1.6.13
SECRETION AND SIGNAL DELIVERY TO PLANT CELLS OF PSEUDOMONAS SYRINGAE AVR PROTEINS BY A CLONED ERWINIA CHRYSANTHEMI HRP (TYPE III) SECRETION SYSTEM IN ESCHERICHIA COLI AND FURTHER FUNCTIONAL ANALYSIS OF HARPINECH

JH HAM, DW BAUER and A COLLMER

Department of Plant Pathology, Cornell University, lthaca, NY 14853-4203, USA

Background and objectives
Many Gram-negative bacterial pathogens of plants are equipped with the type III (Hrp) protein-secretion system, which allows a variety of proteins to traffic through bacterial cell membranes and the plant cell wall [1]. This Hrp secretion system is responsible both for pathogenicity in hosts and for hypersensitive response (HR) elicitation in non-hosts. In Pseudomonas syringae and Erwinia amylovora, the proteins trafficking the Hrp system have been delineated into three groups, harpins, the HrpA pilin, and Avr proteins. Harpins and HrpA pilin are encoded by two of the clustered hrp genes and can be secreted into the medium in an Hrp-dependent manner. In contrast, Avr proteins, which are thought to contribute collectively to bacterial pathogenicity in hosts lacking a cognate R gene and are responsible individually for the R-gene-specific HR, appear not to be secreted into the medium but to be delivered directly into plant cells in an Hrp-dependent manner [1]. Harpins (encoded by hrpN in E. amylovora and by hrpZ in P. syringae) have been shown to elicit the HR when infiltrated into tobacco leaves. However, their actual function is unclear except for some evidence of an interaction with plant cell walls. The HrpA pilus, encoded by hrpA, is essential for Avr signal delivery into plants [1]. Erwinia chrysanthemi, a necrotrophic broad-host-range pathogen, also has an hrp cluster, which is responsible for HR elicitation on tobacco leaves even though its virulence depends primarily on plant cell-wall-degrading enzymes such as pectate lyases. E. chrysanthemi hrp mutants are reduced in infectivity in host plants and unable to elicit the HR in tobacco leaves. This study has been done with the following objectives: (i) isolation and functional analysis of the E. chrysanthemi hrp cluster; (ii) comparison of the hrp genes of E. chrysanthemi, a host-promiscuous pathogen, with those of narrow-host-range pathogens such as P. syringae or E. amylovora; and (iii) investigation of the role of harpinech and its unusual features.

Results and conclusions
Three cosmid clones (pCPP2156, pCPP2157 and pCPP2158) were isolated from an E. chrysanthemi cosmid library by probing with E. amylovora hrp genes [2]. Unlike pCPP430 and pHRII, which carry hrp genes from E. amylovora and P. s. pv. syringae 61, respectively, none of these cosmids enabled E. coli to elicit an HR in tobacco leaves. This could result either from an incomplete hrp cluster, or from the absence of an Avr signal to interact with a cognate R gene product in tobacco. avrB from P. s. pv. glycinea, fused with the FLAG epitope (avrB-FLAG), was transformed into E. coli carrying each of the three cosmid clones, and the transformants were infiltrated into Nicotiana clevelandii leaves which react with the AvrB signal. Interestingly, two cosmid clones (pCPP2156 and pCPP2158) could make E. coli elicit an HR in N. clevelandii leaves. Further analysis by DNA sequencing and restriction mapping revealed that pCPP2157 lacks the C-terminal part of hrcU, which is at one border of the hrp cluster. Tn5Cm mutagenesis of pCPP2156 in the hrpJ operon enabled proof of the Hrp-dependent nature of AvrB signal delivery by pCPP2156. Surprisingly, pCPP2156 could secrete AvrB into the medium, and this secretion ability was also Hrp-dependent. Additional secretion tests with AvrPto, an Avr protein from P. s. pv. tomato, indicates that E. chrysanthemi hrp genes can secrete into the medium multiple Pseudomonas Avr proteins.

This is the first report of the secretion of known Avr proteins by bacteria. We are now making a mutation in hrpNEch, which encodes HarpinEch. This will enable us to examine the role in Avr secretion and signal delivery of HarpinEch, which is unusual among harpins in being extracellular (in an Hrp-dependent manner) but not detectable in cell-free medium.

References
1. Alfano JR, Collmer A, 1997. Journal of Bacteriology 179, 5655-5662.
2. Bauer DW, Wei ZM, Beer SV, Collmer A, 1995. Molecular Plant-Microbe Interactions 8, 484-491.