INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The involvement of bacterial plasmids in increasing the survival of their hosts following irradiation with UV wavelengths was first reported in 1965 by Howarth (13), who was working with the ColIb-P9 plasmid in Salmonella enterica serovar Typhimurium LT2. Howarth also noted that the frequency of mutants in irradiated cultures of serovar Typhimurium LT2(ColIb-P9) was increased (14). These two initial observations have been followed by the discovery that a large number of bacterial plasmids from many incompatibility groups confer phenotypes of increased UV survival and mutability (43, 52). Genes conferring the UV mutability phenotype are also chromosomally located in some cases, an important example being the umuDC operon of Escherichia coli.

The umuDC operon is one component of the SOS regulon of E. coli, a set of approximately 20 unlinked genes which are coordinately regulated in response to DNA damage (42). The umuDC operon is regulated by the recA and lexA gene products, with the UmuDC and RecA proteins alone required for UV mutability (42). LexA functions as a repressor through binding to a conserved DNA sequence (SOS box) located within the promoter region of SOS regulon genes (3, 25). The irradiation of E. coli cells with UV wavelengths results in the occurrence of DNA lesions of which the cyclobutane pyrimidine dimer and the pyrimidine(6-4)pyrimidinone photoproduct are the most typical (8); these lesions can result in a blockage of DNA polymerase activity, leading to a stalling of replication. Cellular perception of DNA damage is thought to occur through the binding of RecA to single-stranded DNA immediately downstream of a DNA lesion; during this process, RecA is converted to an activated form (RecA*) (42). RecA* then mediates a self-cleavage reaction of LexA, resulting in the removal of LexA and allowing the expression of the SOS response genes (24).

Following the expression of umuDC, a posttranslational self-modification of UmuD (removal of the first 24 amino acids of UmuD) mediated by RecA* is required to generate the active UmuD' protein (5, 40). UmuD' forms a homodimer (UmuD'₂) which complexes with UmuC, resulting in the mutagenically active UmuD'₂C complex (4). The function of the UmuD'₂C complex in mutagenic DNA repair (MDR) occurs as a result of translesion DNA synthesis (42, 50). The model for replicative lesion bypass includes DNA pol III holoenzyme complex, the UmuD'₂C complex, and RecA* (48). However, recent evidence has shown that the UmuD'₂C complex itself may be sufficient in translesion synthesis with the possibility of either the UmuD'₂C complex or UmuC itself functioning as a DNA polymerase (36, 49).

To date, five distinct plasmid-carried umuDC homologs have been characterized at the sequence level. These include impCAB, mucAB, rulAB, rumAB_R391, and samAB (19, 26, 31, 33, 46). Each of these sequences contains a consensus LexA-binding site within the respective promoter regions and a conserved internal cleavage site within the umuD homolog. With the exception of rulAB, the other known MDR systems were isolated from enterobacteria and have not been well characterized in terms of their contribution to the ecological fitness of their hosts. The rulAB operon was originally cloned from pPSR1, an indigenous plasmid from Pseudomonas syringae pv. syringae A2, and was initially characterized for its role in UV radiation (UVR) tolerance (46). Although not absolutely required for survival, the UVR tolerance phenotype conferred by rulAB was subsequently shown to increase P. syringae populations by 10- to 30-fold in its leaf surface (phyllosphere) habitat (47). rulAB is the most distantly related to umuDC of the other plasmid-carried umuDC homologs, as rulA and rulB share only 30.9% and 41.5% amino acid similarity to umuD and umuC, respectively (42). However, rulAB does share the important features of this group, including its function in UVR tolerance and its lack of expression and activity in a P. syringae recA background (46). Recent evidence has also shown that rulAB is widely distributed among strains within and among pathovars of P. syringae (39, 47), suggesting the importance of this determinant to a wide range of genotypes.

