XopM, An FFAT Motif‐Containing Type III Effector Protein From Xanthomonas, Suppresses MTI Responses at the Plant Plasma Membrane Authors: Charlotte Brinkmann, Jennifer Bortlik, Margot Raffeiner, Manuel González‐Fuente, Linus F Börnke, Suayib Üstün, Frederik Börnke Published: 2024 ABSTRACT ABSTRACTMany gram‐negative pathogenic bacteria use type III effector proteins (T3Es) as essential virulence factors to suppress host immunity and to cause disease. However, in many cases the molecular function of T3Es remains unknown. The plant pathogen Xanthomonas campestris pv. vesicatoria (Xcv) is the causal agent of bacterial spot disease on tomato and pepper plants and is known to translocate around 36 T3Es into its host cell, which collectively suppress plant defence and promote infection. XopM is an Xcv core T3E with unknown function that has no similarity to any other known protein. We found that XopM interacts with vesicle‐associated membrane protein (VAMP)‐associated proteins (VAPs) in an isoform‐specific manner. The endoplasmic reticulum (ER) integral membrane protein VAP is a common component of membrane contact sites involved in both tethering and lipid transfer by binding directly to proteins containing an FFAT (two phenylalanines [FF] in an acidic tract [AT]) motif. Sequence analyses revealed that XopM displays two FFAT motifs that cooperatively mediated the interaction of XopM with VAP. When expressed in plants, XopM supported growth of a nonpathogenic bacterial strain and dampened the production of reactive oxygen species, indicating its ability to suppress plant immunity. Further analyses revealed that the interaction with VAP and the ability to suppress microbe‐associated molecular pattern‐triggered immunity (MTI) are structurally and functionally separable, although XopM requires localisation to the host membrane system for full MTI suppression activity. We discuss a working model in which XopM uses FFAT motifs to target the membrane to interfere with early MTI responses.Keywords: FFAT motif, plant defence, type III effector, VAP proteins, Xanthomonas, XopM Introduction Plants can detect attempted invasion by microbial pathogens by perceiving microbe‐associated molecular patterns (MAMPs) through transmembrane cell surface‐localised pattern recognition receptors (PRRs). Upon activation by MAMP binding, PRRs initiate a number of cellular responses, including the generation of reactive oxygen species (ROS), the initiation of MAP kinase signalling, production of defence hormones such as salicylic acid (SA), and large‐scale transcriptional reprogramming followed by the production of antimicrobial proteins and metabolites, as well as cell wall fortification by callose deposition at the cell–pathogen interface (DeFalco and Zipfel 2021). Collectively, these responses are sufficient to prevent multiplication and spread of a broad range of potential pathogens and thus result in MAMP‐triggered immunity (MTI). Plant bacterial pathogens mostly multiply within the apoplastic space of the infected tissue, and therefore, an effective defence response is heavily dependent on the extracellular release of defence‐related molecules, such as pathogenesis‐related proteins, secondary metabolites and cell wall materials. Thus, it does not come as a surprise that secretory and endocytic membrane trafficking pathways lie at the heart of plant's immune system (Jeon and Segonzac 2023; Gu, Zavaliev, and Dong 2017; Yun and Kwon 2017). The plant's early secretory pathway comprises the endoplasmic reticulum (ER) and the numerous cisternal stacks of the Golgi apparatus. As such, the ER plays a pivotal role in plant immunity by serving as a key site for the synthesis, folding, and quality control of proteins involved in defence responses, thus ensuring the proper functioning of immune receptors and defence molecules (Schäfer and Eichmann 2012). The ER establishes direct membrane contact sites with the plasma membrane (PM) at so‐called ER–plasma membrane contact sites (EPCS). In plants, EPCSs serve as critical hubs for phospholipid homeostasis and cell integrity (Schapire et al. 2008; Ruiz‐Lopez et al. 2021), are involved in endocytosis (Stefano et al. 2018) and autophagy (Wang et al. 2019), and regulate cell–cell transport at plasmodesmata (Levy, Zheng, and Lazarowitz 2015; Ishikawa et al. 2020). Vesicle‐associated membrane protein (VAMP)‐associated proteins (VAPs) are type II ER integral proteins that are conserved among phylogenetically distinct organisms and work as tethers between the PM and EPCSs (Stefan et al. 2011). Type II ER integral proteins VAPs are inserted into the ER membrane via their C‐terminus; however, they lack a domain for making contact in trans with the opposing membrane. Thus, it is assumed that VAPs rely on the interaction with other proteins to form multisubunit complexes and hence function as hubs to recruit numerous proteins to the ER (Murphy and Levine 2016). The architecture of VAPs is modular and composed of three conserved domains: an N‐terminal major sperm protein (MSP) domain followed by a central amphipathic domain that is predicted to comprise a coiled‐coil domain, and a C‐terminal hydrophobic transmembrane domain (TM). One primary class of VAP‐binding partners are cytoplasmic proteins containing two phenylalanines (FF) in an acidic tract (the FFAT motif), which binds specifically to the MSP domain (Loewen and Levine 2005; Kamemura and Chihara 2019). The canonical FFAT motif has a core of seven residues, EFFDAxE in single letter amino acid code, and an adjacent upstream region with multiple acidic or phosphorylated residues that additionally contribute to the motif (Loewen and Levine 2005; Murphy and Levine 2016; Slee and Levine 2019). However, most verified FFAT motifs deviate from the canonical motif in their core and/or upstream residues, indicating that a certain degree of sequence flexibility does not necessarily affect FFAT motif functionality (Slee and Levine 2019). In yeast, four VAP‐interacting proteins have experimentally been shown to contain FFAT motifs, and bioinformatics analysis suggest that the yeast proteome might contain up to 55 FFAT‐containing proteins capable of VAP binding (Slee and Levine 2019). Although a number of VAP‐interacting proteins have recently been identified in plants (Codjoe et al. 2022; Greer et al. 2020; Fox et al. 2020; Stefano et al. 2018; Saravanan et al. 2009; Wang et al. 2019, 2014), the role of FFAT motifs in mediating the binding of plant VAPs to their alleged protein partners is unknown as none of these proteins has yet been reported to contain a functional FFAT motif. In Arabidopsis, the VAP family comprises 10 members (VAP27‐1 to VAP27‐10) divided into three distinct phylogenetic clades (Wang et al. 2016). Members of clades I and III possess a single TM domain that determines their localisation to ER, while clade II members lack this domain and are assumed to localise to the PM (Wang et al. 2016). Clade I VAPs VAP27‐1 and VAP27‐3 were shown to localise to the ER as well as to EPCSs (Wang et al. 2014, 2016) and demonstrated to bind to clathrin at endocytic membranes (Stefano et al. 2018). Although VAPs appear central to membrane processes involved in pathogen defence, a general role for VAPs in plant immunity has so far not been established. To overcome MTI and to cause disease on a given host plant, adapted pathogens have evolved virulence factors that act to suppress plant immunity and promote infection. The injection of type III effector (T3E) proteins into the host cell is an efficient mechanism employed by many gram‐negative bacterial pathogens to overcome MTI and establish a favourable environment (Büttner 2016; Khan et al. 2018; Macho and Zipfel 2015). T3E proteins have diverse biochemical activities that interfere with host cellular processes on all levels of the plant's immune response, including ER‐mediated processes or vesicle trafficking (Jeon and Segonzac 2023; Yun et al. 2023). The