AvrPto from the Pseudomonas cell. To examine this, avrPto deletions C25, C41, and C74 were placed into pBI121 and tested with the Agrobacte- rium transient assay. Agrobacterium EHA105 con- taining avrPto induced an HR in 2 days, whereas EHA105 containing the avrPto deletion C25 in- duced an HR after 4 days; the other deletions did not elicit an HR ( X. Tang and G. Martin, unpublished results). This suggests that the carboxyl terminal 25 amino acids of AvrPto are not required for secretion from the bacterial cell; this portion of AvrPto may serve as an activation domain, interact with other components in the signaling pathway, or have a role in AvrPto stability. 28. A. Ellingboe, in Active Defense Mechanisms in Plants, R. Wood, Ed. (Plenum, New York, 1980), pp. 179 –192. 29. I. Brown, J. Mansfield, I. Irlam, J. Conrads-Strauch, U. Bonas, Mol. Plant-Microbe Interact. 6, 376 (1993). S. Young, F. White, C. Hopkins, J. Leach, ibid. 7, 799 (1994). 30. L. Rahme et al., Science 268, 1899 (1995). 31. E. Galyov, S. Hakansson, A. Forsberg, H. Wolf- Watz, Nature 361, 730 (1993); J. Bliska and S. Falkow, Trends Genet. 9, 85 (1993). 32. Levels of protein expression were determined with the use of antibody to LexA (a gift from E. Golemis) for LexA fusion proteins (in pEG202) and antibody to the hemagglutinin (HA) epitope tag (Boehringer-Man- heim) for the AvrPto::HA fusion protein (in pJG4-5). 33. We thank L. Dunkle, A. Friedman, S. Gelvin, and K. Perry for helpful comments on the manuscript. Sup- ported, in part, by a Purdue Research Foundation Doctoral Fellowship (D.H.), National Science Foun- dation grant MCB-93-03359 (G.B.M.) and a David and Lucile Packard Foundation Fellowship (G.B.M.). 8 August 1996; accepted 22 November 1996 Molecular Basis of Gene-for-Gene Specificity in Bacterial Speck Disease of Tomato Steven R. Scofield,* Christian M. Tobias,* John P. Rathjen, Jeff H. Chang, Daniel T. Lavelle, Richard W. Michelmore, Brian J. Staskawicz† Transient expression of the Pseudomonas syringae avirulence gene avrPto in plant cells resulted in a Pto-dependent necrosis. The AvrPto avirulence protein was observed to interact directly with the Pto resistance protein in the yeast two-hybrid system. Mutations in the Pto and avrPto genes which reduce in vivo activity had parallel effects on asso- ciation in the two-hybrid assay. These data suggest that during infection the pathogen delivers AvrPto into the plant host cell and that resistance is specified by direct interaction of Pto with AvrPto. In plants, resistance to a variety of patho- gens is determined by the action of comple- mentary pairs of resistance (R) genes in the host and avirulence (avr) genes in the patho- gen. These gene-for-gene interactions have been observed between plants and a diverse array of pathogens, including viruses, bacte- ria, fungi, nematodes, and insects (1). From genetic analysis it has been proposed that R genes recognize an elicitor produced directly or indirectly by the pathogen’s avr gene, which leads to a resistance response in the infected plant (2). Bacterial speck disease of tomato is caused by Pseudomonas syringae pv. tomato (Pst). In tomato, resistance to strains of Pst that contain the avr gene avrPto is con- ferred by the Pto gene (3). The Pto locus encodes a family of related serine-threonine kinases. Among these, Fen is active in a parallel pathway that confers sensitivity to the insecticide fenthion (4). R and avr proteins may interact directly, thereby activating plant defenses. For the protein product of Pto, this binding would likely occur intracellularly because of its predicted cytoplasmic localization. The ac- tivity of many avr genes, including avrPto, depends on an hrp secretion pathway which is similar to the type III secretory systems of Yersinia, Shigella, and Salmonella (5). These pathogens translocate a set of virulence pro- teins into host cells. Therefore, we consid- ered the possibility that the bacterial AvrPto protein moves across the plant cell wall and plasma membrane where it directly interacts with the tomato Pto protein. Evidence indicating that AvrPto acts in- side the plant cell was obtained by tran- siently expressing avrPto in transgenic Nico- tiana benthamiana plants transformed with Pto (6) (Fig. 1). This resulted in necrosis similar to the Pto-mediated HR elicited by P. syringae expressing avrPto and indicated that AvrPto was active within the plant cell. Deletion of 30 amino acids from the COOH-terminus of AvrPto did not elimi- nate this activity, whereas deletion of 59 amino acids destroyed activity. Activity of the deletion derivatives in the transient expression assays correlated with biological activity in P. syringae. These results suggest- ed that the products of the avr and R genes may interact directly. We employed the yeast two-hybrid sys- tem to directly test this hypothesis (7). Pto, Fen, and avrPto coding sequences were ex- pressed as fusions to GAL4 DNA binding (BD) and transcriptional activating (AD) domains. Reciprocal combinations of BD and AD fusions were tested for -galactosi- dase reporter gene activity in yeast. Inter- action was only observed when the BD::Pto and AD::AvrPto fusions were coexpressed (Fig. 2). Controls did not show any inter- action. Furthermore, no interaction was de- tected between BD::Fen and AD::AvrPto (Figs. 2 and 3A). To test the biological relevance of the interaction, inactive alleles of Pto and avrPto were tested. Three inactive Pto al- leles, pto6, pto7 and pto11, were previously identified through mutagenesis of resistant tomato plants (8). Sequence analysis re- vealed single amino acid changes in each mutant allele (Fig. 3B). The mutant Pto sequences showed no detectable interaction with AvrPto in yeast. Thus, mutant alleles that confer susceptibility to Pst also fail to interact with AvrPto in the two-hybrid sys- tem. The two deletions of AvrPto tested in S. R. Scofield, J. P. Rathjen, J. H. Chang, D. T. Lavelle, NSF Center for Engineering Plants for Resistance Against Pathogens (CEPRAP), University of California, Davis, CA 95616. C. M. Tobias, Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720. R. W. Michelmore, NSF Center for Engineering Plants for Resistance Against Pathogens and Department of Veg- etable Crops, University of California, Davis, CA 95616. B. J. Staskawicz, NSF Center for Engineering Plants for Resistance Against Pathogens and Department of Plant and Microbial Biology, University of California, Davis, CA 94720. * These authors made equal contributions to this report. †To whom correspondence should be addressed. E- mail: stask@garnet.berkeley.edu A B Fig. 1. Agrobacterium tumefaciens strain A281 (sites 3 to 6) or the nontumorogenic strain C58C1 (sites 1 and 2) containing avrPto gene constructs (10) were infiltrated into leaves of a Pto trans- formed plant (A) or leaves of untransformed N. benthamiana (B). Expression of avrPto (sites 1 and 4), and avrPto deletions of 30 (site 6), and 59 (site 3) amino acids off the COOH-terminus was controlled by the CaMV 35S viral promoter in a binary plan transformation vector. An avrPto construct (pPtE6) lacking left and right T-DNA borders or a plant promoter (11) was also infil- trated (sites 2 and 5). Leaves were photographed 3 days after inoculation. REPORTS SCIENCE  VOL. 274  20 DECEMBER 1996 2063 on January 13, 2017 http://science.sciencemag.org/ Downloaded from