A multidisciplinary approach to investigate plant-pathogen interactions

José Sebastián Rufián Plaza TESIS DOCTORAL 2015 A multidisciplinary approach to investigate plant-pathogen interactions Universidad de Málaga Dpto. Biología Celular, Genética y Fisiología Área de Genética Departamento de Biología Celular, Genética y Fisiología TESIS DOCTORIAL A multidisciplinary approach to investigate plant-pathogen interactions Autor: José Sebastián Rufián Plaza Directores: Carmen R. Beuzón y Eduardo Rodriguez Bejarano Málaga, 2015 AUTOR: Jose Sebastián Rufián Plaza http://orcid.org/0000-0002-3871-3706 EDITA: Publicaciones y Divulgación Científica. Universidad de Málaga Esta obra está bajo una licencia de Creative Commons Reconocimiento-NoComercialSinObraDerivada 4.0 Internacional: http://creativecommons.org/licenses/by-nc-nd/4.0/legalcode Cualquier parte de esta obra se puede reproducir sin autorización pero con el reconocimiento y atribución de los autores. No se puede hacer uso comercial de la obra y no se puede alterar, transformar o hacer obras derivadas. Esta Tesis Doctoral está depositada en el Repositorio Institucional de la Universidad de Málaga (RIUMA): riuma.uma.es 3 Área de Genética Departamento de Bilogía Celular, Genética y Fisiología Dra. Carmen R. Beuzón, Profesora titular del Área de Genética del Departamento de Biología Celular, Genética y Fisiología de la Universidad de Málaga, y Dr. Eduardo Rodriguez Bejarano, Catedrático del Área de Genética del Departamento de Biología Celular, Genética y Fisiología de la Universidad de Málaga, CERTIFICAN: Que Don José S. Rufián Plaza ha realizado bajo su dirección y supervision en la Unversidad de Málaga el trabajo titulado “A multidisciplinary approach to investigate plant-pathogen interactions”, presente en esta memoria, que constituye su tesis doctoral. Y para que así conste y tenga los efectos que corresponden, en cumplimiento con la legislación vigente, extienden el presente informe. En Málaga, a 6 de noviembre de 2015. Carmen R. Beuzón Eduardo R. Bejarano 4 5 PROPUESTA DE TRIBUNAL Presidente Dr. Miguel Ángel Botella Mesa Departamento de Biología Molecular y Bioquímica Universidad de Málaga Secretario Dr. Pablo Tornero Feliciano Instituto de Biología Molecular y Celular de Plantas Universidad Politécnica de Valencia-CSIC Vocales Dr. Francisco Ramos Morales Departamento de Genética Universidad de Sevilla Dra. Natalia Requena Sánchez Department of Molecular Phytopathology Karlsruhe Institute of Technology Dra. Marta Martín Basanta Departamento de Biología Universidad Autónoma de Madrid Suplentes Dr. David Posé Padilla Departamento de Biología Molecular y Bioquímica Universidad de Málaga Dr. Francisco Manuel Cazorla López Departamento de Microbiología Universidad de Málaga 6 7 “When you make the finding yourself – even if you’re the last person on Earth to see the light – you’ll never forget it” Carl Sagan ! 8 15 INDEX General Introduction 19 Pseudomonas syringae: an archetypal plant-pathogen 21 The type III secretion system: a molecular weapon of Gram-negative bacteria 24 The plant immune system: An unespecializad specific system 25 The Plant - P. syringae interaction: A molecular battle between effectors, their targets and the NLR proteins 28 Objectives 31 Experimental procedures 33 Chapter 1: Dynamics of heterogeneous populations of Pseudomonas syringae within plant tissues reveal a diversity of interactions 47 Introduction 48 Results 52 Discussion 63 Chapter 2: A bistable switch controls virulence of bacterial plant pathogen Pseudomonas syringae 67 Introduction 68 Results 73 Discussion 82 Chapter 3: Auto-acetylation on K289 is not essential for HopZ1a-mediated plant defense suppression 87 Introduction 88 Results 92 Discussion 105 Chapter 4: The bacterial effector HopZ1a acetylates MKK7 to suppress plant defense responses 111 Introduction 112 Results 120 Discussion 135 Concluding Remarks 141 Conclusions 147 Resumen en castellano 151 References 163 16 Abbreviations, CDPK:!Calcium-Dependent!Protein!Kinase.! CFP:!Cyan!Fluorescent!Protein.! Cfu:!Colony-formin!units.! CI:!Competitive!Index.! COI:!Cancelled-Out!Index.! Col-0:!Arabidopsis*thaliana!cultivar!Columbia.! DC3000:!Pseudomonas*syringae!pv.!tomato!DC3000.! DEX:!Dexamethasone.! Dpi:!Days!Post!Inoculation.! ETI:!Effector-Triggered!Immunity.! GFP:!Green!Fluorescent!Protein.! HGT:!Horizontal!Gene!Transfer.! Hop:!Hrp-outer!protein.! HR:!Hypersensitive!Response.! JA:!Jasmonic!Acid.! MAPK:!Mitogen-Activated!Protein!Kinase.! NLR:!Nucleotide!binding!Leucin!Rich!repeat.! PAMP:!Pathogen-Asociated!Molecular!Pattern.! Pf55:!Pseudomonas*fluorescens!55.! Pph:!Pseudomonas*syringae!pv.!phaseolicola.! PR1:!Pathogenesis-Related!1.! PRR:!Pattern-Recognition!Receptor.! Psy:!Pseudomonas*syringae!pv.!syringae.! PTI:!Pattern-Triggered!Immunity.! Pto:!Pseudomonas*syringae!pv.!tomato.! Pv:!Pathovar.! RLCK:!Receptor-Like!Cytoplasmic!Kinase.! RLK:!Receptor-Like!Kinase.! ROS:!Reactive!Oxygen!Species.! SA:!Salycilic!Acid.! SAR:!Systemic!Acquired!Resistance.! T3E:!Type!Three!Effectors.! T3SS:!Type!Three!Secretion!System.! YFP:!Yellow!Fluorescent!Protein.! ZAR1:!HopZ1a-Activated!Resistance.! ZED1:!HopZ1a-ETI-Deficient.! !, 18 ! , ! ! ! ! ! ! ! General'introduction! ! !, General introduction 20 !, General introduction 21 The estimation of human global population from the United Nations indicates that the world total population could rich 9.15 billion in 2050 (Nikos Alexandratos and Bruinsma, 2012). This increase on the world population must be accompanied by an increase on food production. A major restriction to food production is crop losses due to plant diseases. Plant pathogens, including fungi, oomycetes, bacteria and viruses, are a major constraint to production with a strong economic impact. Understanding plantpathogen interactions is therefore paramount for the development of robust and sustainable strategies to control disease and improve crop production. This thesis focuses on the interaction between the plant-pathogenic bacteria Pseudomonas syringae and its hosts. It includes molecular and cellular studies at both sides of this interaction, canvasing from the bacterial virulence mechanisms to the plant defense response pathways. Pseudomonas syringae: an archetypal plant pathogen Pseudomonas syringae is a bacterial plant pathogen that has been extensively researched from the 1980s. P. syringae strains have been ranked as number 1 of plantpathogenic bacteria based on its scientific and economic importance (Mansfield et al., 2012). P. syringae is a gram-negative bacterial pathogen that colonizes the aerial part of the plant, including leaves and fruits. It has a dual lifestyle, with an initial epiphytic phase on the surface of the plant, and an endophytic phase inside the plant apoplast, where the bacterial population survives and proliferates establishing a hemibiotrophic interaction with the host plant. The apoplast (the intercellular space between the cells of the parenchima) is the environment where the bacterial population grows better and this growth does not require killing of the plant cells for nutrients, as necrotrophs do. However, the interaction cannot be classified as fully biotrophic since it involves host cell death at later stages of the infection process, either due to the virulence action of the bacteria or to the defenses triggered by the plant in response to the attack. 1. Pseudomonas syringae classification Pseudomonas syringae is a complex of strains that possess a wide host range, including many economically important crops, woody plants and weeds, such as the model plant Arabidopsis thaliana. Plants are generally resistant to most strains, and the General introduction 22 ability of a given strain to cause disease in a given host is considered the exception rather than the rule. On the basis of this more limited host range of the different isolates, the P. syringae complex is divided into more than 50 pathovars (Young, 2010). Furthermore, strains belonging to the same pathovar can still differ on their interactions with different cultivars or ecotypes of the same plant species, giving raise to an additional subdivision into races. The sequencing annotation of the whole genomes three model strains belonging to different pathovars of the P. syringae complex: pathovar tomato, Pto DC3000 (Buell et al., 2003), pv. phaseolicola, Pph 1448A (Joardar et al., 2005), and pv. syringae, Psy B728a (Feil et al., 2005) provided a large amount of information and tools and determined an qualitative leap for the field. 2. Pseudomonas syringae pv. tomato, phaseolicola and syringae Pseudomonas syringae pv. tomato is the causal agent of bacterial speck in tomato plants. In (1986), Cuppels generated the strain DC3000, a rifampicin resistant derivative of a wild-type strain, which was used for auxotrophy and pathogenicity studies. Some years later, Whalen et al. (1991) showed the ability of DC3000 to produce disease in the model plant Arabidopsis thaliana. The ability of DC3000 to infect both tomato and Arabidopsis plants made this strain of great interest for the plant pathogen interactions field. DC3000 is a weak epiphyte (Boller and Felix, 2009) compared with other strains such as B728a; while B728a can maintain a high epiphytic population for several days, most of the epyphitic population of DC3000 dies in less than 48h. Thus, DC3000 needs to enter into the apoplast in order to survive. One of the mechanisms that DC3000 uses to effectively enter into the host tissue is mediated by the production of a polyketide toxin called coronatine, a molecular mimic of the planthormone methyl-jasomate (Weiler et al., 1994). Upon perception of bacteria entering the plant apoplast, the plant induces closure of stomata in order to prevent further bacterial entry. Coronatine activates cellular pathways that result in the reopening of stomata, thus allowing high numbers of bacteria to invade the apoplast (Melotto et al., 2008). Pseudomonas syringae pv. phaseolicola is the etiological agent of halo blight disease in common bean (Phaseolus vulgaris). This disease is characterized by the appearance of water-soaked lesions in leaves and pods, often surrounded by a chlorotic halo (W.H., 1926), and it is prevalent worldwide. Since bacteria can colonize and General introduction 23 survive into the dry seeds, fields infected by P. syringae pv. phaseolicola are usually destroyed to prevent dissemination. One of the control strategies to prevent halo blight is the rotation of resistant cultivars. Based on the resistance/susceptibility interactions between eight bean cultivars and 175 different strains of P. syringae pv. phaseolicola, the pathovar was divided in nine races (Taylor et al., 1996). The strains belonging the race 6, including 1448A, are able to produce disease in all cultivars tested. Pseudomonas syringae pv. syringae is the most heterogeneous group among the P. syringae pathovars. It includes strains that produce from brown spot in bean, to blossom blight in pear, or apical necrosis in mango trees, among other diseases. One of the characteristics of many of the strains of this group, including B728a, is the production of the phytotoxin Syringolin A (Ramel et al., 2009). This toxin acts inside the plant cell as a proteasome inhibitor, suppressing defense responses and promoting bacterial proliferation (Schellenberg et al., 2010). Furthermore, Syringolin A allows bacteria to move from the primary infection site through the xylem (Misas-Villamil et al., 2011). Figure 1. Diseases caused by different P. syringae strains. (A) Bacterial speck of tomato produced by P. syringae pv. tomato. (B) Halo blight of bean produced by P. syringae pv. phaseolicola. (C) Bacterial brown spot of lima (top panel) and apical necrosis of mango, produced by P. syringae pv. syringae. Fotograph authorship: A, Top panel: A. Collmer. Lower panel: C. Smart. B, Top panel: H. Schwartz. C, Top panel: R. Mulrooney. General introduction 24 The Type III Secretion System: a molecular weapon of gram-negative bacteria. 1. Structural and regulatory components of the T3SS Many P. syringae strains possess toxins that contribute to bacterial virulence. However, all strains require a Type Three Secretion System (T3SS) to be pathogenic. The T3SS is a complex nanomachine that exports proteins across the bacterial inner and outer membrane into the host cell cytosol. The components of the T3SS are very conserved among gram-negative bacteria. The genes encoding such conserved proteins are named as hrc (hrp conserved). The hrp/hrc genes are clustered in pathogenicity islands located in the chromosome within a single gene cluster called the hrp locus (hypersensitive response and pathogenicity). The T3SS is not constitutively expressed in the bacteria, but induced under certain conditions, such as in some minimallaboratory medium, or the plant apoplast. The central regulatory element of the hrp/hrc genes is HrpL, a member of the ECF family of alternative sigma factors that activates the expression of genes containing a consensus sequence, the hrp box, within their promoters (Xiao and Hutcheson, 1994; Fouts et al., 2002). The expression and assembly of the T3SS is an intricate process regulated by positive and negative feedbacks (Ortiz-Martin et al., 2010a; Ortiz-Martin et al., 2010b) in which structural proteins can also play a role in regulation, e.g. HrpA, the main component of the T3SS pillus, which positively regulates expression of hrpL (Preston et al., 1998). The structural proteins of the T3SS form a needle composed of an innermembrane ring, an outer membrane ring and a pillus, through which the proteins are secreted. In addition to regulatory and structural proteins, T3SS also include harpins or helper proteins, chaperones, and effectors. Harpins are secreted but not translocated into the host cell and play an auxiliary role in the penetration of the pillus into the plant cell wall, and in effector delivery (Kvitko et al., 2007). Chaperons are small proteins essential for the appropriated folding of the proteins of the system. These proteins are bound to the corresponding effector within the bacterial cytoplasm, protect them from aggregation or degradation, and may direct them to the needle complex when required. Finally, effectors are the only proteins translocated into the plant cell cytoplasm, where they collectively contribute to virulence by modifying host cellular processes to allow bacterial survival and proliferation Objectives 1. To characterize the impact and source of diversity within pathogen populations during the interaction with the plant host, using single-cell technology and the model bacterial pathogen Pseudomonas syringae. 1.1 To analyze how strains of P. syringae differing in their virulence relate or interfere with each other during colonization of the host. 1.2 To look for a mechanistic explanation for the phenotypic heterogeneity observed in the development of P. syringae populations during colonization of the plant apoplast. 2. To characterize new molecular mechanism of plant defense suppression, through the analysis of how P. syringae effector HopZ1a suppresses plant defenses. 2.1 To evaluate and characterize the role of lysine 289 on HopZ1a function. 