gram‐negative bacterium Xanthomonas campestris pv. vesicatoria (Xcv; synonymously designated as Xanthomonas euvesicatoria ) is the causal agent of bacterial spot disease in pepper and tomato plants (Jones, Stall, and Bouzar 1998). The ability of Xcv to cause disease is largely dependent on a suite of approximately 36 T3Es, several of which are conserved between different Xcv strains (Potnis et al. 2011; Schwartz et al. 2015). The Xcv effectors XopB and XopJ localise to host cell membranes and have been shown to specifically inhibit MAMP‐induced callose deposition and block secretion of a green fluorescent protein (GFP) reporter fused to a signal peptide (Bartetzko et al. 2009; Schulze et al. 2012; Priller et al. 2016). The protein XopM has been identified as a type III secreted effector in Xcv (Schulze et al. 2012) and was shown to interfere with MAMP‐induced reporter gene expression in a protoplast system (Popov et al. 2016). XopM does not share significant sequence similarity with other effector proteins or with any protein found in databases and its biochemical activity, target protein in the host cell, and contribution to Xcv virulence remains unknown. In this study, we show that XopM interacts with plant VAPs in an isoform‐specific manner via two FFAT motifs within the N‐terminal part of the effector. XopM suppressed MTI responses when ectopically expressed in plants; however, the ability to affect plant defence was shown to be structurally separable and functionally independent from the VAP‐binding activity. Further protein–protein interaction studies suggest that XopM interacts with key players of membrane trafficking regulation. Thus, we discuss a model in which XopM uses FFAT motifs as targeting signals to reach certain subcompartments of the host endomembrane system to interfere with vesicular transport processes. Results XopM Suppresses Plant Immunity To test the contribution of XopM to Xcv virulence under laboratory conditions, we generated a ΔxopM deletion mutant strain in the Xcv 85–10 background and subsequently compared bacterial growth relative to the wild‐type strain in susceptible pepper plants. Six days after inoculation, the ΔxopM knockout strain reached similar cell titres as the control (Figure S1), indicating that the removal of this single effector protein from the overall Xcv effector repertoire does not affect bacterial virulence under the chosen experimental conditions. To test the effect of XopM expression in plant cells, stable transgenic Arabidopsis lines expressing a XopM‐GFP fusion protein under the control of a β‐estradiol‐inducible promoter were generated and expression of the fusion protein was verified by immunoblotting (Figure S2). Independent transgenic Arabidopsis lines supported enhanced bacterial growth of a nonpathogenic ΔhrcC Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) strain, strictly depending on the induction of XopM‐GFP expression by β‐estradiol treatment (Figure 1). This indicates that XopM‐GFP interferes with MTI responses in Arabidopsis. Bacterial growth in XopM‐GFP‐expressing Arabidopsis plants. XopM expression supports growth of the type III effector (T3E)‐deficient Pseudomonas syringae pv. tomato (Pst) ΔhrcC. Two‐week‐old Col‐0 and XopM‐GFP transgenic Arabidopsis seedlings were either treated with 50 μM β‐estradiol (+) to induce XopM‐GFP production or with 0.1% ethanol (−) as a control. The following day, the seedlings were infected by flooding with Pst ΔhrcC (5 × 106 cfu/mL). The cfu per mg of plant tissue was measured 3 days postinoculation (dpi). Individual values are plotted with the bar representing the mean of n ≥ 10. Error bars indicate the standard deviation (± SD). Letters indicate statistical significance (p < 0.05) calculated by a one‐way analysis of variance with Tukey's multiple comparison test. The experiment was carried out three times with similar results. Immune responses generally triggered by MAMPs include a burst of reactive oxygen species (ROS) (Kadota et al. 2014). When transiently expressed in leaves of Nicotiana benthamiana using Agrobacterium tumefaciens (hereafter, Agrobacterium), XopM fused to GFP at its C‐terminus suppressed an ROS burst triggered by flg22 treatment to a similar extent as the positive control AvrPto from Pst DC3000 (Figure 2A). In a complementary approach, we tested the ability of XopM to suppress the activity of extracellular peroxidase (POX) enzymes that are induced upon MAMP perception (Mott, Desveaux, and Guttman 2018). Similar to the observation before, XopM was able to suppress POX activity almost down to the level of the negative control (Figure 2B). Expression of all proteins was verified by immunoblotting (Figure S3). XopM interferes with reactive oxygen species (ROS) production and suppresses flg22‐induced peroxidase (POX). (A) flg22‐induced production of ROS was measured in leaf discs transiently expressing effector proteins using a luminol‐based assay. Leaf discs were treated with flg22 and ROS production measured as relative light units (RLUs) over the time course of an hour. Leaves infiltrated with MgCl2 were not treated with flg22 as a negative control (−flg22). At least 12 samples were measured (n ≥ 12), except for the negative control where only four samples were measured (n = 4). Individual values are plotted with bars representing the mean and error bars give the standard deviation (±SD). Statistical significance was calculated by one‐way analysis of variance (ANOVA) with Tukey's multiple comparison test. The different letters indicate significant differences (p < 0.05). (B) Relative POX activity of flg22‐treated leaf discs transiently expressing GFP‐tagged effector proteins. Leaf discs were treated with 100 nm flg22 for 16 h and POX activity was measured. Discs infiltrated with MgCl2 were not treated with flg22 (−flg22). At least 16 leaf discs (n ≥ 16) were measured, except for the negative control where n = 8. POX activity of free GFP was set to 100 and the relative activity of the others adjusted to it. Different letters indicate a significant difference (p < 0.05). One‐way ANOVA with Tukey's multiple comparison test was used. Boxes show the lower quartile, median and upper quartile values. Error bars represent a 95% confidence interval. The experiments were carried out three times with similar results. Taken together, these findings indicate that despite the fact that the individual deletion of XopM from Xcv does not affect bacterial growth under laboratory conditions, XopM possesses substantial MTI suppression activity in different plant species. Therefore, it appears highly likely that XopM targets are conserved in different plants irrespective of their susceptibility towards Xcv. XopM Interacts With VAP Proteins in Yeast and in Planta We next aimed to identify the cellular targets of XopM in the plant cell. To this end, we screened a yeast two‐hybrid (Y2H) cDNA library from tobacco ( Nicotiana tabacum ) for proteins that interact with XopM. Although tobacco is not an Xcv host species, the library has proven very useful to identify Xcv T3E target proteins (Üstün, Bartetzko, and Börnke 2013; Leong et al. 2022; Raffeiner et al. 2022). A screen of < 106 cfu yielded two candidate XopM interaction partners and a sequence analysis revealed that both encode VAP proteins. One was identical to VAP12 from Nicotiana tomentosiformis (GenBank acc. XM_009796889.1), and the other one matched VAP12 from Nicotiana sylvestris (GenBank acc. XP_009627005.1) when compared to the protein database. Thus, both sequences probably represent homologous genes derived from the two parental genomes of N. tabacum , which is an allotetraploid. Interaction assays in yeast between XopM and both VAPs confirmed the interaction (Figure 3A). Both proteins share 97% similarity on the amino acid sequence level with each other, and thus, we decided to proceed with only one of the clones for further analyses (acc. XM_009796889.1), which is subsequently