12 is essential for interaction with MdNIMIN2.

Figure 5.

The CFEM domain is required for the functions of CfEC12 protein and MdNIMINs regulate apple resistance to Colletotrichum fructicola. (a) The CfEC12 without CFEM domain or with amino acid mutation (C26A, C30A) lose the inhibition capacity of plant cell death induced by BAX protein. (b) The CfEC12 without CFEM domain or with amino acid mutation (C26A, C30A) could not interact with host target MdNIMIN2 in yeast. (c) The interaction between CfEC12 mutants and MdNIMIN2 was determined by luciferase complementation assays. The fluorescence signal in image indicates the protein–protein interaction. The combination of CfEC12‐cLuc with MdNIMIN2‐nLuc was used as a positive control. (d) MdNIMIN1, MdNIMIN2 and MdNIMIN3 did not interact with each other in yeast. (e) MdNIMIN1, MdNIMIN2 and MdNIMIN3 interacted with MdNPR1 in yeast respectively. (f) Yeast triple hybridization showed that the interaction of MdNPR1 and MdNIMIN3/MdNIMIN2 was repressed in the presence of MdNIMIN2 or MdNIMIN1. (g) Expression patterns of MdNIMIN1 and MdNIMIN3 during C. fructicola infection. (h) Rot symptoms on apple fruit overexpressing mCherry, MdNIMIN1‐mCherry, MdNIMIN2‐mCherry or MdNIMIN3‐mCherry after C. fructicola infection. (i) Diameters of rot lesions on inoculated apple fruit. Values represent means ± SD (n = 9). Different letters indicate significant differences (P < 0.01, ANOVA with Tukey's comparisons).

MdNIMIN1, MdNIMIN2 and MdNIMIN3 show differential interactions with MdNPR1 and resistances to C. fructicola

Considering that NIMINs are the important regulators of NPR1 in plants, we tested the effects of MdNIMINs on C. fructicola infection. In the public annotated Malus domestica genome file, we found three NIMINs, MdNIMIN1, MdNIMIN2 and MdNIMIN3, which are homologous to AtNIMINs and NbNIMINs (Figure S8). In Arabidopsis, NIMINs regulate the NPR1 balance in response to pathogen‐induced SA signal by sequentially interacting with NPR1 (Weigel et al., 2001). Therefore, we determined the relationships among MdNIMIN1, MdNIMIN2, MdNIMIN3 and MdNPR1. Based on Y2H assay, MdNIMIN1, MdNIMIN2 and MdNIMIN3 could not interact with each other (Figure 5d). Whereas MdNIMIN1, MdNIMIN2 and MdNIMIN3 interacted with MdNPR1 (Figure 5e). Next, in the yeast three‐hybrid system, the colonies expressing MdNPR1, MdNIMIN2 and MdNIMIN3 developed slowly on SD‐Leu‐Trp‐His‐Met plates than colonies expressing MdNPR1 and MdNIMIN3 in which the production of MdNIMIN2 was inhibited by adding 1 mm Met in the medium (Figure 5f). Furthermore, a competitive LCI assay showed that MdNIMIN2 suppressed the fluorescence produced by MdNIMIN3‐MdNPR1 interaction (Figure S14a,c). These results suggested that MdNIMIN2 and MdNIMIN3 competitively interacted with MdNPR1. Similarly, MdNIMIN2‐MdNPR1 interaction was suppressed by MdNIMIN1 in yeast three‐hybrid and LCI assays (Figure 5f; Figure S14b,d).

We further examined the expressions of MdNIMIN1 and MdNIMIN3 in response to C. fructicola infection. In comparison with MdNIMIN2 (Figure 4a), MdNIMIN1 and MdNIMIN3 were up‐regulated at 8 hpi, which indicated that MdNIMIN1, MdNIMIN2 and MdNIMIN3 were differentially expressed during C. fructicola infection (Figure 5g). Next, MdNIMIN1, MdNIMIN2 and MdNIMIN3 were transiently expressed in apple fruit respectively. After inoculated with C. fructicola WT strain, the lesion diameter of fruit expressing MdNIMIN1‐mCherry or MdNIMIN2‐mCherry was significantly smaller than that in control fruit expressing mCherry (Figure 5h,i). By contrast, in fruit expressing MdNIMIN3‐mCherry, the lesion diameter was larger than in fruit expressing mCherry (Figure 5h,i). The above results inferred that MdNIMINs functioned in apple resistance to C. fructicola through a different interaction with MdNPR1.

A conserved region of MdNIMIN2 interacting with both CfEC12 and MdNPR1

To dissect region of MdNIMIN2 interacting with CfEC12, the amino acid sequence of MdNIMIN2 was divided into six segments, and corresponding deletion mutants were fused with pGADT7 separately (Figure 6a). MdNIMIN2‐m1 and MdNIMIN2‐m6 showed obvious interactions with CfEC12. However, MdNIMIN2‐m2, MdNIMIN2‐m3, MdNIMIN2‐m4 and MdNIMIN2‐m5 failed (Figure 6b). We examined the region of MdNIMIN2 required for the interaction with MdNPR1 and found that only MdNIMIN2‐m1, MdNIMIN2‐m5 and MdNIMIN2‐m6 interacted with MdNPR1, but not MdNIMIN2‐m2, MdNIMIN2‐m3 and MdNIMIN2‐m4 (Figure 6b).