2.2 To evaluate plant MAP Kinase Kinase 7 as putative virulence target of HopZ1a. ! ! ! ! ! ! ! Experimental,procedures! Experimental procedures 34 Experimental procedures 35 Bacterial strains and growth conditions Bacterial strains used in this work are listed in the Table 1 of each chapter. All bacterial strains were grown in Lennox Broth (LB, Lennox (1955)), unless otherwise stated. Escherichia coli DH5α (Hanahan, 1983) derivatives were grown at 37ºC. Pseudomonas syringae strains and Agrobacterium tumefaciens C58C1 (Deblaere et al., 1985), derivatives were grown at 28°C. Antibiotics were used at the following concentrations: for E. coli DH5a, ampicillin (Amp), 100 mg/ml, kanamycin (Km), 50 mg/ml; gentamycin (Gm), 10 mg/ml and chloramphenicol (Cm), 6 mg/ml. For P. syringae and P. fluorescens strains, Km 15 mg/ml, Gm 10 mg/ml and nitrofurantoin (Nf), 50 µg/ml. For A. tumefaciens, Km 50 mg/ml, tetracycline (Tc) 5 µg/ml and rifampicin (Rf) 50 µg/ml. All plates used to grow plant-extracted bacteria contained cycloheximide (2 µg/ml) to prevent fungal contamination. Plant material Phaseolus vulgaris bean cultivar Canadian Wonder plants were grown at 23°C, 95% humidity, with a controlled photoperiod of 16h light/ 8h dark with a light intensity of 200 µmol/m2/s. Arabidopsis thaliana Col-0 and derivatives were grown in soil, or for disease development assays, in jiffy-7 (Jiffy Products Ltd, Norway). In either case, they were grown in temperature-controlled chambers, at 21ºC with a controlled photoperiod of 8h light/ 16h dark with a light intensity of 200 µmol/m2/s. Nicotiana benthamiana was grown in soil in temperature-controlled chambers, at 21ºC with a controlled photoperiod of 16h light/ 8h dark with a light intensity of 200 µmol/m2/s. Arabidopsis zar1-1 (Lewis et al., 2010) and DEX-MKK7 (Zhang et al., 2007) plants were crossed to obtain homozygous zar1-1/DEX-MKK7 lines. Bacterial inoculation and recovery from plant leaves For P. syringae inoculum preparation, bacterial lawns were grown on LB plates for 48 h at 28°C, collected and suspended in 2 mL of 10 mM MgCl2. The OD600 was adjusted to 0.1, corresponding to 5 x 107 colony forming units (cfu/mL) and serial dilutions made to reach the desired inoculum concentration. Plant inoculation by infiltration to be used for either microscopy, bacterial growth assays, symptoms development on bean leaves, or PR1 accumulation, in either bean Experimental procedures 36 plants or Arabidopsis were carried out as follows: one fully expanded leaf from a 10days old bean plant, or three fully expanded young leaves from 5-week-old Arabidopsis plants were pressure infiltrated using a 1-mL syringe without needle. The inoculum concentration for microscopy varied from 5 x 107 cfu/mL to 5 x 104 cfu/mL, depending on the experiment. For bacterial growth assays, the inoculum dose used was 5 x 104 cfu/mL, unless otherwise stated, and for PR1 accumulation experiments 5 x 105 cfu/mL. For standard growth assays, three 10-mm-leaf discs were taken from either the inoculated or outside the inoculated area and ground in 1 mL MgCl2. Serial dilutions were plated and bacteria enumerated. Dip-inoculation for microscopy or growth assays was carried out by dipping leaves for 30 seconds in a 5 x 107 cfu/ml mixed bacterial suspension in 10 mM MgCl2 and 0.02% Silwett L-77 (Crompton Europe Ltd, Evesham, UK). Infiltration of bean leaves to be analyzed for flow cytometry was carried out after dipping a whole leaf into a 5 x 105 cfu/ml bacterial solution in 0.01% Silwett L-77 (Crompton Europe Ltd, Evesham, UK), using a pressure chamber. Four days post inoculation (dpi) bacteria were recovered from the plant by an apoplast fluid extraction. This extraction was carried out by pressure infiltrating a full leaf with 10 ml of a 10 mM MgCl2 solution inside a 20 ml syringe. Following 5 cycles of pressure application, the flow-through was removed and placed into a fresh 50 ml tube, and the leaf retained within the syringe introduced into another. Both tubes were centrifuged at 4ºC for 30 min at low speed (900 g). Pellets were resuspended into 1 ml of MgCl2 and analyzed by flow cytometry. For Arabidopsis symptom visualization, 3-week-old plants were sprayed with a bacterial suspension containing 5x107 cfu/ml in 10 mM MgCl2 containing 0.02% Silwet-L77 (Crompton Europe Ltd, Evesham, UK). Plants were kept covered for 24h to keep humidity high. Competitive index and cancelled-out assays A detailed protocol for Competitive Index (CI) assays is attached as Apendix 1 (Macho et al., 2016). To calculate LBCIs, 500 µl of a 5x104 cfu/ml mixed inoculum, containing equal cfu of wild type and derivate strains, was inoculated into 4.5 ml of LB medium and grown for 24h at 28ºC with aeration. Serial dilutions were then plated onto LB agar Experimental procedures 37 and LB agar with the corresponding antibiotic, to determine the precise ratio between the co-inoculated strains. In plants, competitive index and cancelled-out (COI) assays were performed as previously described (Macho et al., 2007). Using a blunt syringe, 4to 5-week old plants were inoculated with a 5x104 cfu/ml mixed bacterial suspension, containing equal numbers of wild type and derivative strains. Serial dilutions of the inoculum were plated onto LB agar and LB agar with kanamycin to confirm dose and relative proportion between the strains, which should be close to one. At 4 dpi, three 10mm-diameter leaf discs were homogenized into 1 ml of 10 mM MgCl2, by mechanical disruption. Bacteria were enumerated by plating serial dilutions onto LB agar with cycloheximide, and LB agar with kanamycin and cycloheximide, to differentiate the strains within the mixed infection. Bacterial enumeration was carried out in the dilution displaying between 50 and 500 colonies per plate. The CI is defined as the mutant-towild type ratio within the output sample divided by the mutant-to-wild type ratio within the input (inoculum) (Freter et al., 1981; Taylor et al., 1987). The cancelled-out index (COI) is calculated dividing the output ratio between the strain expressing two effectors and the strain expressing one effector, by their input ratio (Macho et al., 2010a). Competitive and cancelled-out indices shown are the mean of three replicates displaying typical results from at least three independent experiments. Errors bars represent standard error. Each CI or COI was analyzed using a homoscedastic and 2-tailed Student’s t-test and the null hypothesis: mean index is not significantly different from 1, or from other mean value (P value < 0.05). Fluorescent labelling of bacterial strains Constitutively expressed fluorescent reporter genes (eCFP or eYFP [enhanced cyan, and yellow fluorescent proteins, respectively]) were introduced into the chromosome of Pph strains 1448A and 1449B using a Tn7 delivery system (Lambertsen et al., 2004): plasmids used are listed in Table 3 of chapter 1. Plasmids were introduced into Pph strains by tetraparental mating, as previously described (Lambertsen et al., 2004). PCR using primers Tn7-GlmS and Tn7R109 (Lambertsen et al., 2004), as well as Southern blot analysis using aacC1 (GmR) as a probe, were used to confirm the correct and unique insertion of the transposon within the genome. Bacterial strains carrying chromosome-located transcriptional fusions of the hrp genes hrpL, hrcU and hopAB1 to a promoterless gfp gene were generated using an adaptation Experimental procedures 38 of Zumaquero et al. (2010). The hrpL and hopAB1 genes are encoded as monocistronic units, whereas hrcU is the last gene of an operon (Rahme et al., 1991; Xiao and Hutcheson, 1994; Jackson et al., 2000). For each gene, two fragments of approximately 500 pb were amplified from Pph 1448A genomic DNA using iProof High-Fidelity DNA Polymerase (Bio-Rad, USA); one fragment corresponding to the 3’ end of the ORF, including the STOP codon, and the other corresponding to the sequence immediately downstream the STOP codon. The reverse primer for the first fragment and the forward primer for the second share a 16 bp overlapping region, including the T7 primer sequence and an EcoRI restriction site, providing homology and a cloning site between both fragments. All primers used are listed in Table 2 of chapter 2. Each reaction was carried out at 94ºC for 3 min, followed by 20 cycles at 94ºC for 20s, 55ºC for 30s, and 72ºC for 50s, followed by 7 min at 72ºC, and the reaction mixture contained 0.64 mM deoxynucleoside triphosphate (dNTP) mix, 5% dimethylsulfoxide (DMSO), 0.4 ng of each primer, 1 ng of genomic DNA, the appropriate enzyme buffer, and commercial ultrapure water (Nalgene, Rochester, NY, USA). Five µl of each gelpurified PCR product were used, without additional primers or template, in a PCR reaction consisting of 8 cycles at 94ºC for 30s, 52ºC for 1 min, and 72ºC for 1 min, finishing with 7 min at 72ºC. Five µl of this reaction, containing a single fragment of approximately 1 Kb, was used as a template for an additional amplification with the forward primer from the first amplified fragment and the reverse from the second (0.4 M of each), 0.64 mM dNTP mix, 5% DMSO, the corresponding buffer, and ultrapure water, for a reaction consisting of 20 cycles at 94°C for 20 s, 53°C for 30 s, and 72°C for 1 min, finishing with 7 min at 72°C. The resulting fragments, including the end of each ORF and its downstream sequence separated by an EcoRI site, were A/T cloned into pGEM-T (Promega, USA) and fully sequenced to discard mutations, giving raise to pDLM3 (phopAB1-EcoRI), pDLM4 (phrcU-EcoRI), and pDLM5 (phrpL-EcoRI). Plasmid pZEP07 (Hautefort et al., 2008) was used as template for PCR-amplification of a fragment containing a promoterless gfp gene carrying its own ribosomal-binding site (Willmann et al., 2011), followed by an EcoRV site and a chloramphenicol resistance cassette. This fragment was A/T cloned into pGEM-T (Promega, USA) generating pDLM1. A fragment containing the nptII kanamycin resistance gene, flanked by FRT sites (Flipase Recognition Target), was PCR-amplified using iProof High-Fidelity DNA Polymerase (Bio-Rad, USA), pDOC-K (Lee et al., 2009) as a template, and the Experimental procedures 39 corresponding primers, and clone into the EcoRV site from pDLM1, to generate pDLM2. Plasmid pDLM2 was used as a template to amplify a fragment containing the promoterless gfp gene with its RBS, the kanamycin resistance gene, and the chloramphenicol resistance gene, and this fragment cloned into pDLM3, pDLM4 and pDLM5, digested with EcoRI and blunt-ended by treatment with the Klenow polymerase fragment (Takara, Japan) generating plasmids pDLM6, pDLM7 and pDLM8, respectively. These resulting plasmids were introduced by electroporation into Pph 1448A, as previously described (Zumaquero et al., 2010), and the transformation plated into LB plates supplemented with kanamycin. Replicas of the resulting colonies were carried out on LB plates supplemented with ampicillin (300 µg/ml) to determine which clones were the result of plasmid integration (a single recombination event), and which the result of allelic exchange (a double recombination event). Southern blot analysis, using the nptII gene as a probe, was used to confirm that allelic exchange occurred at a single and correct position within the genome. Microscopy Sections of inoculated P. vulgaris leaves (approximately 5 mm2) were excised with a razor blade, and mounted onto slides in double-distilled H2O (lower epidermis toward objective) under a 0.17 mm coverslip. Images of the leaf mesophyll were taken using the Leica SP5 II confocal microscope (Leica Microsystems GmbH, Germany). Variable AOTF filters were used for the visualization of the following fluorophores (excitation/ emission): eYFP (514 nm/ 525 to 600 nm), eCFP (458/ 465 to 505 nm), plant autofluorescence (458/ 605 to 670 nm) and VENUS (515 nm/ 525 to 600 nm). Z series imaging were taken at 1 mm or 10 mm intervals when using 40x or 10x objectives respectively. Images were processed using Leica LAS AF (Leica Microsystems). CCID analyzes were performed as described in Godfrey et al. (2010) using Fiji distribution of ImageJ software. Apoplast-extracted bacteria were stained with FM4-64 following instructions from the provider (Life Technologies), and analyzed using the Leica DMR fluorescence microscope (Leica Camera). Experimental procedures 40 For Bimolecular Fluorescence Complementation (BiFC), leaf sections were excised with a razor blade at 20 hours after Agrobacterium tumefaciens inoculation (described below) and mounted into slides in distillated water. Flow Cytometry and Cell Sorting Five hundred µl of an overnight P. syringae culture in LB was washed twice into MgCl2, and added to 4.5 ml of Hrp-inducing medium (HIM, with 10 mM fructose and pH 5.7; (Huynh et al., 1989)). Bacterial cultures at the indicated growth stage, as well as apoplast-extracted bacterial suspensions, were analyzed using a Cytomics FC500-MPL cytometer (Beckman Coulter). Stationary cultures were sorted using a MoFloTM XDP cytometer (Beckman Coulter). Cultures of wild type Pph 1448A or Pph hopAB1::gfp were analyzed, and based on this analysis, gates were drawn to separate the cells displaying fluorescence levels overlapping those of a 1448A non-GFP bacterial population used as a negative control, from cells expressing higher GFP levels. From each gate, cells were collected into a sterile tube. Immediately before sorting, cells were spun at 12,000 g for 10 min, and the resulting pellets resuspended into 10mM MgCl2, and bacterial concentration adjusted to 106 cfu/ml, and inoculated into bean plants. An aliquot of sorted cells was analyzed again at the cytometer to visualize the differences in expression of the separated populations. Data were analyzed with FlowJo Software. Generation of point mutations HopZ1aK289R point mutation was generated following the instructions of the QuikChange Lightning Multi Site-Directed Mutagenesis Kit (Agilent Technologies) using the vector pAME30 (Macho et al, 2010) as template. The primers used were Z1aM1 (CCGGTGGATTTTTATAGGCATGGCGCTTCGCTG) and Z1aM2 (CAGCGAAGCGCCATGCCTATAAAAATCCACCGG). The mutation was verified by sequencing. To generate MKK7K74R and MKK7K167R, primers used are listed in Chapter 4, Table 3, and NZYMutagenesis kit (NZYtech, Portugal) was used, following the instructions provided, and pENTR-MKK7 as template. Point mutations were verified by sequencing. ! ! ! ! ! ! ! Chapter(1:(Dynamics(of(heterogeneous( populations*of*Pseudomonas*syringae! within&plant&tissues&reveal&a&diversity& of#interactions! José S. Rufián, Alberto P. Macho, David S. Corry, John W. Mansfield, Dawn L. Arnold, Carmen R. Beuzón. Chapter 1: Dynamics of heterogeneous populations of P. syringae 48 Introduction, Plants are continuously under attack by a myriad of microorganisms, often by a mixture of them simultaneously. The different defense mechanisms that plants use to resist such attacks have been the subject of extensive study for years, where much of this research has focused on the interaction between bacterial pathogens and the plant host (Dodds and Rathjen, 2010). These studies, mostly carried out in laboratory conditions, analyze the interaction in a one-to-one basis, one pathogen at a time. However, field infections are likely to be more complex. The interaction between the plant and a given pathogen can be influenced by additional interactions of the plant with the same or different pathogens. Contact with some pathogens, for instance, can induce systemic acquired resistance or SAR (Cameron et al., 1994). Induction of SAR determines a restriction of growth of a newly incoming pathogen in systemic tissues, by determining the pre-activation of defense responses in uninfected tissues, such as accumulation of salycilic acid (SA), or pathogenesis-related proteins (PRs). Plantassociated bacteria can also determine pre-activation of defenses in systemic tissues in a SA-independent, jasmonic acid (JA)-dependent process known as induced systemic resistance (ISR) (van Loon et al., 1998). Even in the context of a seemingly single infection, genomic changes within the population have been shown to give raise to genetic variants, which can differ in their interaction with the plant. Only within the P. syringae complex the strain diversity is overwhelming, with many variants of a pathogen originated through horizontal gene transfer (HGT), mutation, chromosomal re-organisation, and unknown events giving raise to phase variation within the populations, and to the differentiation of variants with important differences on host adaptation (Ma et al., 2006; Godfrey et al., 2011). At the center of the process of host adaptation of P. syringae is the T3SS and the T3Es (Alfano and Collmer, 1997). Evolution of the T3SS and that of its effectors has been mostly driven by HGT (Rohmer et al., 2004; Sarkar et al., 2006). An example of how variants of a given pathogen with different virulence capabilities can meet within the plant is found in the bean pathogen P. syringae pv. phaseolicola (hereafter referred to as Pph). Inoculation of Pph 1302A into resistant bean plants activates the HR, which in turns triggers the excision of the bacterial PPHGI-1 locus, a genomic island encoding a T3E, HopAR1 (formerly AvrPphB), responsible for triggering the HR in this cultivar (Pitman et al., 2005). Once excised from the Introduction 49 chromosome PPHGI-1 is maintained as a circular episome in which gene expression is severely reduced (Godfrey et al., 2011). As excision of PPHGI-1 takes places in a number of cells during an incompatible interaction, the bacterial population bifurcates into two subpopulation differing in whether they express, or not HopAR1. In a host carrying the resistance gene against HopAR1, these two subpopulations have a very different interaction with the host, since those carrying a genomic copy of the island will induce a strong HopAR1-triggered HR, and those carrying its episomal version will not (Godfrey et al., 2011). Little is known about how mixed populations of a given pathogen interact and develop with the host plant. Work in the 1960s showed that inoculation of a P. syringae strain capable of triggering strong plant defenses (avirulent strain) may result in the limitation of a co-inoculated virulent strain (Klement and Lovrekovich, 1961; Omer and Wood, 1969). However, a population of a non-pathogenic strain may also benefit from co-inoculation with a pathogenic one, reaching a larger population size than if inoculated alone (Young, 1974). More recently, following the lead of research in animal pathogens, where mixed infections are used to analyze bacterial virulence (Beuzón and Holden, 2001), our laboratory set up the conditions to use competitive index in mixed infections as a means to assay virulence (Macho et al., 2007). This work showed that co-inoculated P. syringae strains affected each other’s growth or not, depending on the concentration of the inoculum. Thus, when high inoculation doses were used, it was possible to detect dominant negative effects of avirulent strains on the growth of virulent, or complementation of growth defects of non-pathogenic by pathogenic coinoculated strains, such as those previously described. These interferences were not detected at lower inoculation doses, like those expected to take place within a natural infection. In this work, we apply fluorescent confocal microscopy to analyze how mixed bacterial populations of P. syringae, with key differences in their virulence capacity, develop within the plant host. We take advantage of the improvements in the biophysical properties of fluorophores and confocal microscopic imaging technology, which allows to minimize the background auto-fluorescence detection from both healthy, and most importantly, plant tissue undergoing the HR (Godfrey et al., 2010). These advances make possible the