The Pseudomonas syringae type III effector HopG1 triggers necrotic cell death that is attenuated by AtNHR2B

The plant pathogenic bacterium Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) has become a paradigm to investigate plant-bacteria interactions due to its ability to cause disease in the model plant Arabidopsis thaliana. Pst DC3000 uses the type III secretion system to deliver type III secreted effectors (T3SEs) directly into the plant cytoplasm. Pst DC3000 T3SEs contribute to pathogenicity by suppressing plant defense responses and targeting plant’s physiological processes. Although the complete repertoire of effectors encoded in the Pst DC3000 genome have been identified, the specific function for most of them remains to be elucidated. Among those effectors, the mitochondrial-localized T3E HopG1, suppresses plant defense responses and promotes the development of disease symptoms. Here, we show that HopG1 triggers necrotic cell death that enables the growth of adapted and non-adapted pathogens. We further showed that HopG1 interacts with the plant immunity-related protein AtNHR2B and that AtNHR2B attenuates HopG1- virulence functions. These results highlight the importance of HopG1 as a multi-faceted protein and uncover its interplay with AtNHR2B.

up the minimal functional set of effectors needed for Pst DC3000 to be virulent in N. benthamiana 23 . Among those effectors, HopG1 was previously implicated in the suppression of PTI and ETI responses [24][25][26] . Further research showed that HopG1 localizes to mitochondria 24 , where it interacts with the mitochondrial-localized kinesin motor protein to modulate actin cytoskeleton and promote the development of chlorotic symptoms 27 .
We previously identified AtNHR2B (Arabidopsis thaliana nonhost resistance 2B) as a protein that functions in plant immunity [28][29][30][31] . Atnhr2b mutant plants are immunocompromised, and as a result, they are susceptible to pathogens that normally do not cause disease in wild-type plants, called non-adapted pathogens 31 . We further showed by live-cell imaging that AtNHR2B tagged with the green fluorescent protein (GFP) localizes to chloroplasts and compartments of the endomembrane system 31 , and through proteomics approaches, we also showed that AtNHR2B interacts with proteins localized to chloroplasts, mitochondria and nucleus 32 .
Here, we show that the Pst DC3000 T3E HopG1 interacts with AtNHR2B and that AtNHR2B interferes with HopG1 virulence functions. We also expand the range of functions for HopG1 with the discovery that transient or stable expression of HopG1 in N. benthamiana and Arabidopsis, respectively, triggers necrotic cell death that enables the growth of adapted and non-adapted pathogens. This necrotic cell death is a virulence function that confers advantage to the pathogen, rather than an HR-type of programmed cell death that confers resistance to the plant. These results highlight the importance of HopG1 as a multifaceted protein and uncover the interplay between HopG1 and AtNHR2B.

Results
Adapted pathogens changed the pattern of localization and abundance of AtNHR2B-GFP. We previously showed that AtNHR2B is induced by pathogens and pathogen-derived elicitors, and that AtNHR2B-GFP transiently expressed in N. benthamiana localizes to the cytoplasm and to highly dynamic structures (punctae) reminiscent of subcellular compartments of the endomembrane system 31 . AtNHR2B is induced by nonadapted bacterial pathogens (that are unable to cause disease in a particular host), as well as, adapted bacterial pathogens (that cause disease in a specific host); however, we do not know if pathogens' lifestyle (non-adapted vs adapted) have an effect on the function of AtNHR2B. Since AtNHR2B is an immune-related protein, we hypothesized that adapted pathogens could target this protein for parasitism. To start testing that hypothesis, we evaluated the localization of AtNHR2B-GFP by transient expression in N. benthamiana. N. benthamiana plants transiently expressing AtNHR2B-GFP were mock-treated with water, inoculated with P. syringae pv. tabaci (Pstab) (adapted pathogen of N. benthamiana), or with Pst DC3000 (non-adapted pathogen of N. benthamiana).

