UvrD303, a Hyperhelicase Mutant That Antagonizes RecA-Dependent SOS Expression by a Mechanism That Depends on Its C Terminus

Abstract

Genomic integrity is critical for an organism's survival and ability to reproduce. In Escherichia coli, the UvrD helicase has roles in nucleotide excision repair and methyl-directed mismatch repair and can limit reactions by RecA under certain circumstances. UvrD303 (D403A D404A) is a hyperhelicase mutant, and when expressed from a multicopy plasmid, it results in UV sensitivity (UV^s), recombination deficiency, and antimutability. In order to understand the molecular mechanism underlying the UV^s phenotype of uvrD303 cells, this mutation was transferred to the E. coli chromosome and studied in single copy. It is shown here that uvrD303 mutants are UV sensitive, recombination deficient, and antimutable and additionally have a moderate defect in inducing the SOS response after UV treatment. The UV-sensitive phenotype is epistatic with recA and additive with uvrA and is partially suppressed by removing the LexA repressor. Furthermore, uvrD303 is able to inhibit constitutive SOS expression caused by the recA730 mutation. The ability of UvrD303 to antagonize SOS expression was dependent on its 40 C-terminal amino acids. It is proposed that UvrD303, via its C terminus, can decrease the levels of RecA activity in the cell.

All organisms must undergo DNA replication to pass their genetic information on to their offspring. Even for a simple organism such as Escherichia coli, this is not a trivial task. The coordinated action of many proteins is required to make up a replisome, which must commence DNA replication at the proper time during the cell cycle and at the correct location on the genome. Once this process begins, further challenges must be overcome to continue replication with high processivity and accuracy. Lesions to the template DNA, such as nicks, single-stranded DNA (ssDNA) gaps, abasic sites, double-stranded DNA (dsDNA) breaks, covalent base modifications, or proteins bound to the DNA, may result in stalled or collapsed replication forks, which must be repaired and restarted (13, 33).

The E. coli RecA protein is important for recombination and survival after DNA damage. RecA has at least two roles in survival after DNA damage (reviewed in reference 12). First, RecA is needed to induce the SOS response by acting as a sensor and allosteric effector of autoproteolysis of the LexA transcriptional repressor, whose degradation leads to the SOS response. Second, RecA is required for strand invasion during recombinational repair. Both functions require RecA to polymerize onto ssDNA and form a protein-DNA helical filament. This filament acts to increase the rate of LexA autoproteolysis and to perform recombinational repair. It is known that the SOS response increases the rates of transcription of at least 40 genes (11). Two SOS-regulated genes important for this paper are recA and uvrD.

UvrD (helicase II) is a 3′-5′ helicase with ssDNA-dependent ATPase activity (20, 30). It is part of the SF1 superfamily of helicases that includes Rep. UvrD functions in both nucleotide excision repair (NER) and methyl-directed mismatch repair (MMR) (17, 40, 43). In both processes, the helicase activity of UvrD is required to unwind the DNA strand containing the lesion such that a polymerase and DNA ligase may finish the repair process. Failure to complete these activities in a uvrD mutant results in sensitivity to UV light and an increased frequency of spontaneous mutagenesis.

A role for UvrD in regulating homologous recombination has also been established both in vivo and in vitro. In vivo, uvrD-null mutants lead to more RecA-green fluorescent protein (GFP) foci per cell in a population, and these foci are, on average, brighter than those in wild-type cells (6). UvrD is also required for the removal of RecA from arrested replication forks in temperature-sensitive dnaE and dnaN mutants (16). In vitro, UvrD has been shown to disrupt recombination intermediates and to remove RecA from ssDNA (46). This function seems to be conserved, since the Saccharomyces cerevisiae homolog of UvrD, Srs2, is capable of removing its cognate recombinase, Rad51, from DNA (24, 47). Failure to perform this antirecombination function in both E. coli uvrD and S. cerevisiae srs2 mutants results in a hyperrecombination phenotype (2, 38, 55).

Previously, Kushner and colleagues constructed a novel allele of uvrD called uvrD303 (53). This mutant has two point mutations (D403A and D404A) in a region of the protein that is thought to regulate helicase activity (5, 7, 26). In vitro, UvrD303 was shown to have a specific helicase activity as much as 10-fold higher than that of the wild-type protein, depending on the substrate (53). When uvrD303 was expressed from a multicopy plasmid, the cells were UV sensitive (UV^s) and recombination deficient (Rec⁻) and had a decreased frequency of spontaneous mutagenesis (53).