named NtVAP12. The type III effector XopM interacts with VAP12. (A) XopM fused to the GAL4 DNA‐binding domain (BD) was co‐expressed with two tobacco VAP12 proteins fused to the GAL4 activation domain (AD). The empty vector containing the AD (pAD) was used as a negative control. – LT, yeast growth on medium without Leu and Trp. – HLT, yeast growth on medium lacking His, Leu and Trp, indicating expression of the HIS3 reporter gene. LacZ, activity of the lacZ reporter gene. (B) Confirmation of XopM VAP12 interaction in planta via co‐immunoprecipitation. Using Agrobacterium‐mediated transient expression, NtVAP12‐myc was transiently co‐expressed with free GFP or XopM‐GFP. After 24 h, the total protein content (Input) was treated with GFP‐beads (IP). Recombinant proteins were detected by immunoblotting using anti‐GFP or anti‐myc antibodies. Ns, Nicotiana sylvestris ; Nt, Nicotiana tabacum . To determine if XopM interacts with NtVAP12 in planta, a GFP pull‐down assay was performed. To this end, we transiently expressed either XopM‐GFP or free GFP each combined with or without NtVAP12‐myc in N. benthamiana. One day after infiltration with agrobacteria, we performed a pull‐down of XopM‐GFP using GFP‐Trap beads and analysed the eluates by immunoblotting with anti‐GFP and anti‐myc antibodies. XopM‐GFP, but not free GFP, was able to pull‐down NtVAP12‐myc, verifying the specific interaction of both proteins in planta (Figure 3B). XopM and VAP12 Colocalise to the ER and EPCS To study XopM protein intracellular localisation relative to the localisation of NtVAP12, XopM and NtVAP12 were tagged with mCherry and GFP, respectively, and transiently expressed in N. benthamiana leaves. The NtVAP12‐GFP signal decorated a reticular network, characteristic of the ER (Figure 4A). In addition, NtVAP12‐GFP fluorescence was also visible in immobile punctuated structures, reminiscent of the putative EPCS observed for AtVAP27 localisation in previous studies (Wang et al. 2014, 2016). The XopM‐mCherry signal showed a very similar distribution as the NtVAP12‐GFP signal, mostly highlighting the reticular ER in addition to punctate structures located within the reticular network (Figure 4A). When co‐expressed in the same cell, the XopM‐mCherry and the NtVAP12‐GFP signals displayed considerable overlap, including the regions representing possible EPCS (Figure 4B). Co‐expression with an ER marker protein (HDEL‐mCherry) confirmed that the reticulate localisation pattern of XopM and NtVAP12 indeed represents the ER (Figure 4B). Furthermore, co‐expression of either XopM or NtVAP12 with XopJ, an Xcv T3E that has previously been shown to associate with the PM (Bartetzko et al. 2009), suggests at least a partial localisation of both proteins to the PM, although the fluorescence signal is not as homogenously distributed as it is for XopJ (Figure 4B). This might further suggest a localisation of both proteins to specific subcompartments at the PM. Finally, both XopM and NtVAP12 displayed strong colocalisation with AtVAP27‐1, suggesting that both proteins colocalise with VAP27 at equivalent ER–PM contact sites in plants (Wang et al. 2014, 2016). XopM‐mCherry colocalises with NtVAP12‐GFP in plant cells. (A) Confocal microscopy pictures of transiently expressed GFP‐NtVAP12 and XopM‐mCherry in Nicotiana benthamiana epidermal cells. (B) Confocal microscopy pictures of transiently co‐expressed GFP‐NtVAP12, XopM‐GFP and YFP‐AtVAP27‐1, with XopM‐mCherry, endoplasmic reticulum marker HDL‐mCherry, plasma membrane marker XopJ‐RFP and RFP‐AtVAP27‐1 in N. benthamiana epidermal cells. Images were taken 30–48 h after infiltration. Colocalisation was measured using Pearson's correlation coefficient (ρ). The scale bar represents 10 μm. The expression of all fusion proteins used in the above experiments was verified by immunoblotting (Figure S4). Taken together, both fusion proteins localise to similar compartments of the cell within the endomembrane system, including EPCS. This further supports the notion of an in vivo interaction of both proteins in plant cells. XopM Interacts With VAPs From Different Plant Species The Arabidopsis genome encodes 10 VAP isoforms that phylogenetically fall into three different clades (Wang et al. 2016). Based on sequence similarity, the clade I VAPs AtVAP27‐1 and AtVAP27‐3 are the most closely related orthologues relative to NtVAP12 and both interact with XopM in yeast (Figure S5A,B). In addition, the clade III Arabidopsis VAP isoform AtVAP27‐4 also binds to XopM in yeast, while another clade I isoform, AtVAP27‐4, does not interact (Figure S5B). We also tested VAP12 homologues from other species for the ability to bind XopM in yeast. As shown in Figure S5B, N. benthamiana NbVAP12 and SlVAP12 from tomato also interact with XopM. In a similar vein, N. tabacum NtVAP2‐1, a clade III member most closely related to AtVAP27‐4 also binds XopM in yeast (Figure S5B). The data suggest that XopM interacts with VAPs from different species from at least two of the three phylogenetic clades. XopM–VAP Interaction Depends on the FFAT‐Binding Domain of VAP Structural predictions as well as mutation experiments have shown that the majority of VAP interaction partners bind via conserved residues within the MSP, especially those that carry an FFAT motif (Loewen and Levine 2005; Slee and Levine 2019). To map NtVAP12 residues critical for its interaction with XopM, an NtVAP12 truncation variant limited to the predicted MSP domain ranging from residues 8 to 129 of the protein (MSP NtVAP8‐129; Figure 5A) was tested for binding XopM in yeast. As shown in Figure 5B, the NtVAP12 MSP domain was sufficient for interacting with XopM. Multiple residues within the yeast Scs2 VAP protein, including T41 and T42, critical for binding protein partners, especially via their FFAT motif, were previously identified (Loewen and Levine 2005). Based on sequence comparison, the corresponding residues in NtVAP12 are T46 and T47. Alanine substitution of T47 abolished the interaction of NtVAP12 with XopM, while the substitution of T46 had no apparent effect on binding between the two proteins in yeast (Figure 5C). Similar to the T47A substitution, a T46/47A double exchange also abolished the interaction (Figure 5C). Control experiments for the interaction assays with empty vector combinations are summarised in Figure S6A. Structure–function analysis of the NtVAP12–XopM interaction. (A) Schematic representation of NtVAP12. NtVAP12 has a major sperm (MSP) domain at the N‐terminal, in the centre a coiled‐coiled domain (CCD) and a transmembrane domain (TMD) at the C‐terminus. (B) The MSP domain fused to the GAL4 activation domain (AD) was co‐expressed with the GAL4 binding domain (BD) fused to XopM. Empty BD vector (pBD) was used as a negative control. (C) Potential residues for binding XopM of NtVAP12 were substituted and co‐expressed with XopM in yeast. Empty AD vector (pAD) was used as a negative control. –LT yeast grown on a selective medium lacking Leu and Trp. –LTH, yeast grown on a selective medium lacking Leu, Trp and His, indicating HIS3 reporter gene activity. LacZ, activity of the lacZ reporter gene. Nt, Nicotiana tabacum. These data suggest that the binding of XopM to NtVAP12 might be mediated by a similar mechanism previously described for the interaction of VAPs with FFAT motif‐containing proteins (Slee and Levine 2019). XopM Contains Two FFAT Motifs That Collectively Mediate XopM–VAP Interaction To further explore the structural requirements for the XopM binding to VAPs, an algorithm designed by Murphy and Levine (2016) was used to predict the presence of possible FFAT motifs within the XopM polypeptide sequence. The analysis revealed the presence of two putative FFAT motifs located 23 amino acids apart within the N‐terminal region of the XopM protein (Figure 6A). One FFAT motif [G60Y61E62T63A64N65D66] is quite degenerated when compared to the consensus FFAT sequence [E1(F/Y)2(F/Y)3D4A5x6E7] and can be