Figure 6.

CfEC12 competitively binds to a MdNIMIN2‐interacting region of MdNPR1. (a) Scheme shows the strategies for MdNIMIN2 deletion assay to identify the key interaction region between CfEC12 and MdNPR1. (b) Y2H assays show that the region of 13–80 aa in MdNIMIN2 was required for the interaction with CfEC12 and the region of 13–63 aa was required for interaction with MdNPR1. Yeast co‐expressing CfEC12/MdNPR1 and MdNIMIN2 mutants (MdNIMIN2‐m1 to ‐m6) were grown on DDO and QDO/A/X medium. (c) Key region of MdNIMIN2 was determined by screening different combinations of MdNIMIN2 deletion mutants and CfEC12 using firefly luciferase assays. The fluorescence signal on leaf shows the protein–protein interaction. The combination of MdNIMIN2‐nluc with CfEC12‐cluc was used as a positive control (P) and the combination of nLuc and cLuc served as negative control (N). (d) Detection of firefly luciferase activity using Luminometer; the error is the standard deviation of the mean value (n = 8). (e) The fluorescence signal on leaf shows the protein–protein interaction. The combination of MdNIMIN2‐nluc with MdNPR1‐cluc was used as a positive control. (f) Firefly luciferase activity was quantified using luminometer; the error is the standard deviation of the mean value (n = 8). (g) Yeast hybrid assay showed CfEC12 could not interact with MdNPR1. (h) Yeast hybrid interaction of MdNPR1 and MdNIMIN2 proteins in the absence and presence of CfEC12. CfEC12 was expressed from the Met25 promoter which is repressed in presence and de‐repressed in absence of methionine. Yeast containing pBridge‐GFP‐MdNIMIN2 and AD‐MdNPR1 was used as control. (i) Protein–protein interaction in the absence or presence of Met was determined in quantitative assays. The values are given as averages ± SD of three independent assays. (j) The interaction between MdNIMIN2 and MdNPR1 was reduced by co‐expression of CfEC12‐GFP in LCI assay. (k) Quantification of luciferase activity in the leaves shown in (i). Bars represent mean ± SD (n = 8). (l) The effect of CfEC12 on the MdNIMIN2‐MdNPR1 interaction in Co‐IP assay. Flag‐MdNIMIN2‐nLuc and HA‐MdNPR1‐cLuc were co‐expressed with increasing amounts of CfEC12‐GFP. The expressed proteins were immunoprecipitated with anti‐Flag antibodies, and crude and immunoprecipitated proteins were then analysed with anti‐GFP, anti‐Flag or anti‐HA antibodies. Asterisks indicate statistically significant differences (an unpaired one‐tailed t‐test was performed. P < 0.01).

Additionally, we performed LCI assays to confirm the region required for the interaction relationship with CfEC12 and MdNPR1 in planta. Different fragments of MdNIMIN2‐nLuc and CfEC12‐cLuc or MdNPR1‐cLuc were co‐expressed in N. benthamiana leaves by Agrobacterium tumefaciens. Strong fluorescence signals in the combinations of MdNIMIN2‐m1‐nLuc and CfEC12‐cLuc, MdNIMIN2‐m6‐nLuc and CfEC12‐cLuc, MdNIMIN2‐m1‐nLuc and MdNPR1‐cLuc, MdNIMIN2‐m6‐nLuc and MdNPR1‐cLuc, and MdNIMIN2‐m5‐nLuc and MdNPR1‐cLuc were detected in N. benthamiana (Figure 6c–f). However, the expressions of MdNIMIN2‐m2‐nLuc and CfEC12‐cLuc, MdNIMIN2‐m3‐nLuc and CfEC12‐cLuc, MdNIMIN2‐m4‐nLuc and CfEC12‐cLuc, MdNIMIN2‐m5‐nLuc and CfEC12‐cLuc, MdNIMIN2‐m2‐nLuc and MdNPR1‐cLuc, MdNIMIN2‐m3‐nLuc and MdNPR1‐cLuc, and MdNIMIN2‐m4‐nLuc and MdNPR1‐cLuc could not restore the luciferase in plant cells respectively (Figure 6c–f). Deletion of the 13–63 aa region abolished MdNIMIN2‐CfEC12 and MdNIMIN2‐MdNPR1 interactions. Therefore, this region was involved in interaction with both CfCE12 and MdNIMIN2.

