The powdery mildew effector Eae1 targets a SAMS enzyme within the ethylene biosynthesis pathway to disrupt plant immunity

Abstract

Phytopathogens secrete effector proteins that disrupt plant immunity. Phytohormones, such as ethylene, function as immune signals that regulate plant responses to phytopathogens. However, the mechanisms by which powdery mildew fungi utilize effectors to manipulate ethylene signaling remain poorly understood. This study reports that Eae1, an effector from the powdery mildew fungus Erysiphe quercicola, can reduce ethylene levels and attenuate the immune response in rubber trees (Hevea brasiliensis), thereby promoting fungus infection. Notably, Eae1 shares homologs with effectors from other powdery mildew fungi. Our findings indicate that the Eae1 transcription can be induced by enhanced ethylene signaling. Eae1 is translocated to plant chloroplasts, where it destabilizes Hevea brasiliensis S-adenosyl-L-methionine synthetase, a key enzyme in ethylene biosynthesis, leading to reduced ethylene production. The chloroplast-localized protein Toc159/AIG1 (AIG1) may facilitate this interaction. Overall, our study reveals a mechanism by which powdery mildew fungi disrupt ethylene-mediated resistance in host plants.

Background

Powdery mildew fungi (Ascomycota, Erysiphales) constitute a large group of obligate biotrophic phytopathogens that cause disease in thousands of plant species, resulting in great economic losses (Glawe and Dean 2008). These fungi depend on living host tissues to complete their life cycle. When the fungal conidia attach to the host tissue surface, they germinate and penetrate the tissues by forming specialized structures called haustoria, which facilitate nutrient uptake and molecular exchange with the plant (Hückelhoven and Panstruga 2011). At later stages of infection, discernible fungal colonies developed on the surface of the plant. To understand the pathogenic mechanisms of these fungi, an investigation was conducted into the functions of genes that may contribute to their pathogenesis.

Targeted gene expression has been effectively diminished in various powdery mildew fungi through the application of RNA interference techniques, including host-induced gene silencing (HIGS) and spray-induced gene silencing (SIGS) (Martínez-Cruz et al. 2018; Ruiz-Jiménez et al. 2021). The inability of powdery mildew fungi to grow on artificial culture media precludes the application of conventional transformation methods typically utilized for model eukaryotic organisms. To facilitate RNA interference in these fungi, RNA molecules specifically designed to target fungal RNAs are either produced by transgenic host plants or synthesized in vitro; subsequently, these RNA molecules are delivered into fungal cells through the host plants or via external spraying (Wang et al. 2016; Qiao et al. 2021; Cao et al. 2024). The RNA molecules that are delivered undergo processing by Dicer enzymes, resulting in the formation of small interfering RNAs. These small RNAs subsequently participate in the RNA-induced silencing complex, where they facilitate the degradation of target RNAs.

Many phytopathogens secrete effector proteins that manipulate the plant immune system to facilitate parasitism (Lo Presti et al. 2015). Utilizing whole-genome sequencing in conjunction with large-scale proteome analysis, candidate secreted effector proteins (CSEPs) of powdery mildew fungi were identified based on the criteria that necessitate the presence of a signal peptide in each CSEP, while also ensuring the absence of homologs in non-mildew fungi (Bindschedler et al. 2016). The genomes of barley and wheat powdery mildew fungi (Blumeria graminis f. sp. hordei and Blumeria graminis f. sp. tritici) encoded 722 and 734 CSEPs, respectively (Pedersen et al. 2012; Bourras et al. 2018). The powdery mildew fungus Erysiphe quercicola (formerly classified as Oidium heveae) infects rubber trees (Hevea brasiliensis), which are the primary source of natural rubber, a crucial industrial material (Liyanage et al. 2017). E. quercicola encodes 133 CSEPs, which are lower than that of B. graminis, likely due to a slower evolutionary rate of resistance in the host plant (Liang et al. 2018). Recently, an E. quercicola CSEP (CSEP01276) was found to disrupt the function of the 9-cis-epoxycarotenoid dioxygenase (NCED) enzyme, which plays a critical role in the biosynthesis of abscisic acid (ABA) (Li et al. 2020).

Plant hormones, including salicylic acid (SA), ethylene, and jasmonate (JA), regulate immunity against pathogenic microbes; however, certain phytopathogens have demonstrated the ability to circumvent hormone-mediated plant defenses (Pieterse et al. 2012; Shigenaga et al. 2017). In plants, the biosynthesis of ethylene commences with the conversion of L-methionine (L-Met) and adenosine triphosphate (ATP) into S-adenosyl-L-methionine (SAM). This reaction is catalyzed by the enzyme S-adenosyl-L-methionine synthetases (SAMSs). Subsequently, the conversion of SAM to 1-aminocyclopropane-1-carboxylic acid (ACC) is catalyzed by ACC synthase (ACSs). Finally, ACC is transformed into ethylene by ACC oxidases (Broekaert et al. 2006; Gong et al. 2014; Zhao et al. 2021). The production of ethylene is directly correlated with the activity of these enzymes (Shigenaga et al.2017). Moreover, ethylene increases the expression of ethylene-responsive transcription factors (ERFs), resulting in stress responses, including immune activation in plants (Groen et al. 2013; Ravanbakhsh et al. 2018). Nevertheless, studies on how powdery mildew effectors manipulate hormones are lacking.

Here, we claim that ethylene signaling plays an important role in disease resistance against powdery mildew and report that an E. quercicola CSEP called ethylene-associated effector 1 (Eae1) reduces ethylene level in H. brasiliensis. Our results show that Eae1 functions to affect plant ethylene levels during infection. Eae1 interacts with the chloroplast translocase, HbAIG1, in chloroplasts. Furthermore, this interaction facilitates the degradation of HbSAMS5, which may contribute to the mechanism by which Eae1 suppresses ethylene accumulation.