considered a noncanonical (nc)FFAT motif, while the other predicted motif [Q90F91Y92D93A94E95D96] matches the consensus sequence to a higher degree. Both motifs are preceded by a stretch of serine residues that, if phosphorylated, could provide the acidic environment typically located upstream of the consensus FFAT sequence (Murphy and Levine 2016). A comparison of nonduplicated XopM sequences lodged with GenBank (185 sequences in total) indicates conservation of both FFAT motifs across Xanthomonas species and strains (Figure 6B). Structure–function analysis of XopM. (A) Schematic representation of XopM. XopM has two FFAT (two phenylalanines [FF] in an acidic tract) at its N‐terminus. The noncanonical (ncFFAT) and canonical (cFFAT) motifs are highlighted in red and the acidic tract in grey. (B) Representation of the level of conservation of individual residues of the FFAT motifs of XopM. Amino acid properties: green = polar, purple = nonpolar, black = aromatic, red = positive, blue = negative. (C) The seven core residues of each motif are shown and the amino acids and the residue that was substituted in each experiment is marked red. The tobacco VAP12 (NtVAP12) fused to the GAL4 DNA‐binding domain (BD) was co‐expressed with different XopM proteins fused to the GAL4 activation domain (AD). –LT yeast grown on a selective medium lacking Leu and Trp. –LTH, yeast grown on a selective medium lacking Leu, Trp and His, indicating HIS3 reporter gene activity. LacZ, activity of the lacZ reporter gene. (D) In planta co‐immunoprecipitation confirming loss of interaction following mutation of the FFAT motifs. Free GFP, XopM‐GFP and XopMY61A/F91A‐GFP were co‐expressed in Nicotiana benthamiana with NtVAP12‐myc. The total protein content (Input) and after treatment with GFP‐beads (IP) were detected by immunoblotting using anti‐GFP and anti‐myc antibodies. NtVAP12, Nicotiana tabacum . Previous data suggest that a phenylalanine or a tyrosine residue at position 2 of the FFAT motif is critical for VAP binding, while amino acid substitutions are tolerated at most other positions (Murphy and Levine 2016). To test the functional importance of the putative FFAT motifs in XopM, amino acid substitutions were introduced at different positions within the motifs, including the essential position 2 of the ncFFAT (Y61A), the cFFAT (F91A), or both FFAT motifs (Y61A/F91A), as well as a conserved alanine residue at position 5 of the cFFAT (A94D). An interaction assay in yeast between NtVAP12 and the respective XopM substitution variant revealed that altering only one of the FFAT motifs does not affect the ability of XopM to bind NtVAP12 (Figure 6C), while the simultaneous change of both FFAT motifs abolishes the interaction. This indicates that at least one intact FFAT motif is required for the XopM/VAP12 interaction to occur while both motifs appear to be functionally equivalent. Control experiments for the interaction assays of XopM variants with empty vector combinations are summarised in Figure S6B. To further corroborate these findings, an in planta immunoprecipitation experiment was carried out where NtVAP12‐myc was co‐expressed in combination with either XopM‐GFP or XopMY61A/F91A‐GFP in leaves of N. benthamiana. NtVAP12‐myc was pulled down by XopM‐GFP only but not by the XopMY61A/F91A‐GFP variant (Figure 6D), confirming the functional relevance of the FFAT motifs for XopM binding to NtVAP12 also in planta. Structure Prediction of the XopM/VAP Protein–Protein Interaction AlphaFold, an artificial intelligence (AI) tool created by DeepMind for predicting protein structures, has been recognised as an exceptionally accurate method when compared to other established in silico approaches (Fowler and Williamson 2022). To obtain further functional insights into the interaction between XopM and VAP, the AlphaFold Colab was used to predict the 3D structure of XopM in isolation as well as complexed with VAP. When considered in isolation, the N‐terminal region of XopM, harbouring the two FFAT motifs, was predicted to be unfolded (Figure S7A). Bacterial T3Es typically contain intrinsically disordered regions in the N‐terminus that can fold in a stimulus‐dependent manner (e.g., during protein–protein interactions) (Marin, Thallmair, and Ott 2013). The remaining part of XopM was predicted to consist of two β sheets and 16 α helices in a bundle‐like structure and a comparison to known protein structures using the DALI server (Holm 2022) did not reveal any further structural clues related to a possible biochemical activity of the effector. However, when XopM was modelled in complex with VAP, the N‐terminal region was predicted with high confidence to be associated with the MSP of VAP (Figure S7B). Intriguingly, one of the two XopM residues determined to be critical for VAP binding by protein–protein interaction studies (F91) was modelled to be positioned in proximity to T47 of the MSP of VAP, which in turn is an essential residue for the interaction between the two proteins (Figure S7C). These predictions further corroborate a model in which FFAT motifs in XopM mediate the interaction with VAP through conserved T‐residues within the MSP. The N‐Terminal Region of XopM Is Necessary and Sufficient for VAP Binding To further narrow down the structural requirements for binding of XopM to VAP, truncation versions either comprising the XopM N‐terminus ranging from residue 1 to residue 105 and including both predicted FFAT motifs, or encompassing the C‐terminal portion of the effector from amino acid 106 to 519 were tested for their ability to interact with NtVAP12 in yeast. As shown in Figure 7A, the N‐terminal 105 amino acids of XopM were sufficient to mediate an interaction with NtVAP12 in a Y2H assay, while the XopM fragment comprising residues 106 to 519 failed to interact with VAP in yeast. This finding was confirmed by an in planta pull‐down experiment where a myc‐tagged NtVAP12 was co‐expressed with either XopM1‐106‐GFP or XopM106‐519‐GFP in leaves of N. benthamiana. A subsequent immunoblot demonstrated that NtVAP12‐myc was precipitated by XopM1‐106‐GFP but not by XopM106‐519‐GFP (Figure 7B). Structure–function studies with truncated XopM. (A) Schematic representation of the two truncated variants of XopM: the N‐terminal variant (XopM1‐105), which contains the two FFAT (two phenylalanines [FF] in an acidic tract) motifs (red dashes), and a C‐terminal variant (XopM106‐519). Proteins were tagged with the GAL4 binding domain (BD) and each co‐expressed with NtVAP12 tagged with the GAL4 activation domain (AD). –LT yeast grown on a selective medium lacking Leu and Trp. –LTH, yeast grown on a selective medium lacking Leu, Trp and His. LacZ, activity of the lacZ reporter gene. (B) Co‐immunoprecipitation confirms that only the N‐terminal XopM variant that contains the two FFAT motifs interacts with NtVAP12 in planta. XopM1‐105 and XopM106‐519 were transiently co‐expressed with NtVAP12‐myc in Nicotiana benthamiana leaves. 