CfEC12 blocks the interaction between MdNIMIN2 and MdNPR1

Y2H showed there is no interaction detected between CfEC12 and MdNPR1 (Figure 6g). We thus employed the yeast three‐hybrid system to monitor the interaction relationships among MdNIMIN2, MdNPR1 and CfEC12. When CfEC12 was expressed in yeast cells together with MdNIMIN2 and MdNPR1, the colonies grew slowly without activation of X‐α‐Gal on the defective medium SD‐Leu‐Trp‐His‐Met (Figure 6h,i). However, repression of CfCE12 expression by addition of 1 mM Met to SD‐Leu‐Trp‐His‐Met medium recovered the MdNIMIN2‐MdNPR1 interaction (Figure 6h,i). These results indicate that CfCE12 can inhibit the interaction between MdNIMIN2 and MdNPR1. Furthermore, we took advantage of LCI assay to verify the inhibition effect in planta. The expression of CfEC12 reduced the fluorescence signals of luciferase being recovered by MdNIMIN2‐nLuc and MdNPR1‐cLuc compared to that without the CfEC12 expression (Figure 6j). Quantification analysis showed that when both MdNIMIN2‐nLuc and MdNPR1‐cLuc were present, the catalytic activity of luciferase in leaves expressing CfEC12 decreased significantly than that without CfEC12 (Figure 6k). Western blots determined that the CfEC12 did not affect the expressions of MdNIMIN2‐nLuc and MdNPR1‐cLuc (Figure 6l). Finally, competitive Co‐IP assay confirmed that the protein levels of MdNPR1 co‐immunoprecipitated by MdNIMMIN2 were notably decreased with the amount of CfEC12 increased (Figure 6l). Together, these results indicated that CfEC12 blocked the interaction between MdNIMIN2 and MdNPR1 in a competitive manner.

MdNIMIN2‐mediated immunity depends on MdNPR1

When MdNPR1 was overexpressed and silenced in apple fruit, it was found that the diameters of lesion on MdNPR1‐overexpressed fruit were smaller, and the same on MdNPR1‐silenced fruit were larger in comparison with cLuc control (Figure 7a,b), indicating that MdNPR1 contributed apple resistance to C. fructicola infection. Additionally, MdPR1, MdPR2 and MdPR5 were significantly up‐regulated in MdNPR1‐overexpressed fruit than in MdNPR1‐silenced or cLuc‐expressing fruit (Figure 7c). The silencing deficiency of MdNPR1 was determined by RT‐qPCR (Figure S15a). The cLuc and MdNPR1‐cLuc expression levels were examined by Western blot (Figure S15b).

Figure 7.

CfEC12 suppresses the host resistance regulated by MdNIMIN2‐MdNPR1 module. (a) Apples were infiltrated with Agrobacterium tumefaciens carrying pCambia1300‐MdNPR1‐cLuc and pTRV‐MdNPR1. Apples infiltrated with pTRV and cLuc were used as control. C. fructicola was inoculated on the apples. (b) Diameters of lesions in apple fruit were quantified after 3 days of inoculation. Bars represent mean ± SD (n = 5). (c) The expressions of MdPR1, MdPR2 and MdPR5 in MdNPR1‐overexpressing or ‐silencing apple fruit. (d) A. tumefaciens containing nLuc and MdNIMIN2‐nLuc were agroinfiltrated into the pTRV and pTRV‐MdNPR1 apple fruit respectively. C. fructicola were inoculated onto agroinfiltrated apples. (e) Lesion diameters were measured at 5 days after infection. Bars represent mean ± SD (n = 5). (f) RT‐qPCR data showed relative expression of pathogen‐related genes activated by MdNPR1 including MdPR1, MdPR2 and MdPR5. The apple UBQ gene expression was used for normalization in RT‐qPCR assays. Gene expression levels were relevant to pTRV and nLuc fruit. (g) Effect of salicylic acid (SA) on apple Glomerella leaf spot resistance. SA at different concentrations (0, 0.05, 0.1, 0.2 and 0.3 mm) was sprayed on apple leaves. After 24 h, the conidia of C. fructicola were sprayed on the treated leaves. The lesion symptoms on apple leaves were photographed at 72 hpi. (h) The disease severity was assessed by fungal biomass versus foliar biomass using quantitative PCR analysis of C. fructicola β‐tubulin gene versus apple UBQ gene. Different letters indicate significant differences. (i) A working model shows the virulence mechanism of CfEC12 in suppressing host immunity. Generally, upon pathogen attacking, NIMIN2 (N2) binds to C‐terminus of NPR1 to promote defence gene expressions. However, during initial infection, C. fructicola secretes and delivers CfEC12 into cell nucleus. CfEC12 competitively binds to apple NIMIN2 with NPR1, which prevents the interaction between NIMIN2 and NPR1. As a result, CfEC12 reduces NPR1‐depending defence responses during C. fructicola infection. In (b, c, e, f and h), different letters indicate significant differences (ANOVA test with Tukey's comparisons, P < 0.01).

