A high throughput near saturating screen for type III effector genes from Pseudomonas syringae Jeff H Chang Jonathan M Urbach Terry F Law Larry W Arnold An Hu Saurabh Gombar Sarah R Grant Frederick M Ausubel and Jeffery L Dangl Departments of Biology CB 3280 and Microbiology and Immunology CB 7290 and Curriculum in Genetics Carolina Center for Genome Sciences University of North Carolina Chapel Hill NC 27599 Department of Molecular Biology Massachusetts General Hospital Cambridge MA 02139 and Syngenta Biotechnology Inc P O Box 12257 3054 Cornwallis Road Research Triangle Park NC 27709 Pseudomonas syringae strains deliver variable numbers of type III effector proteins into plant cells during infection These proteins are required for virulence because strains incapable of delivering them are nonpathogenic We implemented a whole genome highthroughput screen for identifying P syringae type III effector genes The screen relied on FACS and an arabinose inducible hrpL factor to automate the identification and cloning of HrpLregulated genes We determined whether candidate genes encode type III effector proteins by creating and testing full length protein fusions to a reporter called 79AvrRpt2 that when fused to known type III effector proteins is translocated and elicits a hypersensitive response in leaves of Arabidopsis thaliana expressing the RPS2 plant disease resistance protein 79AvrRpt2 is thus a marker for type III secretion system dependent translocation the most critical criterion for defining type III effector proteins We describe our screen and the collection of type III effector proteins from two pathovars of P syringae This stringent functional criteria defined 29 type III proteins from P syringae pv tomato and 19 from P syringae pv phaseolicola race 6 Our data provide full functional annotation of the hrpL dependent type III effector suites from two sequenced P syringae pathovars and show that type III effector protein suites are highly variable in this pathogen presumably reflecting the evolutionary selection imposed by the various host plants host microbe interaction plant pathogenesis Arabidopsis FACS T he commonly studied plant bacterial pathogen Pseudomonas syringae is subdivided into pathovars based on their ability to cause disease on one or more distinct host species During infection P syringae and other Gram negative pathogens deliver type III effector proteins via a type III secretion system TTSS hrp hrc genes in P syringae ref 1 from the bacterium into host cells 2 P syringae strains incapable of delivering type III effectors are nonpathogenic 3 Thus the type III effectors each strain delivers are required for pathogenicity In contrast if just one of the type III effectors is recognized by the plant immune system s surveillance machinery disease resistance R proteins a battery of host responses is triggered including localized programmed cell death termed the hypersensitive response HR 4 In this case the pathogen is rendered avirulent its multiplication is limited and it does not cause disease As a consequence some type III effector genes have been functionally defined as avirulence avr genes Recognition of type III effector proteins by corresponding R proteins may therefore limit the particular host range of individual P syringae pathovars P syringae type III effector genes share several characteristics Their expression is coordinately regulated with the TTSS encoding genes by the alternative factor HrpL 5 Genes encoding both the TTSS and type III effector proteins also share a cis element hrp box in their promoters 6 Finally delivery of type III effector proteins into the host cell depends on the TTSS and a loosely defined N terminal signal in the type III effector protein that includes 10 serine or proline in the first 50 aa an aliphatic www pnas org cgi doi 10 1073 pnas 0409660102 amino acid or proline at position 3 or 4 and the absence of negatively charged amino acids in the first 12 residues 7 9 Nonsaturating genetic screens have relied on these characteristics Guttman et al 9 identified 15 protein fusions defining genes whose N termini were sufficient for TTSS dependent translocation These included 12 previously uncharacterized type III effector proteins Boch et al 10 identified genes that were induced by in planta conditions and regulated by HrpL in vitro Six of these encoded previously identified type III effector proteins Bioinformatic approaches relying on the predicted shared characteristics of type III effector genes have also been used to identify candidate type III effector genes The genomes of three P syringae pathovars tomato race DC3000 Pto www pseudomonas syringae org phaseolicola race 6 1448a Pph6 www pseudomonas syringae org and syringae race B728a Psy www jgi doe gov index html have been sequenced 11 Genes encoding putative type III effector proteins and helpers referred to as hop hrp dependent outer protein genes were identified in these genomes by the presence of putative hrp boxes 12 13 and ORFs with N terminal amino acid compositions consistent with known Hops 8 9 14 15 Helper proteins are secreted from the bacterium via the TTSS but likely do not function within the host cell Their presumptive roles are to assist the TTSS in delivery of type III effector proteins 15 From these efforts Pto was estimated to deliver 40 different Hops 7 10 12 15 Far fewer Hops were predicted in Psy B728a 14 Thirty Hops were predicted from Pph6 based on homology to known and predicted Hops from other strains 16 The exact number of type III effector proteins based on these predictions has not been experimentally validated We modified a technique termed differential fluorescence induction DFI 17 for high throughput discovery of HrpLregulated genes from pathovars of P syringae Genomic libraries of P syringae were screened for clones expressing GFP in a HrpLdependent manner by using FACS Identified clones were sequenced and assembled Resulting contiguous DNA sequences contigs were examined for the characteristics of type III effector genes mentioned above We amplified and cloned full length candidate genes in frame to the coding region of the C terminal 177 aa of the type III effector protein AvrRpt2 from which the N terminal 79 aa are deleted 79AvrRpt2 refs 9 18 and 19 Cells expressing these fusion proteins were infiltrated into plants expressing RPS2 the R protein that recognizes AvrRpt2 to determine which of the candidate genes encode proteins that can