Our laboratory is interested in further elucidating the role of the rulAB system in the population biology of P. syringae. To that end, we are fully characterizing the functional roles of rulAB (UVR tolerance and MDR), along with analyzing the regulation of this determinant so that we can ultimately address the biological significance of the rulAB system to P. syringae in its natural environment. Our studies reported here utilize UV-B (290 to 320 nm) wavelengths, which is in contrast to most analyses of DNA repair and UVR-induced mutagenesis in microorganisms, where higher-energy UV-C (254 nm) wavelengths are used. In nature, UV-C wavelengths are screened by the stratospheric ozone layer and do not reach the earth's surface; thus, the use of UV-B wavelengths is relevant from an ecological standpoint. The most important types of UV-B-induced DNA lesions appear to be cyclobutane pyrimidine dimers and pyrimidine(6-4)pyrimidinone photoproducts (30). Ecological studies have shown that solar UV-B radiation has profound deleterious effects on microorganisms dwelling in locations with high sunlight exposure, including surface aquatic habitats and the phyllosphere (11, 15, 45).

In this study, we report an analysis of the MDR capacity of rulAB and an analysis of mutability in a P. syringae strain containing a stable insertional mutation within rulAB. We also examined the MDR potential of rulAB-containing P. syringae in planta and show that UVR-induced mutability occurs in leaf surface populations at rates similar to those observed in vitro. We further demonstrate that rulAB complements an E. coli umuDC mutant and compare the MDR activity of rulAB with that of mucAB and umuDC in E. coli and P. aeruginosa by using defined genetic constructs. Finally, we analyzed the regulation of rulAB in response to increasing doses of UV-B (290 to 320 nm) radiation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and culture conditions. The bacterial strains, plasmids, and specific oligonucleotides utilized which were relevant to this study are listed in Table 1. All bacterial strains were grown in Luria-Bertani (LB) medium (Difco) or King's medium B (KB) (17); E. coli strains and Pseudomonas aeruginosa PAO1 were grown at 37°C, and Pseudomonas fluorescens Pf5 and P. syringae pv. syringae B86-17 were grown at 28°C. Plasmid transfer from E. coli to Pseudomonas strains was accomplished by triparental mating using the helper plasmid pRK2013. The exconjugants were selected on MG medium (16) or Pseudomonas isolation agar (Difco) supplemented with appropriate antibiotics. Antibiotics were added to the media in the following concentrations: ampicillin, 75 µg ml¹; carbenicillin, 150 µg ml¹; cycloheximide, 100 µg ml¹; gentamicin (Gm), 20 µg ml¹; kanamycin, 50 µg ml¹; rifampin, 100 µg ml¹; and spectinomycin, 40 µg ml¹.

Molecular biology techniques. Plasmid isolation from P. syringae was accomplished using the technique of Crosa and Falkow (6). Restriction digestions, isolation of DNA fragments from agarose gels, PCR amplifications, ligations, and Southern transfer to nylon membranes were performed using standard techniques (38). DNA fragments used as probes were labeled with digoxigenin-11-dUTP (Genius kit; Boehringer Mannheim, Indianapolis, Ind.) following the instructions of the manufacturer. Hybridizations at 65°C followed by high-stringency washes were performed as described previously (44).

Construction of P. syringae pv. syringae GWS242 (B86-17 rulB::Km). The rulAB gene disruption on the native plasmid pB8617A of strain B86-17 was created via gene replacement using a homologous recombination strategy. The intact rulAB sequence on pB8617A was initially cloned into pBluescript II SK(+) as a 6.4-kb BamHI fragment (pJJK1) and subcloned as a partial 3.8-kb EcoRI fragment into pGem 7zf, creating pJJK12. This plasmid was then digested with BssHII, resulting in the removal of a 0.6-kb segment from within rulB. Following BssHII digestion, the ends were made blunt by incubation with the Klenow fragment of DNA polymerase (Life Technologies), and the 1.2-kb kanamycin resistance (Km^r) gene cassette (released from pBSL86 as a HincII fragment) was ligated in. The entire DNA region was excised as a 4.4-kb XbaI-BamHI fragment and ligated into the suicide gene replacement vector pJQ200SK, creating pJJK16. pJJK16 contained approximately 1.0 kb of flanking DNA sequences on each side of the rulAB (rulB::Km) sequence. Following transfer of pJJK16 into P. syringae pv. syringae B86-17, those P. syringae cells in which a plasmid integration event had occurred were selected on MG supplemented with Gm. Several isolated Gm^r colonies were then cultured in LB broth containing Km. After 2 days of incubation, 0.1-ml aliquots were plated onto LB containing Km and 5% sucrose to counterselect against the sacB gene encoded on pJQ200SK. The sucrose-resistant (Suc^r) colonies recovered were subsequently tested for sensitivity to Gm as a phenotypic assay for loss of the vector sequences. Since there are no BamHI sites within rulAB or the Km^r cassette, the increased size of pB8617A rulB::Km was visualized by comparing the sizes of the BamHI fragment from pB8617A and pB8617A rulB::Km, which hybridized to an internal rulAB gene probe consisting of the 0.7-kb HindIII-PstI fragment from pGWS140. Further confirmatory hybridizations were done using the Km^r cassette from pBSL86 as a probe both to the BamHI plasmid digests and to BamHI digests of total genomic DNA from P. syringae pv. syringae B86-17 and GWS242 to ensure that there was only one insertion of the Km^r cassette within the GWS242 genome.