To determine the effects of MdNIMIN2 interaction with MdNPR1 to mediate plant immunity, we examined the disease severity on the MdNPR1‐silenced and/or MdNIMIN2‐overexpressed apple fruit during C. fructicola infection. It was shown that in pTRV fruit, the lesion diameter was significantly reduced to about 32% in MdNIMIN2‐nLuc‐expressing fruit compared to the control nLuc‐expressing fruit (Figure 7d,e). However, in pTRV‐MdNPR1 fruit, the lesions were reduced to about 13% in MdNIMIN2‐nLuc‐expressing fruit (Figure 7e). The nLuc and MdNIMIN2‐cLuc expression levels were determined by Western blot (Figure S15c).

Since NIMIN2 interacts via binding to NPR1, promoting SAR marker gene PR1, PR2 and PR5 expressions, we detected the transcripts of MdPR1, MdPR2 and MdPR5 in all treatments. The differential transcript levels of MdPR1, MdPR2 and MdPR5 in pTRV fruit expressing nLuc and MdNIMIN2‐nLuc were much larger than that in pTRV‐MdNPR1 fruit expressing nLuc and MdNIMIN2‐nLuc (Figure 7f). These results indicated that MdNIMIN2 enhanced plant resistance depending on MdNPR1. Furthermore, considering that plant NIMINs‐NPR1 regulation cascade is involved in SA, we sprayed SA on apple leaves and then inoculated with conidial suspension. At 5 dpi, the lesions on SA‐treated leaves and the relative biomass of C. fructicola decreased significantly with the increase in the SA concentration (Figure 7g,h).

Discussion

Colletotrichum is one of the most important groups of fungal plant pathogens in the world, causing serious yield and economic losses (Crouch et al., 2014; Dean et al., 2012). Up to date, genome and transcriptome analyses reveal that Colletotrichum spp. possess a large amount of effector proteins that are induced during infection of host (Kleemann et al., 2012). So far, however, the virulence mechanisms of only a few effectors have been determined (Azmi et al., 2018; Irieda et al., 2014). In this study, we demonstrated that C. fructicola delivers a nucleus‐localized effector CfEC12 to target host NIMIN2 protein that is a positive regulator of NPR1‐dependent immunity. CfEC12 competes with NPR1 for NIMIN2 binding, resulting in the reduction in plant defences (Figure 7i). These results significantly advance our understanding of mechanisms underlying Colletotrichum pathogenesis.

The CFEM domain proteins are uniquely and widely distributed in the fungal kingdom (Zhang et al., 2015). Some CFEM‐containing proteins function as secretory effectors in phytopathogenic fungi, such as Podosphaera xanthii, Sclerotinia sclerotiorum, Puccinia triticina and Botrytis cinerea (Seifbarghi et al., 2017; Vela‐Corcía et al., 2016; Zhao et al., 2020; Zhu et al., 2017). In our study, we also found that CfEC12 contains a typical CFEM domain. Deletion of CFEM domain did not rescue the deficiency in suppression of cell death in N. benthamiana, as well as in the interaction with MdNIMIN2. The mutation of CFEM domain in CfEC12 also led to loss of function in suppression of plant immunity. The founding supports that the CFEM domain is essential for CfCE12 functions. The importance of CFEM domain is also reported in Magnaporthe oryzae. Knockout of CFEM domain in Pth11, a cell membrane location protein, resulted in defects in the appressorium differentiation and ROS homeostasis regulation (Kou et al., 2017).

The original function of CFEM proteins locate in the cell membrane or anchor to the surface of cell membrane or wall, associating with the functions of surface sensing and signal transduction (Zhang et al., 2015). Extracellular receptor ACI1 participates in cAMP signal transduction and further influences appressorium differentiation in M. grisea (Kulkarni and Dean, 2004). Some CFEM domain proteins also anchor to the outer layer of the cell membrane through a GPI anchor site, functioning other various roles. For example, Rbt5, Pga7 and Csa2 participate in haem‐iron acquisition for Candida species (Kuznets et al., 2014; Nasser et al., 2016). In addition, PstCFEM1 as an apoplastic virulence effector inhibits wheat basal immunity (Bai et al., 2022). Recently, it has been reported that Fusarium graminearum secreted an effector CFEM1 into host apoplast to compromise ZmWAK17‐mediated resistance (Zuo et al., 2022). However, in our study, we revealed a novel CFEM effector, CfCE12, was secreted and delivered into plant nucleus. Further transient expression of CfCE12 could suppress BAX‐induced PCD and ROS burst in N. benthamiana leaves, suggesting that CfCE12 functions as an effector to suppress defence responses. CFEM candidate effectors of Colletotrichum graminicola have been identified in multiple plant subcellular compartments including cell membrane, nucleus, cytosolic bodies and the whole cell, which is probably in connection with their functions (Gong et al., 2020). All these suggest that CFEM‐containing protein can utilize diverse subcellular locations to achieve function differentiation. Systematic functional analysis of the distinct location category of CFEM effectors will deepen our understanding of this kind of secretory proteins in pathogenic process.