Results

Ethylene contributes to the resistance of H. brasiliensis against E. quercicola

To determine whether ethylene production is associated with the resistance of H. brasiliensis to E. quercicola. The leaves were exogenously treated with ethephon solutions, an agent that slowly releases ethylene at pH > 4.0 (Stotz et al. 2000), followed by inoculation with E. quercicola. Treatment with ethephon at concentrations of 25, 50, 75, and 100 μM significantly reduced the infection by E. quercicola at 7 days post-inoculation (dpi), with the 50 μM concentration exhibiting the most pronounced inhibitory effect (Fig. 1a, b). Similarly, treatment with 1 μM ACC, a direct precursor of ethylene, enhanced the resistance of H. brasiliensis to E. quercicola. In contrast, treatment with 1 μM aminoethoxyvinylglycine (AVG) (Zhai et al. 2022), an ethylene biosynthesis inhibitor, resulted in an increased susceptibility to infection. A higher density of hyphae and haustoria was observed on the leaves treated with AVG compared to those treated with a mock solution (distilled water) at 3 and 5 dpi (Additional file 1: Figure S1).

Fig. 1

Exogenously applied ethylene positively regulates H. brasiliensis resistance to E. quercicola. a Exogenously applied ethephon reduced H. brasiliensis infection. Ethephon at different concentrations was applied to the leaves for 24 h, followed by inoculation with E. quercicola conidia. Mock is distilled water used as the solvent of ethephon. Photographs were captured at 7 days post-inoculation (dpi). b Lesion area in leaves treated with different concentrations of ethephon was quantified at 7 dpi. Statistical differences are marked with asterisks compared to the mock (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test). c Reactive oxygen species (ROS) accumulation was visualized by 0.1% 3,3’-diaminobenzidine (DAB) staining in the infected leaf tissues with mock and 50 μM ethephon treatments. Photographs were captured at 48 h post-inoculation (hpi). Scale bars: 50 µm. Red arrows indicate the represented ROS accumulation. d Callose deposition was visualized by aniline blue staining in H. brasiliensis leaves treated with mock and 50 μM ethephon. Scale bars: 50 µm. Red arrows indicate the represented callose deposition. e, f The proportion of infected sites with ROS accumulation and callose deposition, respectively. Statistical differences are marked with asterisks compared to the mock (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test). g Abscisic acid (ABA), salicylic acid (SA), jasmonate (JA), and 1-aminocyclopropane-1-carboxylic acid (ACC) elicited by EqTub2-silenced strain in H. brasiliensis (FW: Fresh weight). Statistical differences are marked with asterisks compared to the samples at 0 hpi. (n = 9 leaf discs from 3 independent experiments. * P < 0.05, ** P < 0.01, Student’s t-test)

In leaves treated with 50 μM ethephon, the accumulation of reactive oxygen species (ROS) and callose deposition, the hallmarks of plant immune response, were significantly induced at 48 h post-inoculation (hpi) (Fig. 1c–f). ROS accumulation and callose deposition were detected using 0.1% 3,3’-diaminobenzidine (DAB) and aniline blue staining, respectively, as described previously (Xiao et al. 2003). Consistently, ACC treatment promoted ROS accumulation and callose deposition (Additional file 1: Figure S2).

Taken together, both exogenously applied and endogenous ethylene can activate the resistance to E. quercicola.

Ethylene responds to infection by E. quercicola

To investigate which phytohormones respond to the infection by E. quercicola, we attempted to use a strain with reduced pathogenicity for inoculation because the wild-type strain may effectively impact the production of phytohormones. β-tubulin plays a critical role in microtubule cytoskeleton assembly, which is essential for cellular growth and differentiation (Steinberg and Fuchs 2004). We generated a strain in which the β-tubulin gene (EqTub2) was silenced through the application of dsRNA corresponding to the 1–282 bp region of EqTub2 cDNA (Additional file 1: Figure S3a). Using liquid chromatography to analyze the phytohormone contents, we found that the EqTub2-silenced strain strongly induced the elevation of SA and ACC at various time points (Fig. 1g). ACC content peaked at 5 dpi, and SA content peaked at 48 hpi. Moreover, SA and ACC contents remained relatively high at 7 dpi. In contrast, the elevation of ABA and JA levels was not induced (Fig. 1g). These results support the involvement of ethylene in the regulation of plant resistance against E. quercicola.

CSEP01552 plays a regulatory role in plant ethylene

We attempted to identify the key CSEPs that affect ethylene production in E. quercicola. Previous studies showed that oomycete elicitor INF1 can upregulate genes for ethylene production and signaling in the tobacco plant Nicotiana benthamiana (Yang et al. 2016). Here, INF1 was co-expressed with a tested CSEP (GFP-fusion version) in this plant to determine which CSEP can suppress the hypersensitive response (HR) induced by INF1. Meanwhile, CSEP01276, a previously reported CSEP from E. quercicola known to suppress HR (Li et al. 2020), was used as the positive control. Following the co-expression, we found that several CSEPs, including CSEP01552, suppressed the HR (Additional file 1: Figure S4a and Additional file 2: Table S1). Additionally, when tested individually, these CSEPs did not induce ROS accumulation or HR ( Additional file 1: Figure S4b and S5; Additional file 2: Table S1). Quantitative reverse transcription (qRT)-PCR results showed that CSEP01552-GFP expression significantly inhibited the upregulation of ethylene-associated genes, including NbACO1a, NbACS1, NbEIN2, and NbERF5 (Imano et al. 2022). SA-associated genes, including NbPR1 and NbNPR1 (Tateda et al. 2014), were also upregulated by INF1, but co-expression of CSEP01552-GFP with INF1 did not drastically alter NbPR1 and NbNPR1 levels (Additional file 1: Figure S4c).

Treatment with the bacterial flg22 peptide or the Pseudomonas syrinage pv. tomato (Pst) DC3000 has been reported to induce ethylene production in Solanum lycopersicum (Xiao et al. 2007). We infiltrated the flg22 peptide or DC3000 into N. benthamiana, resulting in elevated levels of ethylene and ACC (Additional file 1: Figure S4d, e). When CSEP01552-GFP was expressed 24 h after the infiltration, we found that CSEP01552-GFP expression inhibited ethylene and ACC accumulation in N. benthamiana (Additional file 1: Figure S4d, e).