UV-B irradiation and MDR assays. Bacterial strains were grown to late log phase in LB medium containing the appropriate antibiotics; 1 ml of the cultures was pelleted, washed with an equal volume of sterile saline (0.85% NaCl) solution, and resuspended in an equal volume of saline. The cell suspension was then diluted with an additional 9 ml of saline and placed in a sterile glass petri dish. Cell suspensions were irradiated with UV-B (peak, 302 nm) wavelengths using an XX-15M model UV-B lamp (UVP Products, San Gabriel, Calif.) filtered through cellulose diacetate (Kodacel; Eastman Kodak, Rochester, N.Y.) to eliminate any UV-C wavelengths (<290 nm) given off. The UV-B lamp was turned on 15 min prior to use to allow for stabilization of the UVR output. The energy output of the lamp was monitored with a UV-X radiometer (UVP Products) and determined to be 2.8 J m² s¹. Cell suspensions were mixed continuously while receiving the UV-B dose to eliminate survival as a result of shading. After irradiation, surviving cells were enumerated by dilution plating conducted under conditions that minimized photoreactivation. For the MDR assays, 0.1 ml of untreated cells and cells from all UV-B treatments were diluted in 0.9 ml of sterile saline and added to 1 ml of 2× LB broth. Following overnight incubation under dark conditions, appropriate dilutions of cell suspensions were plated on LB alone and LB containing rifampin. The frequency of mutation to rifampin resistance (Rif^r) was calculated as the number of Rif^r mutants per 10⁸ cells.

Analysis of MDR in phyllosphere populations of P. syringae pv. syringae B86-17. P. syringae pv. syringae B86-17 was grown on LB agar medium for 18 h before inoculation. The cells were washed and then resuspended in 0.1 M potassium phosphate buffer (pH 7.0), and cell suspensions were adjusted turbidimetrically to approximately 5 × 10⁹ cells ml¹. Bush bean plants (Phaseolus vulgaris cv. "Blue Lake 214") were grown under controlled conditions in a growth chamber (24°C, 80% relative humidity, 240-µmol m² s¹ light intensity, and 12-h photoperiod) with care taken during watering to prevent nonspecific leaf surface colonization. Experiments consisted of spraying inoculum of strain B86-17 onto the primary bean leaves using an air brush (Badger model 350; Badger Co., Franklin Park, Ill.) until the leaf surfaces were uniformly coated with the cell suspension. Immediately and at 24, 72, and 120 h after inoculation, four of the leaves were excised and individually irradiated with a 500-J m² dose of UV-B by using the XX-15M lamp as described above. An additional four leaves were also removed, and these leaves were not irradiated as a control. The leaves were placed onto sterile, moist paper towels in a covered sterile Bio-Assay dish (Nunc Products) and incubated under dark conditions for 12 h at 25°C. After incubation, each leaf was sonicated for 7 min in 20 ml of sterile washing buffer (0.1 M potassium phosphate [pH 7.0] and 0.1% peptone). The combined washing buffer from four leaves per treatment was reduced to 1 ml by centrifugation, and appropriate serial dilutions of the leaf washings were plated onto KB agar medium amended with cycloheximide alone (KBc) and KBc containing rifampin. A total of six independent experiments were conducted. The data were normalized based on total leaf surface populations and reported as the number of Rif^r cells per 10⁸ total cells. A one-way analysis of variance based on UV-B irradiation was done to compare the mean number of Rif^r cells per 10⁸ total cells at each sampling time; means were differentiated using a t test.