Salicylic acid (SA) signalling is crucial for plant resistance to C. fructicola. Upon attack by C. fructicola, SA accumulation is quickly induced and triggers expressions of downstream pathogenesis‐related (PR) genes in strawberry (He et al., 2019). The exogenous application of SA enhances the activities of catalase (CAT), superoxide dismutase (SOD), peroxidase (POD), phenylalanine ammonia‐lyase (PAL) and polyphenol oxidase (PPO) and induces the expression levels of PR genes, conferring resistance to C. fructicola in apple (Zhang et al., 2016). In addition, apple MKK4‐MPK3‐WRKY17 module can mediate SA degradation, resulting in enhanced susceptibility to C. fructicola (Shan et al., 2021). Non‐expressor of pathogenesis related‐1 (NPR1) is the SA receptor, performing a central role in downstream resistance signalling, but is monitored strictly by NIMINs depending on progressing threat by pathogens to maintain appropriate defence gene activation (Dong, 2004; Hermann et al., 2013; Mukhtar et al., 2009; Wang et al., 2006; Wu et al., 2012). In normal conditions, avoiding excessive activation of plant immunity, plant NIMIN3 binds to NPR1 by the N‐terminus of NPR1, repressing NPR1‐dependent PR1 expression. Upon pathogen attacking, NIMIN2 is induced at low levels of SA to relieve NIMIN3 repression by binding to the NPR1 C‐terminus. The interaction of NIMIN2 and NPR1 is transient and NIMIN1 is replaced by NIMIN2 via a common motif binding to the C‐terminal part of NPR1. NIMIN1 action on NPR1 is transient and removed from the NPR1 complex by SA in respond to late SAR genes (Fu et al., 2012; Maier et al., 2011; Wu et al., 2012). In this study, MdNIMIN1, MdNIMIN2 and MdNIMIN3 interacted with MdNPR1, but they did not interact with each other. Yeast hybrid assays revealed that MdNIMIN2 competes with MdNIMIN3 to bind MdNPR1, and MdNIMIN2 has a stronger affinity with MdNPR1. Similarly, MdNIMIN1 has a stronger bond with MdNPR1 than MdNIMIN2. Additionally, MdNIMIN1, MdNIMIN2 and MdNIMIN3 had up‐regulated expressions during the early stage of C. fructicola infection and differential resistance to C. fructicola. These findings have similarities to the AtNIMINs‐AtNPR1 cascade regulation that NIMIN2 plays an indispensable role of intermediate bearer in NPR‐dependent defence gene expression. In this study, CfEC12 restrains the interaction between MdNPR1 and MdNIMIN2. CfEC12 function as an NPR1 competitor to interact with NIMIN2 blocks the next combination of NPR1 and NIMIN1 and reduces NPR1‐dependent immune responses, promoting infection of C. fructicola on apple.

The SA receptor NPR1 (non‐expressor of pathogenesis related‐1) is a transcriptional co‐regulator playing a central role in disease resistance signalling, the activity of which is tightly controlled (Dong, 2004; Mukhtar et al., 2009; Wang et al., 2006; Wu et al., 2012). In unchallenged plant cells, NPR1 is largely S‐nitrosylated as inactive oligomers in the cytoplasm. Upon pathogen infection, thioredoxins catalyses NPR1 monomerization that allows the NPR1 monomers to translocate into cell nucleus (Mou et al., 2003; Tada et al., 2008). Then, NPR1 monomers interact with TGA transcription factors to activate PR gene expression (Fan and Dong, 2002; Zhang et al., 1999; Zhou et al., 2000). In nucleus, NPR1 is also regulated by various co‐factors including NIMINs, TOPLESS and CBNAC‐SNI1 (Seyfferth and Tsuda, 2014). The attack of pathogen effectors on NPR1 or NPR1 co‐regulator becomes an important way to suppress SAR. Bacterial type III effector AvrPtoB mediates NPR1 degradation by 26S proteasome, blocking SA perception to reduce plant resistance (Chen et al., 2017). Cochliobolus victoriae victorin binds to the active site of TRX‐h5, suppressing NPR1 monomerization in the cytosol (Lorang et al., 2012; Tada et al., 2008). Puccinia striiformis f. sp. tritici effector PNPi blocks the interaction of NPR1 with TGA transcription factors (Wang et al., 2016). Phytophthora capsici effector RxLR48 directly interacts with NPR1 and suppresses plant immune responses (Li et al., 2019). Our research found a novel mechanism that C. fructicola attacks NPR1‐dependent basal immunity by the specific effecter CfEC12 by interacting with NPR1 co‐activator MdNIMIN2. Based on the mechanism, application of SA significantly enhanced the resistance to GLS. Additionally, the positive immunity regulator MdNIMIN2 can be used as a gene resource for apple molecular breeding or gene editing breeding in the future.

