Citrus exocortis viroid (CEVd) and tomato mosaic virus (ToMV), which produce a systemic non-necrotizing infection in tomato (Lycopersicon esculentum cv. Rutgers), strongly induced the accumulation of a phenolic compound that we have characterized as 2,5-dihydroxybenzoic acid (gentisic acid, GA) by nuclear magnetic resonance, following purification by high-performance liquid chromatography. Levels of free and total GA increased more than 150-fold in response to CEVd and ToMV infections. Unlike these non-necrotizing infections, the necrotizing reaction elicited by Pseudomonas syringae pv. syringae in this host did not produce any accumulation of GA. It is also shown that, in healthy leaf tissues, benzoic acid (BA) and salicylic acid (SA) were rapidly converted to GA, SA being the immediate precursor of GA, according to radiolabeling studies. Interestingly, exogenous GA elicited accumulation of the previously described CEVd-induced antifungal pathogenesis-related (PR) proteins P23, P32, and P34. These proteins were not induced by exogenous SA, which is able to elicit other CEVd-induced PR proteins in tomato. These results suggest that GA acts as a pathogeninduced signal, additional to SA, for activation of plant defense genes in tomato.

In the present work we have studied the accumulation of gentisic acid (2,5-dihydroxybenzoic acid, a metabolic derivative of salicylic acid, SA) in the plant-pathogen systems, Cucumis sativus and Gynura aurantiaca, infected with either prunus necrotic ringspot virus (PNRSV) or the exocortis viroid (CEVd), respectively. Both pathogens produced systemic infections and accumulated large amounts of the intermediary signal molecule gentisic acid as ascertained by electrospray ionization mass spectrometry (ESI-MS) coupled on line with high performance liquid chromatography (HPLC). The compound was found mostly in a conjugated (beta-glucoside) form. Gentisic acid has also been found to accumulate (although at lower levels) in cucumber inoculated with low doses of Pseudomonas syringae pv. tomato, producing a nonnecrotic reaction. In contrast, when cucumber was inoculated with high doses of this pathogen, a hypersensitive reaction occurred, but no gentisic-acid signal was induced. This is consistent with our results supporting the idea that gentisic-acid signaling may be restricted to nonnecrotizing reactions of the host plant (Bellés et al. in Mol Plant-Microbe Interact 12:227-235, 1999). In cucumber and Gynura plants, the activity of gentisic acid as inducing signal was different to that of SA, thus confirming the data found for tomato. Exogenously supplied gentisic acid was able to induce peroxidase activity in both Gynura and cucumber plants in a similar way as SA or pathogens. However, gentisic-acid treatments strongly induced polyphenol oxidase activity in cucumber, whereas pathogen infection or SA treatment resulted in a lower induction of this enzyme. Nevertheless, gentisic acid did not induce other defensive proteins which are induced by SA in these plants. This indicates that gentisic acid could act as an additional signal to SA for the activation of plant defenses in cucumber and Gynura plants.

No abstract