Phenotypic comparison of rulAB with mucAB and umuDC. The individual coding sequences of the mucAB, rulAB, and umuDC MDR operons were amplified via PCR using the following oligonucleotide primers and source DNA: (i) mucAB Nde 5', mucAB Bam 3', and pRW144 for mucAB; (ii) rulAB Nde 5', 3' rulB TAA BamHI, and pJJK12 for rulAB; and (iii) umuDC Nde 5', umuDC Eco 3', and pRW154 for umuDC. The amplified product from pRW144 was confirmed as having the correct size on an agarose gel, digested with the appropriate restriction enzymes, and ligated into pET-5a, creating pJJK5. A 0.75-kb region upstream of umuDC on the E. coli chromosome was next amplified from pRW154 using the primers umu Pro SphI 5' and umu Pro 3', digested, and ligated upstream of the mucAB coding sequence in pJJK5, creating pJJK22. The other genetic constructs were made by first ligating the umuDC promoter into pET-5a, creating pJJK20, and then ligating the respective rulAB or umuDC amplified coding sequence into pJJK20. In each case, the translational start site of the MDR coding sequences was the ATG sequence within the NdeI sequence of pET-5a, allowing optimal spacing from a Shine-Dalgarno sequence present on the pET-5a vector. The final constructs, including the umuDC promoter sequences, Shine-Dalgarno sequence from pET-5a, and the respective MDR coding sequences were ligated into pJB321, creating pJJK25, pJJK26, and pJJK27. Each of these plasmids was transferred into E. coli RW120 and P. aeruginosa PAO1 and utilized in MDR assays as described above.

Determination of rulAB promoter activity using the inaZ reporter gene. The intergenic sequences between rulAB and repA on pB8617A were fused to the promoterless ice nucleation gene (inaZ) in pPROBE KI' in order to quantify rulAB expression following UV-B irradiation. The primers T7 (Promega) and rul PX were used to amplify a 1.1-kb fragment from pJJK12; the PCR product was directly ligated into pCR2.1, creating pJJK40. A 0.9-kb HindIII-EcoRI fragment was subsequently excised from pJJK40 and ligated into pPROBE KI', resulting in a transcriptional fusion of the rulAB promoter region with inaZ in the final construct, pJJK41. Ice nucleation activity (INA) was chosen as a reporter because of the sensitivity with which INA can be detected and the large range of INAs which can be quantified (28). INA is also easily quantified in situ, and our future experiments would be geared toward rulAB expression analysis in the phyllosphere. Unfortunately, most strains of P. syringae produce the ice nucleation protein naturally; however, we felt that the sensitivity of the reporter would enable us to detect differences in expression resulting after small differences in the UV-B dose and that it justified the use of inaZ even though a different Pseudomonas species would be used in the assay.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction and analysis of a rulB insertional mutant of P. syringae pv. syringae B86-17. The rulAB determinant was initially isolated from the indigenous plasmid pPSR1 from P. syringae pv. syringae A2, a pathogen of ornamental pear trees (46). In this and subsequent studies, it was of interest to work with the rulAB determinant of a P. syringae strain which was pathogenic on bean, a plant host which is more easily utilized in growth chamber studies. A bean-pathogenic strain, P. syringae pv. syringae B86-17, was chosen for use; this strain has a phenotype of UV tolerance and encodes rulAB on its large indigenous plasmid (47). For our experimental analyses, the rulAB determinant from pB8617A, the indigenous plasmid harbored by P. syringae pv. syringae B86-17, was subcloned as a 3.8-kb XbaI-BamHI fragment in the broad-host-range low-copy-number vector pJB321, creating the plasmid pJJK17.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   (A) Survival of P. syringae pv. syringae B86-17 () and GWS242 (); (B) survival of GWS242(pJB321) () and GWS242(pJJK17) () after UV-B irradiation. Each data point represents the mean (±the standard error of the mean) from three replicate UV sensitivity experiments. (C) Effect of the rulB::Km mutation in GWS242 on UV-inducible mutagenesis. Rifs strains were irradiated with different doses of UV-B, samples were removed to initiate cultures that were incubated in LB for 18 h, and the number of Rifr colonies was determined. The number of spontaneous mutations conferring Rifr in the absence of UV-B irradiation has been subtracted. Each data point represents the mean (± the standard error of the mean) from three replicate experiments. Symbols: , P. syringae pv. syringae B86-17; , GWS242(pJJK17); , GWS242(pJB321).