24 h after infiltration the total protein content was extracted (Input). After treatment with GFP‐beads (IP), detection was carried out by immunoblotting using anti‐GFP and anti‐myc antibodies. NtVAP12, Nicotiana tabacum . These data suggest that the N‐terminal region of XopM harbouring the two predicted FFAT motifs is necessary and sufficient for binding VAPs. Virus‐Induced Gene Silencing of VAP12 Does Not Affect Xcv Growth in N. benthamiana The data obtained so far suggest that XopM specifically interacts with VAP12; however, it is unclear whether this interaction is required to exert its virulence function because a role for VAP12 in plant immunity has so far not been established. To further study if VAP12 is involved in the interaction of Xcv with its host plant, virus‐induced gene silencing (VIGS) with tobacco rattle virus (TRV) of VAP12 in N. benthamiana roq1 plants was followed by infection with Xcv. ROQ1‐deficient plants fail to recognise the Xcv T3E XopQ, rendering them susceptible to Xcv infection (Schultink et al. 2017; Gantner et al. 2019). For design of the silencing construct, we employed the VIGS tool on the Sol Genomics Network website (Fernandez‐Pozo et al. 2015) with default settings and using the entire NbVAP12 coding sequence (CDS) as a query. As a result, the VIGS tool suggested using a fragment covering 300 bp of the NbVAP12 CDS to achieve high silencing specificity. Strong downregulation of NbVAP12 of about 90% could be confirmed in N. benthamiana roq1 plants 2 weeks after infiltration with the VIGS vectors as compared with the pTRV2‐GFP silencing control (Figure S8A). Six plants per construct were selected and used in a bacterial growth assay to investigate if the downregulation of NbVAP12 affects Xcv multiplication. To this end, leaves of pTRV2‐NbVAP12 and pTRV2‐GFP plants were infected by pressure infiltration with wild‐type Xcv (105 cfu/mL), and the bacterial titre was monitored 5 days postinoculation. As shown in Figure S8B, no significant difference in bacterial growth could be observed between NbVAP12‐silenced plants and the GFP silencing control. This suggests that NbVAP12 does not play a major role in the interaction between Xcv and N. benthamiana roq1 plants under the experimental conditions applied. VAP Binding and MTI Suppression Activity of XopM Are Functionally and Structurally Separable The data shown above suggest that the N‐terminal 105 amino acids containing both predicted FFAT motifs are necessary and sufficient for XopM to bind VAP. To investigate how the ability of XopM to bind VAP relates to its MTI suppression activity, XopM‐GFP truncation variants were transiently expressed in leaves of N. benthamiana. Measurement of ROS production in response to flg22 revealed that XopM1‐105‐GFP was not able to suppress an flg22‐induced ROS burst (Figure 8A), although the biochemical data suggest that the same N‐terminal fragment retained the capacity to bind VAP. In turn, XopM106‐519‐GFP effectively suppressed ROS production, despite the loss of VAP binding (Figure 8A). Subcellular localisation of XopM106‐519‐GFP suggests that the fusion protein appears still to be associated with the plant cell's endomembrane system, although the overall localisation pattern drastically changes relative to the XopM‐GFP wild‐type protein (Figure 8B). XopM106‐519‐GFP is additionally localised to larger vesicular structures that might resemble endocytotic vesicles. To further delineate whether membrane association per se is required for XopM‐GFP to suppress MAMP‐induced ROS production, a nuclear localisation signal (NLS), forcing the protein into the plant cell nucleus, was introduced into the XopM106‐519‐GFP sequence. Subcellular localisation of the NLS‐XopM106‐519‐GFP protein confirmed the functionality of the targeting signal in planta (Figure 8B). However, the NLS‐XopM106‐519‐GFP protein lost the ability to suppress flg22‐induced ROS production below the levels of the GFP control (Figure 8A). Expression of all protein variants was confirmed by immunoblotting (Figure S9). Suppression of flg22‐induced reactive oxygen species (ROS) production requires XopM membrane localisation but not VAP12 binding. (A) flg22‐induced production of ROS was measured in leaf discs transiently expressing effector proteins using a luminol‐based assay. Leaf discs were treated with flg22 and ROS production measured as relative light units (RLUs) over the time course of an hour. Leaves infiltrated with MgCl2 were not treated with flg22 and served as a negative control (−flg22). At least 10 samples were measured (n ≥ 10), except for the negative control where only four samples were measured (n = 4). Individual values are plotted with bars representing the mean and error bars give the standard deviation (±SD). Statistical significance was calculated by one‐way analysis of variance with Tukey's multiple comparison test. The different letters indicate significant differences (p < 0.05). The experiment was repeated three times. (B) Subcellular localisation of XopM‐GFP variants in Nicotiana benthamiana indicated proteins were transiently expressed in the leaves of N. benthamiana. The localisation of the proteins was detected 48 h after infiltration using a confocal laser scanning microscope. The scale bar represents 20 μm. These findings suggest that, at least in a transient CaMV 35S‐driven overexpression scenario, VAP binding and ROS suppression are functionally independent and structurally separable properties of XopM. However, membrane association appears to be essential for ROS suppression activity of XopM. Identification of XopM Interactors Associated With Vesicle Trafficking The experiments described above suggest a physical interaction between XopM and VAP in plant cells; however, it appears that XopM virulence function is probably mediated by other cellular interactions. To determine whether XopM has a wider range of interaction partners, we designed a pull‐down assay using transient expression of either GFP‐tagged NtVAP12, XopM or XopMY61A/F91A in leaves of N. benthamiana. We queried the interactome of the GFP‐tagged proteins by co‐immunoprecipitation (co‐IP) and liquid chromatography–tandem mass spectrometry (LC/MS/MS). Similar to previous reports (Stefano et al. 2018), VAP12 pulled down cytoskeleton components such as actin and proteins involved in ER membrane trafficking, namely coatomer subunits, dynamin and syntaxin isoforms (Figure 9, Tables S1 and S2), supporting the validity of the approach. A number of proteins found in the N. benthamiana VAPome were annotated to be associated with membrane protein maturation, such as ER membrane protein complex subunits, Sec61 subunits or calumenin‐like proteins. In addition, NtVAP12‐GFP associated with other VAP isoforms, indicative for a multimerisation of VAPs in planta. Of note, the NtVAP12 interactome contained a number of Rab‐GTPases (Figure 9, Tables S1 and S2), which have previously been associated with endocytotic and exocytotic vesicle transport (Nielsen 2020). A comparison between the VAPome and the XopM interactome revealed a large overlap, although the relative abundance of individual partner proteins varied (Figure 9, Tables S1 and S2). For example, the monomeric GTPase RAC5, a known regulator of the cytoskeleton (Klahre et al. 2006) was approximately 10 times more abundant in NtVAP12 pull‐downs relative to those with XopM or its variant. In turn, VAP isoforms were generally much more abundant in the XopM pull‐down relative to the NtVAP12 pull‐down, while, for example, calumenin was considerably depleted in the XopM interactome (Figure 9, Table S1). In contrast to the XopM wild‐type protein, the XopMY61A/F91A variant was not able to pull‐down VAP isoforms (Table S1, Figure 9), confirming the results obtained from direct interaction assays described above and further supporting the essential role of the XopM FFAT motifs in mediating the XopM–VAP interaction. However, other than the loss of VAP binding, mutation of the two XopM FFAT motifs had a surprisingly little effect on the overall interactome composition when compared to that of the XopM wild‐type protein. XopMY61A/F91A still interacted with proteins involved in membrane trafficking, including the group of Rab‐GTPases that were found to interact with NtVAP12 as well as with XopM. This indicates that mutation of the XopM FFAT motifs does not qualitatively affect the interaction with other proteins than VAPs and thus might provide an explanation as to why deletion or mutation of the VAP‐binding region of XopM does not affect XopM's ability to suppress MTI responses. On a cautionary note, the experimental approach used in the experiments described above involves gross overexpression of XopM in plant cells driven by the CaMV 35S promoter, which does not translate well into a natural infection situation where the number of T3E molecules translocated into the host cell is probably less. Thus, overexpression might mask