To determine whether CSEP01552 affected ethylene production in H. brasiliensis-E. quercicola interaction, the CSEP01552 gene was silenced through the application of dsRNA corresponding to the 23–286 bp region of cDNA (Additional file 1: Figure S3b), and the ethylene content in the infected sites of the strain was measured at 5 dpi (Fig. 2a). The results indicated that the silencing of CSEP01552 caused higher ethylene content than did wild-type strain and the strain treated with GFP-dsRNA (corresponding to 207–494 bp region of GFP cDNA). Treatment with AVG fully restored the growth and infection ability of the CSEP01552-silenced strain (Fig. 2b, c). The lesions were expanded, and more haustoria were produced in the sites inoculated with this strain under AVG treatment (Fig. 2d, e). Meanwhile, this treatment resulted in a minimal restoration of growth and infection capability of the EqTub2-silenced strain, likely attributable to the more significant roles of β-tubulin in cellular physiology (Fig. 2b, c). Thus, CSEP01552 (hereafter referred to as candidate ethylene-associated effector 1, Eae1) is a critical factor for suppressing ethylene production in plants.

Fig. 2

Silencing CSEP01552 affects ethylene in H. brasiliensis. a Ethylene content at the infected sites inoculated with CSEP001552-silenced (-CSEP001552) strain was elevated compared to that of the sites inoculated with the wild-type (WT) fungi. The samples were collected at 5 dpi and subjected to high-performance liquid chromatography-mass spectrometry (HPLC–MS/MS) analysis. Statistical differences are marked with asterisks (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test). b The pathogenicity of EqTub2- and CSEP01552-silenced strains in H. brasiliensis leaves treated with mock (distilled water) or aminoethoxyvinylglycine (AVG). Photographs were captured at 7 dpi. c Lesion sizes were determined at 7 dpi. Statistical differences are marked with asterisks when two groups are compared (n = 9 leaf discs from three independent experiments. ** P < 0.01, Student’s t-test). d Microscopic analysis of haustoria produced by strains in the leaves with mock and AVG. Zoom-in panels are magnified views of the sites labeled with white boxes. Scale bars: 50 µm. e The quantification of the haustoria. Statistical differences are marked with asterisks when two groups are compared (n = 9 leaf discs from three independent experiments. ** P < 0.01, Student’s t-test)

Eae1 can suppress plant immune response

Utilizing the Basic Local Alignment Search Tool (BLAST) within the National Center for Biotechnology Information (NCBI) database, we identified that Eae1 possesses homologs in 25 additional species of powdery mildew fungi (Additional file 2: Table S2). To investigate the transcript levels of Eae1 in response to ethylene-mediated immunity, we extracted RNA from E. quercicola, which was inoculated onto leaves with or without ethephon treatment, and quantified the transcript levels using qRT-PCR analysis. The results indicated that Eae1 expression was significantly upregulated at 48 hpi and 5 dpi in both the mock (distilled water) treatment and the ethephon treatment (Additional file 1: Figure S6a). Notably, ethephon treatment resulted in a higher level of Eae1 expression compared to the mock treatment at these time points (Additional file 1: Figure S6a). Additionally, the expression levels of the three other CSEP genes, which were found to inhibit INF1-induced HR (Additional file 2: Table S1), were also assessed in this experiment (Additional file 1: Figure S6b–d). These genes exhibited elevated expression levels at 24 or 48 hpi, but not at 5 dpi, and ethephon treatment did not significantly alter the transcript patterns of these CSEPs. These findings suggest that Eae1 expression is positively correlated with the strength of ethylene-mediated immunity.

The plant immune response to Eae1-silenced (-Eae1) strain was examined. The degree of pathogenicity in H. brasiliensis leaves inoculated with Eae1-silenced stains was significantly decreased, compared with mock (RNase-free water) and GFP-dsRNA treatment stains.(Fig. 3a, b). The ROS production and callose deposition were increased in the -Eae1 strain by DAB and aniline blue staining (Fig. 3c–e). These results indicated that the -Eae1 strain has defects in suppressing the immune response.

Fig. 3

Silencing of Eae1 decreased E. quercicola pathogenicity. a Eae1 silencing by dsRNA application reduced infection. Photographs were captured at 7 dpi. Mock is RNase free water used as the solvent of dsRNA. b Lesion area in leaves inoculated with E. quercicola conidia treated with mock, GFP- and Eae1-dsRNA were quantified at 7 dpi. Statistical differences are marked with asterisks compared to the mock (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test). c Representative photographs show DAB and aniline blue staining to visualize the infected sites with ROS and callose accumulation, respectively. Scale bars: 50 µm. Red arrows indicate the represented ROS accumulation and callose deposition, respectively. d, e The proportions of the infected sites with ROS and callose accumulations were quantified. Statistical differences are marked with asterisks compared to the mock samples. (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test)

Eae1’s secretory characteristics and chloroplast localization

Powdery mildew fungal CSEPs were predicted to possess an N-terminal signal peptide, which is a signature of secreted proteins. The signal peptide of Eae1 composed of 23 amino acids (aa 1–23) was predicted using the SignalP 6.0 server. We detected the activity of the signal peptide using the Saccharomyces cerevisiae YTK12, whichlacks invertase enzyme and Trp biosynthesis (Jacobs et al. 1997). We reconstructed the pSUC2 vector by combining the signal peptide with invertase and introduced this construct into YTK12. pSUC2 contains the Trp1 gene that permits YTK12 to grow on CWD-W culture medium lacking Trp. Furthermore, this strain was able to convert raffinose into sucrose, thereby facilitating its growth on YPRAA culture medium. Additionally, it converted 2,3,5-triphenyltetrazolium chloride (TTC) into a red-colored compound (Additional file 1: Figure S7). Therefore, the signal peptide of Eae1 is able to induce the secretion of invertase, similar to the signal peptide of the previously reported secreted effector protein Avr1b (Dou et al. 2008; Yin et al. 2018). These results suggested that Eae1 is a secreted protein.

We investigated the subcellular localization of Eae1 by expressing Eae1-GFP in N. benthamiana leaves. Eae1-GFP was found to localize to both the chloroplasts and cytoplasm (Fig. 4a). Following the isolation of protoplasts from these leaves, the localization of Eae1-GFP in the chloroplasts was also confirmed (Fig. 4b). In contrast, the control protein GFP was found to have minimal localization within the chloroplasts (Fig. 4a, b). Furthermore, a recombinant Eae1 (Eae1_NbPR1-SP), which used the PR1 signal peptide from N. benthamiana to substitute the native signal peptide of Eae1, significantly impaired chloroplast localization compared with Eae1 containing the native signal peptide (Fig. 4a, b). These results suggest that the chloroplast translocation peptide may be associated with the N-terminal region of Eae1.