Mock-treated plants and plants inoculated with
Pst DC3000 showed the expected localization to cytoplasm and punctae. Interestingly, after inoculation with Pstab, the signal from AtNHR2B-GFP was limited to the cytoplasm (Fig. 1a).
The change in AtNHR2B-GFP pattern of localization in N. benthamiana associated with inoculation with the adapted pathogen was validated in transgenic Arabidopsis plants expressing AtNHR2B-GFP from its native promoter 31 . In such plants, the fluorescent signal was very low in mock-treated plants, but significantly increased upon inoculation with Pst DC3000 and Pstab. Similar to the results obtained in N. benthamiana, the pattern of AtNHR2B-GFP localization differed depending on the pathogen used for the inoculation; inoculation with the non-adapted pathogen of A. thaliana, Pstab, caused the localization of AtNHR2B-GFP to cytoplasm and small punctae as previously described 31 (Fig. 1b). However, inoculation with the adapted pathogen of A. thaliana, Pst DC3000, changed the pattern of AtNHR2B-GFP localization and the fluorescence signal was detected in the cytoplasm only and appeared diffused and distorted (Fig. 1b).
To evaluate if the diffusion of the fluorescence signal was due to adapted pathogens interfering with protein abundance, we collected leaves from N. benthamiana and Arabidopsis plants expressing AtNHR2B-GFP that were mock-treated, or inoculated with their corresponding adapted and non-adapted pathogens. Collected leaves were used for protein extraction and quantitative Western blot analysis using anti-GFP antibodies to detect AtNHR2B-GFP and anti-actin antibodies for normalization. N. benthamiana expressing AtNHR2B-GFP and mock-treated, showed low levels of AtNHR2B-GFP. However, inoculation with both adapted (Pstab) and non-adapted (Pst DC3000) pathogens increased AtNHR2B-GFP concentration in comparison with mock-treated plants, but the amount of AtNHR2B-GFP was reduced by 50% in plant samples treated with Pstab in comparison with plants treated with Pst DC3000 (Fig. 1c, Supplementary Fig. 1). In a similar way, AtNHR2B-GFP transgenic Arabidopsis lines did not accumulate any protein after mock treatment and showed ~ 40% less AtNHR2B-GFP after inoculation with the adapted pathogen Pst DC3000 than plants inoculated with Pstab ( Fig. 1d. Supplementary Fig. 2). Altogether, these results revealed that adapted pathogens alter AtNHR2B-GFP protein localization and abundance.
The Pst DC3000 mutant lacking HopG1 is able to grow better in the Atnhr2b mutant background. The finding that the localization and protein abundance of AtNHR2B is altered by adapted pathogens, prompted us to investigate whether AtNHR2B could be a target for Pst DC3000 T3E. The finding that HopG1 localizes to mitochondria 24,27 , together with our previous result that AtNHR2B interacts with proteins localized to mitochondria 32 , led us to hypothesize that HopG1 could target AtNHR2B. To test that hypothesis, we initially evaluated the growth of wild-type Pst DC3000 and the Pst DC3000 ΔhopG1 mutant in wild-type Col-0 and in Atnhr2b mutant plants at 3 days post-inoculation (dpi) (Fig. 2). As an adapted pathogen of Arabidopsis, the wild-type strain Pst DC3000 is able to grow in wild-type Col-0 to 10 7 CFU/cm 2 at 3 dpi. In contrast, the growth of the Pst DC3000ΔhopG1 was ~ 10-fold lower than the growth of wild-type Pst DC3000 in wild-type Col-0. In the Atnhr2b mutant plants, the growth of the wild-type Pst DC3000 was equivalent to its growth in wild-type Col-0 plants. Interestingly, the Pst DC3000ΔhopG1 mutant grew to higher levels (~ 10 7 CFU/cm 2 ) in Atnhr2b mutant plants, and those levels were equivalent to the growth of Pst DC3000 in wild-type Col-0. These The Pst DC3000 effector HopG1 interacts with AtNHR2B. To start dissecting the functional relationship between HopG1 and AtNHR2B, we tested the physical interaction between AtNHR2B and HopG1 with the yeast two-hybrid system using HopG1 as a bait and AtNHR2B as a prey. Yeast co-transformed with pDEST32::HopG1 and pDEST22::AtNHR2B grew on Triple Drop Out (TDO) media (-leu,-trp,-his) containing 15 mM 3-AT indicating the transcriptional activation of histidine biosynthetic genes, as a result of the interaction between HopG1 and AtNHR2B. That interaction is not the result of auto-activation, because yeast transformed with the empty vector pDEST32 and pDEST22::AtNHR2B did not grow on TDO + 15 mM 3-AT (Fig. 3a).