application of confocal microscopy to follow distribution and growth of P. syringae strains labeled with eYFP or eCFP inside the Chapter 1: Dynamics of heterogeneous populations of P. syringae 50 leaf. We analyze the cellular basis that determine whether co-existing P. syringae bacteria with different virulence capabilities affect each other’s growth, and discuss the potential implications for host adaptation. Table 1. Strains used and generated in this work. Strain Genotype Reference 1448A P. syringae pv. phaseolicola wild-type strain race 6 (Teverson, 1991) 1449b P. syringae pv. phaseolicola wild-type strain race 7 (Teverson, 1991) IOM1 1448A Δ hrcV, KmR (Macho et al., 2007) JRP8 1448A Tn7-eGFP, GmR This work JRP9 1448A Tn7-eYFP, GmR This work JRP10 1448A Tn7-dsRFP, GmR This work JRP11 1448A Tn7-eCFP, GmR This work JRP12 1448A Δ hrcV Tn7-eYFP, GmR This work JRP15 1449b Tn7-eYFP, GmR This work JRP17 1449b Tn7-eCFP, GmR This work RW60 1449b Vir–, pAV511–, RifR (Jackson et al., 1999) JRP18 RW60 Tn7-eYFP, GmR This work Table 2. Plasmids used in this work. Name Description Reference pUXBF13 Helper plasmid, providing the Tn7 transposase proteins, AmpR Bao et al. (1991) pRK2013 Conjugation helper plasmid, KmR Figurski et al. (1979) AKN132 dsRFP, AmpR, GmR Lambertsen et al. (2004) AKN100 eGFP, AmpR, GmR Lambertsen et al. (2004) AKN069 eYFP, AmpR, GmR Lambertsen et al. (2004) AKN033 eCFP, AmpR, GmR Lambertsen et al. (2004) Chapter 1: Dynamics of heterogeneous populations of P. syringae 52 Results, Development of mixed versus single micro-colonies within the plant apoplast Since our previous work showed that interference between co-inoculated largely depended on the concentration of the inoculum (Macho et al., 2007), ,we set out to apply confocal microscopy to analyze the cellular basis for that observation. We generated fluorescently labeled derivatives of the bean pathogen P. syringae pv. phaseolicola 1448A wild type strain (Table 1), and followed their distribution and growth within the leaf apoplast after inoculation at different concentrations. All fluorescent derivatives expressed the fluorophores, and in keeping with previous reports (Godfrey et al., 2010), the clearest differences in mixed cultures between labeled strains were detected when combining eCFP and eYFP-labeled bacteria (data not shown). Examination by light and fluorescent microscopy of mixed cultures of 1448A eYFP and 1448A eCFP showed that every bacteria tagged with a given fluorophore emitted the corresponding fluorescence at a similar level, both in laboratory medium and in planta (Figures 1A and 1B). Growth of 1448A eYFP as well as symptom induction were confirmed to be wild type-like in all conditions tested (Table 3). Growth of 1448A eCFP showed a slight delay compared to wild type, both in laboratory medium and in planta (Table 3), but virulence was not affected, since the induction of symptoms was similar to that of 1448A eYFP (Figure 1D), and the number of colonies observed for these two strains was similar following mixed inoculation (Figure 1C). The size of the 1448A eCFP colonies in leaves was somewhat smaller than that of 1448A eYFP colonies, in keeping with the observed delay of growth observed both within the plant and in laboratory medium (Figure 1C and Table 3). Thus, confocal microscopy of leaves inoculated with 1448A eYFP and 1448A eCFP can accurately reflect wild type growth dynamics within the host. Table 3. CI analysis of fluorescent P. syringae strains. Mixed Strains LBCI (±SE) in plantaCI (±SE) 1448a wt vs. 1448a YFP 0,97 ± 0,06 - 1448a wt vs. 1448a CFP 0,61 ± 0,1* - 1448a YFP vs. 1448a CFP 0,59 ± 0,03* 0,64 ± 0,04* RW60 YFP vs. 1449b CFP 0,74 ± 0,06* - 1449b YFP vs. 1449b CFP - 0,82 ± 0,02* Results 53 Figure 1. Expression of either eYFP or eCFP in Pph 1448A is constitutive and does not affect virulence. (A). Left panel: Fluorescence microscopy image showing constitutive expression of eYFP and eCFP from bacteria grown within rich medium. Right panel: Merged fluorescence (eYFP, eCFP) and phase-contrast microscopy images. (B) Confocal fluorescence microscopy image of bean leaf sections 3 days post-inoculation (dpi) with a 5x105 cfu/ml suspension of 1448a eYFP (Left) or 1448a eCFP (Right). Bacterial microcolonies and even individual bacteria can be visualized within the plant apoplast. (C) Confocal fluorescence microscopy image of a bean leaf sections 2 days post-inoculation (dpi) with a 5x105 cfu/ml bacterial suspension containing equal amounts of 1448a eYFP and 1448a eCFP. Differently labeled wild type bacteria can be distinguished within singleor mixed-colored colonies. Red in images in B and C corresponds to auto-fluorescence emitted by chloroplasts. (D) Symptoms developed by either 1448A eCFP and 1448a eYFP X dpi with 5x105 cfu/ml. Scale bar: A and B: 25 µm. C: 100 µm. Chapter 1: Dynamics of heterogeneous populations of P. syringae 54 Derivatives of 1448A expressing either eYFP or eCFP were co-infiltrated into bean leaves at different concentrations (5x107, 5x106, 5x105 and 5x104 cfu/ml), and leaf sections were taken at different days post-inoculation (1 to 5 dpi) and observed by confocal microscopy (Figure 2). Numerous colonies could be observed as early as 1 dpi in leaves inoculated with 5x107 cfu/ml (Figure 2A). The majority of these colonies displayed a mixture of eYFP and eCFP. By 2 dpi, many of the colonies merged as they grew in size. Longer time points showed tissue damage too extensive to be examined by microscopy, as severe disease symptoms had already started to develop. In leaves inoculated with 5x106 cfu/ml, colonies were clearly visible by 2 dpi, and although a considerable proportion of them showed both eYFP and eCFP, single-colored colonies could be detected as well (Figure 2B). As time increased, this trend became clearer despite the fact that confluence led to the appearance of merged colonies. To quantify this trend, we counted single-colored versus bicolored colonies, in three independent experiments (three fields per sample and experiment), and calculated the ratio between them. The ratio singleto bicolored colonies was 0.68±0.25, 2 days post-inoculation with 5x107 cfu/ml, supporting the observation of a minority of single-colored colonies at the highest inoculum concentration. However, this ratio was almost 20-fold higher (12.25+/-4.35) 3 days post inoculation with 5x106 cfu/ml, indicating an increase in single-colored colonies associated to lower inoculum concentrations. In short, a 10-fold decrease in the concentration of the inoculum led to a 20-fold increase in the ratio between single-colored versus bicolored colonies, from a minority of single-colored colonies, to a majority. Following this trend, leaves inoculated with 5x105 cfu/ml contained a vast majority of single-colored colonies by 3 dpi (Figure 2C). In fact, almost no bicolored colonies could be observed regardless of the time analyzed. In these conditions, equal numbers of both types of single-colored colonies could be observed, although the average size of the eCFP colonies was generally smaller than that of the eYFP colonies, as expected with the growth delay observed for the eCFP strain. As before, a later time point (4 dpi) was associated to an increment in the number of colonies where confluent growth had lead to large single-colored colonies merging (image shown in the back cover of this thesis). Lastly, when leaves were inoculated with 5x104, the inoculation dose most commonly use for competitive assays (Macho et al., 2007; Macho et al., 2016), we were Results 55 unable to observe any bicolored colonies by 4 dpi, and only a few merged, seemingly due to confluent growth of close single-colored colonies colonies by 5 dpi (Fig. 2D). Figure 2. Co-inoculation of 1448a eYFP and 1448a eCFP at different concentrations determines different dynamics of colony development. Wild-type Pph 1448A eYFP (yellow) or Pph 1448A eCFP (cyan) were co-inoculated in equal amounts at the indicated dose (left numbers, cfu/ml) and leaf sections taken and visualize at different dpi. Some time points were not displayed either because only a few and scattered bacteria could be observed, or the inoculated leaf tissue was too damaged due to the progress of the infection. Scale bars: 100 µm Chapter 1: Dynamics of heterogeneous populations of P. syringae 56 Close proximity to a pathogenic strain promotes growth and spreading of a nonpathogenic strain. Previous work including our own have shown that when inoculated at high concentratetions pathogenic strains can complementation growth of a non-pathogenic strain (Young, 1974; Macho et al., 2007). We co-infiltrated 1448A eCFP and its nonpathogenic mutant derivative, ΔhrcV (eYFP) (Table 1) into bean leaves at different concentrations (5x107, 5x106 and 5x105 cfu/ml). Leaf sections were taken at 3 dpi and observed by confocal microscopy (Figure 3). The hrcV gene encodes an essential component of the Hrp T3SS (Cornelis and Van Gijsegem, 2000), required for development of infection in compatible hosts (Alfano and Collmer, 1997). As previously seen in Figure 2, images display differences in the rate of bicolored versus single-colored colonies depending of the concentration of the inoculum used: higher concentrations (5x107 cfu/ml) lead to a higher rate of bicolored colonies (Figure 3). Three days post-inoculation with 5x107 cfu/ml, growth of ΔhrcV eYFP, estimated as the amount of yellow fluorescence, was very similar to that of pathogenic 1448A eCFP (blue fluorescence) (Figures 3A and B). We confirmed these results using CI assays (Figure 3C). CIs result from dividing the output ratio between the two strains by their input ratio (should be close to 1:1), thus a CI close to 1.0 indicates that the coinoculated strains are growing similarly.. In these experiments, yellow fluorescence associated to the mutant non-pathogenic strain (ΔhrcV eYFP) was rarely detectable in single-colored colonies and almost only found in association to blue (1448A eCFP), in bicolored colonies (Figure 3A).. As the inoculum concentration decreased, the rate of bicolored colonies decreased too (Figure 3A), as did growth of ΔhrcV eYFP (Figures 3B and C) reaching its maximal difference with pathogenic wild type 1448A eCFP, over 100-fold, in leaves inoculated with 5x104 cfu/ml. These results indicate that close proximity to pathogenic 1448A promotes growth of the non-pathogenic ΔhrcV strain Discussion 63 Discussion, In this study we have used confocal microscopy to analyze how P. syringae pv. phaseolicola distributes and grows within the plant apoplast, and to show an array of different scenarios taking place when strains displaying differences on their virulence meet within the plant. We confirm our previous results that the outcome of the individual interaction between a given strain and the host may have a strong impact on the development of a co-inoculated strain in certain conditions. We demonstrate that the interferences observed between strains co-inoculated at high bacterial concentrations, are due to close proximity, rather than to the overall high bacterial load on the inoculated area. This conclusion is supported by the fact that similar bacterial concentrations within the apoplast at the onset of the experiment, display different distribution within the inoculated area leading to differences in growth depending on the means of inoculation (i.e. infiltration versus dip-inoculation). In trans defense suppression within a mixed infection Interestingly, although complementation of non-pathogenic strains can only be clearly observed in mixed or merged colonies, interferences between virulent and avirulent strains (i.e. dominant-negative effects on wild type growth, or defense suppression) can also be observed between bacterial colonies located near, but not in direct contact. Thus, different types of interference seem to require a different degree of proximity to take place. In support of this conclusion, while no complementation of growth of the ΔhrcV mutant can be detected following inoculation by infiltration at 5x104 cfu/ml, a mild dominant negative effect (five-fold decrease of growth) can be detected on a wild type strain when co-inoculated with its avirulent derivative RW60 at the same concentration (Macho et al., 2007). These results could reflect that cell-to-cell signaling between host cells may have a stronger role in the establishment of ETI than in the establishment of other type of defense responses, for which a more direct contact may be necessary both for its activation and its suppression. Another interesting result in regard to defense suppression within mixed infections is that in trans complementation of the ΔhrcV by the co-inoculated wild type allows the mutant to spread beyond the inoculated area. It has been reported that ETI in Chapter 1: Dynamics of heterogeneous populations of P. syringae 64 Arabidopsis and PTI in N. benthamiana stop vascular transport through the xylem (Oh and Collmer, 2005; Freeman and Beattie, 2009), which P. syringae may use to colonize distant tissues (Misas-Villamil et al., 2011). On these bases, it has been proposed that vascular defenses may play a role in restricting pathogen spread (Misas-Villamil et al., 2011). Our results indicate that close proximity between pathogenic and non-pathogenic bacteria allows the later to colonize distant tissues, and this could be achieved through in trans suppression of either local or vascular defenses involved in restricting ssystemic colonization. Here, we have analyzed interference between co-inoculated strains in the leaf apoplast of Canadian Wonder bean plants, but the relation between bacterial proximity and each type of interference detected could vary in specific microenvironments within different plant tissues, or in different hosts, depending on the intensity of the defenses involved. In relation to this, we previously observed that mixed populations of the pathogen Ralstonia solanacearum have very different dynamics depending on where they were inoculated within the plant: the same dose of inoculum triggered strong complementation in the stem, but not in the leaf (Macho et al., 2010b). One possible explanation was that in inoculation within the stem, rendered higher local concentration of bacteria, but it is also possible that differences in plant defenses in vascular versus parenchyma cells allow for differences in complementation. Bacterial entry into the plant apoplast during dip-inoculation Infiltration forces bacteria into the apoplast through the applied pressure, whereas, as in nature, bacteria have to gain entry through their own means when spray or dipinoculated. We have observed that different means of inoculation lead to differences in how bacteria distribute within the apoplast upon entry. Whereas dip-inoculation leads to a clear concentration of bacteria around trichomas, which could constitute a weak structural point prone to tissue damage, facilitating bacterial entry through the associated wounds, it did not lead to bacterial concentration around stomata. Stomata are largely considered to be the main port of entry of P. syringae into the plant apoplast (Melotto et al., 2006). However, this conclusion is largely based on the study of the interaction between P. syringae pv. tomato and other coronatine-producing P. syringae strains, and Arabidopsis and/ or tomato plants (Melotto et al., 2008). These studies have shown that despite inmediate closure of the stomata upon detection of conserved Discussion 65 patterns in the pathogen, such as flagelling, the production of the phytotoxin coronatine allows these bacteria to re-open stomata, allowing further and more efficient invasion of the leaf. However, how other P. syringae strains may reach the apoplast is less clear. P. syringae pv. tabaci has also been shown to re-open stomata even though it does not produce coronatine (Melotto et al., 2006), suggesting other virulence factor(s) encoded by this strain can either inhibit stomata closure or re-open them. However, it has also been suggested that pathogens lacking the ability to re-open stomata may favor other ports of entry, such as is the case for the foliar pathogen Xanthomonas campestris pv. campestris, which mainly enters the leaf through hydathodes, water-exuding pores at the edge of leaves (Hugouvieux et al., 1998). Thus, since P. syringae pv. phaseolicola does not produce coronatine, the absence of bacterial concentration around stomata in the dip-inoculated bean leaves could be an indication of this pathogen lacking the ability to re-open stomata. In support of this notion, we observed a clear concentration of bacterial colonies located close to the edges of the leaves, were hydathodes are located (Figure 4E). Interestingly, hydathodes has been recently shown as an alternative route of entry for the coronatine-producing P. syringae pv. tomato DC3000 in Arabidopsis (Yu et al., 2013). Nonetheless, our observation could reflect a bacterial preference for P. syringae pv. phaseolicola entry through hydathodes rather than stomata, or be simply a consequence of the distribution of the inoculating solution (and perhaps rainwater) on the leaf surface. But the most relevant result obtained when dip-inoculating bean leaves is that coinoculated strains interfere with each other at lower intra-apoplast concentrations than needed following infiltration. These results supports that the incidence of these type of interferences can be relevant for adaptation to the host in nature, where inoculation occurs mostly through