The Sainsbury Laboratory, Norwich Research Park, Colney, Norwich NR4 7UH, UK

Background and objectives
We are interested in the interaction of bacterial lipopolysaccharides (LPS) with plants and the role that such interactions may play in bacterial phytopathogenesis. As a major constituent of the outer membrane, LPS is thought to contribute to bacterial survival in planta by acting as a barrier to diffusion into the bacterial cell of toxic plant components. However other roles of LPS in bacterial-plant interactions have also been suggested. LPS can promote symbiosis in Rhizobium-legume interactions. In contrast, there are a number of reports that LPS from plant pathogens can trigger defence related responses in plants. Some of the earliest work in this area showed that inoculation of tobacco plants with LPS or LPS-protein complexes from a number of bacteria, including non-pathogens, can prevent the hypersensitive response (HR) caused by subsequent infection by plant pathogenic bacteria in incompatible plant/pathogen interactions and reduce symptoms in compatible interactions [1]. This response to LPS is a light-independent process, usually localised to the site of inoculation, requiring several hours to become established suggesting that the protective mechanism depends on a plant response to LPS. We have studied this localised induced response (LIR) in pepper in an attempt to understand its molecular basis.

Results and conclusions
LPS both from pathovars of Xanthomonas campestris (Xc) and from enteric bacteria (Salmonella and E. coli) can elicit LIR. The levels of mRNA for several plant defence-related or stress-induced genes were estimated in LPS pre-treated and untreated (control) pepper leaves after inoculation with avirulent [X.c. pv. campestris (Xcc)] and avirulent X.c. pv. vesicatoria strain 75.21 (Xcv). Pre-treatment with both types of LPS modified the patterns of gene expression seen in response to subsequent bacterial inoculation, while LPS alone did not induce (enteric bacteria) or was only a weak inducer (Xanthomonas) of these genes.

Levels of salicylic acid (SA), an effector of defence-gene activation, were measured after inoculation of Xcc (avirulent) or Xcv (avirulent) in LPS pre-treated and control pepper tissue. Inoculation of pepper leaves with Xcv, water or LPS alone did not lead to an increase in the levels of SA. Inoculation of Xcc into healthy tissue caused a marked increase in the levels of SA at 24 h after inoculation. Pretreatment of the pepper tissue with LPS changed the amount of SA accumulating in response to subsequent Xcc inoculations; levels of SA in response to Xcc were five-fold lower at 24 h after bacterial inoculation.

In the course of measuring SA accumulation, changes in other phenolics were detected. High levels of feruloyltyramine and coumaroyltyramine were revealed in LPS pre-treated tissue 1-2 h after inoculation of Xcc, but not in response to LPS alone and not until 24 h after inoculation of Xcc in the untreated tissue (control). Neither of these compounds were detected in the control (untreated) plants, or in plants inoculated with Xcv. However the conjugates were detectable (at low levels) in LPS pre-treated tissue in response to Xcv 4 h after inoculation of the bacteria. We have synthesised the compounds and so far shown that feruloyltyramine inhibits growth of Xcc in liquid culture. We are continuing to analyse these tyramine conjugates looking at levels of wall-bound feruloyltyramine and coumaroyltyramine in LIR exhibiting tissue compared to levels in control tissue. The antimicrobial activity of the tyramine conjugates and the behaviour of bacteria in planta suggest that prevention of HR results from inhibition of bacterial growth and hence inhibition of the generation of the HR-inducing factor(s)in plants.

1. Sequeira L, 1983. Annual Review of Microbiology 37, 51-79.