In order to survey naturally occurring sequence variation in PopP2, we first selected 20 R. solanacearum strains isolated from commercially grown pepper or tomato plants showing wilting symptoms in the Republic of Korea, on the basis of their geographic location, the host plant they were collected from (Pepper, strains ‘Pe_’ and Tomato, strains ‘To_’) and the year of collection (Table 1 and Supplementary Figure 1). Using gene-specific primers, we could amplify and confirm the presence of popP2 in 17 of the 20 R. solanacearum strains (Table 1). The genomic sequence encoding the C-terminal region of PopP2 that is necessary and sufficient for avirulence in Arabidopsis, amino acids 149–488, was analyzed in the 17 popP2-harboring strains and compared to the GMI1000 reference (Salanoubat et al., 2002; Sohn et al., 2014). 11 strains (Pe_2, Pe_3, Pe_18, Pe_24, Pe_27, Pe_42, Pe_45, Pe_56, To_1, To_7, and To_42) harbored four SNPs resulting in the following amino acid residue changes: G156D, S288N, G396E, and V465M. These 11 strains also harbored four synonymous mutations at A169, A186, V291, and V406. Five other strains (Pe_1, Pe_26, Pe_28, Pe_40, and To_63) harbored the four aforementioned non-synonymous SNPs as well as two additional SNPs resulting in S167C and Q179R. These five isolates all harbored the same aforementioned synonymous SNPs except for the mutation at A169. Finally, the Pe_13 strain harbored G156D, S167C, and Q179R as well as the synonymous mutation at A186. In addition to this, Pe_13 harbored a SNP resulting in A234G and a single nucleotide insertion, which resulted in a premature stop codon, E241^∗ (Figure 1). Therefore, our survey identified three novel polymorphic PopP2 variant groups, which we termed PopP2^Pe_2 (four non-synonymous SNPs, present in 11 strains), PopP2^Pe_1 (six non-synonymous SNPs, present in five strains) and PopP2^Pe_13 (four non-synonymous SNPs and a frameshift insertion, present in only one of the selected strains) (Figure 1). The majority of strains analyzed here (11 of 17) harboring the “Pe_2” popP2 allele were isolated from regions spanning the length of the Republic of Korea from Hwacheon in the north down to the southern coastal region, Haenam (Supplementary Figure 1). Among the five strains harboring the “Pe_1” allele, four were isolated from pepper fields in western regions; the other was isolated from tomato in eastern Bongwha. The Pe_13 strain carrying the truncated PopP2 variant was isolated from Imsil in the south–west (Supplementary Figure 1). The specific host cultivar genotypes are unknown. Thus, no obvious correlation between the host or the location of the isolated strains and the presence of a specific popP2 allele could be revealed by this survey.

Nonetheless, closer analysis of the PopP2 coding sequence highlighted the presence of a putative LxLxL ethylene-responsive element binding factor-associated amphiphilic repression (EAR) motif; a motif involved in transcriptional repression via recruitment of transcriptional co-repressors (Hiratsu et al., 2003; Kagale and Rozwadowski, 2011). The reported motif comprises three leucine residues with amino acid spacers; however, PopP2 possesses an additional fourth leucine residue (LxLxLxL) at amino acids 323–329, which is almost adjacent to C321, one of the three conserved catalytic residues, H260, D279, and C321 (Tasset et al., 2010). Similarly to the catalytic residues, the putative EAR motif is fully conserved in all the sequenced PopP2 variants (Figure 1).

