INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The pPT23A plasmid family encompasses the majority of native plasmids identified in the plant-pathogenic bacterium Pseudomonas syringae; these plasmids share a gene (repA) required for plasmid replication and, in most cases, additional areas of homology (16, 30, 35, 38). pPT23A-type plasmids are diverse in size, and multiple plasmids sharing large regions of repeated sequences may be present in the same cell (3, 8, 30, 35). Plasmids in the pPT23A family can encode determinants of importance in host-pathogen interactions such as the coronatine biosynthesis locus, which increases virulence, and the avirulence genes avrD, avrPphC, and avrPphF, which affect host range (1, 21, 46). Additional sequences known to be encoded on plasmids of the pPT23A family include the stbCBAD locus involved in plasmid stability (18), copper resistance determinants (6) and the streptomycin resistance transposon Tn5393 (39), and insertion sequence elements including IS51, IS801, IS870, and IS1240 (1, 18, 30). A common feature of all of these determinants is that functional loci encoding these traits are typically limited in distribution to small groups of P. syringae pathovars.

Given the distribution of the pPT23A plasmid family throughout P. syringae pathovars, it is likely that these plasmids encode a "backbone" of traits of general importance to the P. syringae species. We are interested in the evolution of the pPT23A plasmid family in P. syringae, including determining the range of pathovars encompassed by particular plasmid lineages, characterizing genes encoded on these plasmids, and delineating the importance of horizontal transfer in pPT23A plasmid biology. In a previous study, we suggested that subgroups of pPT23A-like plasmids formed stable cohesive lineages as defined by Riley and Gordon (32), and further genetic evidence implied that individual plasmids had resided within their host strain for long time periods (39). Thus, to further understand the biology of the pPT23A plasmid family, it was desirable to study a locus that is widely distributed among P. syringae pathovars and might encode a trait of general importance.

One such locus is the rulAB operon, a homolog of the umuDC mutagenic DNA repair system first described in Escherichia coli (37). This operon encodes tolerance to UV radiation (UVR) and was recently cloned and characterized from a pPT23A-like plasmid from P. syringae pv. syringae (41). In contrast to other known pPT23A family loci, functional copies of the rulAB determinant are widely distributed and have recently been described in at least 14 pathovars of P. syringae (42). rulAB⁺ P. syringae strains vary widely in their tolerance to UVR (42), and we wished to determine if specific sequence alterations accounted for these observations. A functional rulAB locus is critical for the maintenance of population size on leaf surfaces that are irradiated with UV-B radiation (42). The importance of epiphytic population growth on leaf surfaces to the epidemiology of most P. syringae-host interactions (19) may explain the wide distribution of the rulAB determinant.

Our objectives in this study were (i) to compare sequences of rulA both within the pathovar syringae and among six other pathovars and (ii) to determine if the observed sequence differences affect the contribution of rulA to rulAB-mediated UVR tolerance.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. The bacterial strains and plasmids utilized in this study are listed in Tables 1 and 2, respectively. Escherichia coli and Pseudomonas aeruginosa strains were grown at 37°C on Luria-Bertani medium (Difco Laboratories, Detroit, Mich.) and Pseudomonas isolation agar (PIA) (Difco), respectively. P. syringae strains were grown at 28°C on King's medium B (24) or Luria-Bertani medium. When necessary, media were supplemented with the following antibiotics at the indicated concentrations (in micrograms per milliliter): ampicillin, 75; carbenicillin, 200; gentamicin, 50; kanamycin, 25; and rifampin, 100. Triparental matings, using the helper plasmid pRK2013, were done to mobilize plasmid constructs into P. aeruginosa PAO1.

UV sensitivity characterization. We used either UV-B or UV-C radiation in our UV sensitivity experiments. The UV-B sensitivity of the P. syringae strains following a dose of 590 J m² (biologically effective dose calculated using the DNA damage action spectrum of Setlow [36]) was determined; this survival value can be compared to those of P. syringae strains assayed previously (42). UV-C radiation also was used because the higher-energy UV-C wavelengths more readily distinguish differences in the UV sensitivity of individual strains.