To evaluate the HopG1/AtNHR2B interaction in the appropriate biological context, we transiently coexpressed HopG1 fused to the FLAG epitope (HopG1-FLAG) with AtNHR2B-GFP for co-immunoprecipitation. HopG1-FLAG and free GFP were also co-infiltrated and used as control. Immunoprecipitation of AtNHR2B-GFP using GFP Nanobody/V H H coupled to agarose beads (GFP-Trap ® Agarose, Chromotek), co-immunoprecipitated HopG1-FLAG as detected by Western Blot using Anti-FLAG antibodies (Fig. 3b, Supplementary Fig. 3). In contrast, immunoprecipitation of free GFP did not co-immunoprecipitate HopG1-FLAG, demonstrating that the physical interaction between HopG1 and AtNHR2B also occur in planta and it is not an artifact of the GFP tag. We further evaluated the in-situ interaction of AtNHR2B and HopG1 by co-expressing AtNHR2B fused to the N-terminal half encoding domain of the enhanced yellow fluorescent protein gene (nEYFP), and the HopG1 AtNHR2B-GFP were infiltrated with water (mock), or inoculated with either the adapted pathogen Pstab or the non-adapted pathogen Pst DC3000 at 1 × 10 6 CFU/ml. Treated leaf samples were collected after 24 hpi and imaged by laser scanning confocal microscopy. Confocal images were taken at ×20 magnification and using an excitation and emission wavelengths of 488 nm and 500-535 nm, respectively. White arrows show punctate bodies. Scale bar = 10 µm. (b) The host adapted pathogen Pst DC3000 alters AtNHR2B-GFP localization in Arabidopsis. Arabidopsis transgenic plants expressing AtNHR2B-GFP were mock-treated or inoculated with the adapted pathogen Pst DC3000 and the non-adapted pathogen Pstab at 1 × 10 6 CFU/ml. Treated leaf samples were collected after 24 hpi and imaged by laser scanning confocal microscopy. Confocal images were taken at ×20 magnification and using an excitation and emission wavelengths of 488 nm and 500-535 nm, respectively. White arrows show punctate bodies. Scale bar = 10 µm. (c) The host adapted pathogen Pstab alters AtNHR2B-GFP protein abundance in tobacco plants. N. benthamiana plants transiently expressing AtNHR2B-GFP were infiltrated with water (mock) or inoculated with either the adapted pathogen Pstab or the non-adapted pathogen Pst DC3000 as described for (a). Treated leaf samples were collected after 24 hpi to evaluate protein abundance by Western blot using anti-GFP and anti-actin antibodies. Protein quantification was done by ImageJ software (http:// rsb. info. nih. gov/ ij/), using actin signal for normalization. Number below the protein bands represent the AtNHR2-GFP/Actin ratios. (d) The host adapted pathogen Pst DC3000 alters AtNHR2B-GFP protein abundance in Arabidopsis plants. A. thaliana transgenic plants expressing AtNHR2B-GFP were infiltrated with water (mock), or inoculated with either the adapted pathogen Pst DC3000 or the non-adapted pathogen Pstab as described for (b). Treated leaf samples were collected after 24 hpi to quantify protein abundance by Western blot using anti-GFP and anti-Actin antibodies. Protein quantification was done as described for (c).  33 . The in-situ interaction between AtNHR2B and HopG1 was revealed by the restoration of the yellow fluorescence and shown to occur in the cytoplasm; importantly, amino acids 1-195 in AtNHR2B are important for the interaction as deletion of the region encoding those amino acids abolishes the interaction with HopG1 (Fig. 3c).  6). Co-expression of AtNHR2B-GFP with HopG1-FLAG did not alter the AtNHR2B-GFP protein abundance ( Supplementary Fig. 7). Collectively, these results indicate that HopG1 transiently expressed in N. benthamiana induces cell death that is attenuated by AtNHR2B degrading HopG1.