certain effects that are protein‐level sensitive. Heat map of proteins identified as potential binding partners by affinity mass spectrometry. Interaction partners pulled down by GFP EV, VAP12‐GFP, XopM‐GFP or XopMY61A/F91A double mutant. Heat map is illustrated as double gradient colourmap from largest value dark blue (70) to 0 (white). VAP12 values were set to baseline value 1.0 (light blue) for relative abundance comparison. Outside the defined range was the vesicle‐associated protein 2–2 pulled down by XopM (over 200) marked with black. Each cell is labelled with its relative value compared to VAP12. After transient protein expression in Nicotiana benthamiana, each construct was pulled down by GFP affinity trap followed by LC–MS/MS analysis. Each value represents the mean of the protein abundance of three technical replicates relative to VAP12. Discussion The ability of Xcv to cause disease on a given host plant depends on a suite of approximately 36 T3Es, which collectively are essential for bacterial virulence but whose individual contribution to disease progression remains unknown in many cases (Schwartz et al. 2015). Although cellular targets and modes of action are known for only a limited number of Xcv effectors, it appears that suppression of MTI responses in the host plant is a major mechanism of effector function (Popov et al. 2016; Büttner 2016). XopM is a T3E widely distributed among xanthomonads with homologues found in Pseudomonas species as well as in Acidovorax (Schulze et al. 2012). However, the XopM polypeptide shares no similarity to any other known protein and its function during infection is unknown. An Xcv xopM deletion mutant showed no difference in bacterial growth in susceptible pepper plants relative to the Xcv wild‐type strain. It is well known that deletion of individual effector genes often does not result in a change in the virulence phenotype, which may reflect the aspects of the experimental context and/or redundancy of effector function (Castañeda et al. 2005). The overlapping activities of bacterial effectors in phytopathogens by targeting the same host targets have been extensively reported (Kvitko et al. 2009; Badel et al. 2006; Cunnac et al. 2011; Munkvold et al. 2009). Therefore, the functional characterisation of Xcv effectors can benefit from expression of individual effectors either transiently or in stable transgenic plants, rather than studying single effector deletions. Several studies using heterologous expression of individual effectors have revealed their contribution to virulence (Gimenez‐Ibanez et al. 2018; Hann and Rathjen 2007; Popov et al. 2016). Inducible expression of XopM in Arabidopsis supports the in planta growth of a nonpathogenic Pst DC3000 strain, which is normally limited due to MTI. This demonstrates that XopM is able to interfere with MTI responses and can make a significant contribution to bacterial virulence. Bacterial growth assays in XopM‐expressing plants do not reveal whether a specific MTI response is repressed by the effector; however, transient XopM expression in N. benthamiana suggests that the T3E suppresses MAMP‐induced ROS production. ROS generation during MTI occurs through PM‐resident nicotinamide adenine dinucleotide phosphate (NADPH) oxidase family members, which are activated through phosphorylation by multiple protein kinases, including BIK1 (Kadota et al. 2014). Upon flg22 recognition, the activated FLS2 complex phosphorylates and releases BIK, providing a direct link between MAMP recognition and elicitation of a ROS burst. Thus, it is conceivable that XopM directly targets ROS NADPH oxidase family members or their direct activation pathway to specifically dampen this branch of the immune response. However, the fact that XopM also inhibits induction of POX activity points more into the direction of a broader effect of the effector on MTI that could also involve inhibition of defence gene induction or secretion of antimicrobial molecules. This is in line with the observation that XopM inhibits the induction of a flg22‐induced defence reporter gene in an Arabidopsis protoplast system (Popov et al. 2016). Several bacterial T3Es have been shown to inhibit PRR activation or downstream signalling and thus impact multiple immune responses including ROS production and transcriptional activation (Büttner 2016). Thus, the inhibition of different defence outputs by XopM might point towards targeting an early step in the signalling cascade, for example, PRR signalling, at the PM. An at least partial localisation of XopM at the PM and/or the endomembrane system is in good agreement with the subcellular localisation pattern of the XopM‐GFP fusion protein. Several effectors contain consensus sites for myristoylation or palmitoylation at their N‐terminus and fatty acid modification has been shown to essential for their localisation to the host cell membrane system as well as for their respective virulence function (Bartetzko et al. 2009; Lewis et al. 2008; Nimchuk et al. 2000). Apparently, XopM does not carry a consensus sequence for fatty acid modification and thus its membrane association must be achieved through different means. Here, we report the binding of XopM to the ER membrane tethering protein VAP12 and both proteins display an overlapping localisation pattern when expressed in plant cells. VAP proteins are ER‐resident tail‐anchored membrane proteins and the Arabidopsis orthologues of NtVAP12, VAP27‐1 and VAP27‐3, have been shown to form ECPS through the interaction with PM‐localised proteins (Wang et al. 2014; Saravanan et al. 2009). The data shown here suggest that the localisation pattern of XopM and NtVAP12 overlaps within ECPS, and thus, the effector might target a process specific to this particular membrane subcompartment. The XopM interaction is specific for certain VAP isoforms and requires the VAP MSD, whereas the CCD and TMD appear to be dispensable for XopM binding. Mutational analysis within the VAP MSD domain suggests that the XopM interaction depends on residues that have previously been shown to be involved in binding proteins containing FFAT motifs (Loewen and Levine 2005). Further experiments confirmed that XopM contains two eukaryotic FFAT motifs within its N‐terminus that are functionally required for the interaction of XopM with VAPs. Intriguingly, both XopM FFAT motifs can substitute for each other for VAP binding, suggesting that the presence of two FFAT motifs in XopM is an evolutionary adaptation to increase the interaction with VAPs or represents a possibility to interact with two VAPs simultaneously. The sequence of the XopM cFFAT (QFYDEAD) closely matches the consensus FFAT motif, and key residues that have been shown to mediate the FFAT/VAP interaction, such as the aromatic amino acid residue at position 2, are conserved (Loewen and Levine 2005; Furuita et al. 2021). Structural predictions of the XopM/VAP interaction suggest that the phenylalanine at position 2 of the XopM cFFAT is positioned in proximity of critical FFAT interaction residues with the VAP MSP. The ncFFAT of XopM (GYETANS) is more divergent from the FFAT consensus sequence. However, it also carries an aromatic residue at position 2 (Y) and position 4 harbours a threonine residue that could be phosphorylated and therefore mimic the aspartate residue present in the canonical sequence (Di Mattia et al. 2020). Furthermore, both XopM FFAT motifs are directly preceded by multiple serine residues, which, if phosphorylated, could impart additional negative charges to the motif, which would increase VAP binding (Furuita et al. 2021). The presence of functional FFAT motifs in XopM is intriguing for several reasons. First, FFAT motifs are considered eukaryotic motifs generally not present in prokaryotes (Slee and Levine 2019), and second, functional FFAT motifs in VAP‐binding proteins from plants have, to our knowledge, so far not been described. Recent work led to the identification