To investigate the effect of natural polymorphism on the avirulence activity of PopP2, the three newly identified PopP2 variants lacking N-terminal region were translationally fused with AvrRps4 N-terminal domain (AvrRps4N) and delivered by the P. syringae pv. tomato (Pto) DC3000 type 3 secretion system into the resistant Ws-2 Arabidopsis accession, carrying functional RPS4 and RRS1-R, to assay for HR (Sohn et al., 2014). At 1 day post-infiltration (dpi) Pto DC3000-delivered AvrRps4N:PopP2^Pe_1 and AvrRps4N:PopP2^Pe_2 variants triggered a strong HR (Figure 2A). Conversely, the truncated AvrRps4N:PopP2^Pe_13 could not trigger HR in Arabidopsis. This lack of avirulence activity was expected as the catalytic residues required for acetyltransferase activity and RPS4/RRS1-mediated recognition are absent in this variant due to a premature stop codon (Tasset et al., 2010; Sohn et al., 2014) (Figure 1). Similar recognition events were observed when PopP2 natural variants were co-expressed with RRS1 and RPS4 in tobacco leaf cells after Agrobacterium-mediated transient transformation (hereafter, agroinfiltration) (Figure 2B). Expression of all three protein variants was confirmed by immuno-detection in Nicotiana benthamiana leaf extracts (Figure 2C). Notably, none of the six identified SNPs present in the avirulent PopP2^Pe_1 allele affected recognition by RPS4/RRS1 despite the close proximity of two polymorphisms to catalytic residues (S288N and G396E). This suggests that these polymorphisms do not impair PopP2 acetyltransferase activity and that RPS4/RRS1-mediated recognition can accommodate significant variation in the PopP2 sequence.

To test if PopP2^149-488 (hereafter referred to as PopP2) requires an LxLxL motif (hereafter; EAR motif) for avirulence function, we performed site-directed mutagenesis on the PopP2 LxLxLxL sequence to generate the LxAxAxL variant (hereafter, PopP2^LAAL). Mutation of the two central leucine residues ensured that both LxLxL sequences were disrupted. Additionally, we fused a synthetic EAR motif, SRDX (LDLDLELRLGFA, derived from the SUPERMAN repressor domain) or a mutated version of the artificial EAR motif, srdx (FDFDFEFRLGFA) to the C-terminus of PopP2^C321A and PopP2^LAAL to test the specificity of EAR motif function (Hiratsu et al., 2003).

To assay for HR in Arabidopsis, we used a modified Pseudomonas fluorescens strain, Pf0-1, which carries a functional type 3 secretion system [hereafter Pf0-1(T3S)] for delivery of the PopP2 variants (Thomas et al., 2009). As was previously reported, AvrRps4N:PopP2 triggered HR at 24 hours post-infection (hpi) in the resistant Ws-2 accession, but not in the susceptible Col-0 accession when delivered by Pf0-1(T3S) (Sohn et al., 2014). The catalytic cysteine mutant, PopP2^C321A, was unable to trigger HR in Ws-2 due to the loss of acetyltransferase activity required for RPS4/RRS1-mediated recognition, and this was unaffected by fusion of C-terminal SRDX or srdx (Supplementary Figure 3) (Tasset et al., 2010; Sohn et al., 2014). Interestingly, the PopP2^LAAL variant with a disrupted EAR motif also failed to elicit HR in Ws-2. This suggests that PopP2 requires a functional EAR motif to trigger RPS4/RRS1-mediated HR. Indeed, fusion of the SRDX motif to the C-terminus of PopP2^LAAL partially restored the ability of PopP2 to trigger HR in Ws-2, suggesting that PopP2-triggered HR is dependent on a functional EAR motif. Fusion of the mutated artificial EAR motif, srdx, to PopP2^LAAL had no effect as expected (Figure 3A). PopP2 variants (WT, C321A, and LAAL) were delivered in planta as demonstrated by a secretion assay in N. benthamiana leaves (Supplementary Figure 5). In addition, we measured ion leakage to quantify the macroscopic HR symptoms and found that PopP2^LAAL induced ion leakage to the same extent as the negative control, PopP2^C321A. In agreement with the HR data, PopP2^LAAL-SRDX induced more ion leakage than PopP2^LAAL to a level intermediate between PopP2^LAAL and PopP2^WT (Figure 3B).