Experimental procedures

Plant materials and microbial strains

Nicotiana benthamiana were planted in a greenhouse for 6 weeks at 22–25 °C. Healthy apple leaves and fruits (Malus domestica cv. Gala) were sampled from an orchard in Yangling, Shaanxi Province, China. Colletotrichum fructicola 1104‐7 strain was cultured routinely on potato dextrose agar (PDA) medium at 25 °C with dark. Phytophytora nicotianae was cultured on V8 medium at 25 °C.

Gene expression assay

The total RNA was extracted using the RNAprep Pure Plant Kit (TIANGEN BIOTECH CO. LTD, Beijing, China). cDNA synthesis was performed using One‐Step gDNA Removal and cDNA Synthesis SuperMix (TRANSGEN BIOTECH CO. LTD, Beijing, China). The RT‐qPCR was carried out using RealStar Green Power Mixture (GenStar, Beijing, China) in an Applied Biosystems StepOnePlus Real‐Time PCR Systems. Gene expression was quantified using the $2^{-\mathrm{\Delta \Delta }{C}_{\mathrm{t}}}$ method. C. fructicola β‐tubulin gene was used as an endogenous control for quantification of CfEC12 expressions and the actin of N. benthamiana was used as reference gene to assess gene silencing efficiency. All primers used in this study are listed in Table S1.

Gene deletion and complementation

The CfEC12 mutants were generated using split‐marker strategy as described in previous study (Shang et al., 2020b). For complementation, the ORF of CfEC12 with approximate 1.5‐kb upstream fragment was inserted into G418‐resistant pHZ‐100 vector and then transformed into ΔCfEC12‐5 mutant protoplasts. Diameters of vegetative colony of WT and mutant strains were measured after cultured at 25 °C in dark for 7 days.

Cellophane and plant infection assays

Conidia preparation, cellophane inoculation and pathogenicity assay were performed by Shang et al. (2020b). For SA treatment assay, different concentrations of SA (0. 0.05, 0.1, 0.2 and 0.3 mm) were sprayed on apple leaves for 24 h, and then conidia suspension was sprayed on the treated leaves. Disease lesions on leaves were photographed and assessed by qPCR after 60 or 72 hours inoculation. After 5 days post‐inoculation, lesions on fruit were recorded based on the diameter. Conidia germination and appressoria formation on leaves inoculated at 24 and 60 hpi were quantified.

DNA extraction and qPCR analysis

In order to evaluate the biomass of fungi and leaves, 50 μg genomic DNA was used to amplify C. fructicola β‐tublin (CfβTUB) and M. domestica UBQ (MdUBQ) fragments. The value of fungi biomass versus leaf biomass was applied to estimate the pathogenic variation in WT and mutant strains.

Bioinformatic analysis

The putative signal peptide was predicted by SignalP 4.1 (Nielsen, 2017). The domain was identified based on the conserved domain database and SMSRT websites (Letunic and Bork, 2018). Homologous protein sequences of CfEC12 were searched by BlastP program and downloaded from the NCBI database. Proteins sequences were aligned using ClustalW (Thompson et al., 1994) and edited by JalView (Clamp et al., 2004). Evolution tree was constructed using MEGA 11 (Kumar et al., 2016).

Secretion assay and confocal observation

The predicted signal peptide coding sequence of CfEC12 was amplified and inserted into pSUC2 vector. The constructed plasmid pSUC2‐SP, the positive control plasmids pSUC2‐Avr1b and the negative pSUC2‐Mg87 were introduced into the invertase negative yeast strain YTK12 respectively. Then single colonies of transformed YTK12 were tested for secretion activity (Oh et al., 2009).

To detect the secretion of CfEC12, the CDS of CfEC12 was fused with mCherry driven by TrpC promoter in PHZ‐100 plasmid to be introduced into WT strain, and the obtained CfEC12‐mCherry strain was cultured in 100 mL PDB liquid medium for 5 days. Fungal tissues were collected by centrifuging at 8000  g for 5 min in order to extract protein. The supernatant was filtered through 0.45‐μm‐diameter filter and then concentrated to 50 μL for immunoblotting by Amicon Ultra‐15 (3 kDa NMWCO) (Merk Millipore Ltd., Tullagreen, Carrigtwohill, Co. Cork, Ireland). For observation of CfEC12 location, the CfEC12‐mCherry and mCherry strains were inoculated on the surface of onion inner epidermal layer for 2 days.

Cell death suppression assays

The coding sequences of CfEC12 (with and without signal peptide) were fused into vector pGR107. CfEC12 mutants including deletion and site‐directed mutation CFEM domain were synthesized into pGR107 by Sangon Biotech (Shanghai, China) Co., Ltd., generating pGR107‐CfEC12^∆CFEM and pGR107‐CfEC12^C26A/30A respectively.

Agrobacterium tumefaciens cell suspension carrying pGR107‐CfEC12, pGR107‐CfEC12^ΔSP, pGR107‐CfEC12^ΔCFEM or pGR107‐CfEC12^C26A/30A were suspended to a final density of OD₆₀₀ = 0.6 in infiltration buffer (10 mm MgCl₂, 10 mM MES and 150 μm acetosyringone) and infiltrated initially. A. tumefaciens cells carrying pGR107‐BAX or pGR107‐INF1 were infiltrated into the same site 24 h later. The symptom was observed and photographed 5–7 days post‐inoculation. A. tumefaciens cell carrying empty pGR107 vector was infiltrated as control treatment. For each plasmid, three leaves from three plants were infiltrated.