of the Chlamydia trachomatis protein IncV, which constitutes a structural component that tethers the ER membrane to the membrane of the bacteria‐containing vacuole inside the host cell through interaction with VAP (Stanhope et al. 2017). The IncV–VAP interaction relies on the presence of two FFAT motifs in the C‐terminal cytosolic tail of IncV. In analogy to XopM, IncV contains a cFFAT motif as well as an ncFFAT motif with multiple preceding serine residues (Stanhope et al. 2017). Thus, pathogenic bacteria with fundamentally different lifestyles and virulence strategies appear to employ similar molecular mechanisms to achieve VAP interaction and FFAT motifs in XopM add to the growing arsenal of eukaryotic‐like functions in T3Es from plant pathogens as a mechanism to mimic host proteins and exert control over host cellular pathways (Büttner 2016). Effectors generally operate by binding or enzymatically modifying host molecules. These host molecules are classified into two major functional classes: (1) “targets” directly modified by effectors to modulate host processes; (2) “helpers” or “facilitators” co‐opted by effectors to execute their activities, for instance to traffic within host cells or to facilitate interaction with target proteins (Win et al. 2012). Binding of XopM to VAP is supported by several lines of experimental evidence; however, VAPs do not appear to represent the actual virulence target of the effector. First, VIGS of XopM‐binding VAPs in N. benthamiana has no effect on infection with virulent bacteria and thus, there is currently no substantial evidence that XopM‐interacting VAPs play an active role in immunity. Second, and more importantly, the abilities of XopM to either bind VAP or to suppress MTI responses appear to be functionally independent and structurally separable. While the N‐terminal part of XopM, harbouring the two FFAT motifs, by itself is able to bind VAP, it has no effect on flg22‐induced ROS production when transiently expressed in N. benthamiana. In turn, the C‐terminal portion of XopM fails to interact with VAP but retains its ability to suppress ROS in an MTI assay. The N‐terminally truncated XopM variant still associates with the endomembrane system, although the overall localisation pattern deviates from that of the full‐length protein. Only when the T3E is forced into the plant cell's nucleus is MTI suppression lost. Thus, XopM probably acts at the membrane of the host cell but in at least in a transient overexpression scenario virulence activity does not require VAP binding. Thus, the question remains what the actual virulence target of XopM might be? A group of proteins that binds to NtVAP12 but also to XopM irrespective of the presence of FFAT motifs, are different isoforms of Rab GTPases. These regulate intracellular transport and the trafficking of proteins between different organelles via endocytotic and secretory pathways (Nielsen 2020). Several members of this family have been identified as regulating crucial processes of plant immunity. In Arabidopsis, the expression of dominant‐negative mutants of a number of RABA members affected different steps of endocytosis of FLS2 and thus interfered with MTI activation (Choi et al. 2013). Furthermore, the highly conserved effector protein RxLR24, secreted by the oomycete Phytophthora brassicae, interacts with the RABA GTPase family member RABA1a in vesicles and PMs, inhibiting the function of RABA GTPase in vesicle secretion and reducing vesicle secretion of antimicrobial agents (Tomczynska, Stumpe, and Mauch 2018). Thus, it is conceivable and in line with the observations presented here that XopM could target Rab GTPases to interfere with immunity‐related endocytic or exocytic vesicle trafficking. We currently have no experimental data on the possible biochemical activity of XopM. In principle, XopM binding could interfere with Rab regulation, with the association to the cytoskeleton or with the binding to specific proteins located in target membranes. Future experiments will have to address the functional relevance of the XopM–Rab interaction for the virulence function of the effector. Intriguingly, the XopM FFAT motif defective mutant still binds to Rab GTPases in an unbiased pull‐down experiment, while the ability to interact with VAPs is completely lost. This is in line with the observation that loss of VAP binding does not affect XopM virulence function under the experimental conditions but question the relevance of VAP binding of the effector. The fact that FFAT motifs are widely conserved in XopM homologues from different xanthomonads argues for considerable evolutionary pressure for their preservation. Possibly, confined localisation of VAP12 in certain membrane subcompartments, such as EPCS, enables XopM to target specific subpopulations of Rab GTPases with immunity‐related functions. It has to be borne in mind that ectopic overexpression of effector proteins, as in our study, can led to misleading results by causing promiscuous interactions and regulation of pathways that are associated with the degree of overexpression rather than the function of the protein. Thus, future experiments will have to aim at the functional characterisation of XopM protein–protein interactions under more natural conditions. Nevertheless, our study provides novel insights into T3E membrane targeting mechanisms and it will be interesting to determine whether FFAT motifs serve as targeting signals in other Xanthomonas T3Es and in effector proteins from other phytopathogens and how this relates to their respective virulence function. Experimental Procedures Plant Materials and Growth Conditions Pepper ( Capsicum annuum ‘Early Cal Wonder’ [ECW]) and N. benthamiana and N. benthamiana roq1 plants were grown in soil in a growth chamber with daily watering, and subjected to a 16 h light:8 h dark cycle (25°C:20°C) at 240–300 μmol m−2 s−1 light and 75% relative humidity. Transgenic β‐estradiol (Sigma‐Aldrich) inducible XopM‐GFP Arabidopsis thaliana lines were generated by transformation of Col‐0 wild‐type plants with a plasmid containing the entire XopM‐coding region inserted into binary vector pABindGFP (Bleckmann et al. 2010) using the floral dip method (Clough and Bent 1998). For the selection of transgenic plants, seeds of T0 plants were sterilised and sown onto Murashige and Skoog (MS) medium (Sigma‐Aldrich) supplemented with 20 μg/mL hygromycin B (Roth). Primary transformants were allowed to self‐fertilise and propagated into the homozygous T3 generation. Arabidopsis seedlings were grown on ½ × MS in 10 h light:14 h darkness (22°C/18°C) at a light intensity of 150 μmol m−2 s−1 and 70% humidity. Transgene expression was induced 2 weeks after germination by flooding the Petri dishes with 50 μM β‐estradiol and 0.1% vol/vol Tween 20. Control plants were flooded with 0.1% vol/vol ethanol and 0.1% vol/vol Tween 20. Bacterial Strains and Plant Infection Assays Strains used in this study were as follows: Xanthomonas campestris pv. vesicatoria 85–10 (Xcv wild type), Xcv ΔxopM deletion strain, Pseudomonas syringae pv. tomato DC3000 Pst ΔhrcC (type III secretion‐deficient mutant). The Xcv ΔxopM deletion strain was constructed as described (Üstün, Bartetzko, and Börnke 2013) using the primers listed in Table S2. Bacterial infection assays of pepper and N. benthamiana roq1 plants were carried out exactly as detailed in Raffeiner et al. (2022). The bacterial infection assay of Arabidopsis seedlings was adapted from Ishiga et al. (2011) with a bacterial density of 5 × 106 cfu/mL. At least three biological replicates were used per experiment unless otherwise stated. Transient Expression of Proteins in N. benthamiana Transient expression of proteins using infiltration of leaves of N. benthamiana was carried out as described (Raffeiner et al. 2022). Measurement of ROS Production and Peroxidase Activity ROS production upon flg22 treatment was monitored using the luminol‐based