Investigation of the Mechanism of Action of a Chlorosis-Inducing Toxin Produced by Pseudomonas phaseolicola

A toxin that induced chlorotic haloes (typifying haloblight disease) on primary leaves of Phaseolus vulgaris L. (var. Canadian Wonder) was partially purified from culture filtrates of the causative agent Pseudomonas phaseolicola (Burkh.) Dowson. This material was used to investigate chlorosis induction. Haloes could only be induced in those bean leaves that were expanding and synthesizing chlorophyll (Chl); the toxin, therefore, does not promote Chl breakdown. Chl, carotene, and xanthophyll synthesis were inhibited in sections of greening barley (Hordeum vulgare L.) leaves, irrespective of the irradiance level. In parallel experiments, the toxin decreased the level of 5-aminolevulinic acid by amounts sufficient to account for toxin-inhibition of Chl synthesis. Electron microscopy revealed no difference between the transformation of etioplasts into chloroplasts in toxin-treated and control tissue, despite a 60% reduction in Chl in the former. The incorporation of [¹⁴C]acetate into lipid by greening barley leaf sections and by isolated Pisum sativum chloroplasts in the light and the dark was inhibited about 60% by the toxin. The distribution of radioactivity among the spectra of acyl residues was the same in the control and toxin-treated material. It is suggested that the toxin interferes with an early process common to the synthesis of different lipids, including Chl.     

The Release of Alginate Lyase from Growing Pseudomonas syringae pathovar phaseolicola

Pseudomonas syringae pathovar phaseolicola, which produces alginate during stationary growth phase, displayed elevated extracellular alginate lyase activity during both mid-exponential and late-stationary growth phases of batch growth. Intracellular activity remained below 22% of the total activity during exponential growth, suggesting that alginate lyase has an extracellular function for this organism. Extracellular enzyme activity in continuous cultures, grown in either nutrient broth or glucose–simple salts medium, peaked at 60% of the washout rate, although nutrient broth-grown cultures displayed more than twice the activity per gram of cell mass. These results imply that growth rate, nutritional composition, or both initiate a release of alginate lyase from viable P. syringae pv. phaseolicola, which could modify its entrapping biofilm. Received: 14 April 2000 / Accepted: 11 August 2000     

In vitro Production of Halo Blight Inducing Toxin by Pseudomonas phaseolicola (Burkh.) Dowson

Three pathogenic strains of Pseudomonas phaseolicola (strain 1 and 3 virulent and strain 5 weakly virulent) were tested for their toxic activity. All three strains produced detectable amounts of toxin in vitro. Cultural conditions and length of incubation greatly influenced toxin production. Maximum amount of toxin was produced at 20°C and pH 6.5. Glycerol served as the best carbon source and 1-cysteine as the best amino acid for toxin production.     

Mode of Action of the Toxin from Pseudomonas phaseolicola: I. Toxin Specificity, Chlorosis, and Ornithine Accumulation

The specificity of the Pseudomonas phaseolicola toxin for enzyme inhibition and its relationship to toxin-induced chlorosis in bean leaves (Phaseolus vulgaris L.) was examined. The toxin showed no significant inhibitory activity against glutamine synthetase, glutamine transferase, carbamyl phosphate synthetase, aspartate carbamoyltransferase, or arginase at concentrations 100-fold higher than that needed to inhibit ornithine carbamoyltransferase by 50%.     

Relationship between peroxidase activity and the amount of fully N-methylated compounds in bean plants infected by Pseudomonas savastanoi pv. phaseolicola

Changes in the level of endogenous formaldehyde (HCHO), some N-methylated compounds (choline and trigonelline) and peroxidase activity were examined in the leaves of bean genotypes (Phaseolus vulgaris L.) with different disease-sensitivity during ontogenesis in the stressfree condition and after natural infection by Pseudomonas savastanoi pv. phaseolicola (until the appearance of lesions). HCHO, as its dimedone adduct, and fully N-methylated compounds were determined by overpressured layer chromatography (OPLC) in different developmental stages and in the infected leaves/leaf discs. Peroxidase activity was measured by a spectrophotometric method. HCHO level decreased with ageing of the primary leaf and accordingly in the leaves at different developmental stages, then increased again in both cases due to the demethylation and methylation processes. Concentration of choline and trigonelline as potential HCHO generators decreased considerably while peroxidase activity increased with ageing of the plants. Comparing the symptomless and the Pseudomonas infected leaf discs (with watersoaked lesions) we found a decrease in the level of HCHO, choline and trigonelline and there was detectable increase in the peroxidase activity in the infected leaf tissues. Our findings are in accordance with previously published results that peroxidases play an important role in oxidative demethylation processes. Our hypothesis is that the high level of HCHO in the old leaves can originate from methylated components as the result of peroxidase activity and this high level may lead to the old leaf being resistant to pathogen. This conclusion is supported by the fact that the leaves of susceptible bean genotypes became resistant to Pseudomonas while growing older.     