2

2

2

2

View larger version (22K): [in this window] [in a new window]   FIG. 2.   (A) Survival of P. aeruginosa PAO1(pJB321) () and PAO1(pJJK17) () after UV-B irradiation. Each data point represents the mean (±the standard error of the mean) from three replicate UV sensitivity experiments. (B) Analysis of rulAB-mediated MDR in PAO1. Rifs strains were irradiated with different doses of UV-B, samples were removed to initiate cultures that were incubated in LB for 18 h, and the number of Rifr colonies was determined. The number of spontaneous mutations conferring Rifr in the absence of UV-B irradiation has been subtracted. Each data point represents the mean (±the standard error of the mean) from three replicate experiments. Symbols: , PAO1(pJB321); , PAO1(pJJK17).

In planta UV-B mutability of P. syringae pv. syringae B86-17. The role of the rulAB determinant in enabling P. syringae strains to maintain and increase population size in their natural leaf surface (phyllosphere) habitat has been previously established (47). However, we have not previously assessed UV-B mutability in planta and, to our knowledge, the occurrence of MDR has not been previously shown using any bacteria in their natural habitat. Our studies were done by inoculating populations of P. syringae pv. syringae B86-17 onto bean leaves, excising the leaves at various sampling times after inoculation, irradiating the leaves with UV-B at 500 J m², and subsequently incubating the excised leaves for 12 h in a sterile, high-humidity environment. Following the incubation, bacteria were removed from the leaves by sonication and enumerated through plating on KBc alone and KBc containing rifampin. A total of six experiments were done, with the frequency of mutation to Rif^r shown to be significantly greater from irradiated leaves than from nonirradiated control leaves at all sampling points (Table 2). Over time, the ratio of Rif^r cells recovered from irradiated leaves to those recovered from nonirradiated leaves decreased slightly (4.8- to 2.9-fold) (Table 2). The levels of increase in mutability (at 500 J m²) observed immediately and at 24 h following inoculation were similar to those observed in the previous experiments performed in vitro (Fig. 1C).

Phenotypic comparison of rulAB with mucAB and umuDC. Since rulAB was found to be distantly related to umuDC at the amino acid level (42), it was of interest to determine if rulAB was functionally similar to other well-characterized MDR operons and especially to determine if rulAB could restore mutability functions to an E. coli umuDC mutant. In order to facilitate phenotypic-analysis studies of rulAB, it was desirable to choose hosts in which there was minimal interfering background activity in terms of mutability. We chose two hosts for these experiments, E. coli RW120, in which the chromosomal source of MDR, the umuDC determinant, was deleted, and P. aeruginosa PAO1, a strain which does not encode an MDR determinant. Comparisons of rulAB with its homologs mucAB and umuDC were done using the defined genetic constructs pJJK25, pJJK26, and pJJK27. These constructs only differed by the inclusion of the individual MDR systems; the promoter, Shine-Dalgarno sequences, and vector were identical among them.

View larger version (27K): [in this window] [in a new window]   FIG. 3.   Ability of defined genetic constructs encoding rulAB, mucAB, and umuDC to restore mutagenesis functions to an E. coli umuDC strain and to induce mutagenesis in P. aeruginosa PAO1. (A) E. coli strain RW120 [lexA+ recA+ (umuDC)595::cat]. The number of spontaneous mutations conferring Rifr in the absence of UV-B irradiation has been subtracted. Symbols: , rulAB (pJJK25); , mucAB (pJJK26); , umuDC (pJJK27). (B) P. aeruginosa PAO1. As with E. coli RW120, the number of spontaneous mutations conferring Rifr in the absence of UV-B irradiation has been subtracted. Symbols: , rulAB (pJJK25); , mucAB (pJJK26); , umuDC (pJJK27). In panels A and B, each data point represents the mean (±the standard error of the mean) from three replicate UV sensitivity experiments.