rainwater, with high local epiphytic concentration of bacteria taking place (Kinkel et al., 2000; Lindow and Brandl, 2003). Mixed infection assays reveal a new phenomenon We initially set out to analyze how meeting within the plant would affect the fate of bacteria with different virulence capabilities, and chose to do so using mixed infection assays, normally used as a sensitive assay to measure virulence attenuation (Beuzón and Holden, 2001; Macho et al., 2007; Macho et al., 2010b; Feng et al., 2012). However, through these assays we detected such a variety of interactions between co- Chapter 1: Dynamics of heterogeneous populations of P. syringae 66 inoculated bacteria within a single experiment that lead us to explore the possible source for such phenotypic heterogeneity. These results (Figure 5) show that different types of interferences do take place between virulent and avirulent bacteria within the same infection. The dominant negative effect previously detected in growth assays (Macho et al., 2007) would thus be the net resultant of all of these types of interferences. These different scenarios between the co-inoculated strains could arise from differences in the timing of the arrival and establishment of bacteria to a given tissue location. But they could also take place if the bacterial population was not homogeneously expressing the genes involved in determining the interaction with the plant (e.g. the genes encoding the effectors triggering the HR, or the genes encoding the T3SS): if there were cell-to-cell differences in the expression of a virulence gene(s) within each bacterial clonal population. ! ! ! ! ! ! ! Chapter(2:(A"bistable"switch"controls" virulence)of)bacterial)plant)pathogen) Pseudomonas*syringae! José S. Rufián, María A. Sánchez-Romero, Diego López-Márquez, Alberto P. Macho, John W. Mansfield, Dawn L. Arnold, Javier Ruiz-Albert, Josep Casadesús, Carmen R. Beuzón. Chapter 2: Bistability on the T3SS of P. syringae 68 Introduction, Infectious processes involve spatial and temporal gene expression changes that follow the migration of bacterial pathogens from the point of invasion to the target tissues of the host. Pathogen progression through the host is accompanied by adjustments in transcription and phenotype to respond to different stimuli and microenvironments. However, the sources of variation may not always be deterministic, directly correlated to stimuli. Stochastic events may cause cell-to-cell transcriptional and phenotypic differences between genetically identical individuals, and may take place within the same microenvironment. This leads to a probabilistic determination of gene transcription and phenotype generally known as phenotypic heterogeneity or variation. Phenotypic heterogeneity has been known to take place within microbial clonal populations for decades (Bigger, 1944; Novick and Weiner, 1957). Moreover, a heterogeneous unimodal pattern of gene expression may become bimodal, bifurcating into two distinct patterns, in a process known as bistability that leads to the formation of bacterial subpopulations or lineages. Genetically identical bacterial subpopulations formed without the involvement of environmental cues may be even viewed as a programmed event. Bistability on gene expression can be generated within a heterogeneous population by a positive feedback loop, as described by Novick and Weiner (1957) in the E. coli lac operon, or by a double negative feedback loop as in the lysis/lysogeny decision of bacteriophage lambda. The literature on bacterial bistable switches has been enriched with interesting examples in the last decade (Davidson and Surette, 2008; van der Woude, 2011; Sanchez-Romero and Casadesus, 2014; van Vliet and Ackermann, 2015). For instance, bistability is known to occur in the acquisition of competence for DNA uptake and in the formation of a sporulating subpopulation in Bacillus subtilis (Chai et al., 2008). Despite these examples, microbiological studies of pathogens growing in tissue sites have largely relied on averaging microbial responses for entire populations. The recent advent of single cell analysis has enabled a shift in perspective allowing detection and analysis of cell-to-cell non-genetic variation. The importance of analyzing phenotypic heterogeneity in pathogen populations has been very recently highlighted in the context of antibiotic exposure for several animal and human pathogens including Salmonella enterica (Helaine and Holden, 2013; Arnoldini et al., 2014; Campbell-Valois et al., 2014; Claudi et al., 2014; Diard et al., Introduction 69 2014; Sanchez-Romero and Casadesus, 2014; Manina et al., 2015). The formation of distinct bacterial lineages has also been proposed to play a role in the colonization of animals by some bacterial pathogens. In Salmonella enterica, phenotypic heterogeneity has been observed at several stages of host colonization. During intestinal infection, flagella are necessary for swimming and also facilitate invasion of intestinal cells (Stecher et al., 2004). However, intestinal populations of Salmonella are a mixture of flagellated and non-flagellated bacteria (Saini et al., 2010b; Stewart and Cookson, 2012). Formation of a non-flagellated subpopulation may be viewed as a stealth strategy because flagellin is highly immunogenic (Josenhans and Suerbaum, 2002). Another phenotypic bifurcation during Salmonella infection is observed in systemic infection: upon entry into macrophages the Salmonella population splits into two subpopulations, one of which replicates while the other enters a dormant-like state (Helaine and Holden, 2013). Colonization of the gall bladder by Salmonella involves also lineage formation: a subpopulation invades the gall bladder epithelium, while another subpopulation remains in the gall bladder lumen (Baumler et al., 2011). The potential relevance of phenotypic heterogeneity in the development of bacterial infections is further supported by recent reports showing within-host bistable expression of major virulence traits, such as the cholera toxin in Vibrio cholerae (Nielsen et al., 2010), the SPI1 type III secretion system (T3SS) in S. enterica (Stecher et al., 2008; Saini et al., 2010a; Diard et al., 2013), or NO-detoxification in Yersinia pseudotuberculosis (Davis et al., 2015). Despite the increasing evidence supporting the notion of bacterial pathogens exploiting non-genetic variation to adapt to mammalian hosts, little is known about the occurrence or potential impact of these processes in the adaptation of bacteria to nonanimal niches. In this work, we have addressed this question using the plant pathogen Pseudomonas syringae, an archetype considered by many the most relevant plant pathogenic bacteria (Mansfield et al., 2012). P. syringae is relevant both academically, as a model pathogen, and economically, for its increasing impact in agriculture, with the recent resurgence of old diseases (Shenge, 2007), and the emergence of new infections of worldwide importance (Green et al., 2010). In plants, P. syringae can be found in the surface of leaves, from where it enters the leaf through natural openings and wounds to reach the intercellular spaces of the leaf parenchyma, the apoplast, where it replicates causing disease. The plant presents a two-tiered response against P. syringae (Jones and Dangl, 2006). Chapter 2: Bistability on the T3SS of P. syringae 70 In this study, we analyzed the molecular mechanism behind the phenotypic heterogeneity observed during bacterial colonization of the plant host. We demonstrate that expression of key structural and regulatory components, as well as expression of a translocated effector of the P. syringae T3SS is bistable within the plant apoplast. Bistability also takes place within the homogeneous environment of the laboratory, during exponential growth in nutrient-limited, T3SS-inducing medium, and it is a dynamic, reversible and non-heritable process. Furthermore, it has consequences for bacterial adaptation to the plant host since populations sorted by their T3SS expression levels display differences in virulence. Genetic analysis of the regulatory elements controlling expression of the T3SS in P. syringae identifies the HrpV/HrpG double negative regulatory loop as the bistable switch involved in turning heterogeneity into bistability, requiring the transcriptional activator HrpL, and enhanced by the contribution of HrpA, the main component of the T3SS pilus, involved in a positive feedback loop. To our knowledge, this is the first example of bacterial heterogeneity and bistability of a major virulence determinant shown within a non-mammalian host. , , Results 71 Table 1. Strains used and generated in this work. Strain Genotype Reference 1448A P. syringae pv. phaseolicola wild-type strain race 6 (Teverson, 1991) JRP9 1448A Tn7-eYFP, GmR This work JRP11 1448A Tn7-eCFP, GmR This work DLM1 1448A hrpL::gpf, KmR This work DLM2 1448A hrcU::gpf, KmR This work DLM3 1448A hopAB1::gpf, KmR This work IOM49 1448A ΔhrpA (Ortiz-Martin et al., 2010a) JRP-F1 1448A ΔhrpA; hrpL::gpf, KmR This work JRP-F2 1448A ΔhrpA; hopAB1::gpf, KmR This work IOM57 1448A ΔhrpG (Ortiz-Martin et al., 2010b) IOM48-F 1448A ΔhrpV Ortiz-Martín et al., 2010b IOM58 1448A ΔhrpG ΔhrpV Ortiz-Martín et al., 2010b JRP-F3 1448A ΔhrpG; hopAB1::gpf, KmR This work JRP-F4 1448A ΔhrpV; hopAB1::gpf, KmR This work JRP-F5 1448A ΔhrpG ΔhrpV; hopAB1::gpf, KmR This work Table 2. Plasmids used in this work. Name Description Reference pAME8 pAMEX derivative, contains the avrRpt2 effector expressed from the nptII promoter (Macho et al., 2009) pIOM22 pBBR1-MCS-4 derivative, contains a promotorless hrpL gene expressed from the lacZ promoter (Ortiz-Martín et al., 2010a) pIOM92 pBBR1-MCS-4 derivative, contains a (Ortiz-Martín et al., Chapter 2: Bistability on the T3SS of P. syringae 72 promotorless hrpG gene expressed from the lacZ promoter 2010b) pIOM53 pBBR1-MCS-4 derivative, contains a promotorless hrpV gene expressed from the lacZ promoter (Ortiz-Martín et al., 2010b) Table 3. Primers used in this work. Name Description Restriction site HrpL A1 CGGTATCCGTCAACTGACGG NA HrpL A2 GAATTCTATCCACTCAGGCGAACGGG EcoRI HrpL B1 TGAGTGGATAGAATTCTCTGTCTGGAACCAAC TCGC EcoRI HrpL B2 ATGGGCGACCATCGGATCC NA HrcU A1 GTGATTCTGGGGTTGCTGC NA HrcU A2 GAATTCAGCTCCCAGCTTAAAGCTCC EcoRI HrcU B1 AGCTGGGAGCTGAATTCGCAAGCCAGGCGTA ACAGG EcoRI HrcU B2 TTCTACTACAACGTCGCTGC NA HopAB1 A1 GCATCCTTTATAACTGACCC NA HopAB1 A2 GAATTCCTGAAATCAGTTCAGCTTAACG EcoRI HopAB1 B1 CTGATTTCAGGAATTCTCGTTGTAGTGGCCGG EcoRI HopAB1 B2 GGACAGGTCGTAGTAGAGCG NA Zep07F GAATTCTAAGAAGGAGATATACATATGAG NA Zep07F GAATTCTTATCACTTATTCAGGCGTA NA Results 79 Bistability of the T3SS genes is established at the level of HrpL by the action of HrpA, and HrpV/HrpG regulatory loops HrpL regulates expression of hopAB1 (Xiao and Hutcheson, 1994). Thus, the bistability displayed by hopAB1 is likely to originate from that affecting hrpL expression. To test this hypothesis, a plasmid carrying a copy of hrpL under the control of the lacZ promoter, constitutive in P. syringae (Ortiz-Martin et al., 2010a) was introduced in the strains carrying the hrpL::gfp or hopAB1::gfp fusions and its effect in gfp expression analyzed by flow cytometry. Figure 5A shows how the bimodal distribution of hrpL::gfp becomes unimodal in the presence of constitutively expressed HrpL, and that of hopAB1::gfp strongly reduced, displaying a shift towards a more activated state, supporting the central role of HrpL in the establishment of bistability in the system. The establishment of bistability usually requires the action of at least one feedback loop (either a positive or a double negative regulatory loop) to turn quantitative differences originated through heterogeneous expression, into qualitative differences and a bimodal distribution of expression (Veening et al., 2008). Two such feedback loops regulate the expression of the T3SS genes in P. syringae a positive feedback loop carried out HrpA the main component of the T3SS pilus (Roine et al., 1997 2000), and a double negative feedback loop carried out by HrpV, an anti-activator of the T3SS genes that binds to HrpS the enhancer-binding protein required for HrpL expression, and HrpG, which binds HrpV acting as an anti-anti-activator (Wei et al., 2005). We analyzed the roles of these two regulatory loops in the establishment of bistability using mutants and/ or plasmids encoding the corresponding genes. Bistability in both hrpL::gfp and hopAB1::gfp was reduced but not abolished in an ΔhrpA mutant background (Figure 5B), supporting a non-essential contribution for HrpA to the bistable phenotype. Bistability of hopAB1::gfp was completely abolished in a ΔhrpG (Figure 6A), where in the absence of its repressor, HrpV establishes a negative feedback loop expected to dampen noisy and heterogeneous expression. In the absence of HrpV, expression increased and bistability was also increased, regardless of the presence of HrpG, since the phenotype of the ΔhrpV mutant was identical to that of a double mutant ΔhrpV ΔhrpG (Figure 6A). The epistatic effect of the ΔhrpV mutation on the phenotype of ΔhrpG supports the notion of HrpG exerting its role in regulation of the expression of the T3SS genes mainly through its role as repressor of HrpV. Constitutive expression of Chapter 2: Bistability on the T3SS of P. syringae 80 either regulator from a plasmid lead to reciprocal results, with a stronger bistable phenotype associated to overexpression of HrpG, and the absence of bistability in cells overexpressing HrpV (Figure 6B). Plasmid-expression of these regulators also caused these effects on the expression of hrpL::gfp. Figure 5. Bistability of hrpL::gfp and hopAB1::gfp is abolished by constitutive expression of HrpL, and mildly reduced in the absence of HrpA. Flow cytometry analysis of HIM-growing bacterial strains carrying chromosome-located transcriptional fusions to the T3SS genes hrpL, hrcU or hopAB1, in different genetic backgrounds. Histograms show GFP fluorescence distribution in strains growing at 24h. Black histograms show fluorescence of the fusions in the wild type 1448A background. Coloured histograms show fluorescence of the fusions in: (A) A strain carrying a plasmid encoding HrpL under the control of a lacZ promoter, and (B) A strain carrying a Δ hrpA mutation. All data were collected for 100,000 events per sample. Results 81 ! ! Figure 6. Bistability of hrpL::gfp and/ or hopAB1::gfp is abolished by deletion of hrpG or constitutive expression of HrpV, and enhanced by deletion of hrpV or constitutive expression of HrpG. Flow cytometry analysis of HIM-growing bacterial strains carrying chromosome-located transcriptional fusions to the T3SS genes hrpL, hrcU or hopAB1, in different genetic backgrounds. Histograms show GFP fluorescence distribution in strains growing at 24h. Black histograms show fluorescence of the fusions in the wild type 1448A background. Coloured histograms show: (A) Fluorescence of hopAB1::gfp in strains carrying mutations in either hrpG, hrpV, or both; (B) Fluorescence of either hopAB1::gfp or hrpL::gfp in a strain carrying a plasmid encoding either HrpG or HrpV, under the control of a lacZ promoter. All data were collected for 100,000 events per sample. Chapter 2: Bistability on the T3SS of P. syringae 82 Discussion, This study was undertaken to investigate the molecular mechanisms underlying the heterogeneity observed within the process of colonisation of the plant apoplast. We found that expression of the T3SS genes, the major virulence determinant of P. syringae, is heterogeneous and bistable during bacterial growth within the apoplast. The plant apoplast is a complex environment where different factors may alter bacterial responses. Thwaites and collaborators (2004) reported that following leaf extraction, bacteria that remained attached to the plant cell wall displayed higher amounts of T3SS gene transcripts than those found within the apoplastic fluid. However, the heterogeneity hereby described was detected in apoplastic bacteria. The apoplast is a nutrient-limited environment (Rico and Preston, 2008), where unsuppressed plant defenses can still be detected (Mitchell et al., 2015), and can therfore be considered stressful for bacteria. Bacterial stress is accompanied by a reduction in both transcription and translation that has been proposed to increase phenotypic heterogeneity in some systems (Veening et al., 2008). Interestingly, the apoplastic environment has been shown to activate recombination-mediated genomic rearrangements in P. syringae that lead to phase variation upon activation of the HR (Lovell et al., 2011). The fact that phenotypic heterogeneity is enhanced in some stressful situations suggests a potential adaptive value. Indeed, theoretical studies proposed that such adaptative advantage may become apparent in changeful and/ or hostile environments (Kussell et al., 2005; Kussell and Leibler, 2005), while later reports have provided experimental evidence supporting this notion for several animal and plant pathogens (Srikhanta et al., 2010; Lovell et al., 2011; Hernandez et al., 2012; Claudi et al., 2014; Sanchez-Romero and Casadesus, 2014; Manina et al., 2015). The differences in virulence established in this work for bacterial subpopulations sorted according to T3SS expression levels further suggests that non-genetic phenotypic heterogeneity can be relevant for bacterial adaptation to plant hosts. On these grounds, a tentative interpretation of our results is that the stressful conditions found