Fig. 4

Eae1 localization in plant cells. a Microscopic analysis of GFP, Eae1-GFP, and Eae1_NbPR1-SP-GFP in N. benthamiana. White arrows indicate the chloroplast-localized Eae1-GFP. Scale bars: 50 µm. b Microscopic analysis of the Eae1-GFP and Eae1_NbPR1-SP-GFP in the protoplasts isolated from the N. benthamiana leaves. Scale bars: 50 µm. c The His-tagged Eae1 (Eae1-His) was incubated with the protoplasts of H. brasiliensis mesophyll cells for 6 h, and then the chloroplast faction was extracted for western blotting. Input, the His-tagged proteins detected pre-incubation; Protoplast, the total proteins of the protoplasts after incubation; Chloroplast, chloroplast fractions extracted from the protoplasts after incubation. Anti-RuBisCO was used as a loading control. A truncated GFP protein (168 aa) fused with His-tag (GFP₁₆₈-His) and Eae1_NbPR1-SP-His were used as controls. The experiments were conducted twice with similar results

We further investigated whether Eae1 can be translocated into the chloroplast of H. brasiliensis after secretion. Detecting Eae1 translocation during H. brasiliensis infection is difficult due to the low abundance of native Eae1. Therefore, we utilized an incubation assay involving His-tagged Eae1 (Eae1-His), which was purified in vitro. Eae1-His was incubated with the protoplasts isolated from H. brasiliensis mesophyll cells for 6 h. Following incubation, we isolated chloroplast fraction from the protoplasts and conducted western blotting with the chloroplast fraction (Fig. 4c). Additionally, a truncated GFP protein (168 aa) fused with His-tag (GFP₁₆₈-His) and Eae1_NbPR1-SP-His were used as controls. The results indicated that Eae1-His was detected in the chloroplast fraction, whereas the control proteins were not detected. Therefore, Eae1 may be localized to the chloroplast following its secretion.

The interactions among Eae1, HbAIG1, and HbSAMS5

We hypothesized that Eae1 interacts with plant proteins that are localized in the chloroplasts and have a role in ethylene biosynthesis. To test this hypothesis, we allowed Eae1-His to bind to Ni–NTA beads and conducted a pull-down assay to isolate the H. brasiliensis proteins that exhibit binding affinity for Eae1-His. These proteins were subsequently identified using liquid chromatography-tandem mass spectrometry (LC–MS/MS) analysis, which revealed the presence of a chloroplast translocase, HbAIG1 (XM_021794555.2), and a SAMS enzyme, HbSAMS5 (XM_058144837.1) (Additional file 2: Table S3). HbAIG1 is homologous to Toc159, which contains an AIG1 domain (Pfam: PF11886) and is localized to the outer membrane of the chloroplast (Nakai 2018). SAMSs serve as the key regulators of ethylene biosynthesis (Li et al. 2011; Gong et al. 2014).

We assessed the interaction between Eae1 and HbAIG1 or HbSAMS5 utilizing yeast two-hybrid (Y2H), bimolecular fluorescence complementation (BiFC), co-immunoprecipitation (Co-IP), and pull-down analyses. These results revealed a significant interaction between Eae1 and HbAIG1 (Fig. 5). Through BiFC, YFP signals indicative of this interaction were detected in the chloroplasts (Fig. 5b). Conversely, only Co-IP results indicated an interaction between Eae1 and HbSAMS5 (Additional file 1: Figure S8), which may suggest an indirect interaction between these two proteins. Additionally, the interaction between HbAIG1 and HbSAMS5 was confirmed by all the approaches employed (Additional file 1: Figure S9a–d). Notably, BiFC results also indicated that this interaction partially occurs within chloroplasts (Additional file 1: Figure S9b).

Fig. 5

The interaction between Eae1 and HbAIG1. a Yeast two-hybrid (Y2H) assays detecting the interaction between the Eae1 and HbAIG1. Positive colonies displayed a blue-color on the SD-LWHA medium with X-α-GAL. The combination of AD-T and BD-p53 was used as a positive control. -LW, SD medium lacking Leu and Trp; -LWHA, SD medium lacking Leu, Trp, His, and Ala. Red asterisk indicates the combination of Eae1 and HbAIG1. b Bimolecular fluorescence complementation (BiFC) verifies the interaction between Eae1 and HbAIG1 in N. benthamiana. Scale bars: 100 µm. White arrows indicate the represented interacted region. c The interaction between Eae1-GFP and HbAIG1-FLAG was determined by co-immunoprecipitation (Co-IP) assays. Input, total proteins extracted from N. benthamiana leaves co-expressing two tested proteins. IP, the proteins eluted from anti-GFP beads. Red asterisks indicate the interested bands. d The interaction between Eae1-His and HbAIG1-GST was determined by the pull-down assay. A truncated GFP protein (168 aa) fused with His-tag (GFP₁₆₈-His) was used as controls. Red asterisks indicate the interested bands. The experiments were conducted twice with similar results

We hypothesized that HbAIG1 may mediate the connection between Eae1 and HbSAMS5. To evaluate this hypothesis, we assessed the interaction between Eae1 and HbSAMS5 by incorporating HbAIG1 into the BIFC assay. When HbAIG1-FLAG was expressed in the leaves, YFP signals indicating the interaction between Eae1 and HbSAMS5 were observed (Additional file 1: Figure S8e). Therefore, HbAIG1 is able to facilitate this interaction.

HbSAMS5 is important for ethylene biosynthesis

A recent study found that H. brasiliensis can absorb in vitro-synthesized dsRNA applied to leaf surfaces, and the dsRNA homologous to H. brasiliensis MLO12 gene can transiently silence this gene (Li et al. 2023). To further validate the efficacy of SIGS in H. brasiliensis, the H. brasiliensis phytoene desaturase (HbPDS3) gene (XP_021691516.2) was successfully silenced using dsRNA corresponding to the 1428–1689 bp region of cDNA. This silencing led to a chlorophyll-deficient phenotype (Additional file 1: Figure S10). In Arabidopsis thaliana, PDS3 is known to be essential for chloroplast biogenesis (Foudree et al. 2010). Additionally, the H. brasiliensis leaves were treated with norflurazon (Foudree et al. 2010), which inhibits phytoene desaturase activity in carotenoid biosynthesis, or were cultured in the absence of light, both of which resulted in a chlorophyll-deficient phenotype (Additional file 1: Figure S10).