Given the hyperhelicase activity of UvrD303 in vitro, it is possible that more RecA removal from DNA and more-efficient MMR activity account for the Rec⁻ and antimutator phenotypes, respectively, of uvrD303 mutants. The reason for the UV^s phenotype of these cells, however, is not clear and can be explained by at least two models. One model predicts that despite its hyperhelicase activity, UvrD303 is deficient in NER, perhaps due to failure to interact with other essential upstream proteins in the repair pathway (i.e., UvrB) (1). Alternatively, a second model predicts that through its hyperhelicase activity, UvrD303 inappropriately removes RecA filaments that are required for initiating the SOS response or for recombinational repair of UV-induced DNA damage, or both. One goal of this work was to determine which, if either, of these two models would explain the UV^s phenotype of uvrD303 mutants.

In this study, the uvrD303 allele was transferred from a plasmid, where it had been studied previously, to the chromosome in order to avoid copy number effects. The results show that chromosomally encoded uvrD303 is still Rec⁻, an antimutator, and UV^s. uvrD303 is epistatic with recA and additive with uvrA for the UV^s phenotype, suggesting a defect in RecA-dependent repair rather than in NER. The uvrD303 mutant is modestly defective in induction of the SOS response after UV treatment, and its UV sensitivity can be suppressed by introducing a lexA null mutation. It is shown that uvrD303 is able to fully suppress the constitutive SOS expression of recA730 cells. Both of the SOS-antagonistic phenotypes of uvrD303 mutants are dependent on the protein's 40 C-terminal amino acids.

MATERIALS AND METHODS

Strains and media.

All bacterial strains are derivatives of E. coli K-12 and are described in Table 1. The protocol for P1 transduction has been described previously (50). All P1 transductants were selected on 2% agar plates; the agar was made with either Luria broth (LB) or 56/2 minimal medium (50) supplemented with 0.2% glucose, 0.001% thiamine, and specified amino acids. The antibiotic used for selection was either kanamycin (50 μg/ml), chloramphenicol (25 μg/ml), or tetracycline (10 μg/ml). All transductants were grown at 37°C and purified on the same type of medium on which they were selected. Assays for conjugal recombination and UV survival have been described elsewhere (8, 41).

TABLE 1.

Strains and plasmids used in this work

Strain or plasmid	recA	uvrD	Other relevant genotype	Source or derivation
Strains
CAG18491	+	+	metE3079::Tn10	44
GJ1989	+	+	lexA3 malB::Tn9	39
JC13509a	+	+	+	Lab stock
JC19108	+	+	lexA71::Tn5 malE::Tn10	Lab stock
JC19328	306	+	+	Lab stock
JJC145	+	+	PO66-HfrPK191 srlC::Tn10	4
JJC2457	+	cath	+	16
MV1138	+	+	srlC300::Tn10 recAo281	M. Volkert
SK9041	+	294	+	53
SS775	+	+	lexA3 malE::Tn10-9	GJ1989→JC13509e
SS996	+	+	del(attB)::psulA-gfp	31
SS1054	+	+	metE3079::Tn10	CAG18491→JC13509d
SS1465	+	+	gal76::Tn10 del(attB)::psulA-gfp	Lab stock
SS2533	+	cat	del(attB)::psulA-gfp	JJC2457→SS996e
SS3085	4136, 4155b	+	+	37
SS3368	4136, 4155b	cat	+	6
SS4414	+	+	lexA3 malE::Tn10-9	SS775/pKD46
SS4629	730	+	del(attB)::psulA-gfp	Lab stock
SS4952	100c	+	+	3
SS4976	4142	+	recAo1403 del(attB)::psulA-gfp	Lab stock
SS5112	730	+	metE3079::Tn10 del(attB)::psulA-gfp	CAG18491→SS4629d
SS5129	+	+	del(uvrA)100::kan	3
SS5141	+	303, 705	lexA3 malE::Tn10-9	This study
SS5145	+	701, 706	lexA3 malE::Tn10-9	This study
SS5191	+	701, 706	+	SS5145→JC13509f
SS5402	+	701, 706	+	SS5191/pRC35
SS5428	+	701	+	This study
SS5450	+	701	+	SS5428→SS1054g
SS5482	+	303, 705	+	SS5141→JC13509f
SS5489	+	303, 705	+	SS5482/pRC40
SS5496	+	303	+	This study
SS5704	+	303	+	SS5496→SS1054g
SS5711	4136, 4155b	303	+	SS3085→SS5704f
SS5737	+	701	gal76::Tn10 del(attB)::psulA-gfp	SS1465→SS5450d
SS5756	306	303	+	JC19328→SS5704d
SS5764	+	303	del(uvrA)100::kan	SS5129→SS5704f
SS5766	+	+	del(uvrA)100::kan	SS5129→JC13509f
SS5769	306	+	del(uvrA)100::kan	SS5756→ SS5766d
SS5801	730	303	del(attB)::psulA-gfp	SS5496→SS5112g
SS5802	+	+	metE3079::Tn10 del(attB)::psulA-gfp	CAG18491→SS996d
SS5803	+	303	del(attB)::psulA-gfp	SS5496→SS5802g
SS5807	4142	+	metE3079::Tn10 recAo1403 del(attB)::psulA-gfp	CAG18491→SS4976d
SS5808	4142	cat	recAo1403 del(attB)::psulA-gfp	JJC2457→SS5807g
SS5809	4142	303	recAo1403 del(attB)::psulA-gfp	SS5496→SS5807g
SS5907	100c	+	+	SS4952→JC13509f
SS5912	100c	+	+	SS5907/pLH29
SS5915	200c	+	+	SS5912/FLP
SS5959	+	+	lexA71::Tn5	JC19108→JC13509f
SS5960	+	303	lexA71::Tn5	JC19108→SS5704f
SS5962	+	+	srlC300::Tn10 recAo281	MV1138→JC13509d
SS5963	+	303	srlC300::Tn10 recAo281	MV1138→SS5704d
SS5985	+	303, 705	+	SS5482/pRC49
SS5986	+	303, 701	+	This study
SS5987	+	303, 701	+	SS5986→SS1054g
SS5988	+	303, 701	del(attB)::psulA-gfp	SS5986→SS5802g
SS5989	730	303, 701	del(attB)::psulA-gfp	SS5986→SS5112g
SS5990	730	701	del(attB)::psulA-gfp	SS5450→SS5112g
SS5993	+	294	+	SK9041→JC13509f
Plasmids
pGZK20				53
pGZK31				53
pKD46				14
pKO3				27
pLH29				22
pRC19				This study
pRC32				This study
pRC33				This study
pRC35				This study
pRC40				This study
pRC49				This study
pSJS1488				37