It became apparent that the PopP2 EAR motif was required to trigger HR in Arabidopsis. However, HR does not always correlate with the defense response leading to immune transcriptional reprogramming (Yu et al., 1998; Cawly et al., 2005; Gassmann, 2005). Therefore, we also investigated the requirement of this motif for the induction of defense-related genes and disease resistance. To this end, we analyzed expression of multiple defense marker genes known to be upregulated by bacterial T3S-delivered PopP2: PR1, FMO1, PBS3, and SARD1 (Sohn et al., 2014). The Arabidopsis accession Ws-2 was infiltrated with Pf0-1(T3S) carrying PopP2 variants and samples were taken 8 hpi for RNA extraction and qRT-PCR analysis. Consistent with the loss of HR induction by PopP2^LAAL, the upregulation of all four defense genes was significantly impaired by mutation of the EAR motif. Additionally, fusion of the SRDX motif partially restored defense gene upregulation to a level more similar to PopP2^WT (Figure 3C). To further test the requirement of the EAR motif for PopP2 avirulence function, we assayed bacterial growth of Pto DC3000 carrying PopP2 variants in Arabidopsis (Sohn et al., 2014). As expected from our previous findings, Pto DC3000(PopP2^LAAL) did not exhibit the growth restriction that PopP2^WT did, and grew to a number comparable to Pto DC3000 (PopP2^C321A) (Figure 3D). Furthermore, Pto DC3000 (PopP2^LAAL-SRDX) growth was partially restricted, corroborating our other evidence that PopP2 recognition in planta is dependent on a functional EAR motif. Overall, we have demonstrated that the PopP2 EAR motif is required for HR elicitation, upregulation of defense marker genes and bacterial growth restriction. The specificity of these effects of PopP2 EAR motif was further confirmed by the gain of avirulence observed with PopP2^LAAL fused to a C-terminal synthetic EAR motif.

Delivery of PopP2 but not PopP2^C321Avia P. fluorescens Pf0-1(T3S) has been demonstrated to inhibit PTI as indicated by cell death induction by subsequent Pto DC3000 infiltration (Crabill et al., 2010; Badel et al., 2013; Le Roux et al., 2015). This suggested that PopP2 acetyltransferase activity is required for virulence activity to suppress host PTI (Le Roux et al., 2015). Since discovering that the conserved EAR motif was required for avirulence, we sought to investigate its requirement for PopP2 virulence activity as measured by PTI suppression in N. benthamiana leaves. Infiltration of Pf0-1(T3S) carrying PopP2 variants alone [empty vector (EV), PopP2, PopP2^C321A, PopP2^LAAL, PopP2^LAAL-SRDX, and PopP2^LAAL-SRDX] induced no cell death while infiltration of Pto DC3000 carrying each aforementioned variant induced a cell death. As previously shown, infiltration of Pf0-1(T3S)(PopP2) followed by infiltration of Pto DC3000 resulted in a cell death response due to PopP2 PTI suppression; however, infiltration of Pf0-1(T3S)(EV) or Pf0-1(T3S) (PopP2^C321A) followed by infiltration of Pto DC3000 resulted in no cell death induction due to N. benthamiana PTI-induced inhibition of Pto DC3000-induced cell death. Interestingly, we discovered that Pf0-1(T3S) (PopP2^LAAL) infiltration into N. benthamiana leaves was also unable to suppress host PTI, as demonstrated by the lack of cell death induction by subsequent Pto DC3000 infiltration (Supplementary Figure 4). This suggests that the EAR motif is not only required for avirulence activity but also for PopP2 virulence function. Fusion of the artificial EAR motif, SRDX, partially restored the PTI suppression ability of PopP2, whereas fusion of the mutated version, srdx, had no effect (Supplementary Figure 4).

The EAR motif is known to confer transcriptional repression activity via the recruitment of a co-repressor (Kagale and Rozwadowski, 2011). This led us to hypothesize that PopP2 recognition in Arabidopsis requires the recruitment of a transcriptional co-repressor. Therefore, we screened a library of transcriptional co-repressors for interaction with PopP2 using a LexA-based yeast-two-hybrid (Y2H) assay (Gyuris et al., 1993). The library we built comprises several members of the Groucho/Tup1-like family of co-repressors with known LisH domains and WD repeats (Supplementary Table 1) (Liu and Karmarkar, 2008). LEUNIG (LUG) and the closely related LEUNIG_HOMOLOG (LUH) as well as TPL and its close homologs TPR1–4 are the best characterized of the Gro/Tup1-like co-repressors. SEUSS (SEU) and its close homologs, SEUSS-LIKE (SLK1–2), can interact with LUG or LUH to form a functional repressor complex (Franks et al., 2002; Sitaraman et al., 2008; Grigorova et al., 2011; Shrestha et al., 2014). Finally, high expression of osmotically responsive genes 15 (HOS15) and SIN3-associated polypeptide of 18 kDa (SAP18) were included, as they are known to mediate transcriptional repression in Arabidopsis via chromatin modification (Song and Galbraith, 2006; Hill et al., 2008; Zhu et al., 2008).