quantification (Liang et al. 2013). Measurement of plant peroxidase (POX) was carried out as described (Mott, Desveaux, and Guttman 2018) with slight adjustments using a plate reader (Infinite 200 PRO; Tecan). Y2H Assays Y2H techniques were performed according to the yeast protocols handbook and the Matchmaker GAL4 Two‐hybrid System 3 manual (Clontech) using the yeast reporter strains AH109 and Y187. To identify XopM‐interacting proteins, a GAL4 AD‐domain tagged cDNA library from N. tabacum (Börnke 2005) was screened as detailed before (Üstün, Bartetzko, and Börnke 2013). In Planta GFP Pull‐Down Assay Immunoprecipitation was performed by adding 1 g of homogenised leaf material 50 μL of GFP‐Trap coupled to magnetic beads (ChromoTek). Samples were processed as described (Raffeiner et al. 2022). Immunoblotting Protein extracts were boiled with 5× SDS sample buffer and separated by SDS‐PAGE. Immunoblotting was carried out with anti‐myc antibody (1:2500; Abcam), followed by goat anti‐rabbit IgA alpha chain (HRP) (1:5000) secondary antibody (Abcam). HA‐tagged proteins were detected by anti‐HA peroxidase high‐affinity antibody (Sigma). GFP was detected using a horseradish peroxidase‐conjugated anti‐GFP antibody (1:1000; Santa Cruz Biotechnology Inc.). Signals were visualised using chemiluminescence (Thermo Fisher Scientific) with a ChemiDoc Imaging system (Bio‐Rad). Confocal Microscopy and Image Analysis Fluorescence in N. benthamiana epidermal cells was detected 30–48 h after infiltration with Agrobacterium tumefaciens by confocal laser scanning microscopy (STELLARIS 8; Leica) using a water‐immersed objective (HC PL APO CS2 63×/1.20 W) with a resolution acquisition set to 1024 × 1024 and a line average of 3. For GFP, excitation was set to 488 nm and emission was recorded at 500–560 nm. For YFP, excitation was set to 514 nm and emission was recorded at 530–570 nm. For RFP and mCherry, excitation was set to 587 nm and emission was recorded at 600–640 nm. Chlorophyll autofluorescence was visualised in the far‐red wavelength. After acquisition, images were processed using Application Suite X (LASX; Leica) and ImageJ (Schindelin et al. 2012). For clarity purposes, contrast on the individual channels was manually adjusted. Colocalisation was estimated using Pearson's correlation coefficients calculated with the ‘Colocalisation finder’ plug‐in (C. Laummonerie & J. Mutterer, IBMP Strasbourg, France) on regions of interests that represent cells expressing both fluorophores simultaneously. Virus‐Induced Gene Silencing VIGS was performed as described previously (Üstün, Bartetzko, and Börnke 2013). The mRNA levels of silenced plants were checked via reverse transcription‐quantitative PCR as described in Raffeiner et al. (2022) using the primers listed in Table S3. Large‐Scale GFP Pull‐Down and LC–MS/MS Analysis Pull‐down followed by LC–MS/MS analysis was basically carried out as previously described (Bortlik et al. 2023). In short, on bead trypsin digest of GFP‐tagged proteins were treated previously with 100 μL of 0.1% RapiGest SF (Waters) followed by reduction and alkylation. Trypsin (1:50 wt/wt trypsin/protein) was added and incubated at 37°C overnight. Subsequently, desalting of peptides was carried out according to Witzel et al. (2019). Digested proteins were analysed using the Thermo Fisher Q Exactive high field mass spectrometer by reverse‐phase HPLC‐ESI‐MS/MS using the Dionex UltiMate 3000 RSLC nano system coupled to the Q Exactive High Field (HF) Hybrid Quadrupole Orbitrap MS (Thermo Fisher Scientific) and a nano‐electrospray flex ion source (Thermo Fisher Scientific) as described by Bortlik et al. (2023) with slight modifications. In short, samples were separated on a 25‐cm EASY‐Spray column (75 μm inner diameter) at 40°C. Peptides were separated over a linear 100 min gradient of 5%–44% solvent B at a flow rate of 300 nL/min. To identify proteins the mass spectrometer operated from m/z 375–1500 at a resolution of 140,000. Automated gain control (AGC) target was 3 × 106 with a maximum injection time of 65 ms. Triplicates of each sample were measured. Proteome discoverer software (PD v. 2.4) was used to analyse and align the LC–MS raw data files, with its built‐in MS Amanda, MS Mascot and Sequest HT search engine (Thermo Scientific) (Witzel et al. 2019). The MS data were searched against common contaminants, separated FASTA files of the tagged bait proteins as well as a N. benthamiana database based on the Niben1.0.1 draft genome (Kourelis et al. 2019). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1. Deletion of XopM does not affect the virulence of Xanthomonas campestris pv. vesicatoria in pepper ( Capsicum annuum ) plants. Figure S2. Immunoblot analysis of inducible of XopM‐GFP expression in transgenic Arabidopsis lines. Figure S3. Verification of effector protein expression for reactive oxygen species (ROS) and peroxidase (POX) assays. Figure S4. Verification of fusion protein expression for microscopic studies. Figure S5. XopM interacts with several VAP isoforms. Figure S6. Control experiments for yeast two‐hybrid interaction assays. Figure S7. Structural modelling of XopM and XopM with NtVAP12. Figure S8. Bacterial growth of Xanthomonas campestris pv. vesicatoria is not affected by reduced expression of NbVAP12. Figure S9. Verification of protein expression of XopM protein variants for reactive oxygen species (ROS) assays. Table S1. Selected VAP12 and XopM interacting proteins related to the endomembrane system identified by affinity purification coupled to mass spectrometry (AP‐MS). Table S2. Mother table with significant protein hits from affinity purification coupled to mass spectrometry (AP‐MS). Table S3. Oligonucleotides used in this study. Acknowledgements This study was supported by funds from the Deutsche Forschungsgemeinschaft (DFG; German Research Foundation) to F.B. (BO1916/5‐2 and CRC973 project C6) and the European Cooperation in Science and Technology EuroXanth (CA16107) from the European Union. S.Ü. was supported by an Emmy Noether Fellowship GZ: UE188/2‐1 awarded by the DFG. We thank Kerstin Bieler and Mandy Heinze for their skilful technical help. We also thank the confocal microscopy facility of Faculty of Biology and Biotechnology/Plant Cell Biology/Ruhr University Bochum (DFG project number 523980288, GZ: INST 213/1180‐1 FUGG). Open Access funding enabled and organized by Projekt DEAL. Funding: This work was supported by Deutsche Forschungsgemeinschaft (BO1916/5‐2, INST 213/1180‐1 FUGG, UE188/2‐1). Data Availability Statement Any additional data associated with this study are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Figure S1. Deletion of XopM does not affect the virulence of Xanthomonas campestris pv. vesicatoria in pepper ( Capsicum annuum ) plants. Figure S2. Immunoblot analysis of inducible of XopM‐GFP expression in transgenic Arabidopsis lines. Figure S3. Verification of effector protein expression for reactive oxygen species (ROS) and peroxidase (POX) assays. Figure S4. Verification of fusion protein expression for microscopic studies. Figure S5. XopM interacts with several VAP isoforms. Figure S6. Control experiments for yeast two‐hybrid interaction assays. Figure S7. Structural modelling of XopM and XopM with NtVAP12. Figure S8. Bacterial growth of Xanthomonas campestris pv. vesicatoria is not affected by reduced expression of NbVAP12. Figure S9. Verification of protein expression of XopM protein variants for reactive oxygen species (ROS) assays. Table S1. Selected VAP12 and XopM interacting proteins related to the endomembrane system identified by affinity purification coupled to mass spectrometry (AP‐MS). Table S2. Mother table with significant protein hits from affinity purification coupled to mass spectrometry (AP‐MS). Table S3. Oligonucleotides used in this study. Data Availability Statement Any additional data associated with this study are available from the corresponding author upon reasonable request.