Oxidative burst detection and papillary callose staining assays

Nicotiana benthamiana leaves transiently expressing CfEC12 were treated with 10 μm Flg22 and 1μM Chitin. To measure the accumulation of ROS, leaves were incubated in luminescence buffer (100 mm luminol and 20 mg/mL horseradish peroxidase). Luminescence was recorded with a microplate reader (Thermo Fisher Scientific) for 25 min. For detection of the papillary callose deposition, leaves were cleared in destaining solution (alcohol: acetic acid, 1:1 v/v) for 48 h, transferred to 50% alcohol and stained with aniline blue buffer (0.05% aniline blue in phosphate buffer, pH 9.5) for examination under a fluorescence microscope (Olympus BX‐51).

Protein extracts and immunoprecipitation assays

Total protein of N. benthamiana leaves was extracted using a protein extraction kit (BestBio, Shanghai, China). Proteins were separated by SDS‐PAGE gel and transferred to polyvinylidene fluoride (PVDF) membrane (Millipore, Billerica, MA). The mouse anti‐GFP monoclonal primary antibody (TRANSGEN), mouse anti‐HA monoclonal primary antibody (TRANSGEN) and mouse anti‐BAX monoclonal primary antibody (Beyotime Biotechnology) were dissolved in 1:2000 dilution to detect the CfEC12 fusion proteins and BAX protein. After incubated with primary antibodies at 4 °C overnight, membranes were washed with TBST buffer and transferred into a 1:4000 dilution of horseradish peroxidase‐conjugated anti‐mouse secondary antibody (TRANSGEN) for 2 h at 22 °C. The antigen–antibody complex was visualized using chemiluminescence Western Blot Kit (TRANSGEN). The membrane was exposed to a light film (Kodak).

Yeast hybrid assays

For yeast two hybrid, the CfEC12^ΔSP and MdNPR1 were fused with the bait vector pGBKT7. MdNIMIN1, MdNIMIN2 and MdNIMIN3 were constructed into pGADT7 or pGBKT7 and co‐introduced in pairs into Y2hGold yeast strain. Yeast cells were gradually cultured on SD‐Trp‐Leu (DDO), SD‐Trp‐Leu‐His (TDO) and SD‐Trp‐Leu‐His‐Ade (QDO) (with X‐α‐gal and ABA) plates respectively.

For yeast three hybrids, the pGADT7‐MdNPR1 construct was co‐conformed with pBridge‐CfEC12‐MdNIMIN2, pBridge‐MdNIMIN1‐MdNIMIN2 or pBridge‐MdNIMIN2‐MdNIMIN3 into Y2hGold yeast strain respectively. To examine the relationships among MdNIMIN1, MdNIMIN2, MdNIMIN3 and MdNPR1, the transformants were analysed on SD‐His‐Leu‐Trp media containing X‐ɑ‐gal with or without 1 mM methionine. The activities of LacZ reporter gene were quantified as described (Weigel et al., 2001).

Bimolecular fluorescence complementation (BIFC) assays

The coding regions of MdNIMIN2 were inserted into CYFP vector with the C‐terminal fragment of the green fluorescent protein and CfEC12 was cloned into NYFP vector with the N‐terminal fragment of the green fluorescent protein. The fusion vectors CfEC12‐NYFP and MdNIMIN1‐CYFP or MdNIMIN2‐CYFP or MdNIMIN3‐CYFP were transformed into A. tumefaciens strain GV3101 and transiently co‐expressed in N. benthamiana leaves. After 3 days, fluorescence signals of GFP were examined under an FV3000 confocal laser microscope (Olympus, Tokyo, Japan) at 488 nm. Nuclei were stained with 4,6‐ diamidino‐2‐phenylindole (DAPI).

Co‐immunoprecipitation (Co‐IP) assays

Agrobacterium tumefaciens strain GV3101‐containing pBin‐CfEC12‐GFP, pBin‐CfEC12^ΔCFEM‐GFP, pBin‐CfEC12^C26A/C30A‐GFP and pCambia1307‐MdNIMIN1‐Flag or pCambia1307‐MdNIMIN2‐Flag, or pCambia1307‐MdNIMIN3‐Flag recombinants were co‐agroinfiltrated into N. benthamiana leaves respectively. Total of the protein solution was mixed with GFP‐Trap beads (Sangon Biotech, Shanghai, China) at 4 °C overnight. Subsequently, the precipitates attached to the beads were washed with washing buffer (10 mm Tris‐Cl [pH 7.5], 150 mm NaCl, 0.5 mm EDTA and 0.1% Tween 20). The immune precipitates were boiled in SDS protein loading buffer, separated by PAGE‐SDS gel and detected using the monoclonal anti‐GFP or anti‐Flag antibody.