within the apoplast may favour phenotypic heterogeneity and confer selective advantage to P. syringae populations with heterogeneous and/ or bistable expression of the T3SS and effector genes. Bacteria not expressing the T3SS could act as ‘cheaters’ benefiting from the T3SS-mediated contribution of those ‘cooperator’ bacteria that express it, in a model Discussion 83 similar to that proposed for complex microbial communities with secreting and nosecreting genotypes (Xavier et al., 2011). Plus, ‘cheater’ bacteria could gain a potential fitness advantage, being relieved of the metabolic burden of expressing the T3SS. Moreover, activation of expression of the T3SS genes in P. syringae has been shown to be associated to repression of housekeeping functions (Ferreira et al., 2006). Such a fitness advantage has been described in animal pathogens such as S. enterica, P. aeruginosa and Yersinia spp (Wiley et al., 2007; Kohler et al., 2009 2011). Supporting this notion, Barret and collaborators (2011) found unexpectedly high frequencies of polymorphism affecting several virulence traits, including the T3SS, in natural populations of P. syringae infecting Arabidopsis thaliana, and demonstrated that less aggressive strains increase their growth potential in mixed infections and have a fitness advantage in non-host environments. Hence, reversible phenotypic heterogeneity on the T3SS expression might facilitate adaptation to different agricultural and environmental conditions. Giving strength to this view, Morris and collaborators (2008) found that natural isolates of P. syringae from various agricultural and non-agricultural habitats showed changes in the phenotypic but not the genotypic population structure. Because the apoplast is a complex environment, we turned to an in vitro system for further investigation. Since HrpL is the transcriptional activator of the other two genes, hrcU and hopAB1, and also regulates indirectly its own expression through a positive loop determined by the product of the HrpL-activated hrpA gene (Wei et al., 2000), and a double negative loop determined by the products of the HrpL-regulated hrpG and hrpV genes (Wei et al., 2005), the simplest explanation for our results was that the phenotypic heterogeneity observed for hrcU and bistability of hopAB1 originated from hrpL heterogeneous and bistable expression. The fact that not all three genes display bistability may reflect different thresholds in the response of each gene to HrpL activation. Indeed, different levels of HrpL-regulation have been reported for T3SS-structural and effector genes in P. syringae (Zumaquero et al., 2010). It could also be due to additional regulatory loops differentially affecting these genes, dampening or enhancing heterogeneity, e.g. an additional negative regulation of structural genes such as hrcU. Although no such loops have been described for the hrcQRSTU operon, the hrpC operon encoding HrcC and HrcF structural components of the T3SS has been reported to display a differential regulation (Ortiz-Martin et al., 2010b). In any case, the fact that bistability requires native expression of HrpL, and it is abrogated when HrpL is Chapter 2: Bistability on the T3SS of P. syringae 84 consititutively expressed from a plasmid confirms the central role of this protein in the establishment of this phenotype. Indeed, both the double negative regulation carried out by HrpV and HrpG, and the positive regulation carried out by HrpA, identified as necessary for or contributing to bistability, involved the action of HrpL. The drastic and reciprocal effects of deletion and overexpression of either HrpV or HrpG on bistability of the T3SS genes clearly identifies this pair as the bistable switch involved in turning heterogeneity into bistability. The observation of a link between bistability and growth phase could reflect the need for active cell division to generate stochastic differences in the relative levels of these two proteins through involvement of an additional factor differentially associated to growth phase. Bistable expression of tcpA expression in Vibrio cholerae is established through the combined effect of the positive autoregulation of ToxT production and CRP-cAMP, with this complex involved in restricting bistability to stationary phase (Nielsen et al., 2010). The phenotypic heterogeneity displayed by HrpL could also translate in differences in non-T3SS genes of the HrpL regulon. One system potentially affected by HrpL bistability is the flagella assembly, since HrpL has been shown to downregulate motility (Ortiz-Martin et al., 2010a). In P. syringae, differences in motility correlate with differences in adaptation to the host and other environments (Schroth, 1974; Haefele and Lindow, 1987; Hatterman D.R., 1989). In addition, the flagellum is both a costly organelle to assemble (Macnab, 1996) and a strong activator of plant defenses (Felix et al., 1999). In S. enterica, the expression of genes encoding the SPI1 T3SS, necessary to induce gut inflammation and to overcome competition from commensal microbiota (Kaiser et al., 2012), as well as that of the flagellar genes, display phenotypic heterogeneity during colonisation of the host (Cummings et al., 2006; Saini et al., 2010a), leading to the different phenotypic combinations to cooperate to promote the infection through the division of labour (Stecher et al., 2008; Diard et al., 2013). On the basis of these results, we tentatively propose that phenotypic heterogeneity plays a role in the adaptation of P. syringae to agricultural and perhaps to nonagricultural, ecologically relevant environments. These findings can potentially impact control strategies to protect economically important crops. Our results provide the first proof of phenotypic variation in a plant bacterial pathogen, adding to the very few examples showing phenotypic variation directly affecting a major virulence factor, and identify the molecular mechanisms involved in turning expression heterogeneity into Discussion 85 bistability and non-genetic lineage formation. Although we cannot rule out additional sources of variation, our results also provide a mechanistic explanation for the presence of bacterial microcolonies displaying phenotypic differences in growth and/ or activation of defenses within the plant apoplast. Finally, they show that phenotypic heterogeneity, as a bacterial strategy to generate within-host diversity, is not restricted to mammalian hosts. ! ! ! ! ! ! ! Chapter(3:"Auto-acetylation*on*K289!is# not$essential$for$HopZ1a-mediated' plant&defense&suppression! José S. Rufián, Ainhoa Lucía, Alberto P. Macho, Begoña Orozco-Navarrete, Manuel Arroyo-Mateos, Eduardo R. Bejarano, Carmen R. Beuzón, Javier RuizAlbert The results presented in this chapter have been published in: Rufián JS, Lucía A, Macho AP, Orozco-Navarrete B, Arroyo-Mateos M, Bejarano ER, Beuzón CR and Ruiz-Albert J (2015) Auto-acetylation on K289 is not essential for HopZ1a-mediated plant defense suppression. Front. Microbiol. 6:684. doi: 10.3389/fmicb.2015.00684 Chapter 3: K289 auto-acetylation and HopZ1a function 88 Introduction, Many Gram-negative pathogenic bacteria use a type III secretion system (T3SS) to secrete proteins, known as effectors, directly inside the host cell cytosol. Type III effectors (T3Es) modulate diverse processes inside the host, suppressing plant defense responses triggered upon recognition of the pathogen (Gohre and Robatzek, 2008). One such defense is triggered upon recognition of conserved pathogen-associated molecular patterns (PAMPs) and is known as PAMP-triggered immunity or PTI (Boller and Felix, 2009). T3Es can be directly or indirectly detected by the plant resistance proteins, triggering a second line of defense, a strong response known as effector-triggered immunity (ETI) that is typically accompanied by a type of programmed cell death referred to as the hypersensitive response (HR). The ETI response determines a severe restriction in pathogen growth (Chisholm et al., 2006). Effectors triggering strong immunity were originally named avirulence factors, as their expression by a pathogen determines resistance against the disease (Mansfield, 2009). Effectors can also suppress ETI, cell death and other HR-associated phenomena, thus promoting pathogen growth and the development of disease (Jones and Dangl, 2006). We have shown that HopZ1a from Pseudomonas syringae pv. syringae is one such effector (Macho et al., 2010a). Heterologous expression of HopZ1a from P. syringae pv. tomato DC3000 (hereafter DC3000), suppresses RNA and protein accumulation of pathogenesis related-1 (PR1), triggered in Arabidopsis by this pathogen (Macho et al., 2010a), and partially suppresses the ETI triggered by the expression of the heterologous effectors AvrRpt2, AvrRps4 and AvrRpm1 (Macho et al., 2010a). These defense suppression activities of HopZ1a are similar to those described for the related Xanthomonas effector AvrBsT (Kim et al., 2010; Szczesny et al., 2010; Kim et al., 2013). We have also demonstrated that HopZ1a is capable of suppressing systemic acquired resistance (SAR) triggered by either virulent or avirulent bacteria (Macho et al., 2010a). All these virulence activities are fully dependent on HopZ1a C216 catalytic residue. In turn, HopZ1a triggers SA and EDS1-independent immunity in Arabidopsis (Lewis et al., 2010; Macho et al., 2010a) upon recognition by the ZAR-1 resistance gene (Lewis et al., 2010). HopZ1a is a member of the YopJ/HopZ effector superfamily, whose members share a conserved catalytic triad (C/H/D) and have been shown to perform numerous biochemical activities, mainly as proteases and/or acetyltransferases, with some Results 95 HopZ1a suppression of AvrRpt2-triggered defense responses has also been demonstrated in Arabidopsis by directly comparing the growth attenuation determined by the individual expression of each effector, with the growth attenuation determined by their simultaneous expression (Macho et al., 2010a). Thus, we also analyzed the impact of the K289R mutation on the suppression of AvrRpt2-triggered growth restriction. To do so we performed mixed infections and calculated the cancelled-out index (COI), a modification of the competitive index (Beuzón and Holden, 2001), previously applied to this purpose (Macho et al., 2010a). COIs directly measure the differences in growth, within the same plant, between a strain expressing one of the effectors and a strain coexpressing both effectors, i.e. differences in growth of co-inoculated DC3000 expressing HopZ1a and DC3000 co-expressing HopZ1a and AvrRpt2. Thus, we can directly compare how expression of AvrRpt2 affects growth of DC3000 in the presence of HopZ1a or any of its mutant derivatives, with growth of DC3000 expressing only the HopZ1a version. As HopZ1a is expressed in both strains, the growth reduction it causes in Col-0 is cancelled out as it equally affects both strains (Macho et al., 2010a), and any difference in growth detected between the strain expressing both effectors and the strain expressing only the HopZ1a version, would be due to a growth restriction determined by the unsuppressed defenses triggered against AvrRpt2. A diagram illustrating this analysis is included as supplementary material (Figure S1). Chapter 3: K289 auto-acetylation and HopZ1a function 96 As previously reported (Macho et al., 2009; Macho et al., 2010a) DC3000 coexpressing AvrRpt2 and HopZ1a displayed a small although significant growth attenuation compared to that of co-inoculated DC3000 only expressing HopZ1a Figure 2. HopZ1aK289R partially suppresses AvrRpt2-triggered immunity. (A) Western blot showing PR1 accumulation in Col-0 leaves inoculated with 5 x 105 cfu/ml of DC3000 expressing AvrRpt2 (pAME8) alone or co-expressing AvrRpt2 with HopZ1a (pAME33), HopZ1aC216A (pAME34) or HopZ1aK289R (pJRU10). Ten micrograms of total protein were loaded per sample, and Coomassie staining is shown as loading control. The experiment was repeated twice with similar results. (B) Cancelled-out indices (COIs) measuring growth within a mixed infection of DC3000 co-expressing AvrRpt2 and any of the three HopZ1a variants: wild-type HopZ1a (pAME33), HopZ1aC216A (pAME34) or HopZ1aK289R (pJRU10) , in relation to growth of DC3000 expressing only the corresponding HopZ1a: wild-type HopZ1a (pAME30Gm), HopZ1aC216A (pAME27Gm) or HopZ1aK289R (pMAM1Gm). COIs are calculated as the output ratio between the strain expressing both effectors and the strain expressing just one, divided by their input ratio. Each COI value represents the means of 2 independent experiments with 3 biological replicates each. Error bars represent the standard error. Mean values marked with the same letter are not significantly different from each other as established by Student’s t-test (P<0.05). Results 97 (COI=0.69±0.09) (Figure 2B, Figure S1), despite the fact that AvrRpt2 alone triggers a 50-100 fold growth attenuation when expressed by DC3000 from the same plasmid. This result is expected since HopZ1a is capable of partially suppresing the defense response triggered by AvrRpt2 in Arabidopsis (Macho et al., 2010a). Accordingly, growth of DC3000 co-expressing AvrRpt2 and the HopZ1aC216A catalytic mutant was almost 50 fold lower than the growth of DC3000 expressing HopZ1aC216A alone (COI=0.03±0.01) (Figure 2B, Figure S1). However, growth of DC3000 co-expressing AvrRpt2 and the HopZ1aK289R mutant was only a 10-fold lower than growth of the strain expressing HopZ1aK289R alone (COI=0.09±0.01). These results indicate that mutation K289R decreases, but does not abrogate, HopZ1a ability to suppress AvrRpt2triggered restriction of growth, since co-expression of AvrRpt2 and HopZ1aK289R causes a smaller attenuation of growth than co-expression of AvrRpt2 and the HopZ1aC216A catalytic mutant or expression of AvrRpt2 alone (Figure 2B). Our results (Figure 2B) indicate that, unlike the catalytically inactive HopZ1aC216A mutant derivative, HopZ1aK289R mutant is still able to suppress AvrRpt2-triggered immunity, since it still suppresses AvrRpt2-triggered restriction of growth. The fact that we do not detect suppression of PR1 protein in plants inoculated with DC3000 expressing the HopZ1aK289R mutant may indicate that our assay is not sensitive enough, or that the association between the PR1 accumulation and growth restriction associated to AvrRpt2-triggered immunity is not linear. To this regards, a similar lack of linearity in the association between PR1 accumulation and growth restriction during induction of SAR has been previously shown (Macho et al., 2010a). HopZ1a partially suppresses AvrRpt2-triggered immunity in zar1-1 mutant plants Results presented in Figure 2B are in agreement with our previous report concluding that HopZ1a partially suppresses AvrRpt2-triggered ETI in Arabidopsis (Macho et al., 2010a). However, it has been recently reported that HopZ1a transgenic expression in zar1-1 plants does not interfere with AvrRpt2-induced macroscopic HR (Lewis et al., 2014). HopZ1a-triggered immunity in Arabidopsis is dependent on the ZAR1 resistance protein (Lewis et al., 2010). In the light of this report we decided to analyze the ability of HopZ1a to suppress AvrRpt2-triggered immunity in the absence of HopZ1a-triggered immunity. To this purpose, we analyzed HopZ1a impact on AvrRpt2-triggered restriction of growth in a zar1-1 plant genotype (Lewis et al., 2010). Chapter 3: K289 auto-acetylation and HopZ1a function 98 Using CI assays, we compared growth of DC3000 expressing HopZ1a or AvrRpt2 with growth of DC3000 in zar1-1 plants, to determine the growth restriction associated to ETI responses against each of these effectors in the mutant background (Figure 3). Growth of DC3000 expressing HopZ1a was very similar to growth of DC3000 in zar11 plants (CI=0.72±0.09) (Figure 3). Whereas, as expected since AvrRtp2-triggered immunity is independent of ZAR1, the expression of this effector in DC3000 still determined a strong attenuation of growth (CI=0.03±0.01) (Figure 3). However, coexpression of AvrRpt2 and HopZ1a in zar1-1 plants caused significantly less growth attenuation (CI=0.10±0.03) than that caused by expression of AvrRpt2 alone (Figure 3), demonstrating that HopZ1a suppression of AvrRpt2-triggered immunity takes place in the absence of HopZ1a-triggered immunity, and is not caused by an overlap or interference between the two ETI pathways. Figure 3. HopZ1a suppresses AvrRpt2-triggered ETI in zar1-1 plants. Competitive indices (CIs) measuring growth within a mixed infection of DC3000 expressing HopZ1a (Z1a, pAME30), AvrRpt2 (pAME8) or co-expressing both (Z1a + AvrRpt2, pAME33) in relation to growth of DC3000. CIs are calculated as the output ratio between the strain expressing the effector(s) and DC3000, divided by their input ratio. Each CI value represents the means of 3 independent experiments with 3 biological replicates each. Error bars represent the standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (P<0.05). Results 99 HopZ1aK289R retains the ability to suppress Systemic Acquired Resistance (SAR) triggered by DC3000 infection Both virulent and avirulent bacteria can trigger SAR, a defense response elicited in distal (systemic) tissues as a result of local infection. Activation of SAR determines both systemic accumulation of PR1, and restriction of growth of newly incoming bacteria (Cameron et al., 1994). We have previously shown that HopZ1a expression suppresses SAR triggered by DC3000, and that such suppression requires HopZ1a catalytic cysteine C216 (Macho