2

1

2

1

Genetic and molecular biology techniques. Standard molecular biology techniques were utilized. Indigenous pPT23A-like plasmids were isolated from P. syringae strains as described previously (38). Nucleotide sequencing was done using the Big Dye kit (Applied Biosystems, Foster City, Calif.) following the instructions of the manufacturer; sequence reactions were run at the Genetic Technologies Center, Texas A&M University. Oligonucleotides were obtained from Life Technologies, Gaithersburg, Md. The rulA nucleotide and derived amino acid sequences were aligned using the program Clustal W (43). The derived amino acid sequences were analyzed phylogenetically using the Protpars program of PHYLIP (11) as previously described (23). The derived amino acid sequence of rumA, a closely related homolog of rulA from Klebsiella pneumoniae (25), was included as the outgroup.

Genetic analysis of the rulA locus. rulA was amplified from 14 P. syringae strains via PCR using plasmid preparations as the template and the primers rul2 (5'-CGTTAACTGTACGTCCATACAG-3') and rul4 (5'-CGAATTGCAATCGACCAG-3'). The rul2 primer encompassed the consensus LexA-binding site upstream of the rulA coding sequence, and the rul4 primer is the reverse complement of a sequence within rulB that is highly conserved among rulB homologs from E. coli (41). Standard PCR conditions were utilized (35), except the annealing temperature was lowered to 37°C to account for possible sequence divergence at the primer sites. The size of individual amplified fragments was checked on 1.2% agarose gels, and the fragments were ligated directly into the vector pCR2.1 (Invitrogen, Carlsbad, Calif.). The resulting clones were utilized as source material for nucleotide sequencing.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Map and construction of chimeric rulAB expression clones. (A) Insertion of the umuDC promoter region from pRW154 as a SalI-XbaI fragment into pET-5a creating pJJK20. (B) Insertion of the rulAB coding sequence from P. syringae pv. syringae B86-17 as an NdeI-BamHI fragment, creating pJJK21, and replacement of the rulA allele with rulA alleles from other P. syringae strains inserted as NdeI-HindIII fragments. Restriction sites relevant for the construction are shown: B, BamHI; H, HindIII; N, NdeI; S, SalI; X, XbaI. The position of the Shine-Dalgarno sequence of pET-5a is indicated by the underscored bar.

Nucleotide sequence accession numbers. The individual rulA nucleotide sequences generated in this study have been deposited in GenBank (accession numbers U43696 and AF251481 to AF251493).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Sequence diversity within rulA alleles. We determined a UV tolerance phenotype for each strain except P. syringae pv. lachrymans 1188-1 and P. syringae pv. pisi 1086-2, which were both highly UV sensitive (Table 3).

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Radial cladogram resulting from a Protpars program of PHYLIP (11) analysis based on amino acid dissimilarities of 14 sequences of rulA from P. syringae using the homolog rumA from plasmid R391 as an outgroup. Abbreviations Psl, P. syringae pv. lachrymans; Psm, P. syringae pv. maculicola; Psph, P. syringae pv. phaseolicola; Pspi, P. syringae pv. pisi; Pss, P. syringae pv. syringae; Pst, P. syringae pv. tomato; Psv, P. syringae pv. savastanoi. The letters (arbitrarily beginning with Pss HS191) indicate the amino acid change(s) from the A2 allele at the nearest branch point. Amino acid changes are noted in single letter code for the residue in the A2 rulA allele, followed by the residue in the alleles further away from the branching point: A, M1V, D58N, V74F, P95L, Q103H; B, L91V; C, V96L, S97C; D, R79A; E, L18F; F, F39S; G, D70G; H, C10S, T20S, K80T, P118A, E125D; I, V54A; J, F39S, H106D; K, S97 (unique amino acid to Pss A2, C in other alleles); L, N2K, R8W, Q37H, F139S; M, E140G; N, D59Y; O, V96 (unchanged from A2 allele); P, E126D; Q, I121V; R, A65G; S, P17S, Y19C, D42Y, H120Q.