To investigate the role of HbSAMS5 in ethylene biosynthesis, we employed gene silencing of HbSAMS5 in H. brasiliensis through SIGS with dsRNA corresponding to the 34–296 bp region of cDNA. In leaves treated with HbSAMS5-dsRNA and inoculated with the EqTub2 strain, HbSAMS5 silencing significantly reduced ethylene or ACC content (Fig. 6a, b).

Fig. 6

HbSAMS5 has a function in ethylene biosynthesis. a, b The HbSAMS5 silencing by dsRNA application inhibited ethylene and ACC accumulations. The ethylene and ACC contents were determined at 5 dpi using HPLC–MS/MS (FW: Fresh weight). Statistical differences are marked with asterisks compared to the samples of WT (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test). c DAB staining showed that HbSAMS5-GFP expression and ACC treatment for 48 h induced ROS accumulation in N. benthamiana. d, e HbSAMS5 expression in N. benthamiana strongly enhanced the increase of ethylene and ACC under the condition of flg22 or DC3000 treatments. Statistical differences are marked with asterisks compared to the samples of mock (n = 9 leaf discs from 3 independent experiments. ** P < 0.01, Student’s t-test). f, g Either Eae1-GFP or Eae1-GFP combined with HbAIG1 inhibited ethylene and ACC accumulation in N. benthamiana, which was induced by HbSAMS5-FLAG. Statistical differences are marked with asterisks (n = 9 leaf discs from three independent experiments. ** P < 0.01, one-way ANOVA)

We conducted a further investigation into the role of HbSAMS5 by expressing HbSAMS5-GFP in N. benthamiana. HbSAMS5-GFP expression (Additional file 1: Figure S5), driven by the CAMV 35S promoter, caused ROS accumulation comparable to ACC treatment (Fig. 6c). Additionally, HbSAMS5 expression significantly raised ethylene or ACC levels in leaves treated with mock (MgCl₂ solution to suspend bacterial cells), bacterial flg22, or DC3000 (Fig. 6d, e). Therefore, it can be concluded that HbSAMS5 is crucial for promoting ethylene biosynthesis.

Eae1 promotes the destabilization of HbSAMS5

Given that Eae1 suppresses ethylene biosynthesis, we hypothesized that Eae1 may influence the activity of HbSAMS5. To investigate this, Eae1-GFP and HbSAMS5-FLAG were co-expressed in N. benthamiana. Results indicated that Eae1 inhibited the elevation of ethylene and ACC induced by HbSAMS5. Notably, the co-expression of Eae1-GFP and HbSAMS5-FLAG, along with HbAIG1-FLAG, enhanced this inhibition of ethylene and ACC levels (Fig. 6f, g).

We investigated whether Eae1 impacts HbSAMS5 abundance using N. benthamiana to express the tested proteins. Western blotting indicated that Eae1-GFP reduced the abundance of HbSAMS5-FLAG, whereas the unrelated effector CSEP01276-GFP and Eae1_NbPR1-SP-GFP did not exhibit any significant effect (Fig. 7a).

Fig. 7

Eae1 can reduce HbSAMS5 protein abundance. a Western blotting shows that co-expression of Eae1-GFP and HbSAMS5-FLAG resulted in a relatively low HbSAMS5-FLAG content in N. benthamiana. RuBisCO stained with CBB was used as a loading control. The red arrow indicates the lane with the loading of plant proteins containing Eae1-GFP and HbSAMS5-FLAG. GUS₃₇₀-FLAG, a truncated β-glucuronidase (GUS) protein (370 aa) fused with FLAG. b In vitro protein incubation assay, Eae1-His protein promoted the destabilization of HbSAMS5-GST when leaf cytosolic extract was added. The proteins were analyzed by western blotting pre- and post-incubation. RuBisCO stained with CBB represents the loading of leaf cytosolic extract. The red arrow indicates the lane with the loading of plant proteins containing Eae1-His and HbSAMS5-GST. c Eae1-His did not induce the destabilization of HbSAMS5-GST without the addition of leaf cytosolic extract following incubation. RuBisCO stained with CBB represents the loading of leaf cytosolic extract. d HbSAMS5-GST exhibited a decrease in response to increasing concentrations of Eae1-His. RuBisCO stained with CBB represents the loading of leaf cytosolic extract. Similar results were obtained from two independent experiments, and red asterisks indicate the interested bands. e Eae1_NbPR1-SP-GFP did not suppress INF1-induced HR. The values represent the leaves with HR in 9 tested leaves from 3 independent experiments

To further validate the impact of Eae1 on HbSAMS5 abundance, we conducted in vitro experiments using purified Eae1-His and HbSAMS5-GST proteins. These proteins were incubated with H. brasiliensis leaf cytosolic extract for a duration of 12 h, after which the levels of HbSAMS5-GST were quantified (Fig. 7b). The results indicated that HbSAMS5-GST abundance was reduced in the presence of Eae1-His. In contrast, the incubation with Eae1_NbPR1-SP-His and the His-fused signal peptide of Eae1 (Eae1_SP-His) did not produce any detectable effects on HbSAMS5-GST (Fig. 7b). Notably, when H. brasiliensis leaf cytosolic extract was excluded from the incubation mixture, Eae1-His did not lead to a decrease in HbSAMS5-GST abundance (Fig. 7c). Additionally, Eae1-His was tested at different concentrations, and complete destabilization of HbSAMS5-GST was observed at a concentration of 5 μg/mL of Eae1-His (Fig. 7d).

Based on the above, Eae1 does not function as a protease to directly hydrolyze HbSAMS5. Instead, Eae1 may preferentially facilitate the recognition and destabilization of HbSAMS5 by plant degradation systems.

Since we found that Eae1_NbPR1-SP was unable to promote HbSAMS5 destabilization, we also investigated whether Eae1_NbPR1-SP has the activity in suppressing INF1-induced HR. Unlike Eae1, Eae1_NbPR1-SP-GFP did not exhibit this activity (Fig. 7e). Therefore, the N-terminal region, which includes the native signal peptide, is required for the Eae1 function.

