Pseudomonas andropogonis (synonyms: P. stizolobii and P. woodsii) has been reported to infect as many as 51 plant species belonging to 15 different families. This study focused on the systematics and pathogenicity of a comprehensive collection of P. andropogonis strains, as well as other ribosomal RNA (rRNA) group II pseudomonads. Pathogenicity, phenotypic, genotypic, phylogenetic, serological and chemotaxonomic analysis were carried out to better understand the taxonomic relationships in this group of bacteria.
All isolates of Pseudomonas andropogonis, including a total of 43 authentic strains of P. andropogonis originally isolated from 14 different plants, were pathogenic to carnation under glasshouse conditions, resulting in water soaked symptoms within 4-6 day after wound inoculation. Variation among strains was present in pathogenicity to sorghum and rhubarb. This is the first report that rhubarb is one of the hosts of P. andropogonis. The infection mechanism of P. andropogonis on carnation was investigated under the environmental scanning electron microscope.
Phenotypically, all strains of P. andropogonis were highly similar in most biochemical and physiological characteristics. Phenotypic comparisons were also made with an additional 31 authentic strains of P. caryophylli, P. cepacia, P. gladioli pv. gladioli, P. pseudomallei, P. solanacearum, P. pickettii, P. rubrisulbabicans, P. rubrilineans, P. syringae pv. syringae, Xanthomonas campestris pv. campestris, Escherichia coli and the type strain of P. aeruginosa.
Most rRNA group II pseudomonads tested were phenotypically similar. Pseudomonas andropogonis, P. cepacia, P. caryophylli, P. gladioli pv. gladioli, P. pseudomallei and P. rubrisubalbicans shared many phenotypic properties, and were distinct from the other two species P. solanacearum and P. pickettii. The latter two species are closely related. All rRNA group II pseudomonads were distinct from other bacteria tested including P. aeruginosa, the type species of the genus Pseudomonas.
The Biolog system with MicroLog™ 2 software, exploiting the respiratory activity occurring on oxidation and utilisation of a panel of 95 different organic compounds, was successfully applied to identify, within 24 h, isolates and reference strains of Pseudomonas andropogonis (n= 17), P. solanacearum (n=15) including biovars (bvs.) 1, 2, 3 and 4, P. caryophylli (n=l), P. cepacia (n=5), P. gladioli pv. gladioli (n=4), P. pseudomallei (n=2), P. pickettii (n=2), P. rubrisu/babicans (n=2) and some other bacteria with an accuracy of 94.6% at species level. Numerical analysis of the Biolog data with both Jaccard (SJ) and simple matching (SSM) coefficients showed that bacteria tested formed 15 similar phenons. Each phenon included bacterial strains belonging to a single species or a single subgroup within a species. The rRNA group II pseudomonads tested were grouped into three clusters with Jaccard coefficients (SJ) at more than 66% similarity. The three clusters shared less than 51.9% similarity with P. aeruginosa.
Eighteen strains of P. andropogonis were further subdivided into two subgroups. For oxidation of some organic compounds within 24 h, differences between strains do exist, which reflect their original hosts to some extent. Pseudomonas solanacearum isolates were further separated into two divisions representing biovars 1/2, and 3/4, respectively, which were equivalent to those obtained by other workers using restriction fragment-length polymorphism (RFLP) and restriction enzyme analysis.
Ten stable hybridoma cell lines secreting monoclonal antibodies (MAbs) were produced to P. andropogonis. Results of indirect ELISA and indirect immuno-fluorescence showed that MAb 6B3 was specific for P. andropogonis; MAb 3D5W1 reacted with both P. andropogonis and P. caryophylli ; six other MAbs reacted with all strains of seven species of rRNA group II pseudomonads tested except P. solanacearum and P. pickettii. None of the above eight MAbs cross-reacted with other non-fluorescent or fluorescent pseudomonads, xanthomonads and other bacteria tested. The ten MAbs were isotyped. Further tests showed that P. andropogonis was serologically similar to most other rRNA group II pseudomonads of pseudomallei subgroup. The epitopes of MAb 3D5W1 were clearly located within the P. andropogonis cell by immuno-gold labelling. These epitopes were protein in nature. Using MAb 3D5W1, P. andropogonis was readily detected by combining immunofluorescence and a detached carnation leaf assay with an initial inoculum of 4 x 10° cfu/ml after enrichment at room temperature for 4 days.
Bacterial whole cell protein SDS-PAGE profiles provided sufficient evidence for bacterial identification at species level. Each member species of the rRNA group II pseudomonads showed uniform and distinct protein patterns, with the exception of P. solanacearum which was subdivided into distinct groups. Pseudomonas andropogonis strains exhibit very similar patterns with minor differences. Pseudomonas solanacearum showed different protein patterns correlating with different biovars and divisions. All biovar 1 and 2 strains exhibited similar protein patterns, while bv. 3 strains were similar but distinct. The only biovar 4 strain tested, sharing a major band with bv. 3 group, showed a protein pattern different from all the other biovars. The whole cell protein SDSPAGE profiles of 83 strains belonging to 14 bacteria species were numerically analysed.
The DNA base composition provides a basic feature for bacterial species. The rRNA group II pseudomonads had relatively high G+C% values. A total of 39 strains of P. andropogonis showed DNA base composition within the range of 59%-61%, while other members of the rRNA group II pseudomonads tested had mol% G+C of 63%-72%.
DNA/DNA hybridization showed that P. andropogonis strains had DNA homology of more than 75%. All rRNA group II pseudomonads except for P. solanacearum and P. pickettii shared significant level (25%-45%) of DNA homology with P. andropogonis. None of the other rRNA group II pseudomonads tested shared DNA homology with P. solanaearum, which was reported by other workers to share DNA homology with P. pickettii.
Restriction endonuclease analysis of 22 P. andropogonis strains using Sal I showed that strains of this species exhibited identical patterns in molecular weight range from 2.0 kb to 6.6 kb. Differences among strains were observed which reflected host specificity to a certain extent, but not sufficient to support proposal of two pathovars.
Nearly complete sequences of 16S rRNA molecules were determined for 8 bacterial strains representing 5 species of the rRNA group II pseudomonads. Comparative analysis with published sequence data indicates that P. andropogonis, P. caryophylli, P. gladioli, and P. cepacia aggregate in one coherent cluster at 95.6% sequence similarity; P. solanacearum and P. pickettii shared 96.8% similarity with Alcaligenes eutrophus in another cluster. Both clusters joined at 91.7 similarity, which is similar to that for other genera in the beta subclass of Proteobacteria. On the basis of discerning evolutionary relationships, 4 strains of P. solanacearum representing bvs. 1/2 and 3/4 were subdivided into two divisions at 99.5% sequence similarity in agreement with other phenotypic and genotypic studies. The two divisions may be potentially regarded as subspecies.
Based on this study and some other published data, it is suggested that these bacteria should be separated from authentic pseudomonads, and constitute two related genera to accommodate P. andropogonis, P. caryophylli, P. gladioli, P. pseudomallei, P. rubrisubalbicans and P. cepacia; and P. solanacearum, P. pickettii, and A. eutrophus, respectively.