Department of Biological Sciences, Wye College, University of London, Wye, Ashford, Kent TN25 5AH, UK

Background and objectives
We have examined the responses of Arabidopsis ecotype Columbia to challenge by Pseudomonas syringae pv. tomato DC3000 and P. s. pv. phaseolicola (Pph) race 6 strain (1448A), with or without plasmids harbouring the avirulence genes avrPpiA and avrPphB, which match the RPM1 and RPS5 genes for resistance, respectively. Columbia is susceptible to DC3000, allowing rapid bacterial multiplication. Resistance to 1448A is expressed by restricted bacterial multiplication, but a macroscopic hypersensitive reaction (HR) is not induced, only very patchy necrosis 3 or 4 days after inoculation. However, if the avr genes are expressed in Pph, the HR develops. With avrPpiA interacting with RPM1, the HR is very rapid (10 h a.i.). The reaction controlled by avrPphB and RPS5 is less dramatic, the HR being apparent only 24 h a.i. [1]. An additional strain studied was an hrpL mutant of 1448A which failed to multiply or cause any necrosis. Thus in this set of interactions we were able to examine responses ranging from symptomless resistance to full compatibility. Each of the interactions was examined by electron microscopy. In addition to conventional staining we used immunocytochemistry to localize callose deposition and cerium chloride (CeCI3) staining to detect H2O2 [2].

Results and conclusions
The hrp mutant induced rapid localized changes in the plant cell wall adjacent to attached bacteria; an increase in the electron density of the wall was particularly striking. Various forms of paramural papillae were also produced and these often contained osmiophilic material within a matrix of callose. Wild-type Pph 1448A induced similar responses but in addition, at some sites, plant cells also underwent cytoplasmic collapse. Both papilla deposition and plant cell death were observed in response to 1448A(avrPphB). By contrast, with 1448A(avrPpiA), membrane damage and extensive cytoplasmic vesiculation indicative of the onset of the HR occurred before major deposition of callose. During the early stages of HR, no features characteristic of apoptosis such as nuclear fragmentation were observed, but shrinkage of cytoplasm did appear later leading to a characteristic 'corpse morphology'. Cells of DC3000 multiplied rapidly, initially without causing obvious alterations to the ultrastructure of cells. With time, the plant cell cytoplasm became disorganized. Dead cells were morphologically different in resistant and susceptible reactions.

Development of the HR was associated with a burst of H2O2 within the cell wall at reaction sites. The intensity of staining was greatest after inoculation with 1448A(avrPpiA). Significant, but less marked accumulation of H2O2 was also observed irrespective of the strain of Pph used. Staining was reduced more by inhibitors of peroxidase than of NADPH oxidase. Only traces of cerium perhydroxides were seen during the compatible interaction with DC3000.

Our results emphasize the highly localized responses that occur in challenged plant cells, even though the final result may be tissue collapse, as observed with the HR. The timing of the accumulation of H2O2 was dependent on the gene-for-gene interaction, the response controlled by RPM1 being particularly rapid. How the recognition processes operating between bacteria and plants lead to such localized reactions remains to be determined.

1. Simonich MT, lnnes RW, 1995. Molecular Plant-Microbe Interactions 8, 637-640.
2. Bestwick CS, Brown IR, Bennett MH, Mansfield JW, 1997. Plant Cell 9, 209-221.