HopG1 expression causes virulence-related cell death in N. benthamiana that is attenuated by
Since the cell death phenotype can be an HR defense response or necrosis associated with virulence, we resolved between those two alternatives by inoculating N. benthamiana plants transiently expressing HopG1-FLAG with Pstab to evaluate its growth at 1, 2 and 3 dpi. For these experiments, we took advantage of a version of Pstab expressing the fluorescent protein GFP-TIR 34 (Pstab-GFP-TIR), to be able to discriminate Pstab from the A. tumefaciens strains used for transient expression. As an adapted pathogen of N. benthamiana, Pstab-GFP-TIR progressively grew in wild-type N. benthamiana and reached more than 10 8 CFU/cm 2 at 3 dpi (Fig. 4c). In contrast, N. benthamiana plants that were transiently expressing AtNHR2B-GFP, HopG1-FLAG or co-expressing AtNHR2B-GFP and HopG1-FLAG had significantly lower bacterial populations, in comparison with wild-type plants, likely as a result of Agrobacterium-induced PTI (Fig. 4c). Interestingly, among transiently-expressing plants, those expressing HopG1-FLAG showed the highest levels of Pstab-GFP-TIR populations at 1, 2 and 3 dpi. Moreover, while Pstab-GFP-TIR populations at 1 dpi were equivalent between plants transiently expressing AtNHR2B-GFP alone or in combination with HopG1-FLAG, at 2 and 3 dpi Pstab-GFP-TIR populations in plants co-expressing AtNHR2B-GFP and HopG1-FLAG were intermediate when compared with populations in plants expressing either AtNHR2B-GFP or HopG1-FLAG alone (Fig. 4c). These results are consistent with the reduction in cell death phenotype and HopG1-FLAG accumulation in N. benthamiana plants co-expressing AtNHR2B-GFP with HopG1-FLAG. Taken together, these results show that HopG1 triggers plant cell death that is not an HR, but either contributes to plant susceptibility by suppressing PTI, enhances pathogen' virulence or both. Moreover, this virulence activity of HopG1-inducing cell death that promotes pathogen multiplication, is actively attenuated by AtNHR2B.  AtNHR2B-GFP alone. Importantly, less cell death was observed in transgenic plants co-expressing HopG1-FLAG with AtNHR2B-GFP, suggesting that AtNHR2B-GFP also attenuates the HopG1-triggered cell death in Arabidopsis (Fig. 5a). We confirmed that to be the case, by using the same plants' genotypes and pathogen inoculation conditions to quantify cell death by electrolyte leakage at 5-, 10-, and 20-h post-inoculation (hpi). The results showed that as early at 10 hpi, plants expressing HopG1-FLAG have the highest levels of electrolyte leakage that were significantly different from the electrolyte leakage levels in Col-0 plants, as well as in plants expressing AtNHR2B-GFP and plants co-expressing HopG1-FLAG and AtNHR2B-GFP (Fig. 5b).