Protein fusions were assayed for interaction in yeast with LEU2 and lacZ reporter genes under the control of upstream LexA operators. Yeast cells expressing empty vector controls with fusion proteins were assayed for growth and the development of blue color on the induction medium [-His(H)/-Trp(T)/-Ura(U)/-Leu(L)] + X-Gal, to test for auto-activity. TPL, TPR4, and SAP18 in the pB42-AD vector alone activated LEU2 but not lacZ allowing growth on medium lacking leucine; all other yeast cells did not activate reporter genes (Supplementary Figure 2). As expected, LUG interaction with SLK2, used as a positive control, showed strong interaction (Stahle et al., 2009). However, none of the yeast cells co-expressing PopP2-DBD and a transcriptional co-repressor-AD fusion protein showed clear protein-protein interaction in yeast cells (Supplementary Figure 2).

We have assayed the HR activation in response to PopP2 variants in the native Arabidopsis system; we also used a heterologous tobacco overexpression system to assay for RPS4/RRS1 mediated recognition of PopP2 variants (Figure 4A). Agroinfiltration of PopP2^WT, RPS4, and RRS1 resulted in a robust PCD response at 3 dpi in tobacco. Conversely, co-expression of the inactive PopP2^C321A mutant with RPS4 and RRS1 resulted in significantly reduced PCD (Sohn et al., 2014). Intriguingly, the PopP2^LAAL, PopP2^LAAL-SRDX, and PopP2^LAAL-srdx variants all elicited a strong PCD response when co-expressed with RPS4 and RRS1 (Figure 4A). Thus, loss of PopP2 avirulence activity as a result of EAR motif disruption in Arabidopsis could not be reconstituted in the tobacco overexpression system.

To confirm the expression of PopP2 variants, sequences were fused with a C-terminal YFP epitope tag and transiently expressed in N. benthamiana. Total proteins were extracted and subjected to immuno-detection with an anti-GFP antibody. PopP2^WT and PopP2^C321A accumulated to similar amounts. Surprisingly, PopP2^LAAL protein accumulation was significantly lower as indicated by the low intensity band on the blot (Figure 4B). Furthermore, fusion of the synthetic EAR motif restored the protein level to the WT level while fusion of the non-functional srdx motif, had no effect on protein accumulation. This evidence suggests that the stability of PopP2 inside plant cells is dependent on an EAR motif (Figure 4B).

We surveyed the natural variation of PopP2 sequence in R. solanacearum strains isolated from diseased tomato and pepper fields at different geographic locations across the Republic of Korea. Gene-specific sequencing and subsequent analysis revealed that popP2 is highly conserved across South Korean isolates, as we identified only three polymorphic alleles among the 17 popP2-harboring strains. In fact, all but two of the polymorphisms are conservative; they result in a change to an amino acid with similar properties. However, G156D and G396E result in a change from a single hydrogen radical to a negatively charged side chain. We have shown that, despite these polymorphisms, two of the natural variants, PopP2^Pe_1 and PopP2^Pe_2, retained avirulence activity in Arabidopsis Ws-2. The significantly truncated variant, PopP2^Pe_13, did not trigger HR. This is consistent with the previous report showing that the catalytic triad is required for PopP2 acetyltransferase activity (Tasset et al., 2010).