^aJC13509 has the following genotype: sulB103 lacMS286 φ80dIIlacBK1 argE3 his-4 thi-1 xyl-5 mtl-1 rpsL31 tsx. The lacMS286 φ80dIIlacBK1 genes code for two partial nonoverlapping deletions of the lac operon (23, 54).

^bThe full genotype is recAo1403 ygaD1::kan recA4155,4136::gfp-901 (37). This is abbreviated to recA4155,4136 in the table. recAo1403 is an operator mutation that increases the basal or non-SOS-induced level of transcription twofold (49). ygaD is the open reading frame in front of the recA gene. It has no known function. gfp-901 refers to mut-2 (9) with the additional “monomeric” mutation A206T (52). recA4155 is a mutant allele of recA with an arginine-to-alanine change at codon 28. It does not make storage structures in vivo (37). recA4136 refers to the specific fusion of recA to gfp.

^cThe 100 allele (from the Keio collection [3]) essentially replaces the gene in question with a kan gene between two FRT sites. The 200 allele is a result of removing the kan gene with FLP recombinase, induced by isopropyl-β-d-thiogalactopyranoside (IPTG) and expressed from the chloramphenicol resistance plasmid pLH29 (2). The strain was then grown in the presence of IPTG, and Kan^s Cam^s strains were screened.

^dSelect for tetracycline resistance. Screen by PCR if necessary.

^eSelect for chloramphenicol resistance. Screen by PCR if necessary.

^fSelect for kanamycin resistance. Screen by PCR if necessary.

^gSelect for methionine auxotrophy. Screen by PCR if necessary.

^hWe have sequenced the insertion-deletion mutation originally called del(uvrD-yigB)::cam in reference 16. We find that the mutation is actually a deletion-insertion mutation of uvrD and yigE. The endpoints of the deletion are GAAGAAGACGAAG in uvrD and CAGGCAAAATCATA in yigE. The last sequence in the uvrD gene is GAAGAAGACGAAG and the first sequence in the yigE gene is CAGGCAAAATCATA (all sequences between those two are deleted). The cat gene from pACYC184 (inclusive of the sequences starting with CCAAGCTCGAATT and ending with AAGTTGGAACCT) are in place of the uvrD and yigE sequences.

Plasmid constructions and transfer of alleles to the chromosome.

All DNA-modifying enzymes used (restriction endonucleases, DNA polymerase I (Pol I) Klenow fragment, T4 DNA ligase) were purchased from New England Biolabs and used according to the manufacturer's recommendations.