Discussion

Phytohormones play an important role in the regulation of plant immunity. Our results indicate that both exogenous and endogenous ethylene activate immunity against powdery mildew in H. brasiliensis. Treatment with ethephon that releases ethylene enhanced H. brasiliensis resistance, and the most obvious effect was observed at a concentration of 50 μM. It is probable that elevated concentrations of ethephon (greater than 50 μM) may surpass the threshold necessary for the optimal expression of resistance. Additionally, treatment with ACC, which is a direct precursor of ethylene and increases endogenous ethylene levels, also enhanced the resistance. In contrast, AVG treatment, which inhibits ACS activity within ethylene biosynthesis, resulted in a reduction of the resistance. Furthermore, treatment with ethephon and ACC also induced immune responses, including ROS accumulation and callose deposition, in the plant. When the EqTub2-silenced strain, which exhibits a defect in infection, was inoculated onto H. brasiliensis leaves, there was an accumulation of ACC. These phenomena suggest that ethylene responds to infection by this fungus.

We identified Eae1 (CSEP01552) as an effector that inhibits ethylene biosynthesis in plants. In N. benthamiana leaves co-expressing INF1 and Eae1, the INF1-induced HR was suppressed, and ethylene-associated genes were down-regulated. In contrast, Eae1 did not affect the expression of the SA-associated genes. Consistently, Eae1 inhibited the accumulation of ethylene and ACC in plants treated with flg22 or DC3000. Upon inoculation H. brasiliensis with the Eae1-silenced strain, there was an increased production of ethylene. Additionally, the infection ability of the Eae1-silenced strain was restored by AVG treatment.

Eae1 is likely translocated into the chloroplasts following its secretion. We observed the localization of Eae1 within chloroplasts following its expression in N. benthamiana. We also observed the translocation of Eae1 to chloroplasts following the incubation of Eae1 with leaf protoplasts from H. brasiliensis. The translocation of Eae1 to the chloroplasts may facilitate its interaction with the chloroplast protein HbAIG1. This interaction between Eae1 and HbAIG1 was confirmed through various approaches.

SAMSs play a crucial role in the conversion of Met to SAM with the ethylene biosynthesis pathway. The influence of Eae1 on ethylene biosynthesis is predominantly linked to its association with HbSAMS5. Although a direct interaction between Eae1 and HbSAMS5 was not observed in the Y2H, BiFC, and pull-down assays, Eae1 was found to coprecipitate with HbSAMS5 in the Co-IP assay. This suggests that Eae1 may interact weakly or indirectly with HbSAMS5. Additionally, the interaction between HbSAMS5 and HbAIG1 was identified through various interaction methods. We propose that HbAIG1 is a scaffold protein that connects Eae1 and HbSAMS5.

Our results suggest that HbSAMS5 plays a significant role in ethylene biosynthesis. The expression of HbSAMS5 resulted in an increase in the ACC/ethylene content in N. benthamiana. Furthermore, the transient silencing of HbSAMS5 through the application of dsRNA (SIGS) in H. brasiliensis led to a reduction in ethylene levels when the EqTub2-silenced strain was inoculated. Additionally, the expression of HbSAMS5 in N. benthamiana resulted in ROS accumulation, analogous to the effects observed with ACC treatment. These results suggest that HbSAMS5 acts as a SAMS in plants, thereby facilitating ethylene biosynthesis. Notably, when Eae1 was co-expressed with HbSAMS5 in N. benthamiana, it was observed that Eae1 could inhibit the HbSAMS5 activity by increasing ACC and ethylene levels. This inhibition was enhanced by the addition of HbAIG1. This assay provides additional evidence that Eae1 can interfere with HbSAMS5 activity.

This study also investigates the mechanism by which Eae1 impacts HbSAMS5 function. Co-expression of Eae1 and HbSAMS5 in N. benthamiana resulted in the destabilization of HbSAMS5. Additionally, this destabilization of HbSAMS5 was observed when Eae1 and HbSAMS5 were incubated with leaf extract from H. brasiliensis in vitro. Notably, Eae1 did not induce destabilization of HbSAMS5 in vitro when H. brasiliensis leaf extract was excluded. Consequently, we hypothesize that Eae1 may enhance the sensitivity of HbSAMS5 to undefined proteases rather than directly hydrolyzing HbSAMS5. It is known that at least two conserved degradation pathways, namely the ubiquitin–proteasome system and autophagy, are implicated in protein degradation processes (Schreiber and Peter 2014; Zientara-Rytter and Sirko 2016). Proteins designated for degradation by the 26S proteasome are frequently modified by lysine 48 (K48) ubiquitin, whereas those targeted for degradation through autophagy are typically modified by lysine 63 (K63) ubiquitin or are associated with cargo adaptors (Xu et al. 2009; Schreiber and Peter 2014). The modification facilitates the transport of proteins to the proteasome or vacuole for degradation (Xu et al. 2009; Schreiber and Peter 2014). We hypothesize that the presence of Eae1 enhances the modification of SAMS5, thereby increasing its recognition by protein degradation systems.

Several studies have reported that pathogens can manipulate plant SAMSs. For instance, the rice dwarf virus encodes the Pns11 protein, which specifically interacts with rice SAMS (OsSAMS1) to enhance its SAMS activity, as ethylene is known to facilitate viral infection (Zhao et al. 2017). In addition to SAMSs, other components and precursors of the ethylene biosynthesis pathway have been identified as targets of pathogen virulence effectors. Soybean ACSs (GmACS) are destabilized by the Phytophthora RXLR effector PsAvh238, thereby promoting infection (Yang et al. 2019a, b). Furthermore, plant ACC can be degraded by ACC deaminases produced by both pathogenic and nonpathogenic bacteria (Ravanbakhsh et al. 2018).