It must be considered that the R. solanacearum strains from which the popP2 alleles were isolated were highly virulent on pepper or tomato plants. Genome sequence analysis does not reveal any strong homology of Solanaceae disease resistance (R) genes with RPS4 and RRS1 (Consortium, 2012; Kim et al., 2014). However, the existence of PopP2^Pe_13 (a truncated variant that lacks avirulence) suggests that there might be a selective pressure in natural host plants of R. solanacearum. This hypothesis is further supported by the three R. solanacearum strains that were shown to lack popP2 in our study. In this regard, it would be interesting to survey the presence/absence and sequence polymorphism of popP2 in R. solanacearum strains from other host plants or geographic regions. Furthermore, in addition to RPS4/RRS1, it is plausible that PopP2 may be recognized by other R protein(s) and that natural variants found in our study may show altered avirulence activity.

Our study illustrates the first example of a plant pathogenic effector that is dependent on an EAR motif for avirulence activity. Of note, the Xanthomonas campestris pv. vesicatoria (Xcv) effector XopD possesses two EAR motifs [sequence (L/F)DLN(L/F)(X)P] that are both required for virulence. XopD represses defense gene transcription via the two EAR motifs to enable Xcv growth in tomato (Kim et al., 2008). Likewise, we demonstrated that the PopP2 EAR motif is required for PTI suppression, which may contribute to R. solanacearum virulence. In addition, we showed that disruption of the LxLxLxL amino acid sequence rendered PopP2 unstable, but could be stabilized by addition of the synthetic EAR motif SRDX at the C-terminus. Therefore, PopP2 stability appears to depend specifically on the presence of the LxLxLxL sequence. In the tobacco heterologous system, despite being clearly reduced, accumulation of PopP2^LAAL might nonetheless reach a threshold required for detection by the over-expressed RPS4 and RRS1 receptors. Conversely, in the native Arabidopsis system, we can infer that PopP2 EAR motif could play an important role in protein stability, allowing its accumulation above the necessary amount to trigger RRS1 activation.

A novel mechanism of controlling protein stability via an EAR motif has recently been unveiled. Proteasomal and non-proteasomal degradation of the ZINC FINGER OF ARABIDOPSIS THALIANA12 (ZAT12) TF is controlled by an LxLxL EAR motif (Le et al., 2016). Similarly to the reduced accumulation of PopP2^LAAL, the abundance of the ZAT12 variant carrying a mutation in the EAR motif was lower than the wild-type. Le et al. (2016) hypothesized that the ZAT12 EAR motif is involved in mediating interactions with different partners; at least one of which is H₂O₂ responsive and another that is a factor of proteasomal degradation targeting, such as an E3-ubiquitin ligase. Conversely, an earlier study investigating a poplar (Populus spp.) ortholog of ZAT12, Pti Cys2/His2 zinc-finger protein 1 (PtiZFP1), reported that the PtiZFP1 EAR motif promotes its degradation by the 26S proteasome through MAPK binding (Hamel et al., 2011). This is in contrast to the ZAT12 EAR motif-mediated protein stability model, but it provides additional clues to investigate the mechanism regulating PopP2 accumulation in planta.

PopP2 contributes to R. solanacearum virulence in tomato, eggplant, bean, and Arabidopsis (Macho et al., 2010; Le Roux et al., 2015). It is conceivable that successful R. solanacearum strains have acquired PopP2 variants carrying an EAR motif to stabilize the secreted effector by circumventing in planta degradation. This mechanism could have been selected to enhance the virulence of R. solanacearum on host plants. Indeed, we have shown that PopP2 stability is dependent on a functional EAR motif and that this is associated with both PopP2-mediated PTI suppression and host recognition ability.

The requirement of PopP2 for an EAR motif to trigger an avirulence response coupled with the numerous reports of EAR motif-mediated recruitment of co-repressors led us to generate yeast two-hybrid constructs of known Arabidopsis transcriptional co-repressors to screen for interaction with PopP2 in yeast cells. The Groucho (Gro)/Tup1-like family of co-repressors make up the largest and best characterized family of co-repressors in Arabidopsis with at least 13 members, including LUG and LUH, TPL and TPRs and HOS15 (Liu and Karmarkar, 2008). LUG and LUH are partially redundant transcriptional co-repressors and together with SEU/SLKs are involved in embryo and floral development and abiotic stress responses (Sitaraman et al., 2008; Shrestha et al., 2014). Similarly, HOS15 and SAP18 have not been implicated in plant defense so far, but mediate gene repression in response to cold or salt stress (Song and Galbraith, 2006; Hill et al., 2008; Zhu et al., 2008). On the other hand, the recruitment of the functionally redundant TPL and TPR1–4 by EAR motif-containing proteins in defense signaling is well-known. TPL is involved in the regulation of JA signaling for disease resistance to necrotrophic pathogens and stomatal defense (Pauwels et al., 2010; Pauwels and Goossens, 2011). Similarly, TPR1 and other TPRs associate with the R gene SNC1 to participate in transcriptional repression of negative regulators of defense (Zhu et al., 2010).