Because cell death phenotypes are regulated by reactive oxygen species (ROS), and mitochondria are one of the source of ROS, it was necessary to evaluate how HopG1 contributes to the production of ROS in the mitochondria. For that purpose, a mitochondria-specific ROS sensor, MitoTracker Red CM-H 2 XRos (ThermoFisher Scientific, Waltham, MA) was used to evaluate mitochondrial ROS produced after pathogen infection. Inoculation of Pstab triggered an accumulation of mitochondrial ROS in all plant genotypes. However, the levels of accumulation varied between genotypes. Overall, the lowest levels of mitochondrial ROS were observed in wildtype Col-0 and moderate levels were found in transgenic plants expressing AtNHR2B-GFP and co-expressing HopG1-FLAG and AtNH2B-GFP. Remarkably, at 3 min of the assay, the levels of mitochondrial ROS in plants expressing HopG1-FLAG started to increase and those levels peaked at 7 min when they were ~ 3-fold higher than in wild-type Col-0 and ~ 1.5-fold higher than in plants expressing AtNHR2B-GFP or co-expressing AtNHR2B-GFP and HopG1-FLAG (Fig. 5c, Supplementary Table S1). These results showed that the HopG1-mediated cell death is associated with the production of reactive oxygen species in the mitochondria where HopG1 localizes.
HopG1 interferes with AtNHR2B function to promote disease. Our results demonstrating an interaction between HopG1 and AtNHR2B and a possible interplay between both proteins, prompted us to further investigate the antagonism between HopG1 and AtNHR2B. HopG1 was previously shown to interfere with callose deposition when HopG1-HA expressing plants were inoculated with the Pst DC3000 hrcC mutant 24 . In our assays, we used Pstab that also triggers PTI in Arabidopsis as demonstrated by the callose deposits in the wild-type Col-0 plants and in transgenic Arabidopsis plants expressing AtNHR2B-GFP (Fig. 6a). As previously reported, transgenic plants expressing HopG1-FLAG were devoid of callose deposits and were comparable to mock-treated plants. Similarly, plants co-expressing HopG1-FLAG and AtNHR2B-GFP were also devoid of callose deposits suggesting that AtNHR2B is unable to interfere with the PTI-suppressing activities of HopG1 (Fig. 6a).
Consistent with the activities of HopG1 suppressing PTI and the inability of AtNHR2B to interfere with HopG1-mediated PTI suppression, transgenic plants expressing HopG1-FLAG alone or in combination with AtNHR2B-GFP supported 10-fold higher growth of Pstab at 3 dpi in comparison with Pstab growth in Col-0, or in plants expressing AtNHR2B-GFP (Fig. 6b). Taken together, these results demonstrate that in Arabidopsis, HopG1 interferes with plant defense responses even when AtNHR2B-GFP is overexpressed. Moreover, the increased bacterial growth and lack of callose deposition, suggest that similar to the results obtained in N. benthamiana, the HopG1-mediated cell death is not related to the HR but to a virulence mechanism related to HopG1 function.

Discussion
Pst DC3000 deploys a plethora of effectors into the plant cell to interfere with plant defense responses, alter cellular processes and promote bacterial parasitism 6,7 . The Pst DC3000 T3E HopG1 appears to be of paramount importance in the pathogenicity of Pst DC3000 by being one of five effectors constituting the minimal repertoire that makes Pst DC3000ΔhopQ1-1 pathogenic in N. benthamiana 4,23 . Previous studies also demonstrated that HopG1 suppresses the HR in N. benthamiana based on experimental data showing that only the Pst DC3000 hopG1 mutant triggered the HR, whereas the wild type Pst DC3000, or the Pst DC3000 hopG1 (phopG1) complementation strain did not 26 . The elicitation of the HR by the Pst DC3000 hopG1 mutant, suggests that the absence of HopG1 uncovered the HR eliciting activity of other effectors. In contrast, the presence of HopG1 in wild type Pst DC3000, and in the Pst DC3000 hopG1 (phopG1) complementation strain, suppresses the HR. Furthermore, transient expression of the cell death inducer BAX1 36 in N. benthamiana triggered cell death, as expected, but this cell death was not observed when BAX1 was transiently co-expressed with HopG1 26 .
In addition to suppressing the HR, transgenic expression of HopG1-HA in Arabidopsis also led to suppression of PTI, observed as a reduction in callose deposition after infiltration with the Flg 21 peptide, or after inoculation with the non-pathogen Pseudomonas fluorescens 24 . We confirmed these results by showing that Arabidopsis plants expressing HopG1-FLAG and inoculated with the non-adapted pathogen Pstab are also devoid of callose deposits.