The role of ethylene in plant immunity depends on specific host–pathogen interactions and the interplay among various signaling pathways (Robert-Seilaniantz et al. 2011; Shigenaga et al. 2017). In several plant species, including Arabidopsis thaliana, the activation of the ethylene and JA pathways frequently enhances resistance to necrotrophic pathogens, such as Botrytis cinerea (Nie et al. 2017) and Alternaria brassicae (Ton et al. 2002). Conversely, the activation of the SA pathway is associated with increased resistance to biotrophic pathogens, such as P. syringae (Brooks et al. 2005; Spoel et al. 2007). However, this relationship is not universally applicable. Some studies indicate a synergistic relationship between the SA and ethylene/JA pathways. In addition to the SA pathway, the ethylene/JA pathways induce rice resistance to the hemibiotrophic fungal pathogen Magnaporthe oryzae (Broekaert et al. 2006; Peng et al. 2012). Inoculation of rice with M. oryzae can concurrently induce the accumulation of SA and JA (Iwai et al. 2007). The nucleotide-binding leucine-rich repeat (NLR) immune receptor PigmR enhances rice resistance to M. oryzae by promoting ethylene biosynthesis (Zhai et al. 2022). Furthermore, the transcription factor OsEIL3, which responds to ethylene, preferentially activates the transcription factor OsERF040, which positively regulates the resistance to M. oryzae (Zhu et al. 2024). This study reveals the relationship between a powdery mildew fungus and plant ethylene-mediated immunity. The controlled enhancement of ethylene levels may be utilized as a strategy for the management of powdery mildew disease.

Conclusions

Ethylene enhances the resistance of H. brasiliensis to E. quercicola. The effector Eae1 from E. quercicola inhibits ethylene levels by forming a complex with HbSAMS5, a critical component in the ethylene biosynthesis and HbAIG1. HbAIG1 is a chloroplast-localized protein that functions to facilitate the connection between Eae1 and HbSAMS5, leading to HbSAMS5 degradation within the plant. The mechanism of ethylene accumulation and immune response in host plants may result from the formation of this complex.

Methods

Experimental materials and growth conditions

The seedlings of powdery mildew-susceptible rubber tree cultivar (CATAS 7-33-97) were grown in a greenhouse at 24 ℃ under a 16/8 h light/dark cycle at approximately 60% relative humidity. The 5–14 d-old leaves of the seedlings were inoculated with E. quercicola strain HO-73 conidia to promote fungal growth. Nicotiana benthamiana seeds were vernalized at 4 ℃ for 2 d and transplanted into soil, and the seedlings were grown for approximately 4 weeks at 24 ℃ under a 16/8 h light/dark cycle.

Bioinformatics analysis

Signal peptides and protein domains were predicted using Simple Modular Architecture Research Tool (SMART) analysis (http://smart.embl-heidelberg.de) (Letunic et al. 2004), Pfam (http://pfam.xfam.org/) database, and SignalP 6.0 server (https://services.healthtech.dtu.dk/services/SignalP-6.0/) (Emanuelsson et al. 2007). The target gene sequences were searched on the NCBI database (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and BLAST.

RNA extraction and qRT-PCR analysis

Each RNA sample was extracted from 0.5 mg E. quercicola or 1 mg H. brasiliensis leaves using Fungal RNA extraction kit (OMEGA) or Plant RNA extraction kit (OMEGA), and cDNA was synthesized using the PrimeScript™ II 1st Strand cDNA Synthesis Kit (TaKaRa) for qPCR, following the manufacturer’s instructions. qRT-PCR was performed on the QuantStudio™ 5 real-time PCR instrument (Thermo Fisher Scientific) with the following PCR conditions: 95 ℃ for 5 s followed by 40 cycles of 95 ℃ for 5 s, 58 ℃ for 30 s, and 72 ℃ for 30 s. The E. quercicola EF-1a and H. brasiliensis actin genes (GenBank accession: HO004792) were used as reference genes accordingly.

Pathogenesis assay

For ROS staining, leaves were cultured in DAB solution at 37 ℃ for 8 h in the dark (Xiao et al. 2003). Subsequently, leaves were decolorized in a decolorization solution (ethanol:acetic acid = 3:1, v/v) for 4 h and then incubated with 1% aniline blue solution (0.2 g aniline blue, 1 g K₂HPO₄, and 20 mL ddH₂O) to stain the fungal organism on the leaves. For callose staining, the leaves were treated with the decolorization solution until colorless, stained with the aniline blue solution for 2 h, and washed with 1 × phosphate-buffered saline (PBS) (Xiao et al. 2003). Samples with ROS and callose staining were analyzed using bright-field (Nikon Y-TV55) and 4,6-diamidino-2-phenylindole (DAPI) channel fluorescence microscopy (emission/excitation: 488 nm/340 nm) (Nikon Y-TV55), respectively. To observe haustoria, infected leaves were decolorized with the decolorization solution (ethanol: acetic acid = 3:1, v/v), and then the fungal organism was stained with 1% aniline blue solution.

Determination of plant hormones

Liquid chromatography was used to measure the concentrations of ACC, ABA, SA, and JA in the plant samples using the methods described previously (Liu et al. 2019; Yang et al. 2019a, b; Huang et al. 2023). The plant samples were homogenized, reconstituted with an acetonitrile water solution, and filtered through a 0.22 μm membrane filter. Then, 10 μL from each sample was injected into an HPLC–MS/MS system (Agilent 1290 and AB Sciex QTRAP 6500⁺). The concentrations of the plant hormones in the samples were calculated based on the peak area (Additional file 1: Figure S11a) and the calibration curve (Additional file 1: Figure S12a–d).

The ethylene concentration in the plant samples was measured by gas chromatography using the method as previously described (Wang et al. 2023a, b). The plant samples in gas-tight vials were incubated at 25 °C for 4 h to allow the release of ethylene. The gas samples were injected into a Thermo Scientific™ TRACE™ 1300 gas chromatograph equipped with a flame ionization detector. The concentrations of ethylene in the sample gas were calculated based on the peak area (Additional file 1: Figure S11b) and the calibration curve (Additional file 1: Figure S12e).