Functional analysis of individual rulA alleles. We tested rulA alleles from strains representing six distinct branches of the rulA cladogram (Fig. 2). Since the constructs were identical except for the rulA allele, our assumption was that expression of rulAB from each construct was similar. Each chimeric rulAB construct had greater UV tolerance (P = 0.05) than that of PAO1 containing the vector pJB321 at each of five UV-C doses (Table 5). Significant differences in survival were observed among the alleles at four of the five UV-C doses employed (Table 5). The magnitude of survival differences increased with increasing UV-C dose, and the largest difference observed was a fourfold difference at 60 J m² between the rulA alleles from P. syringae pv. syringae 5D425 and P. syringae pv. savastanoi 0886-21 (Table 5). The rulA allele from P. syringae pv. pisi 1086-2 was functional in these experiments, although its parental host strain is UV sensitive (Table 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We compared rulA sequences from seven pathovars of P. syringae, as a prelude to larger-scale plasmid analyses, in an effort to define plasmid lineages within the P. syringae species. Considerable intra- and interpathovar sequence divergence was evident within the rulA gene of P. syringae, a gene which, together with rulB, encodes tolerance to UVR through a mutagenic DNA repair system. The maximum nucleotide sequence differences observed were 12% within the pathovar syringae (8B48 and HS191) and 15% between two pathovars (P. syringae pv. syringae HS191 and P. syringae pv. pisi 1086-2). The minimum sequence difference observed was 1.2% between the two bean isolates of P. syringae pv. syringae, a value that is relatively high when other interstrain sequence comparisons of plasmid-borne P. syringae genes are considered (4, 44, 46). The percent divergence among rulA alleles observed in our study is similar to that observed for four chromosomal loci of P. syringae examined by Sawada et al. (33). In their study, comparisons of partial sequences of gyrB, hrpL, hrpS, and rpoD from 19 pathovars showed overall nucleotide differences ranging from approximately 4 to 15% (33).

Sequence comparisons of plasmid-borne P. syringae genes have been reported for the avirulence gene avrD (three pathovars), a 650-bp region internal to the coronafacate ligase (cfl) gene within the coronatine biosynthetic cluster (four pathovars), and the efe gene encoding the ethylene-forming enzyme (five pathovars) (4, 23, 44, 46). In contrast to results with gyrB, hrpL, hrpS, rulA, and rpoD, relatively few sequence differences were observed, even among pathovars. For example, the nucleotide sequence difference observed within the cfl gene ranged from <1 to 3%, and the efe alleles were virtually identical (< 1% sequence difference) with one exception (4, 44). Based on these comparisons, we think that the avrD, cfl, and efe genes have been disseminated more recently than rulAB among P. syringae pathovars. Alternatively, the avrD, cfl, and efe loci might be subject to strong selection pressure and unable to tolerate significant sequence alterations.

We utilized UV-B (290 to 320 nm) and UV-C (<290 nm) radiation interchangeably in our UV sensitivity analyses. The amount and spectral quality of UV-B (290 to 320 nm) radiation reaching the earth's surface is affected by geographic and physical factors and may range from 1 to >10 kJ m² day¹ (28). In contrast, higher-energy UV-C wavelengths are completely screened by the stratospheric ozone layer and do not reach the earth's surface; however, UV-C wavelengths more readily distinguish differences in the UV sensitivity of individual strains. In previous experiments, we have shown that rulAB confers a phenotype of UV tolerance to both UV-C and UV-B wavelengths (41, 42); such comparability is predicted because the biological effects of both UV-B and UV-C radiation are due mainly to direct DNA damage (13).