Other previous observations on the virulence function of HopG1, revealed its role in the remodeling of the cytoskeleton demonstrated to occur through the interaction of HopG1 with the mitochondrial-localized kinesin motor protein 27 . The mitochondrial localization of HopG1 previously led to the hypothesis that HopG1 alters mitochondrial function, and in support of that hypothesis, Arabidopsis plants expressing HopG1-HA showed reduced oxygen consumption and enhanced ROS levels 24 37 . Such cell death phenotype was explained as a recognition by the plant (HR) given that the effector was from a non-adapted pathogen 36 . However, a Pst DC3000 strain deleted of all effectors and only harboring HopG1 failed to elicit the HR in N. benthamiana 38 , demonstrating that HopG1 by itself is unable to cause the HR. Our data supports the latter findings and further demonstrates that actually, the cell-death phenotype observed is the result of the virulence function of HopG1.
The cell death phenotypes that effectors trigger can be difficult to assess as they might correspond to completely opposite outcomes: the HR, a plant defense response that restricts pathogen proliferation, or necrosis, a virulence function of the pathogen that sustains pathogen proliferation. Thus, to resolve if the HopG1-dependent cell death was HR or virulence-related necrosis, we inoculated the HopG1-FLAG-expressing N. benthamiana plants with its adapted pathogen Pstab to evaluate bacterial growth as evidence of bacterial virulence. We showed that plants expressing Collectively, our results and the results of others highlight that HopG1 is a multi-faceted protein that can suppress the HR, interfere with plant defense responses and cause necrotic cell death associated with chlorotic symptoms. All these phenotypes associated with HopG1 highlight its function early in infection, suppressing immune responses, and later in infection triggering necrotic cell death. The functions of HopG1 are consistent with the hemi-biotrophic lifestyle of Pst DC3000 that combines an early biotrophic phase suppressing cell death responses with a late necrotrophic phase inducing cell death.
Pstab-GFP-TIR was generated by introducing the plasmid pGFP-TIR 34 into Pstab by electroporation. Pstab-GFP-TIR was grown on KB medium supplemented with gentamycin (20 µg/ml). All the strains were grown at 28 °C. Bacterial strains used in this study are listed in Supplementary Table S2. benthamiana plants. Infiltrated leaves were harvested for protein extraction followed by immunoprecipitation using GFP Nanobody/VHH coupled to agarose beads. Immunoprecipitated samples were separated by SDS-PAGE electrophoresis and transferred to a Polyvinylidene difluoride (PVDF) membrane for Western Blot analysis using anti-FLAG and anti-GFP antibodies. (c) Bimolecular fluorescence complementation analysis shows the interaction between HopG1 and AtNHR2B in planta. HopG1 fused to the C-terminal encoding half of EYFP (HopG1-cEYFP), was transiently co-expressed in N. benthamiana with AtNHR2B fused to the N-terminal encoding half of EYFP (nEYFP-AtNHR2B), or with the truncated version AtNHR2B Δ1-195 fused to the N-terminal encoding half (nEFYP-AtNHR2B Δ1-195 ). At 3 dpi, infiltrated leaves were imaged by laser scanning confocal microscopy using excitation and emission wavelengths of 488 nm and 500-535 nm, respectively. Confocal images were taken at ×63 magnification. Scale bars = 50 µm. Transient expression in N. benthamiana. Agrobacterium tumefaciens GV2260 harboring constructs of interest were induced and infiltrated into fully expanded leaves of 3-week old N. benthamiana plants using a needle-less syringe as previously described 31 . Infiltrated leaves were used to visualize protein localization or in situ protein-protein interaction by laser scanning fluorescence microscopy, or to identify proteins of interest by Western blot.