et al., 2010a). To determine whether the HopZ1aK289R mutant retained HopZ1a ability to suppress SAR, we first analyzed the effect of the mutation K289R in HopZ1a ability to suppress SAR-associated restriction of growth of newly incoming bacteria. We inoculated primary leaves with either 10 mM MgCl2 (mock), DC3000, or DC3000 expressing HopZ1a or the corresponding mutant derivatives HopZ1aC216A or HopZ1aK289R (Figure 4A). Two days after inoculation of primary leaves, distal leaves were inoculated with DC3000, and 4 days after this second inoculation we monitored the growth of DC3000. Figure 4A shows that, as previously described, pre-inoculation of primary leaves with either DC3000 or DC3000 expressing the catalytically inactive mutant HopZ1aC216A triggered SAR to equivalent levels, since growth of DC3000 in distal leaves was similarly attenuated in both cases. In contrast, pre-inoculation with DC3000 expressing HopZ1a did not result in detectable attenuation of growth of DC3000 in distal leaves, since it did not show significant differences with that observed in mock pre-inoculated leaves, thus confirming the reported HopZ1a suppression of SAR (Macho et al., 2010a). Systemic leaves from plants pre-inoculated with DC3000 expressing HopZ1aK289R displayed DC3000 cfu values significantly different to those reached in plants pre-inoculated with DC3000 expressing HopZ1aC216A, but similar to those reached in plants pre-inoculated with DC3000 expressing HopZ1a (Figure 4A), supporting the notion that autoacetylation of HopZ1a on K289 is not required for suppression of SAR in Arabidopsis. Chapter 3: K289 auto-acetylation and HopZ1a function 100 To determine how the HopZ1aK289R mutant ability to suppress SAR correlates with suppression of PR1 accumulation in systemic tissue, we used western blot analysis to analyze accumulation of PR1 in systemic leaves of plants pre-inoculated with DC3000 or DC3000 expressing the different versions of HopZ1a. In keeping with previous results (Macho et al., 2010a), expression of HopZ1a in DC3000 suppresses PR1 accumulation in systemic tissues, since distal leaves of plants pre-inoculated with DC3000 expressing HopZ1a displayed a strong reduction of PR1 accumulation when Figure 4. HopZ1a-mediated suppression of SAR is reduced but not abolished by the K289R mutation. (A) Growth of DC3000 inoculated in secondary leaves of plants pre-inoculated in primary leaves by infiltrating either a 10 mM MgCl2 solution (Mock), or 5 x 105 cfu/ml of DC3000, DC3000 expressing HopZ1a (pAME30), or DC3000 expressing the mutant derivatives HopZ1aC216A (pAME27) or HopZ1aK289R (pMAM1). Two days post-inoculation of primary leaves, secondary leaves were inoculated with 5 x 104 cfu/ml of DC3000, and growth was measured at 4 days post-inoculation of the secondary leaves. The experiment was repeated four times with similar results, and the results shown correspond to a representative experiment. The values shown represent the means of 5 biological replicates. Error bars represent the standard error. Values marked with the same letter were not significantly different from each other as established by Student’s t-test (P<0.05). (B) Western blot analysis for immunodetection of PR1 on distal non-inoculated leaves, 2 days after inoculating primary leaves with either 10 mM MgCl2 (Mock), or 5 x 105 cfu/ml of DC3000, DC3000 expressing HopZ1a (pAME30), or DC3000 expressing the mutant derivatives HopZ1aC216A (pAME27) or HopZ1aK289R (pMAM1). Ten micrograms of total protein were loaded per sample, and Coomassie staining is shown as loading control. The experiment was repeated twice with similar results. Results 101 compared to plants pre-inoculated with DC3000 (Figure 4B). This suppression is dependent on HopZ1a catalytic cysteine, since systemic leaves of plants pre-inoculated with DC3000 expressing HopZ1aC216A displayed PR1 levels that were higher than those observed in plants pre-inoculated with DC3000 expressing HopZ1a (Figure 4B). As previously reported (Macho et al., 2010a) the C216A mutation did not entirely abolish HopZ1a ability to suppress PR1 accumulation, since the systemic levels of PR1 in plants pre-inoculated with DC3000 expressing HopZ1aC216A did not reach the levels observed in plants pre-inoculated with DC3000 (Figure 4B). Interestingly, when primary leaves were inoculated with DC3000 expressing HopZ1aK289R, the accumulation of PR1 in distal leaves reached levels that were intermediate between those elicited by DC3000 expressing HopZ1aC216A and DC3000 expressing wild type version of the effector (Figure 4B). These results indicate that the HopZ1aK289R mutant is still able to partially suppress systemic accumulation of PR1 in response to DC3000. HopZ1aK289R triggers ETI in Arabidopsis and N. benthamiana Inoculation of Arabidposis leaves with a 5x107 cfu/ml of DC3000 expressing HopZ1a induces macroscopic HR symptoms in Arabidopsis leaves, which are absent in leaves inoculated with the same dose of DC3000 expressing the HopZ1aC216A catalytic mutant (Lewis et al., 2008; Macho et al., 2010a). It has been reported that the mutation K289R completely prevents HopZ1a-triggered HR, which cannot be detected when expressing the mutant effector under the control of its own promoter (Lee et al., 2012). However, considering that our results presented above indicate that such mutation does not render the effector entirely inactive, we wondered whether HopZ1aK289R could still be able to trigger immunity in Arabidopsis. To analyze whether the HopZ1aK289R mutant was able to trigger macroscopic HR in Arabidopsis, we inoculated leaves with 5x107 cfu/ml of DC3000, DC3000 expressing HopZ1a, or DC3000 expressing either HopZ1aC216A or HopZ1aK289R mutant derivatives, and monitored HR development by 20–24 hpi (Figure 5A). Development of macroscopic HR requires a rather strong ETI response, which might not be reached by lower levels of effector expression (Macho et al., 2009), thus we expressed HopZ1a and its mutant derivatives under the control of the strong constitutive nptII promoter, to factor in the chance of an stronger effector expression allowing detection of a mild ETI. In keeping with previous reports (Lewis et al., 2008; Macho et al., 2010a), a clear HR Chapter 3: K289 auto-acetylation and HopZ1a function 102 was detected in plants inoculated with DC3000 expressing HopZ1a, while no HR could be detected in leaves inoculated with either DC3000 or DC3000 expressing HopZ1aC216A (Figure 5A). Interestingly, leaves inoculated with DC3000 expressing HopZ1aK289R displayed noticeable macroscopic HR (Figure 5A). As expected from previous reports (Lewis et al., 2010) (Lewis et al., 2014), inoculation of zar1-1 leaves with these strains did not induce any visible cell death symptom (Figure S2). The ETI triggered in Arabidopsis against HopZ1a in Arabidopsis determines a strong restriction of bacterial growth that can be measured using competitive index assays (CIs), in mixed infections of DC3000 co-inoculated with DC3000 expressing HopZ1a (Macho et al., 2009; Macho et al., 2010a). To further investigate the impact of the K289R mutation in HopZ1a activation of ETI in Arabidopsis, we performed CI assays by co-inoculating Arabidopsis plants with DC3000 and DC3000 expressing HopZ1a, HopZ1aC216A, or HopZ1aK289R (Figure 5B). As previously described (Macho et al., 2010a), a clear growth attenuation was measured for DC3000 expressing HopZ1a in comparison with co-inoculated DC3000 (CI=0.03±0.01), while no significant attenuation was detected for DC3000 expressing HopZ1aC216A catalytically inactive (CI=0.91±0.10) (Figure 5B). In contrast, DC3000 expressing HopZ1aK289R displayed a small attenuation of growth (CI=0.46±0.10), significantly smaller than that measured for DC3000 expressing HopZ1a, but significantly different from the absence of attenuation observed for HopZ1aC216A-expressing DC3000 bacteria (Figure 5B). Results 103 In addition to triggering HR in Arabidopsis, HopZ1a has been shown to trigger macroscopic HR in Nicotiana benthamiana leaves when transiently expressed using Agrobacterium (Ma et al., 2006); (Lewis et al., 2008). We expressed HopZ1a and its mutant derivatives HopZ1aC216A and HopZ1aK289R in N. benthamiana leaves, under the control of a constitutive promoter, by using Agrobacterium-mediated transient expression, and monitored HR symptoms at 40 hours after Agrobacterium inoculation. While transient HopZ1aC216A overexpression did not result in HR elicitation whatsoever, HopZ1aK289R overexpression elicited an HR of a similar intensity to that elicited by overexpressing HopZ1a (Figure 5C). Figure 5. HopZ1aK289R triggers ETI. (A) Hypersensitive response (HR) to hand-infiltration of Col-0 leaves with bacterial suspensions containing 5 x 107 cfu/ml of DC3000 alone or DC3000 expressing HopZ1a (pAME30), or each of its mutant derivatives HopZ1aC216A (pAME27) or HopZ1aK289R (pMAM1). Photographs were taken 24 hours post-inoculation. Images are representative of at least 30 inoculated leaves per strain and experiment. The experiment was repeated twice with similar results. (B) Competitive indices (CIs) measuring growth within a mixed infection of DC3000 expressing HopZ1a (Z1a, pAME30), or each of its mutant derivatives HopZ1aC216A (Z1aC216A, pAME27) or HopZ1aK289R (Z1aK289R, pMAM1), in relation to growth of DC3000. CIs are calculated as the output ratio between the strain expressing the effector and DC3000, divided by their input ratio. Each CI value represents the means of 3 independent experiments with 3 biological replicates each. Error bars represent the standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (P<0.05). (C) Development of HR following transient expression of either 6xHis-HopZ1a (pBINZ1) or each of its mutant derivatives 6xHisHopZ1aC216A (pBINZ2) or 6xHis-HopZ1aK289R (pBINZ3). Nicotiana benthamiana leaves were inoculated with Agrobacterium tumefaciens C58C1 carrying binary plasmids encoding the corresponding effector genes. Pictures were taken 48 hours post inoculation. The experiment was repeated 3 times with similar results. Chapter 3: K289 auto-acetylation and HopZ1a function 104 Taken together, results shown in Figure 5 indicate that autoacetylation of HopZ1a in its lysine 289 contributes, but it is not essential, to trigger ETI in Arabidopsis. HopZ1aK289R-triggered defenses effectively protects Arabidopsis against disease development We have previously shown that resistance triggered in Arabidopsis by the expression of HopZ1a efficiently protects plants from DC3000 infection, resulting in the absence of virulence-associated disease symptoms (Macho et al., 2010a). To analyze whether the defense response triggered against HopZ1aK289R mutant is sufficient to stymie DC3000 disease in Arabidopsis, we monitored development of disease symptoms at 4–6 dpi on plants spray-inoculated with DC3000, DC3000 expressing HopZ1a, or DC3000 expressing either HopZ1aC216A or HopZ1aK289R mutant derivatives. As expected from previous results (Macho et al., 2010a), plants sprayed with either DC3000 or DC3000 expressing HopZ1aC216A displayed noticeable disease symptoms, namely chlorosis and stunted growth, while plants sprayed with DC3000 expressing HopZ1a did not (Figure 6). Interestingly, plants sprayed with DC3000 expressing HopZ1aK289R did not display chlorosis and only a slightly reduction in plant growth could be observed (Figure 6). These results clearly show that the HopZ1aK289R mutant triggered-resistance effectively protects Arabidopsis plants from disease. Figure 6. Expression of HopZ1aK289R from DC3000 protects Col-0 plants against disease. Virulence symptoms caused by spray-inoculated DC3000, or DC3000 expressing HopZ1a (pAME30), or each of its mutant derivatives HopZ1aC216A (pAME27) or HopZ1aK289R (pMAM1). Arabidopsis plants were sprayed with bacterial suspensions containing 5 x 107 cfu/ml in 0.02% Silwet L-77, and photographed 6 days post-inoculation. The experiment was repeated three times using 5 plants per strain, and representative images are shown. Discussion 111 Figure S1. Diagram depicting the cancelled-out analysis carried out in this study. A. COI is defined as ratio between cfu of the the strain expressing both AvrRpt2 and any version of HopZ1a, and the cfu of the strain expressing only the corresponding HopZ1a version in the output sample, divided by their ratio within the input inoculum. B. Determination and analysis of COI. A mix inoculum containing an equal bacterial number of both strains is infiltrated into plant leaves. The inoculum is plated onto LB and LB supplemented with antibiotics, to differentiate between the co-inoculated strains, and to establish the input ratio which should be close to 1. Bacteria are recovered from plant leaves at 4 days post inoculation (dpi), and plated into LB and LB supplemented with antibiotics, to differentiate the strains ant to determine their output ration. I and II represent the two control outcomes for the analysis. ! ! ! ! ! ! ! ! Chapter(4:"The$bacterial$effector$ HopZ1a'acetylates'MKK7'to'suppress' plant&defense&responses! José S. Rufián, Diego López-Márquez, Javier Rueda-Blanco, Carmen R. Beuzón, Javier Ruiz-Albert Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 114 Introduction, HopZ1a is a P. syringae T3E with the ability to suppress several layers of plant defense. HopZ1a has been shown to suppress in Arabidopsis (i) basal resistance or PTI triggered by P. syringae pv. tomato DC3000 (Macho et al., 2010a; Lewis et al., 2014) (ii), ETI triggered by the expression of the heterologous effectors AvrRpt2, AvrRps4 and AvrRpm1 (Macho et al., 2010a; Rufian et al., 2015), and (iii) systemic acquired resistance (SAR) triggered by either virulent or avirulent bacteria (Macho et al., 2010a; Rufian et al., 2015). On the other hand, HopZ1a triggers ETI in Col-0 Arabidopsis plants upon recognition by the ZAR-1 resistance protein (Lewis et al., 2010), a defense response that is independent of salicylic acid (SA) and EDS1 (Lewis et al., 2010; Macho et al., 2010a). HopZ1a belongs to the YopJ / HopZ superfamily of T3Es, which includes representatives from both animal and plant pathogens (Ma et al., 2006; Lewis et al., 2010). Many of these T3Es have been described to function as acetyltransferases, among other biochemical activities (Trosky et al.; Zhou et al.; Mittal et al., 2006; Jones et al.; Lee et al., 2012). In fact, HopZ1a has been shown to display acetyltransferase activity, with varying degrees of efficiency, on some of its proposed plant targets (Lee et al., 2012; Jiang et al., 2013) or decoys (Lewis et al., 2013). HopZ1a acetyltransferase activity is completely dependent on the integrity of the catalytic triad cysteine (C216), since a HopZ1aC216A mutant behaves as a catalytically inactive mutant (Lee et al., 2012). Likewise, residue C216 is essential for all described HopZ1a virulence and avirulence functions in planta (Ma et al., 2006; Lewis et al., 2008; Macho et al., 2010a; Lewis et al., 2014; Rufian et al., 2015). HopZ1a has also been described to autoacetylate in two serine residues (S349 and S351) that are required for acetyltransferase activity in vitro, virulence activity in planta, and interaction with the co-factor IP6 (Ma et al., 2015). In turn, lysine residue K289, which was originally postulated to be the only HopZ1a auto-acetylation site and key for effector activity (Lee et al., 2012), might only partially contribute to trans-acetylation (Ma et al., 2015) and does not seem to be essential for HopZ1a virulence or avirulence activities (Rufian et al., 2015). A common theme among the majority of T3Es that constitute the YopJ / HopZ superfamily, particularly for those present in animal pathogens, seems to be the interference with MAPK signaling cascades leading to the activation of the immune response. The archetypal member of the superfamily, Yersinia effector YopJ, acetylates Introduction 115 key serine and threonine residues of a subset of mitogen-activated protein kinase kinases (MAP2Ks or MKKs) and MAP kinase kinase kinases (MAP3Ks) from several animal models, competing with the phosphorylation of said residues, which in turn leads to inactivation of downstream defense signaling, and suppression of the immune response (Mittal et al., 2010 ; Meinzer et al., 2012). YopJ can also acetylate lysine residues of several of its targets, however this modification does not seem to be essential for its inhibitory function (Mukherjee et al., 2006; Paquette et al., 2012). Within the same superfamily, AvrA from Salmonella and VopA from Vibrio acetylate key serine, threonine, and lysine residues of their corresponding target MKKs, resulting in inhibition of kinase activity and the suppression of immune responses (Trosky et al.; Jones et al.). YopJ and VopA can also interfere with MAPK modules in yeast, via acetylation of kinases PBS2, MEK1, or Ste7 (Yoon et al., 2003; Trosky et al., 2004; Hao et al., 2008). In plants, MAPK cascades also constitute signaling networks leading to defense against pathogens, since PRR recognition of PAMPs leads to activation of MAPK modules and ultimately to the corresponding immune response (reviewed by (Pitzschke et al.; Feng and Zhou, 2012)). Interestingly, HopZ1a has been described to suppress MAPK activation in Arabidopsis (Lewis et al., 2014). Further, several other T3Es suppress plant defense signaling by targeting mitogen-activated protein kinase (MAPK) cascades at different levels, as in the case of HopAI1 interfering with several MAP kinases (MPKs) (Zhang et al., 2007a; Zhang et al., 2012), or HopF2 