Gene silencing in N. benthamiana

The fragment of N. benthamiana gene Niben101Scf08347g01002.1 (orthologue of MdNIMIN2) was cloned into pTRV2 vector. A. tumefaciens carrying pTRV1 and pTRV2‐NbNIMIN2 were adjusted to OD₆₀₀ = 0.6, mixed at a 1:1 ratio and then infiltrated into the first pair of leaves of 2‐week‐old N. benthamiana. A mixture of A. tumefaciens harbouring pTRV1 and pTRV2 was used as a control. VIGS efficiency was determined using RT‐qPCR after 3 weeks.

Transient overexpression or silencing of genes in fruit

For overexpression, the full‐length coding sequences (CDS) of MdNIMIN1, MdNIMIN2, MdNIMIN3 and MdNPR1 were cloned into the pCambia‐1300 vector. For gene silencing, the 200 bp segment of MdNIMIN2 was coupled into the pTRV2 vector. Recombinant plasmid‐carrying A. tumefaciens strain GV3101 was introduced into the ‘Gala’ fruit as described by previous report (Lv et al., 2019). Once the apples were kept in the dark for 3 or 15 days, C. fructicola conidia suspension was inoculated on the peel near the injection site.

Firefly luciferase complementation imaging (LCI) assay

The coding sequences of CfEC12, CfEC12^ΔCFEM, CfEC12^C26A/C30A and MdNPR1 were inserted into pCambia1300‐cLuc. MdNIMIN2 and MdNIMIN3 were fused into vectors pCambia1300‐nLuc. The Agrobacterium strain GV3101 carrying nLuc/cLuc fusion plasmids were equally mixed and co‐infiltrated into N. benthamiana leaves. After inoculation for 3 days, chemiluminescence images were captured with a plant living imaging system (Lumazone Pylon2048B, Princeton, NJ, USA), and the activities of firefly luciferase were measured with a microplate reader (Thermo Fisher Scientific). Afterwards, total proteins were extracted from leaves and immunoprecipitation was performed using monoclonal anti‐HA or anti‐His antibody.

Generation of transgenic apple plants

For overexpression, the coding sequence of MdNIMIN2 was cloned into the pCambia1302 vector driven by the 35s promoter. For silencing, a 200‐bp fragment of MdNIMIN2 was cloned into vector pCambia1302‐RNAi in both antisense and sense orientations. The resulting constructs pCambia1302‐MdNIMIN2 and pCambia1302‐RNAi‐MdNIMIN2 were transformed into Malus domestica ‘Gala’ by A. tumefaciens strain EHA105. The kanamycin was used as a selectable marker.

Statistical analysis

All experiments were performed with three biological replicates. Statistics were performed using analysis of variance (ANOVA) tests with Tukey's comparisons or unpaired one‐tailed Student's t‐test (*P < 0.01, and ns, no significance). Data are reported as the mean ± SD.

Conflicts of interest

The authors declare that they have no competing interests.

Author contributions

S.S., X.L., R.Z. and G.S. designed the experiments. S.S., G.L. and S.Z. performed the experiments. S.S., X.L., R.Z. and G.S. wrote the paper.

Supporting information

Appendix S1 Colletotrichum fructicola 1104‐7 CFEM‐containing domain protein sequences.

Figure S1 Bioinformatic analysis of CFEM protein in C. fructicola.

Figure S2 The CFEM domain regular expression of CfEC12 with multiple known CFEM domain proteins.

Figure S3 Phylogenetic analysis of CfEC12 (XP 031892272.1) and its homologue proteins in Colletotrichum.

Figure S4 RT‐qPCR analysis of CfEC12 expression patterns.

Figure S5 Generation of CfEC12 deletion and complementation stains.

Figure S6 The effects of CfEC12 deletion on the growth, sporulation and infection structures of C. fructicola.

Figure S7 Determination of proper protein expression in N. benthamiana leaves.

Figure S8 Characterization of NIMINs from M. domestica, A. thaliana and N. benthamiana.

Figure S9 Determination of NbNIMIN2 and MdNIMIN2 expressions and interactions between NbNIMIN2/MdNIMIN2 and NbNIMIN2.

Figure S10 Overexpression and silencing of MdNIMIN2 in apple.

Figure S11 The strategies for CfEC12 domain deletion and mutation.

Figure S12 Protein expressions were examined in N. benthamiana leaves.

Figure S13 Co‐immunoprecipitation showed that CFEM domain deletion or mutation affected the interaction between CfEC12 and MdNIMIN2.

Figure S14 The interaction of MdNPR1 and MdNIMIN3/MdNIMIN2 was inhibited by MdNIMIN2 or MdNIMIN1.

Figure S15 Detection of gene expression in apple fruit.

Table S1 CfEC12‐interacting clones identified by yeast two‐hybrid library screening.

Table S2 Primers used in this study.