Regulation of the rulAB promoter. We used the ice nucleation gene, inaZ, as a reporter in our regulation analyses because of the sensitivity of ice nucleation as a reporter (28) and the relative ease of measurement of this reporter when used for in planta studies, which will be reported elsewhere. We chose to use P. fluorescens Pf5 for our studies, since Pf5 does not possess INA, and also because Pf5 grows at 25°C, a temperature at which INA is readily expressed (28). Prior to the utilization of P. fluorescens Pf5 in our studies, we determined that the rulAB determinant was active in this strain, functioning in the increase of both UV-B survival and mutability (data not shown).

2

2

1

1

2

2

View larger version (25K): [in this window] [in a new window]   FIG. 4.   Analysis of the activity of a rulAB promoter fusion with a promoterless inaZ gene in P. fluorescens Pf5 following UV-B irradiation. P. fluorescens Pf5(pJJK41) was irradiated with different doses of UV-B, samples were removed to initiate cultures which were incubated in KB, and at the designated time points, cells were removed and analyzed for INA using the droplet-freezing assay. Each data point represents the mean from three replicate experiments. Values in parentheses indicate the fold increase in InaZ concentration at each data point relative to that of the nonirradiated control. Symbols for amounts of UV-B irradiation of Pf5(pJJK41): , 75 J m2; , 150 J m2; , 300 J m2; , nonirradiated control.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The rulAB operon, which is widely distributed among diverse strains of P. syringae isolated from many plant hosts, plays an important functional role in protecting strains from the DNA-damaging effects of UV-B radiation. Exposure of P. syringae strains to UV-B wavelengths present in solar radiation would be predicted to occur with regularity in the phyllosphere habitat of this organism. Analysis of the UV-B mutability of the rulAB operon in E. coli RW120 indicated that rulAB is functionally similar to umuDC and other previously characterized MDR operons, such as mucAB and rumAB_R391. Indeed, rulAB shares many other features with the known MDR systems, including the presence of a binding site for the LexA repressor within its putative promoter region, possession of a conserved internal cleavage site within the rulA sequence, and lack of function in a recA background (46). Each of the known umu-like systems varies in its ability to promote mutagenesis in different cellular backgrounds. For example, the approximately threefold increase in our experiments of mutagenesis of mucAB in E. coli RW120 as compared to that of umuDC has been observed by others as well and is attributed to a more efficient processing of MucA to the truncated MucA' form in an E. coli host (10). It is interesting to note that the difference in mutation frequency between rulAB and mucAB was reduced in the P. aeruginosa PAO1 background (Fig. 3B) due to a larger decrease in MucAB-mediated MDR in P. aeruginosa than in E. coli. This is probably due to a decreased affinity of the MucAB system with P. aeruginosa RecA or DNA polymerase.

In the Pseudomonas backgrounds, rulAB has a large effect on UV-B survival for P. syringae as noted previously (46, 47) and for P. aeruginosa in the current study. Although the rates of survival of P. aeruginosa PAO1(pJJK17) and P. syringae pv. syringae B86-17 (rulAB+) were similar for UV-B doses up to 1,900 J m², the frequency of mutation to Rif^r was approximately sevenfold higher in P. aeruginosa. The increased mutation frequency in P. aeruginosa could be due to the relative contribution of other non-MDR systems and rulAB to the overall repair effort. At the 1,900-J m² UV-B dose, the survival rate of P. aeruginosa PAO1(pJB321) is approximately 20-fold lower than that of P. syringae pv. syringae GWS242(pJB321), suggesting that P. syringae is more efficient in non-MDR than P. aeruginosa.

Through analysis of UV-B mutability in the rulB::Km strain GWS242, we demonstrated a small (10- to 20-fold) increase in mutability to Rif^r, an observation that we have also made when examining other natural P. syringae strains which do not contain rulAB (Kim and Sundin, unpublished). These observations are consistent with the occurrence of another source of UV mutability within the P. syringae genome. Hybridizations of BamHI-digested genomic DNA of the rulAB-negative strain P. syringae pv. syringae FF5 with a rulAB probe under low-stringency conditions resulted in the observation of a single hybridizing band (Kim and Sundin, unpublished). We are currently attempting to clone this chromosomal sequence and determine its role in UV survival and mutability. Other examples of MDR systems cooccurring on plasmids and chromosomes in a single strain have been examined in serovar Typhimurium (31) and more recently in Shigella flexneri (37). In serovar Typhimurium, the chromosomally encoded umuDC_ST determinant is active in UV-inducible mutagenesis, while the plasmid-encoded samAB determinant does not contribute to UV mutability (18). The situation is reversed in S. flexneri, as the plasmid-encoded impCAB determinant is required for UV mutability while the chromosomally encoded umuDC operon is not expressed following UV irradiation and therefore is apparently unable to promote UV-induced mutagenesis (37).