A UV-B dose of 590 J m² differentially affected the survival of the P. syringae strain collection (Table 3). The rulA alleles from the UV-sensitive strains P. syringae pv. pisi 1086-2 and P. syringae pv. lachrymans 1188-1 differed by only three amino acids (2%), and possessed several unique subsititutions compared to the others (data not shown). The rulA from P. syringae pv. pisi 1086-2 could confer UV tolerance in conjunction with the P. syringae pv. syringae B86-17 rulB allele at levels similar to that of the other alleles in the experiment (Table 5). We concluded that the UV sensitivity of P. syringae pv. pisi 1086-2 was due to the presence of a nonfunctional rulB allele, analogous to the weakly functional efe allele borne on a plasmid in P. syringae pv. pisi (44). Alternatively, the UV sensitivity of strain 1086-2 could be due to the presence or absence of another genetic locus that increases the strain's UV sensitivity. Such a situation occurs on the IncJ plasmid R391, which bears the rulAB homolog rumAB_R391 (25) yet sensitizes its host bacterium to UV irradiation (31). Recently, a natural rulB-disruption mutant, containing a 4.5-kb insertion including the avrPpiA1 gene, was discovered in a Race 2 strain of P. syringae pv. pisi (2). We are currently analyzing the rulB alleles from P. syringae pv. pisi 1086-2 and P. syringae pv. lachrymans 1188-1 to determine if similar insertions have occurred.

We found that the six rulA alleles conferred different levels of UV tolerance to the P. aeruginosa PAO1 host. The magnitude of differences in survival increased with increasing UV-C dose, suggesting that some chimeric rulAB alleles were functionally better equipped to handle an increased load of DNA damage. The significant differences in survival observed at higher UV-C doses among the rulA alleles from P. syringae pv. savastanoi 0886-21 and P. syringae pv. syringae 5D425, A2, and B86-17 could be due to unique amino acid alterations (e.g., L91V in 0886-21, E126D in 5D425, C97S in A2, and D58Y and E141G in B86-17) since these alleles were otherwise very similar. However, we did not include the rulB alleles from these strains; thus, it is possible that the differences in survival observed due to differences in rulA could be masked by sequence differences in the rulB alleles. We wish to determine if the differences in UV survival observed affect the area on a leaf surface that can be colonized. It is known that rulAB significantly affects strain survival on leaf surfaces in the presence of UV-B radiation (42). Our current focus is to determine if subtle differences in rulAB activity can significantly impact the phyllosphere fitness of competing rulAB⁺strains.

rulAB appears to be important to the general biology of a wide range of P. syringae strains, and rulAB-mediated UV tolerance enables P. syringae strains to maintain population size in the phyllosphere (42). The attainment of large population sizes on individual leaves is an important prerequisite to disease incidence in some cases (19). Thus, from an ecological standpoint, it is likely that the ability of P. syringae to overcome UV stress is an important feature for its successful occupation of the phyllosphere habitat. That such a trait would be plasmid borne is consistent with Eberhard's hypothesis which states that plasmids tend to bear genes of importance in local adaptations (10). rulAB-mediated UV tolerance may be required only at certain times within the P. syringae life cycle, as strains are not constantly exposed to solar UVR.

Although extensive intrapathovar diversity occurs within P. syringae, and in particular within pathovar syringae (7, 9, 15, 26, 40, 45), some pathovars such as pathovars actinidiae and tomato are strikingly homogenous, with little observed genetic diversity (9, 34). Most P. syringae pathovars are distinct and readily distinguishable (27, 29). Indeed, comparison of sequence relatedness at the genome level has shown some pathovars (e.g., pathovars savastanoi and avellanae) are sufficiently different that reclassification into species separate from P. syringae has been proposed (14, 22). The rulA sequence variation is similar to that seen in previous studies utilizing chromosomal loci; i.e., there was considerable intrapathovar variation within P. syringae pv. syringae, and sequences from P. syringae pv. tomato were very closely related. Thus, we think that the rulAB locus has been evolving for a long period of time within P. syringae, probably mostly borne on plasmids of the pPT23A family. Further large-scale plasmid analyses are needed to unravel the evolutionary history of the pPT23A plasmid family, including identifying lineages and the pathovars that they occupy and examining the effect of residence within distinct pathovars and the effect of host plants on the ultimate composition of the plasmid genome.