To construct pRC19, pGZK20 (53) was digested with BsiWI and subsequently treated with DNA Pol I Klenow fragment to fill in the overhanging ends. The reaction was then run on a 1% agarose gel, and the linear fragment was extracted and purified using the QIAEX II kit (Qiagen). Purified DNA was then religated using T4 DNA ligase. The ligation reaction mixture was used to transform DH5α competent cells. Plasmid DNA from ampicillin-resistant colonies was isolated and screened for the presence of a unique SnaBI restriction site. One plasmid containing the site was named pRC19, and this plasmid essentially recreates the uvrD del(40C) mutation, constructed and characterized previously (32). We have given this mutation the allele number uvrD701.

To construct plasmid pRC32, both pGZK31 (53) and pSJS1488 (37) were digested with BsiWI. Digestion reactions were run on a 1% agarose gel. The linearized plasmid pGZK31 and the small fragment of pSJS1488 containing the kan gene were gel purified using the QIAEX II kit (Qiagen). Purified DNA fragments were then combined in a ligation mixture containing T4 DNA ligase. The ligation reaction mixture was then used to transform DH5α competent cells. Plasmid DNA from kanamycin-resistant colonies was isolated and screened for the proper restriction pattern by digestion with BamHI and SalI. The resulting plasmid was named pRC32. The linear BamHI-SalI fragment of pRC32 containing the kan gene was then transferred to the chromosome by transformation of SS4414 competent cells containing plasmid pKD46 (14). Kanamycin-resistant colonies were purified and screened for ampicillin sensitivity, indicating loss of pKD46. Genomic DNA was then amplified by PCR using primers prSJS708 (5′-TAACAAGCCGCATATCCTGC) and prSJS709 (5′-CGTCGATTAATTCCATAAATCGCTGC) (Sigma). The resulting PCR product was screened for the presence of a BglI site, indicating that the D403A and D404A mutations were on the chromosome. The resulting strain was called SS5141 and contained the uvrD303 mutations plus a kan insertion in the BsiWI site of the uvrD gene. We have given this kan insertion the allele number uvrD705.

pRC33 was generated by digesting pRC19 with BstBI and subsequently treating with DNA Pol I Klenow fragment. pSJS1488 (37) was digested with BsiWI and subsequently treated with DNA Pol I Klenow fragment. DNA fragments were run on a 1% agarose gel. The linear fragment of pRC19 and the small fragment of pSJS1488 containing the kan gene were isolated and purified using the QIAEX II kit (Qiagen). Purified DNA fragments were combined in a ligation mixture with T4 DNA ligase. The ligation mixture was used to transform DH5α competent cells. Plasmid DNA from kanamycin-resistant colonies was purified and screened for the proper size by restriction analysis with BamHI and SalI. One correct plasmid was called pRC33. The linear BamHI-SalI fragment of pRC33 containing the kan gene was then transferred to the chromosome by transformation of SS4414 competent cells containing plasmid pKD46 (14). Kanamycin-resistant colonies were purified and screened for ampicillin sensitivity, indicating loss of pKD46. Genomic DNA was then amplified by PCR using primers prSJS714 (5′-GCGCCAGTTCAGCTACAACG) and prSJS716 (5′-GCTCACCGCTGCCTTCCATATTG) (Sigma). The resulting PCR product was screened for the presence of a SnaBI site, indicating that the uvrD701 mutation was on the chromosome. The resulting strain was called SS5145 and contained the uvrD701 mutation plus a kan insertion in the BstBI site of the uvrD gene. This kan insertion is given the allele number uvrD706.

pRC35 was constructed by digesting both pRC19 and pKO3 (27) with BamHI and SalI. The digested DNA was then run on a 1% gel. The small fragment of pRC35 containing the uvrD701 mutation and the large linear pKO3 fragments were isolated and purified using the QIAEX II kit (Qiagen). These fragments were then combined in a ligation mixture with T4 DNA ligase. The ligation reaction mixture was then used to transform SS5191 competent cells. Plasmid DNA from chloramphenicol-resistant colonies was isolated, and the resulting plasmid was named pRC35. This plasmid was then used to remove the kan gene from the chromosome of this strain (SS5402) by the pKO3-mediated gene replacement method (27). The resulting Kan^s Cam^s strain was called SS5428 and had the uvrD701 allele on the chromosome, as verified by PCR with prSJS714 and prSJS716 and subsequent digestion with SnaBI.