Live-cell imaging. Nicotiana benthamiana plants transiently expressing AtNHR2B-GFP, co-express-
ing AtNHR2B-nEYFP with HopG1-cEYFP, or AtNHR2B Δ1-195 nEYP with HopG1-cEYFP were used to evaluate AtNHR2B-GFP localization, or AtNHR2B/HopG1 interaction by bimolecular fluorescence complementation (BiFC), respectively. GFP and YFP signals were imaged on a Leica SP2 or Leica Stellaris 8 Laser Scanning Confocal Microscope (Leica Microsystems, Buffalo Grove, IL) at 48hpi. Arabidopsis plants expressing AtNHR2B-GFP were also evaluated by laser scanning fluorescence microscopy. GFP and YFP fluorescence was imaged using excitation wavelength of 488 nm and an emission wavelength of 500 to 535 nm.   Bacterial multiplication assays. Leaf disks (0.5 cm 2 ) from inoculated plants were collected, seriallydiluted and plated as previously described 31 . Each experiment was repeated three times.

Bacterial inoculation into
Yeast two-hybrid assay. The  Co-immunoprecipitation. Four-week-old N. benthamiana plants were co-infiltrated with two strains of A.
tumefaciens strains harboring AtNHR2B-GFP and HopG1-FLAG, or A. tumefaciens strains harboring AtNHR2B-GFP and 35S-GFP. Infiltrated leaves were collected at 2 dpi and tissue, ground in liquid nitrogen and homogenized in 1 mL of co-immunoprecipitation extraction buffer (100 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 10 mM MgCl2, 10% Glycerol, 0.2% Nonidet P-40, 1 mM PMSF, 5 mM DTT, 1X Proteinase inhibitor cocktail (Sigma Aldrich, St. Louis, MO). Protein extracts were incubated for 30 min on ice and centrifuged at 4 °C for 30 min at 13,000 rpm. Supernatants containing extracted proteins were collected in a pre-chilled 1 ml Eppendorf tube. Protein concentration was measured by Bradford Assay (BioRad, Hercules, CA) and protein expression in input samples was confirmed by Western blot as described above. One milligram of total protein extract was mixed with 25 µl of GFP Nanobody/V H H coupled to agarose beads (GFP-Trap ® Agarose, Chromotek) and incubated for 3 h and 30 min at 4 °C with end to end rocking. After incubation, protein complexes bound to beads were washed once with 1 × TBS buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.5) and twice with 1 × TBS buffer (50 mM Tris-HCl, 500 mM NaCl, pH 7.5). Protein complexes bound to the beads were eluted in 2 × SDS protein loading buffer, loaded and ran into an SDS-PAGE gel and transferred to PVDF membranes. Proteins were detected by Western Blot as described above.
Cell death assay. Five-week old wild-type Col-0, AtNHR2B-GFP, HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG plants were syringe-inoculated with Pstab at OD 600 = 0.02 (1 × 10 6 CFU/ml). Control plants were inoculated with water only. At 24 hpi, six to nine inoculated leaves were detached from six independent plants for each genotype. Collected leaves were stained with 0.05% trypan blue for 45 min at room temperature and washed twice with PBS 35 . Images were taken on a light microscope using bright field.