Thus, in each of these situations, only one of the MDR operons plays the important role in UV mutability, which leads to the obvious questionwhy are two operons present? Recently, Opperman et al. (32) showed that the E. coli umuDC gene products play an additional role in cell cycle control following DNA damage. These authors demonstrated individual roles for the UmuD and UmuD' proteins and proposed a model in which the UmuD and UmuC proteins were involved in a delay in the recovery of DNA replication following DNA damage (32). RecA-mediated cleavage of UmuD and the formation of the UmuD'₂C complex would then be required for MDR mediated by translesion synthesis; this would be followed by the resumption of DNA replication (32, 35). Cell cycle control and a delay in replication restart following DNA damage are presumably important because they allow for non-MDR processes such as excision repair to occur, ultimately lowering the cellular mutational load. Organisms which encode two MDR systems in which one is functionally dominant in terms of actual repair may utilize the other system in cell cycle control. Alternatively, the formation of chimeric MDR complexes may also be involved in the posttranslational regulation of these systems. We are interested in examining the chromosomal source of UV mutability in P. syringae in order to understand the contribution of this system and rulAB to the overall repair process and UV survival.

The regulation analyses performed in this study indicate that rulAB is expressed in a UV-inducible manner with increased promoter activity in response to increasing UV-B dosage. The rapid induction of expression, maintenance of transcriptional activity at high levels for 4 to 6 h following irradiation, and overall importance of the rulAB determinant to UV-B survival suggest that MDR could play a major role in P. syringae survival in the environment. As more studies show the significance of DNA repair processes to organismal survival in habitats with high solar radiation exposure (e.g., reference 15), it would be important to determine the occurrence and role of MDR in organisms in these habitats.

It was previously shown that rulAB-containing P. syringae strains maintained significantly larger phyllosphere populations following UV-B irradiation (47). Our current results from assays involving phyllosphere populations of P. syringae pv. syringae B86-17 indicate that MDR occurs and is detectable during the period when inoculated cells are establishing an infection on their host. Populations of P. syringae pv. syringae inoculated to bean leaf surfaces typically initiate disease, which is manifested by the occurrence of leaf spots, within 4 to 6 days. In our experiments, we observed leaf spot symptoms at the 5-day sampling time. Our data indicate that UV-B mutability occurs at levels similar to those recorded in vitro when UV-B irradiation is administered immediately after inoculation. Over time, as strain B86-17 established an infection on its bean host, the frequency of UV-B mutability to Rif^r was slightly reduced (from 4.8- to 2.9-fold) but still significantly greater than in a nonirradiated control. As P. syringae strains become established in the phyllosphere, it is thought that a proportion of the cell population colonizes sites on leaves which are protected from external stresses (51). However, our data on UV-B mutability in the phyllosphere and previous data on rulAB-mediated UV-B survival in the phyllosphere (47) also clearly indicate the important contribution of the rulAB determinant to relative fitness during colonization and establishment of infection in planta.

The importance of rulAB to UV tolerance and the potential recurring necessity of this determinant for survival in the phyllosphere seem to distinguish rulAB from the other umu-like operons from an ecological standpoint. Each of the other umu-like operons characterized to date has been isolated from enteric organisms, whose exposure to UV radiation or chemical mutagens would probably be sporadic. One example of the consequence of a limited necessity for MDR in enteric organisms may be an attenuation of activity of the corresponding MDR determinant. For example, when located on its natural plasmid R46, the activity of mucAB is repressed by another gene located approximately 2 kb away (20). The role of some of the enteric-organism umu-like operons in UV protection is also unclear, as inactivation of either or both umuDC_ST and samAB in serovar Typhimurium has no effect on UV survival (18). Thus, the distinguishing features of the rulAB system in terms of activity when present on its native plasmid, function in UV tolerance, and potentially daily expression in response to solar UV damage make this system appropriate for continued analysis of the ecological and evolutionary ramifications of UV-induced mutagenesis.