blocking the phosphorylation of a MAP kinase kinase, MKK5 (Wang et al., 2010). In fact, a considerable number of T3Es interact with plant kinases, altering their function, which results in the interference of plant defense signaling (Block and Alfano, 2011; Macho and Zipfel, 2015). Some T3Es, such as AvrPto, AvrPtoB, or HopF2, target PRRs or its co-receptors, typically acting on them as kinase inhibitors and suppressing early PTI events (Shan et al., 2008; Xiang et al., 2008; Cheng et al., 2011; Wu et al., 2011; Xiang et al., 2011; Zeng et al., 2012a; Zhou et al., 2014). Other T3Es, such as AvrPphB or AvrAC, act on PRR-associated receptor-like cytoplasmic kinases (RLCKs), inactivating their function to suppress PTI (Zhang et al., 2010; Feng and Zhou, 2012). Since RLCKs and MAPKs are important components of plant defense signaling targeted by T3Es, they are accordingly guarded by plant resistance proteins that in resistant plants can detect, directly or indirectly, the alterations induced by T3E action Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 116 on their targets, and trigger the corresponding ETI defense responses, as those described for AvrPphB or AvrAC when acting respectively on the RLCKs PBS1 (Shao et al., 2003) or PBL2/RIPK (Guy et al., 2013), or for HopAI1 while acting on MPK4 (Zhang et al., 2012). In this sense, the ETI triggered by HopZ1a in Arabidopsis seems to be the consequence of the recognition, by the ZAR-1 resistance protein, of HopZ1a acetylation of the RLCK pseudokinase ZED1, which is proposed to act as a decoy (Lewis et al., 2010; Lewis et al., 2013). In addition to the molecular decoy ZED1, a number of plant proteins have been proposed to be the targets of HopZ1a virulence activity (Zhou et al., 2011; Lee et al., 2012; Jiang et al., 2013). Zhou et al. (2011) described the interaction of HopZ1 with Glycine max HID1 (GmHID1), an enzyme involved in the biosynthesis of daidzein, a major soybean isoflavone. HopZ1 activity resulted in the degradation of GmHID1, and the suppression of daidzein biosynthesis, which is induced by P. syringae. However, the authors failed to find any direct inhibitory effect of daidzein on P. syringae growth, therefore the putative role of GmHID1 in PTI was not confirmed, likewise its suppression by HopZ1. It is important to notice that GmHID1 has no putative ortholog in Arabidopsis, the plant model were the majority of HopZ1a virulence or avirulence activities have been characterized. Later on, Lee et al. (2012) described the interaction of HopZ1a with tubulin, using a heterologous in vivo screen in human HEK293T cells. In Arabidopsis, HopZ1a interacted with tubulin at a higher rate than the unrelated control effector HopF2. HopZ1a expression in Arabidopsis seedlings destroyed, with potential consequences on plant basal defense, the microtubule network to a higher extent than expression of the unrelated control effector AvrRpt2, in conditions were both T3Es were triggering ETI to some degree. However, the authors acknowledged that such microtubule destruction might be an indirect effect of HopZ1a acting on a yet unidentified protein, such as a MAPK. Finally, Jiang et al. (2013) described the interaction of HopZ1a with soybean and Arabidopsis JAZ proteins, key negative regulators of jasmonate (JA) signaling. HopZ1a expression induced COI-dependent degradation of JAZ, thus promoting JA-responsive gene expression. The authors hypothesized that, since plant JA signaling is to some Introduction 117 extent antagonistic to SA-dependent plant defense, activation of JA signaling by HopZ1a action might facilitate host defense suppression and pathogenesis. Remarkably, the aforementioned reports (Zhou et al., 2011; Lee et al., 2012; Jiang et al., 2013) addressed only indirectly, if at all, HopZ1a ability to suppress plant basal defenses in regard to the proposed plant targets, while neither addressed the involvement of HopZ1a in the suppression of ETI or SAR. It is conceivable that HopZ1a interferes with many host proteins, with a number of them yet to be identified behind the unexplained suppression phenotypes. Alternatively, given the broad plant defense suppression abilities of HopZ1a, its host target might be a single key positive regulator of defense, participating in the signaling of different branches of the immune response. Considering the pattern of interference with MKKs displayed by many T3Es of the YopJ / HopZ superfamily leading to defense suppression, the existence of multiple T3Es from plant pathogens interfering with host kinases to suppress immunity, the effect on MAPK activation displayed by HopZ1, and the fact that HopZ1a molecular decoy is indeed a plant kinase, we decided to analyze Arabidopsis MKKs as putative targets for HopZ1a. Arabidopsis genome presents only ten genes encoding MKKs, of which only eight are likely to be expressed (Zhang et al., 2008). Among these, MKK3, MKK4/5, and MKK7 have all been identified as positive regulators of plant defense (Doczi et al., 2007; Zhang et al., 2007a; Xu et al., 2008). Interestingly, while neither MKK3 nor MKK4/5 cascades have been associated to positive signaling of SAR, the suppression of which is a trademark of HopZ1a virulence activity, MKK7 has been proved to be essential for SAR activation, since specific silencing of MKK7 blocks SAR induction and promotes growth of P. syringae pv. maculicola and Xanthomonas campestris pv. campestris (Zhang et al., 2007b). All things considered,, we decided to investigate MKK7 as a potential target for HopZ1a interference, such a key target that could singlehandedly account for all the defense suppression phenotypes described for this effector. In this work, we demonstrate that MKK7 participates in the signaling of basal defenses restricting DC3000 growth in Arabidopsis, validating this model pathosystem for the characterization of MKK7 potential interactors. Our results are the first to show MKK7-dependent accumulation of callose, ROS burst response, and MAPK activation, which eventually restrict the growth of both wild-type DC3000 and the corresponding T3SS null mutant. Furthermore, we demonstrate the participation of MKK7 in the Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 118 signaling of the ETI response triggered by AvrRpt2. In regard to HopZ1a interference with MKK7 signaling, we first demonstrate that this effector can suppress the abovementioned MKK7-dependent PTI and ETI defense responses, and continue to prove HopZ1a-MKK7 interaction in planta, as well as HopZ1a-dependent acetylation of MKK7 in vitro. Finally, we determine that MKK7 acetylation occurs in a conserved lysine residue, which we show to be essential for MKK7 activity in vitro and in planta, and that such modification reduces MKK7 self-phosphorylation, and abrogates MKK7 trans-phosphorylation activity on a generic susbtrate , Table 1. Plasmids used in this work Name Promoter Expressed protein Resistance Reference pAME30 nptII HopZ1a Amp, Km Macho et al. (2010a) pAME27 nptII HopZ1aC216A Amp, Km Macho et al. (2010a) pMAM1 nptII HopZ1aK289R Amp, Km Rufián et al., (2015) pAME8 nptII AvrRpt2 Amp, Km Macho et al. (2009) pAME30Gm nptII HopZ1a Amp, Km, Gm Rufián et al. (2015) pENTRTM/DTOPO - - Km Invitrogen pENTR-Z1a - HopZ1a Km This work pENTR-C2 - HopZ1aC216A Km This work pENTR-K2 - HopZ1aK289R Km This work pENTR-GFP - GFP Km This work pENTR-MKK7 - MKK7 Km This work pENTRMKK7K167R - MKK7 K167R Km This work pMD1 35S - Km Tai et al., (1999) pMD-Z1 35S HopZ1a-3xFLAG Km Macho et al. (2010) pMD-C2 35S HopZ1aC216A3xFLAG Km This work pMD-K2 35S HopZ1aK289R3xFLAG Km This work pMD-GFP 35S GFP-3xFLAG Km This work pMD-MKK7 35S MKK7-HA Km This work pMD-MKK7K167R 35S MKK7 K167R -HA Km This work Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 120 pDEST-GWVYNE 35S VENUS-N-term Km Gehl et al., (2009) pDEST-GWVYCE 35S VENUS-C-term Km Gehl et al., (2009) pZ1-VYNE 35S HopZ1a-VENUSN-term Km This work pC2-VYNE 35S HopZ1aC216AVENUS-N-term Km This work pK2-VYNE 35S HopZ1aK289RVENUS-N-term Km This work pMKK7-VYCE 35S MKK7-VENUSC-term Km This work pET28a(+) T7 - Km Novagen (USA) pET28-Z1a T7 HopZ1a Km This work (Chapter 3) pET28-C2 T7 HopZ1aC216A Km This work (Chapter 3) pET28-K2 T7 HopZ1aK289R Km This work (Chapter 3) pGEX-5X-1 tac GST Amp GE Healthcare pGEX-MKK7 tac GST-MKK7 Amp This work pGEX-MKK7K74R tac GST-MKK7K74R Amp This work pGEX-MKK7K167R tac GST-MKK7K167R Amp This work Results 127 HopZ1a suppresses MKK7-dependent defense signaling Taken together, results presented previously validate MKK7 as a potential target for the virulence function of HopZ1a within the plant, since suppression of MKK7dependent defense responses will account for all described HopZ1a-associated virulence phenotypes in the Arabidopsis-DC3000 pathosystem, i.e. suppression of plant defenses associated to PTI (Figures 1 and 2), ETI (Figure 3) and SAR (Zhang et al., 2007b). Therefore, we set out to analyze whether HopZ1a expression interfered specifically with MKK7-dependent defense signaling in Arabidopsis. First, we analyzed whether HopZ1a was able to suppress MKK7-dependent PR1 accumulation (Figure 4A). Overexpression of MKK7 in the de-regulated Arabidopsis mutant bud1 (Dai et al., 2006) has been shown to result in constitutive expression of the molecular marker genes for plant defense responses PR1, PR2, and PR5 (Zhang et al., 2007b). Furthermore, overexpression of MKK7 in induced MKK7-DEX plants has been described to induce PR1 gene expression (Liu et al., 2007). We induced MKK7 expression in MKK7-DEX plants by DEX treatment, 3 hours before inoculation by infiltration with Pseudomonas fluorescens strain Pf55 (hereafter Pf55), a non-pathogenic strain expressing a functional T3SS but only those effector genes purposely cloned and expressed from a plasmid, in this case the HopZ1a effector gene. We then monitored PR1 protein levels by Western blot in plant extracts taken at the inoculation site, 48 hours after bacterial inoculation (Figure 4A). As controls we included mock inoculated plants, and plants inoculated with a Pf55 strain not expressing the HopZ1a gene (Figure 4A). As expected, both control plants displayed local PR1 accumulation in the inoculated tissues as a consequence of transgenic MKK7 expression. PR1 accumulation was higher in Pf55-inoculated than in mock-inoculated plants, since in the former PR1 expression is also induced as a consequence of the mild basal defense triggered by the Pf55 strain. Local PR1 accumulation was almost completely abolished in plants inoculated with Pf55 expressing the HopZ1a effector protein, indicating that HopZ1a is capable of suppressing PR1 accumulation due to Pf55 defense elicitation, and more importantly, capable of suppressesing MKK7-dependent PR1 accumulation. Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 128 To investigate whether HopZ1a suppression of MKK7-dependent PR1 accumulation correlates with higher bacterial growth, we analyzed whether HopZ1a expression from DC3000 was able to abrogate the growth attenuation suffered by this strain in induced MKK7-DEX plants, as a consequence of MKK7 overexpression (Figure 4B). To analyze HopZ1a suppression ability in the absence of the ETI triggered by HopZ1a in Arabidopsis Col-0, we crossed DEX-MKK7 transgenic plants with zar11 knockout plants, lacking the resistance gene ZAR1 that is responsible of the HopZ1atriggered immunity, and selected zar1-1 / DEX-MKK7 homozygous lines. We used Figure 4. HopZ1a interferes in the MKK7 defense activation pathway. (A) Western blot showing PR-1 accumulation in DEX-MKK7 plants induced with dexamethasone. 3 hours after DEX treatment, plants were infiltrated with 10 mM MgCl2 (Mock), or a 5x105 cfu/ml bacterial suspension of Pf55 or Pf55 expressing HopZ1a. Samples were recovered at 48 hours. Ten micrograms of total protein were loaded per sample, and Coomassie staining is shown as loading control. Results presented are representative of two independent experiments. (B) Competitive indices (CI) resulting from mixed infections of DC3000 with DC3000 expressing the effector HopZ1a in zar1-1/DEX-MKK7 plants induced (DEX+) or un-induced (DEX-) with dexamethasone. Index values shown correspond to the mean of three samples and error bars represent the standard error. Experiment was repeated three times with similar results. Results 129 these plants, either DEX-induced for MKK7 overexpression or non-induced, to perform CI analysis comparing the growth of DC3000 versus a co-inoculated DC3000 strain expressing HopZ1a (Figure 4B). As expected (Jiang et al., 2013; Rufian et al., 2015), in non-induced zar1-1 plants growth of DC3000 expressing HopZ1a presents only a slight growth reduction compared with the co-inoculated DC3000 strain (CI=0,55±0,05). In contrast, in DEX-induced zar1-1 plants overexpressing MKK7, DC3000 expressing HopZ1a displays a growth advantage versus DC3000 (CI=1,52±0,27), suggesting that HopZ1a can in fact suppress MKK7-dependent defense responses. HopZ1a interacts with MKK7 in vitro and in planta Data presented in Figure 4 demonstrates that HopZ1a suppresses MKK7dependent defense signaling in Arabidopsis. To investigate whether such HopZ1a defense suppression ability is a consequence of its direct interference with MKK7 function, or rather an indirect effect due to its interaction with another component of the same signaling pathway, we analyzed in planta HopZ1a-MKK7 interaction using coimmunoprecipitation and BiFC assays. Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 130 Figure 5. HopZ1a interacts with MKK7 in planta. (A) Coimmunoprecipitation assay using anti HA beads of MKK7-HA with HopZ1a and derivatives. N. benthamiana plants were agroinfiltrated with mixed inoculum containing the strains encoding MKK7-HA and each one of the 3xFLAG fusions. Protein extraction and immunoprecipitation was performed at 30 hours post inoculation. The output panel show the signal of the 3xFLAG fusion proteins in the elutions. The input panels show immunoblots of 10 ul of total proteins using anti FLAG or anti HA antibodies. (B) Bimolecular fluorocomplementation (BiFC) assay. The N-terminal of the fluorescent protein VENUS was fused to the C-terminal of HopZ1a, and the C-terminal of VENUS to the C-terminal of MKK7. Proteins were transiently expressed in N. benthamiana using Agrobacterium. Leaf sections were analyzed under the confocal 20 hours after inoculation. Results 131 For co-immunoprecipitation assays, we cloned MKK7 and HopZ1a in binary vectors in order to express C-terminal fusion proteins with either HA-tag or 3xFLAGtag, respectively (Table 1). Since catalytically inactive enzymes have been sometimes found to exhibit stabilized interactions with substrates (Blanchetot et al., 2005), we also generated the corresponding plasmids to express the canonical HopZ1a catalytic mutant (HopZ1aC216A-3xFLAG), and the leaky mutant in lysine K289 (HopZ1aK289R-3xFLAG), which might be affected in the transacetylation of the target protein (Ma et al., 2015). Further, HopZ1aC216A does not trigger HR when transiently expressed in planta (Jiang et al., 2013; Lewis et al., 2013) which might improve protein recovery. We co-expressed MKK7-HA with, either HopZ1a-3xFLAG or each of its mutant derivatives HopZ1aC216A-3xFLAG and HopZ1aK289R-3xFLAG, in N. benthamiana leaves, under the control of a constitutive promoter, by using Agrobacterium-mediated transient expression. As a negative control, we used GFP-3xFLAG. We recovered the Agrobacterium infiltrated tissue at 24-30 hours after Agrobacterium inoculation, since transient expression of either MKK7 or HopZ1a, cause cell death in N. benthamiana at later time points (Ma et al., 2006; Lewis et al., 2008; Popescu et al., 2009). We then performed co-immunoprecipitation with protein extracts obtained from the recovered tissues using anti-HA beads, and the corresponding eluates were separated by SDSPAGE and analyzed by western blot using anti-FLAG antibodies. We could not detect interaction of MKK7-HA with either HopZ1a-3xFLAG or the catalytic mutant HopZ1aC216A-3xFLAG. Interestingly, we did detect a hybridization signal indicative of interaction between MKK7-HA and HopZ1aK289R-3xFLAG (Figure 5A), demonstrating that both proteins are indeed interacting partners in planta. To confirm the detected interaction, we performed a bimolecular complementation assay (BiFC). To this purpose, we generated binary plasmids expressing MKK7, HopZ1a, or the corresponding HopZ1a mutant derivatives, as fusions to either the N-terminal or the C-terminal domains of the fluorescent protein VENUS (Table 1). We expressed MKK7 as a fusion protein to the C-terminal domain of VENUS, and expressed HopZ1a and its mutant derivatives as fusion proteins to the N-terminal domain of VENUS. We then co-expressed MKK7-VENUS-C with each of the HopZ1a-VENUS-N fusion proteins in N. benthamiana by using Agrobacteriummediated transient expression, and analyzed by microscopy the samples at 18-20 hours, a time point prior to the development of HR symptoms at the microscopic level. In this Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 132 assay, we detected fluorescence in those samples corresponding to the co-expression of MKK7-VENUS-C with either HopZ1a-VENUS-N or HopZ1aK289R-VENUS-N (Figure 5B). Interestingly, fluorescence was not detectable in the sample corresponding to MKK7-VENUS-C co-expression with HopZ1aC216A-VENUS-N (Figure 5B). Taken together, our co-immunoprecipitation and BIFC results show that HopZ1a and MKK7 do interact in planta, and suggest that such interaction is transient, and