Gene-for-gene interactions between Pseudomonas syringae pv. phaseolicola and Phaseolus.

The gene for cultivar-specific avirulence to Phaseolus vulgaris cv. Tendergreen in races 3 and 4 of Pseudomonas syringae pv. phaseolicola was isolated and sequenced. Genomic clones from libraries of race 3 in pLAFR1 and race 4 in pLAFR3, which altered the phenotype of race 5 from virulent to avirulent in Tendergreen, were found to possess a common approximately 15-kb region of DNA that contained the determinant of avirulence. Subcloning and insertion mutagenesis with Tn1000 located an avirulence gene within a 1.4-kb BglII/HindIII DNA fragment in races 3 and 4. Comparison of the nucleotide sequences of regions of DNA that confer avirulence confirmed that both races have an identical gene for avirulence (designated avrPph3) comprising 801 base pairs (bp) and predicted to encode a cytoplasmic protein of 28,703 Da. A sequence, TGCAACCGAAT, 91% homologous to the motif found in promoter regions of avrB and avrD from P. s. pv. glycinea was located 89-99 bp upstream of the start of the open-reading frame 1. Hybridization experiments showed that avrPph3 was not plasmid-borne and was absent from isolates of P. s. pv. phaseolicola races 1, 2, 5, 6, 7, and 8, P. cichorii, P. s. pvs. coronafaciens, glycinea, maculicola, pisi, syringae, and tabaci. Cosegregation studies of crosses between cultivars resistant (Tendergreen) and susceptible (Canadian Wonder) to races 3 and 4 and transconjugants of race 5 confirmed that a gene-for-gene relationship controls specificity in the interaction between Tendergreen and races 3 and 4 of P. s. pv. phaseolicola.     

Genetic transformation of Pseudomonas phaseolicola by R-factor DNA.

Summary The bacterium Pseudomonas phaseolicola was successfully transformed, for the first time, with R-factors RSF1010 and pBR322 DNA by a calciumshock and heat-pulse technique. Frequency of transformation for RSF1010 ranged from 0.8×10^-7 to 3.1×10^-6 and was ca. 0.4×10^-8 for pBR322.     

Mode of Action of the Toxin from Pseudomonas phaseolicola: II. Mechanism of Inhibition of Bean Ornithine Carbamoyltransferase

A chlorosis-inducing toxin of Pseudomonas phaseolicola was examined for inhibition of ornithine carbamoyltransferease prepared from acetone powder of bean (Phaseolus vulgaris L.) plants. The enzyme has a pH optimum at 8.5, involves a ternary complex reaction mechanism, and shows Michaelis constants of 5.0 mm and 1.7 mm for ornithine and carbamoylphosphate, respectively. Assuming reversible catalysis, Michaelas constants of 11 mm and 3.3 mm are calculated for citrulline and arsenate. Toxin induces allosteric competitive inhibition in relation to carbamoylphosphate and a noncompetitive mode of inhibition in relation to ornithine, except at high toxin concentrations where uncompetitive inhibition is observed. In the backward assay, competitive inhibition is observed for both arsenate and citrulline. Inhibition is increased with preincubation time and shows saturation kinetics with regard to toxin concentration.     

A biochemical difference between healthy bean leaves resistant and susceptible to the halo-blight disease caused by Pseudomonas phaseolicola

Substrate for an endogenous oxidation in homogenates of leaves of bean (Phaseolus vulgaris) was traced to two fractions of lipids, each representing less than 2% of the dry weight of leaves. The substrate lipids, tentatively identified as galactosyl diglycerides, yielded linolenic and linoleic acids on acid hydrolysis. Amounts of linolenic acid in total lipids in resistant and susceptible leaves were similar. Amounts of free linolenic acid in resistant leaves increased eightfold to 408.6 μg and in susceptible leaves fourfold to 130.6 μg/g fresh leaf after homogenization and incubation for 16 min at 4 °C. These quantities are sufficient to account during their lipoxidation for the previously reported oxygen uptakes in homogenates. Differences between resistant and susceptible leaves were traced to the activities of lipase systems which liberate linolenic acid from substrate lipids.     