Transient protein expression in N. benthamiana

Recombinant plasmids were introduced into the A. tumefaciens strain GV3101 (pMP90). Transformed bacteria cells were selected using Luria–Bertani (LB) medium plates containing kanamycin and rifampicin and cultured in liquid LB medium at 28 ℃ for 12 h. These cells were resuspended in 10 mM MgCl₂ to OD₆₀₀ = 0.6, and infiltrated into the leaves of 4-week-old N. benthamiana plants using a syringe. The plants were cultured at 28 ℃ for 48 h, and the target protein expression was determined by fluorescent signal detection and western blotting. To observe Eae1-GFP localization in the N. benthamiana protoplasts, the protoplasts were isolated from the leaves expressing Eae1. The leaves were cut into small discs and incubated with the solution containing macerozyme (R10) and cellulase (R10), as previously described (Fu et al. 2018). In the assays with INF1 expression and the infiltration with DC3000, the leaves were infiltrated with the strain of Eae1-GFP or GFP 24 h prior to infiltrating the same leaf areas with the strain carrying the INF1 vector or with the DC3000 strain (suspended in 10 mM MgCl₂ to OD₆₀₀ = 0.2).

Functional validation of effector signal peptides

The recombinant pSUC2-Eae1^SP vector was transformed into the yeast strain YTK12 and screened on CMD-W medium. The positive colonies were allowed to grow on YPRAA medium to verify invertase secretion activity. The transformants converted TTC to a red-colored compound when the signal peptide drove invertase secretion (Wang et al. 2023a, b).

The dsRNA application and pathogenicity assay

The dsRNA samples used for the gene silencing assays were synthesized using a T7 RNAi transcription kit (Vazyme). All the in vitro-synthesized dsRNA samples were subjected to agarose gel analysis (Additional file 1: Figure S13). The primers used for dsRNA amplification contained the T7 promoter sequence at the 5’ end. The dsRNA was purified using RNA-affinity beads (Vazyme) to remove unincorporated nucleotides and enzymes. To silence Eae1 and EqTub2 in E. quercicola, seven-day-old H. brasiliensis leaves were inoculated with a suspension of fresh powdery mildew conidia (1 × 10⁵ spores/mL) using a pipette. The dsRNA solutions (20 ng/μL) were dropped onto the inoculation sites to treat the fungi twice at 0 and 48 hpi (McRae et al. 2023). To induce the silencing of HbPDS3 and HbSAMS5, 7–10 d-old H. brasiliensis leaves were sprayed with dsRNA solutions (20 ng/μL) twice (0 and 48 h) using a sprayer (Li et al. 2023). In the assay performed to determine the infection rate of Eae1- or EqTub2-silenced strain in H. brasiliensis leaves treated with AVG (Aladdin), 1 μM AVG solution was sprayed onto seven-day-old-leaves for 48 h, followed by inoculation with E. quercicola and treatment with Eae1- or EqTub2-dsRNA.

Protein interaction assays

To obtain the potential proteins binding with Eae1, purified Eae1-His protein (5 μg) was incubated with the Ni–NTA beads (Thermo) for 2 h, followed by incubating with 40 μg H. brasiliensis leaf protein extraction for 2 h. Proteins were eluted from the beads using 100 μL 1 × PBS buffer and subjected to LC–MS/MS analysis (Agilent), which was performed by the protein analysis facility of Beijing Genomics Institute (China). The Y2H assay was performed as previously described (Zhao et al. 2022), while Co-IP assays were conducted using a modified version of the method (Li et al. 2020). Two proteins fused with GFP or FLAG were co-expressed in 4-week-old N. benthamiana by Agrobacterium-mediated transformation. For each sample, the total proteins of the plant tissues (0.5 g) were extracted, and anti-GFP agarose beads (Chromotek) were used to bind the GFP-tagged proteins. Proteins were eluted from the beads using 100 μL 1 × PBS buffer, and 20 μL elution was analyzed by western blotting. BiFC assays were conducted using the method previously described (Li et al. 2020). Two proteins fused with N’YFP or C’YFP were co-expressed in 4-week-old N. benthamiana using Agrobacterium-mediated transformations. Chloroplast detection (emission/excitation: 470 nm/680 nm) and YFP detection were performed 48 h after Agrobacterium infiltration (emission/excitation: 514 nm/525 nm). Pull-down assays were performed using a modified version of the method (Zhao et al. 2022). A His-tagged protein purified from Escherichia coli BL21 (DE3) strain (5 μg) was incubated with the Ni–NTA beads (Thermo) for 2 h, followed by 2 h-incubation of a GST-tagged protein with the beads (Thermo). Proteins were eluted from the beads using 100 μL 1 × PBS buffer, and 10 μL elution was analyzed by western blotting.

Protein extraction and western blotting

To produce proteins in vitro, the pET28a and pGEX-6p-1 vectors carrying the coding sequences were introduced into E. coli BL21 (DE3) strain. The proteins of interest were produced in the transformed bacteria strain induced with 0.8 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 16 ℃ for 12 h. The bacterial cells were resuspended in 1 × PBS and subjected to sonication using a sonicator. The proteins of interest were purified using Ni–NTA and GST resin beads (Thermo). Total proteins of N. benthamiana were extracted from 0.2 g leaf tissue using the Plant Protein Extraction Kit (CWBIO). Chloroplast fractions were isolated from 0.2 g N. benthamiana leaf tissue using the Chloroplast Isolation Kit (Invent), according to the manufacturer’s protocol. The H. brasiliensis leaf mesophyll protoplasts were isolated according to the method as previously described (Zhang et al. 2016). The proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene fluoride (PVDF) membranes, and blocked in skim milk. The membranes were then washed with 1 × PBS buffer to remove unbound primary antibodies and incubated with the secondary antibody solution for 1 h at 16 ℃. The signals were detected using an iBright FL1500 instrument (Invitrogen).

The incubation of purified proteins with H. brasiliensis cytosolic extract

H. brasiliensis leaf cytosolic extract was obtained using the Plant Cytosolic Isolation Kit (Invent) according to the manufacturer’s protocol. The proteins (1, 2, or 5 μg/mL) purified from E. coli BL21 (DE3) were incubated with the leaf cytosolic extract (10 μg/mL) for 12 h at 16 ℃. The mixtures (1 μg per sample) were loaded to SDS-PAGE and analyzed using western blotting.

The primers and gene sequences used in this study

The primers used for qRT-PCR and dsRNA are listed in Additional file 2: Tables S4 and S5. Details of the plasmids are shown in Additional file 2: Table S6. The cDNA sequences are listed in Additional file 2: Table S7.

Data analysis

Student’s t-tests were performed to compare the means between two data groups. One-way analysis of variance (ANOVA) with Tukey’s test was used to compare multiple data groups. P value < 0.01 or 0.05 represents a significant difference.