may be locked-on in the HopZ1aK289R mutant, but not in the catalytic mutant. HopZ1a acetylates MKK7 in vitro on residue K167 HopZ1a has been described to function as an acetyltransferase in vitro, capable of strong autoacetylation, and also of transacetylation of tubulin and the Arabidopsis pseudokinase ZED1 (Lee et al., 2012; Lewis et al., 2013). To determine whether HopZ1a was able to acetylate MKK7, we performed a 14C-labelled-acetyl-coenzyme A (acetyl-CoA) transferase reaction in vitro, in the presence of MKK7 and either HopZ1a, the catalytically inactive mutant HopZ1aC216A, or mutant HopZ1aK289R. As previously described, HopZ1a was strongly autoacetylated, while mutant HopZ1aC216A was not (Figure 6). HopZ1aK289R was acetylated to a similar level of the wild type version of the effector, in agreement with a recent report (Ma et al., 2015). Interestingly, MKK7 was acetylated in the presence of HopZ1a, but not in the presence of HopZ1aC216A or HopZ1aK289R (Figure 6), demonstrating that HopZ1a acetylates MKK7 in vitro. Figure 6. HopZ1a in vitro acetylates MKK7 in the lysine 167. Recombinant GST-MKK7, GSTMKK7K167R or GST were incubated with 6xHis-HopZ1a, 6xHis-HopZ1aC216A or 6xHis-HopZ1aK289R in acetylation buffer containing 14C-Acetyl CoA. Samples were separated in a SDS-PAGE and proteins were transferred to a PVDF membrane. The membrane was exposed to an imaging plate for one week. Results 133 We then set to determine which MKK7 residues were acetylated by HopZ1a, and to that purpose we examined the evidence available regarding HopZ1a acetylation of pseudokinase ZED1. HopZ1a acetylates the Arabidopsis pseudokinase ZED1 on threonine residues located in positions 125 and 177 of its amino acid sequence (Lewis et al., 2013). However, ZED1 pseudokinase is not considered a proper target of HopZ1a virulence activity, but rather believed to function as a decoy to trigger ZAR1-dependent ETI responses (Lewis et al., 2013). In fact, ZED1 is a non-functional kinase, since it lacks a critical proton accepting aspartate within its catalytic loop, a domain that also includes HopZ1a-acetylated threonine 177. Since decoys are expected to mimic the proteins targeted by pathogen effectors, we reasoned that HopZ1a acetylation on its bona fide target kinase was likely to happen on residues situated within the catalytic loop. MKK7 lacks threonine residues in this domain, but presents a lysine residue (K167). We introduced a point mutation on MKK7 substituting the residue K167 with arginine, generating the mutant MKK7K167R, and used it as substrate for HopZ1a acetylation in vitro (Figure 6). The MKK7K167R mutant derivative was not acetylated (Figure 6). These results suggest that K167 is the only, or at least the main, MKK7 residue subject to HopZ1a acetylation. Lysine residue K167 is essential for MKK7 activity in vitro and in planta It has been previously shown that MKK7-dependent activation of the plant defense response requires MKK7 kinase activity (Zhang et al., 2007b). Recombinant MKK7 expressed in E. coli displays in vitro auto-phosphorylation activity, which is absent in the MKK7K74R mutant version of the protein, in which a lysine in the ATP binding site was changed to an arginine (Dai et al., 2006). To determine whether the K167 residue is necessary for MKK7 kinase activity, we carried out in vitro GSTMKK7, GST-MKK7K74R, and GST-MKK7K167R autophosphorilation and trans phosphorylation assays using the generic substrate MBP (Myelin Basic Protein), (Figure 7A). Autophosphorylation of GST-MKK7 is abolished on the GST-MKK7K74R mutant, as expected. Interestingly, the K167R mutation reduces autophosphorylation of MKK7 by an 82%, and completely abolished its trans phosphorylation activity on MBP (Figure 7A) Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 134 To confirm that residue K167 is also essential for MKK7 activity in planta, we expressed MKK7-HA and its mutant derivative MKK7K167R-HA in N. benthamiana leaves, under the control of a constitutive promoter, by using Agrobacterium-mediated transient expression, and monitored cell death symptoms 40 hours after Agrobacterium Figure 7. Lysine 167 is important for MKK7 activity. (A) Recombinant GST-MKK7, GSTMKK7K74R, GST-MKK7K167, or GST were incubated in kinase buffer containing 32P-γ-ATP. Samples were separated in a SDS-PAGE and proteins were transferred to a PVDF membrane. The membrane was exposed to an imaging plate for one day. Coomassie staining is shown as loading control. (B) Agrobacterium transient expression in N. benthamiana of MKK7-HA and MKK7K167R-HA. Pictures were taken two days post inoculation. Results 135 inoculation (Figure 7B). Transient overexpression of MKK7 resulted in the manifestation of macroscopic cell death in N. benthamiana tissues (Figure 7B), most likely as a result of its unregulated activation of the defense responses, as previously suggested by the failure to generate Arabidopsis transgenic plants expressing this kinase from a 35S promoter (Dai et al., 2006; Zhang et al., 2007b). Interestingly, MKK7K167R expression did not elicited macroscopic HR (Figure 7B). These results indicate that the K167 residue, targeted for acetylation by HopZ1a, is essential for MKK7 activity in planta. Acetylation of K167 by HopZ1a interferes with MKK7 kinase activity Since we have demonstrated that HopZ1a acetylates MKK7 in a lysine residue that is essential for MKK7 kinase activity in vitro, and for HR induction in planta, we set out to analyze whether such acetylation indeed affected MKK7 function.. To this purpose, we performed an in vitro acetylation assay using HopZ1a as the means for MKK7 acetylation, followed by a kinase assay on the HopZ1a-modified MKK7 resulting from the previous experiment. We first incubated GST-MKK7 or GSTMKK7K167R with either 6xHis-HopZ1a or 6xHis-HopZ1aC216A in an acetylation buffer containing non-radioactive acetyl-CoA as acetyl group donor. After this incubation, we added the components needed for kinase reaction, including the artificial substrate MBP and 32P-ATP to provide the necessary radioactive signal. The corresponding protein samples were separated by SDS-PAGE and transferred to a PVDF membrane, which was exposed to detect the autoradiographic signal. Data presented in Figure 8 indicates that the levels of phosphorylated MKK7 are reduced in those samples pre-treated with HopZ1a, when compared to those observed when MKK7 was pre-treated with the catalytic mutant HopZ1aC216A (Figure 8). This suggests that acetylation of MKK7 by HopZ1a, which is dependent on the integrity of HopZ1a catalytic site, impairs to some degree the ability of the kinase to autophosphorylate. Furthermore, since the levels of phosphorylated MKK7K167R remain unaffected, the inhibitory effect of HopZ1a seems to pivot on the modification of this particular MKK7 residue. More importantly, MKK7 trans-phosphorylation activity on MBP is completely abolished in those samples pretreated with HopZ1a, but remains unaffected in those samples pre-treated with the catalytic mutant HopZ1aC216A, suggesting that acetylation of MKK7 by HopZ1a, which is dependent on the integrity of HopZ1a catalytic site, completely abrogates the ability Chapter 4: HopZ1a acetylates MKK7 to suppress plant defenses 136 of the kinase to phosphorylate proteins in trans, and therefore is likely to block MKK7dependent signal transduction. !, Figure 7. Acetylation of MKK7 by HopZ1a interferes with kinase activity. Recombinant GSTMKK7, or GST-MKK7K167, were incubated in acetylation buffer with either HopZ1a or HopZ1aC216. After 1 hour, components of kinase buffer were added, including 32P-γ-ATP. Samples were separated in a SDS-PAGE and proteins were transferred to a PVDF membrane. The membrane was exposed to an imaging plate for one day. ! ! ! ! ! ! ! Concluding*remarks! !, Concluding remarks 144 !, Concluding remarks 145 The aim of this thesis was the study of different aspects of the plant-pathogen interaction using the model the bacterial pathogen Pseudomonas syringae. To address this point, we have used different experimental and methodological approaches, from cellular to molecular biology, from genetics to biochemistry. At the beginning of this work, the use of cellular biology techniques and fluorescent proteins allowed us to get new insights in the colonization and infection process of P. syringae within the plant. The phenotypic heterogeneity observed by confocal microscopy on the colonization and responses of plant tissues inoculated with a mixed inoculum of isogenic strains differing on their virulence, lead us to investigate the source of this variation within the bacterial population. Using transcriptional fusions to GFP as a reporter for the expression of different elements of the T3SS, and single-cell analysis techniques, we established that expression of HrpL, the transcriptional activator of the T3SS genes, and that of HopAB1, one of its effectors, is bistable within the plant. Although we cannot rule out additional sources of variation, these findings provide a plausible explanation for the phenotypic diversity observed within infected plant tissues. HopAB1 was the fist effector for which a virulence activity was demonstrated, postulated to be suppression of HopF1-triggered immunity (Jackson et al., 1999). Orthologs of HopAB1 in other P. syringae pathovars have been described to suppress and also trigger defense response in different plant species (Jackson et al., 2002; de Torres et al., 2006). For example, HopAB2 from P. syringae pv. tomato DC3000, acts as an E3 ubiquitin ligase activity that promotes degradation of the tomato Fen kinase and suppresses ETI responses (Rosebrock et al., 2007). Previous results from our laboratory have shown that constitutive expression of HopAB1 in Pph 1448a from a plasmid from either a strong (PnptII) or a medium-to-low promoter (PlacZ) does not complement the attenuation of growth within the plant caused by a ΔhopAB1 mutation. In fact, expression of HopAB1 from these promoters attenuated bacterial growth within the plant to a larger degree that its mutation (Macho et al., 2012). However, plasmid expression of HopAB1 from its native promoter did not cause growth attenuation and complemented growth of a ΔhopAB1 mutant. Constitutive expression of the effector could alter secretion hierarchy or promote unspecific activities of the effector within the plant cell, negatively affecting bacterial growth. However, since Agrobacterium-mediated transient expression of HopAB1 in the bean plants of the same cultivar elicits cell death (Vinatzer et al., 2006), its also possible that an excess of HopAB1 translocation within infected tissues could lead to defense activation and growth attenuation of wild type bacteria. Thus, bistability Concluding remarks 146 on the expression and translocation of HopAB1 could be important to maintain the overall translocated protein levels below the threshold for ETI activation.. The second part of this thesis focuses on another effector that can also activate ETI, through indirect recognition by the NLR protein ZAR1, and suppresses ETI activated against other effectors. In this part of the work, we provide a mechanism for HopZ1a-mediated suppression of PTI, ETI and SAR, through the identification of MKK7 as a target for HopZ1a-mediated acetylation. Unpublished data from our laboratory indicate that P. syringae pv. syringae strain 7B40, which contains a hopZ1a gene, can suppress HopZ1a-triggered ETI response through an unknown mechanism (Rufian et al., in preparation-b). Taking into account that HopZ1a suppresses SAR, bistability of the T3SS genes in this strain could provide an interesting mechanism for colonization of distal tissue. Thus, it is tempting to speculate what would happen if bistability did indeed take place in this strain. If this were the case, a bacterium expressing the effector hopZ1a would be detected by the plant cell, which in response would trigger ETI, causing local cell death and activation of SAR that would, limit its proliferation. However, suppression of SAR by HopZ1a could pave the way for bacteria not-expressing the T3SS to colonize distal tissue, in a cooperative manner similar to that described for Salmonella typhimurium subpopulations generated through bistable expression of one of its T3SS the SPI-1. Interestingly, expression of the SPI-1 T3SS and expression of the flagellum is counter-regulated in Salmonella, also similar to the downregulation of motility reported HrpL (Ortiz-Martin et al., 2010a). Following this model, a bacteria subpopulation not expressing the T3SS could move to distal tissue and colonize distal tissues where activation of SAR would have been suppressed by the ONpopulation whose growth had been restricted through HopZ1a-triggered local defenses. In this case, the T3SS ON subpopulation would have an altruistic behavior since its virulence activity would not benefit itself but the OFF-population, as that demonstrated for the subpopulations of Salmonella generated through SPI1 bistability. Finally, plant defense suppression has been extensively shown as an essential process for the pathogen to proliferate and colonize the plant tissue. Effectors alter a number of plant processes to carry out such suppression, and use a wide array of biochemical activities to do so. Thus, effectors are frequently used as molecular probes to deeper our understanding of the plant immune system, as well as to gain insight into how bacteria generate disease. And our results on the role of MKK7 in plant defense, Concluding remarks 147 revealed during the characterization of HopZ1a activity, are proof of this notion. We have demonstrated the existence of phenotypic heterogeneity on the expression of bacterial virulence factors. However, we cannot rule out that the heterogeneity found in the response of the plant to P. syringae could also originate from phenotypic heterogeneity within neighboring plant cells. This heterogeneity would have a different origin, since plant cells are distributed in tissues, communicated and specialized. One example of phenotypic heterogeneity at this side of the interaction is found during callose deposition experiments. Is frequent to find strong differences on callose desposition triggered by the peptide flg22 within the same leaf. Indeed, the edges of the leaves present less callose deposition than the rest of the tissue (Zhou et al., 2012). Furthermore, the ER bodies, structures composed of cisternae derived from the endoplasmic reticulum, containing antimicrobial compounds, present a seemingly random distribution within the leaf (Nakano et al., 2014; Rufian et al., in preparationa). Moreover, Arabidopsis has been shown to undergo epigenetic changes during P. syringae infection, and these changes have been shown to affect the activation of defenses, and do not take place in the plant cells of the infected tissue (Yu et al., 2013). Although highly speculative, clues provide a tantalizing scenario full of possibilities to study the presence of such heterogeneity, its origin and, most importantly, its implication in plant resistance. 148 !! 149 ! ! ! ! ! ! ! Conclusions! 150 !, 151 1. The use of mixed infections of P. syringae strains differently labeled with fluorescent proteins allows following bacterial colonization of the plant apoplast, revealing information such as dynamics of colony establishment or development, or interaction and interferences between different strains, not accessible through the use of individual infections. 2. Close proximity to wild type bacteria is required to complement growth within the apoplast of a non-pathogenic derivative, lacking a functional type III secretion system. 3. The effector-triggered immunity induced by strain RW60, visualized under a confocal microscope as accumulation of autofluorescent phenolic compounds on the cell outline, can restrict growth of closely located wild type bacteria, but can also be suppressed by them, giving rise to a highly heterogeneous scenario. 4. The expression of hrpL, encoding the type III secretion system transcriptional activator,, as well as that of effector gene hopAB1 is bistable both in planta and within inducing laboratory medium, providing a mechanistic explanation for the phenotypic heterogeneity observed by microscopy during bacterial colonization of the plant apoplast. 5. Bacteria sorted on the basis of differences on the expression of effector gene hopAB1 displays differences in virulence. 6. The HrpV/HrpG double negative regulatory loop acts as a bistable switch required for turning the heterogeneous expression of the type III secretion genes into bistability, in a process that requires the transcriptional activator HrpL, and is enhanced through a positive feedback loop mediated by the pilus protein HrpA. 7. The autoacetylation of residue K289 is important but not essential for HopZ1a suppression of effector-triggered immunity and systemic acquired resistance. 8. The HopZ1aK289R mutant induces a strong defense response in both Arabidopsis and Nicotiana benthamiana plants. 9. The plant MAP Kinase Kinase 7 (MKK7) is a positive regulator of the plant immune system, implicated in the activation of pattern-triggered immunity, effector-triggered immunity and systemic acquired resistance. 10. HopZ1a suppresses accumulation of PR1 and growth restriction triggered by the overexpression of MKK7. 11. HopZ1a interacts with MKK7 and acetylates its lysine 167, which is essential for MKK7 kinase activity in vitro and for defense response activation in planta. 152