Assay and Relationship of HT-2 Toxin and T-2 Toxin Formation in Liquid Culture

Both T-2 toxin and HT-2 toxin can be conveniently quantitated in crude extracts by using a combination of thin-layer chromatography and fluorodensitometry. This technique was used to follow the production of these toxins by liquid cultures of Fusarium poae (NRRL 3287). T-2 toxin was produced prior to HT-2 toxin and hexadeuterio-T-2 toxin was converted by the culture to trideuterio-HT-2 toxin.     

Apparent Involvement of a Plasmid in Phaseotoxin Production by Pseudomonas phaseolicola

Three naturally occurring toxigenic strains (HB-36, G-50, and HB-33), one nontoxigenic strain (HB-20), and one ultraviolet light-induced toxinless mutant (G-50 Tox⁻) of Pseudomonas phaseolicola were examined by dye-buoyant density equilibrium centrifugation for the presence of plasmid deoxyribonucleic acid. All strains contained plasmid deoxyribonucleic acid. Comparison of the plasmid deoxyribonucleic acid of different strains by agarose gel electrophoresis showed that strain G-50 harbored three plasmids, whereas the rest of the strains contained two plasmids each. Irrespective of their toxigenicity, all strains shared the large-sized first plasmid band, but differed with respect to other plasmids. Restriction endonuclease analyses of the plasmids indicated that a 22.50-megadalton plasmid was common to two of the toxigenic strains (HB-36 and G-50). However, strain HB-33, which is also toxigenic, contained a much smaller plasmid (4.23 megadaltons). It is hypothesized that this small plasmid may have arisen by a recombination event from a larger plasmid.     

Inhibition by Agrobacterium tumefaciens and Pseudomonas savastanoi of development of the hypersensitive response elicited by Pseudomonas syringae pv. phaseolicola.

Injection into tobacco leaves of biotype 1 Agrobacterium tumefaciens or of Pseudomonas savastanoi inhibited the development of a visible hypersensitive response to the subsequent injection at the same site of Pseudomonas syringae pv. phaseolicola. This interference with the hypersensitive response was not seen with injection of bacterial growth medium or Escherichia coli cells. Live A. tumefaciens cells were required for the inhibitory effect. Various mutants and strains of A. tumefaciens were examined to determine the genes involved. Known chromosomal mutations generally had no effect on the ability of A. tumefaciens to inhibit the hypersensitive response, except for chvB mutants which showed a reduced (but still significant) inhibition of the hypersensitive response. Ti plasmid genes appeared to be required for the inhibition of the hypersensitive response. The bacteria did not need to be virulent in order to inhibit the hypersensitive response. Deletion of the vir region from pTi had no effect on the inhibition. However, the T region of the Ti plasmid was required for inhibition. Studies of transposon mutants suggested that the tms but not tmr or ocs genes were required. These genes were not acting after transfer to plant cells since they were effective in strains lacking vir genes and thus unable to transfer DNA to plant cells. The results suggest that the expression of the tms genes in the bacteria may inhibit the development of the hypersensitive response by the plant. An examination of the genes required in P. savastanoi for the inhibition of the hypersensitive response suggested that bacterial production of auxin was also required for the inhibition of the hypersensitive response by these bacteria.     

The Effects of Inhibitors Upon Pore Formation by Diphtheria Toxin and Diphtheria Toxin T Domain

The formation of pores by membrane-inserted diphtheria toxin is closely linked to the translocation of its catalytic chain across membranes. In this report a number of aromatic polyanionic molecules were identified that inhibit toxin-induced leakage of molecules from model membrane vesicles. One inhibitor, Cibacron blue, totally blocked pore formation. Aniline blue and Fast Green decreased the size of the molecule released by a given concentration of toxin. Amaranth appeared to reduce the maximal amount of leakage, without greatly affecting the size of the molecule released at a given toxin concentration. Finally, Ponceau S and Cibacron brilliant red appeared to exhibit a mixture of these various types of inhibition. The inhibitors neither prevented the conformational transition of the toxin to form a hydrophobic state at low pH, nor (with the exception of Cibacron Brilliant Red) appeared to strongly inhibit toxin binding to model membranes. Additional experiments showed release of trapped materials from model membranes by isolated T domain of the toxin was similar to that by whole toxin. The effects of inhibitors on T domain induced release was also similar to that they have on whole toxin. Therefore, it is likely that the inhibition of pore formation by whole toxin involves inhibitor interaction with the T domain. The inhibitors identified in this study may be helpful for development of agents that interfere with toxin action in vivo. Received: 10 March 1999/Revised: 22 June 1999