Our yeast-two-hybrid (Y2H) screen data indicate that PopP2 does not interact directly with any of the tested co-repressors in our experimental conditions. Further confirmation of this result could be obtained by testing in planta interaction between PopP2 and transcriptional co-repressors in the future. Nonetheless, based on our findings, we hypothesize that in planta PopP2 stability may be dependent on EAR motif-mediated recruitment of an as yet untested co-repressor or another host component. Identification of the host factor(s) controlling PopP2 stability shall bring new insights into the systems used to monitor pathogen virulence in the plant cell.

Arabidopsis accessions Col-0 and Ws-2 were grown in short day conditions (11 h light/13 h dark) at 22°C. N. benthamiana and N. tabacum W38 plants were grown in long day conditions (14 h light/10 h dark) at 25°C.

For Pseudomonas-mediated delivery, R. solanacearum GMI1000 PopP2^149-488 variants (WT, C321A, LAAL, LAAL-SRDX, and LAAL-srdx) were cloned in the pEDV6 gateway destination vector, which is described in detail in Sohn et al. (2014). The SRDX synthetic EAR motif, LDLDLELRLGFA, is described in Hiratsu et al. (2003). The mutated SRDX, termed srdx, encodes the peptide sequence FDFDFEFRLGFA. SRDX and srdx were fused to the popP2 C-terminus with the cloning primers.

The natural polymorphic popP2 variants (Pe_1, Pe_2, and Pe_13) were amplified from R. solanacearum strains isolated from pepper in Korean fields using gene-specific primers. These were cloned into a modified Golden Gate-compatible pBBR1MCS-5 broad host range vector containing avrRps4 promoter (128 bp) and N-terminus (408 bp), as was used for the pEDV vectors (Sohn et al., 2007). popP2 variants were fused to a C-terminal 6xHA tag.

Golden Gate-compatible pLexA-DBD and pB42-AD vectors were used for the Y2H assay and coding sequences were fused to C-terminal 3xFLAG and 6xHA, respectively. popP2(149–488) was amplified from a previously reported plasmid, pCR8:popP2 (Sohn et al., 2014). Transcriptional co-repressors were amplified from Arabidopsis cDNA (accession Col-0).

The binary vector pICH86988 (provided by Sylvestre Marillonnet) was used for Agrobacterium-mediated delivery into N. tabacum and N. benthamiana. This is a Golden Gate-compatible vector in which popP2 variants, RPS4 (Arabidopsis accession No-0) and RRS1-R (Arabidopsis accession Ws-2) were assembled in fusion with C-terminal YFP, 3xHA and 3xFLAG, respectively.

For assembly into all Golden Gate-compatible vectors, genes were PCR-amplified using gene-specific primers, which also introduced BsaI recognition sequences and specific 4 bp overhangs flanking the sequences. These were cloned into the pICH41021 shuttle vector (modified pUC19 with a mutated BsaI recognition sequence). Golden Gate assembly of the resulting pICH41021 constructs into Golden Gate-compatible destination vectors was carried out to generate final constructs (Engler et al., 2008).

Escherichia coli DH5α was used to maintain and replicate plasmids, as well as for bacterial conjugation. Modified P. fluorescens Pf0-1(T3S) was used for HR assays and ion leakage assays (Thomas et al., 2009). P. syringae pv. tomato (Pto) DC3000 was used for HR assays and in planta bacterial growth assays. To transform Pf0-1(T3S) and Pto DC3000 with the appropriate construct, we used triparental mating with E. coli HB101 (pRK2013) as a helper strain (Sohn et al., 2007) or Pseudomonas electroporation (Choi et al., 2006). Agrobacterium tumefaciens strain AGL1 was transformed by electroporation with appropriate binary constructs for the transient expression assays.