Callose deposition. Five-week old wild-type Col-0, AtNHR2B-GFP, HopG1-FLAG and HopG1-FLAG/AtNHR2B-GFP were sprayed with 20 µM dexamethasone to induce expression of HopG1-FLAG. At 24 h after dexamethasone treatment, plants were syringe-inoculated with Pstab at OD 600 = 0.02 (1 × 10 6 CFU/ml). Ten leaves from six independent plants for each genotype and inoculated with Pstab or infiltrated with water were detached after 24 hpi and stained with 5% aniline blue to visualize callose deposits. Images were taken by Nikon 90i upright scanning laser confocal microscope (Nikon) using a DAPI (4′,6-diamidino-2-phenylindole) filter with excitation wavelength of 405 nm and an emission wavelength of 450-510 nm. Figure 5. Transgenic expression of HopG1 in Arabidopsis causes cell death that is attenuated by AtNHR2B. (a) Cell death phenotype in HopG1 expressing plants is reduced by expression of AtNHR2B. Five-week-old wildtype Col-0, AtNHR2B-GFP, HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG plants were syringe-inoculated with Pstab at a concentration of 1 × 10 6 CFU/mL or infiltrated with water (mock). HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG plants were sprayed with 20 μM DEX 24 h prior bacterial inoculation. At 24 hpi, treated leaves were detached and stained with 0.05% trypan blue. Images were taken on a light microscope using bright field. (b) Ion leakage is enhanced in plants expressing HopG1. Five-week-old wild-type Col-0 and transgenic plants expressing AtNHR2B-GFP, HopG1-FLAG, and AtNHR2B-GFP/HopG1-FLAG were syringe-inoculated with Pstab at a concentration of 1 × 10 6 CFU/ml to evaluate electrolyte leakage at 5, 10, 20 and 24 hpi. HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG were sprayed with 20 μM DEX 24 h prior to bacterial inoculation. Bars represent means and standard deviation of conductivity values (µS/cm). Different letters above bars represent statistically significant difference in electrolyte leakage among plant genotypes using a one-way ANOVA and Fisher's multiple comparison test with P-value ≤ 0.05. (c) Arabidopsis plants expressing HopG1-FLAG generate higher levels of ROS from mitochondrial origin. Wild-type Col-0 and transgenic plants expressing HopG1-FLAG, AtNHR2B-GFP and AtNHR2B-GFP/HopG1-FLAG were flood-inoculated with the non-adapted bacterial pathogen Pstab at a concentration of 1 × 10 6 CFU/ml, or infiltrated with water (mock). HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG were sprayed with 20 μM DEX 24 h prior bacterial inoculation. Leaf disks were incubated with MitoTracker Red CM-H2XRos and mitochondrial ROS fluorescence was measured using excitation/emission wavelengths of 570 nm and 535 nm, respectively. www.nature.com/scientificreports/ Electrolyte leakage assay. Five-week old wild-type Col-0, AtNHR2B-GFP, HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG plants were induced by dexamethasone prior to inoculation with Pstab at OD 600 = 0.02 (1 × 10 6 CFU/ml). Twelve leaves were collected immediately after infiltration and 0.5 cm 2 leaf disks were excise and rinsed with 50 ml for 30 min, to wash ions released during disc excision. After 30 min, water was replaced with 12 ml of fresh deionized water and conductivity was measured at 5, 10, and 20 hpi using the Orion Star A215 conductivity cell (013005MD) (Thermo Fisher Scientific).
Mitochondrial ROS production. Five-week old wild-type Col-0, AtNHR2B-GFP, HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG plants grown on soil were sprayed with a 20 µM dexamethasone solution supplemented with 0.01% silwet, 12 h before inoculation. Leaf disks cut out using a 1.2 cm 2 core-borer and harvested tissue was transferred to clear bottom plates for fluorometric analysis and submerged in Pstab inoculum at a final Five-week-old wild-type Col-0, AtNHR2B-GFP, HopG1-FLAG, and AtNHR2B-GFP/HopG1-FLAG were syringe-inoculated with the non-adapted bacterial pathogen Pstab at a concentration of 1 × 5 /ml. HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG were sprayed with 20 μM DEX 24 h prior to bacterial inoculation. Inoculated leaves were detached at 24 hpi and stained with 5% aniline blue staining to evaluate callose deposition. Images were taken using a confocal microscope under DAPI filter. Scale bar = 20 μm. (b) Transgenic expression of HopG1-contributes to bacterial growth. Wild-type Col-0, AtNHR2B-GFP, HopG1-FLAG, and AtNHR2B-GFP/HopG1-FLAG were syringe inoculated with Pstab at a concentration of 1 × 10 6 CFU/ml. HopG1-FLAG and AtNHR2B-GFP/HopG1-FLAG were sprayed with 20 μM DEX 24 h prior bacterial inoculation. Leaf samples were collected at 0 and 3 dpi. Bars represent means of CFU/cm 2 . Different letters above bars represent statistically significant difference in bacterial growth among plant genotypes using a one-way ANOVA and Tukey's multiple comparison test with P-value ≤ 0.05. All above experiments were repeated three times with similar results.