Microbiology Department, University of Tennessee, Knoxville, TN 37996, USA

Background and objectives
Most pathovars of Pseudomonas syringae produce toxins, many of which have been proposed to have a role in pathogenicity. There is now some information about the genes responsible for toxin biosynthesis, but little is known of the genes and the mechanisms involved in the secretion of these structurally varied secondary metabolites. Isolates of P. syringae pv. syringae typically produce one of three structurally related phytotoxins, syringomycin (SR), syringostatin (SS), or syringotoxin (ST). Previous work established that the syrD gene from an SR-producer strain encodes a protein that is a member of the ATP-binding cassette (ABC) family of transport proteins, suggesting a role for SyrD in the secretion of SR [1]. Furthermore, DNA hybridization studies showed that all toxigenic strains of P. s. pv. syringae carry a syrD gene [2]. The broad goal of our current research is to compare functionally the SyrD proteins from strains that produce different toxins, to evaluate the degree of target specificity that each exhibits.

Materials and methods
The P. s. pv. syringae strains used in this study were (toxin type indicated in parentheses): B301 D (SR), SY1 2 (SS), B457 (ST), and BR346 (non-toxigenic B301 D mutant). Two plasmids were constructed (pNQ 188 and pNQ201) which carry different syrD genes (from strains SY12 and B457, respectively) inserted into the broad-host-range, mobilizable vector pLAFR3. Bacterial matings (conjugations) and in vitro toxin assays were performed as described previously [1, 2].

Results and conclusions
Plasmids pNQ188 and pNQ201 were conjugated from E. coli donor strains to strain BR346. Recombinant BR346 exconjugants were then screened for toxin production. In each case, the heterologous syrD homologue complemented the defective syrD gene in the recipient strain, resulting in restoration of toxin production. This complementation study demonstrated that the SyrD protein of one toxin producer strain had sufficiently relaxed target specificity to allow it to function properly in the secretion of another toxin when expressed in other cells. This result was anticipated because the three syrD genes involved in this study have been sequenced, and were known to share more than 95% sequence identity [NB Quigley, unpublished data]. There are significant differences in the amount of toxin produced by individual strains of P. s. pv. syringae in vitro, but the reason for these differences is unknown. For example, strain B457 produces much more toxin, and strain SY12 produces much less toxin, than strain B301 D [2]. When evaluating the toxigenicity of exconjugants from the above matings, it was apparent immediately that the amount of toxin produced by the recombinant strains matched that of the wild-type strain from which the syrD gene had been cloned, and not that of B301 D, the wild-type parent of strain BR346. Thus only a small amount of toxin was produced by strain BR346 when it carried plasmid pNQ1 88, but a large amount of toxin was produced when it carried plasmid pNQ201. For comparative purposes, toxin production by strains B301 D, SY12 and B457 was evaluated on the same plates. The simplest interpretation of this result is that the three SyrD proteins exhibit different toxin secretion capacities. Therefore, we propose that this simple experiment has revealed a fundamental property of toxin production among strains of P. s. pv. syringae, which is that a strain's toxin-production capacity is limited by the efficiency of its toxin-secretion mechanism. This would account for the range of toxin-production capacities of various strains reported earlier [2], and could have an important impact on the study of other secondary metabolises, including certain clinically important antibiotics, that are secreted by ABC transporters related to SyrD.

1. Quigley NB, Mo Y-Y, Gross DC, 1993. Molecular Microbiology 9, 787-801.
2. Quigley NB, Gross DC, 1994. Molecular Plant-Microbe Interactions 7, 78-90.