For HR assays, bacteria were grown on King’s B (KB) agar containing appropriate antibiotics and harvested in 10 mM MgCl₂. The final concentration of Pf0-1(T3S) suspensions was adjusted to OD₆₀₀ = 0.2; the final concentration of Pto DC3000 suspensions was adjusted to OD₆₀₀ = 0.1. Leaves of 5-week old Arabidopsis plants were hand-infiltrated on the abaxial surface using a blunt-end syringe, and macroscopic symptoms were observed and photographed at 20–24 hours post-infiltration (hpi). For ion leakage assays with Pf0-1(T3S), leaf disks were sampled at 0.5 hpi, washed in water for 30 min (with gentle shaking at room temperature) and transferred to fresh water (0 hpi sample). Ion leakage measurements were taken at 6, 12, 24, 36, and 48 hpi using a conductivity meter (Horiba B-173). For in planta bacterial growth assays, Pto DC3000 strains were grown and harvested as for HR assays; however, bacterial suspensions were adjusted to OD₆₀₀ = 0.001. Leaves of 5-week old Arabidopsis plants were hand-infiltrated on the abaxial surface using a blunt-end syringe, and sampled at 4 days post-infiltration (dpi). Samples were ground in 10 mM MgCl₂, serially diluted and spotted on KB agar containing appropriate antibiotics. These were incubated at 28°C for 2 days prior to counting colonies in order to calculate the number of colony forming units (cfu)/cm² of infected leaf.

The Saccharomyces cerevisiae strains used for the Y2H assays were EGY48 Mat(α) and RFY206 Mat(a). RFY206 carries the pSH18–34 vector, which encodes the lacZ reporter gene under the control of 8 upstream LexA operators. Additionally, pSH18–34 encodes the URA3 selectable marker, allowing growth on media lacking uracil. EGY48 and RFY206(pSH18–34) were transformed with pB42-AD and pLexA-DBD constructs, respectively, using the ‘Frozen-EZ Yeast Transformation II Kit’ according to the manufacturer’s recommendations (Zymo Research). pB42-AD encodes the TRP1 selectable marker, which allows yeast growth on media lacking tryptophan (Trp), pLexA encodes the HIS3 selectable marker, allowing growth on media lacking histidine (His). After transformation of yeast with the appropriate constructs, mating and interaction assays were performed as described in the Yeast Protocols Handbook (Clontech).

Agrobacterium tumefaciens AGL1 carrying the binary constructs were grown in liquid L-medium supplemented with the appropriate antibiotics for 24 h. Cells were harvested by centrifugation and re-suspended in infiltration medium (10 mM MgCl₂ and 10 mM MES-KOH pH 5.6). Suspensions were then adjusted to OD₆₀₀ = 0.1. Bacterial suspensions were mixed in a 1:1 ratio and infiltrated into the abaxial surface of 5-week old N. benthamiana or N. tabacum leaves using a blunt-end syringe. Cell death was observed and photographed at 2–3 dpi.

Plant protein samples were prepared from N. benthamiana 48 h after Agrobacterium-mediated transformation. One full infiltrated leaf was ground in liquid nitrogen and total proteins were extracted in GTEN buffer (10% glycerol, 150 mM Tris-HCl pH 7.5, 1 mM EDTA, 150 mM NaCl) supplemented with 5 mM DTT, plant protease inhibitor tablet (1 tablet/50 ml extraction solution) (Roche) and 0.2% (vol/vol) Nonidet P-40. Lysates were centrifuged for 15 min at 5000 rpm at 4°C. The supernatants were filtered through miracloth mesh (Millipore) and used as input samples. Proteins were separated by SDS-PAGE and immunoblotted using anti-GFP-HRP conjugated antibodies (Santa Cruz). Proteins were detected with a mix of SuperSignal West Pico and SuperSignal West Femto chemiluminescent substrates (Pierce). Membranes were stained with Ponceau S to visualize protein loading.