pRC40 was constructed by digesting pGZK31 (53) and pKO3 (27) with BamHI and SalI. Digestion reactions were run on a 1% agarose gel. The ∼1.7-kb BamHI-SalI fragment of pGZK31 and the large fragment of pKO3 were gel purified using the QIAEX II kit (Qiagen). Purified DNA fragments were then combined in a ligation mixture containing T4 DNA ligase. The ligation reaction mixture was then used to transform DH5α competent cells. Plasmid DNA from chloramphenicol-resistant colonies was isolated and screened for the proper restriction pattern by digestion with BglI or SacII. The resulting plasmid was named pRC40. This plasmid was then transformed into SS5482 and used to remove the kan gene from the chromosome of this strain by the pKO3-mediated gene replacement method (27). The resulting Kan^s Cam^s strain was called SS5496 and had the uvrD303 allele on the chromosome, as verified by PCR with prSJS708 and prSJS709 and subsequent digestion with BglI.

pRC49 was generated by digesting pRC40 (containing the uvrD303 mutations) with BsiWI and subsequently treating with DNA Pol I Klenow fragment. The linear fragment was run on a 1% gel, isolated, and purified using the QIAquick gel purification kit (Qiagen). The purified DNA was then treated with T4 DNA ligase for religation. This ligation mixture was used to transform DH5α competent cells. Plasmid DNA from chloramphenicol-resistant clones was screened for the generation of a new SnaBI site, indicating the presence of the uvrD701 mutation. One correct plasmid was called pRC49 and contained the uvrD303,701 mutations. To transfer these mutations to the chromosome, pRC49 was used to transform SS5482 competent cells, and the pKO3-mediated gene replacement method was performed as described above. The resulting Kan^s Cam^s strain was called SS5986 and had uvrD303,701 on the chromosome, as verified by the presence of a BglI or SnaBI site after PCR using prSJS708 and prSJS709 or prSJS714 and prSJS716, respectively.

Conjugal recombination assays.

Experiments were performed essentially as described previously (4). Briefly, donor (JJC145) and recipient strains were grown to mid-exponential phase in LB medium at 37°C with aeration. They were then diluted to an optical density at 600 nm of 0.1, and the donor was allowed to grow for an additional 20 min before 0.2 ml was aliquoted into a tube for each mating reaction. These tubes were placed at 37°C with no shaking to allow growth of pili for 30 min. At this point, 1.8 ml of the donor was added, and mating reactions proceeded for 20 min with no shaking. Reactions were stopped by vortexing vigorously for 1 min and placing tubes on ice. Serial dilutions were made in 56/2 buffer, and 0.1 ml of dilutions was combined with 0.7% top agar and poured onto minimal agarose plates lacking histidine and supplemented with streptomycin (0.1 mg/ml). Recombination efficiencies are reported as the percentage of his⁺ recombinants per 100 donors.

Spontaneous mutation frequency assays.

Experiments were performed as described previously (18). Results are averages for at least three experiments conducted on three different days.

Preparation of cells for microscopy.

Experiments were performed essentially as described previously (31, 37). Briefly, for RecA4155-GFP experiments, cells were grown to mid-log phase in 56/2 glucose minimal medium and then concentrated 10-fold, and 3 μl was placed on a 1% agarose (dissolved in growth medium) pad on a quartz microscope slide. A coverslip was then placed on top of the cells. Images (phase-contrast and fluorescence) were taken before exposure to UV light (zero time point). Then cells were exposed to 10 J/m² while on the quartz slide, and the same field of view was photographed again 20 min after exposure. At least six different fields of view (three on two different days) for each time point and strain were analyzed. For SOS experiments, cells were grown to mid-log phase in minimal medium, transferred to a sterile petri dish, and either irradiated or mock irradiated with 5 J/m² UV. Cells were then diluted 1:1 in fresh medium and allowed to grow for 1 h with shaking at 37°C. At this point, cells were concentrated 10-fold; 1 μl of internal reference beads (InSpeck Green [505/515] Microscope Image Intensity Calibration kit, 2.5 μm; catalog no. I-7219; Molecular Probes/Invitrogen) was added, and 3 μl of the mixture was added to an agarose pad on a microscope slide. The data shown in this report represent averages of nine fields of view (three on three different days) for each condition and strain. Typically between 1,000 and 3,000 cells are counted.

Microscopy and processing of images.

Microscopy and image processing were performed as previously described (31, 37).

RESULTS

uvrD303 mutants are UV sensitive, recombination deficient, and antimutable.

The uvrD303 allele was transferred to the E. coli chromosome and tested for its ability to participate in NER, homologous recombination, and MMR. Figure 1A shows that the uvrD303 mutant displayed decreased survival compared to that of the wild-type strain after UV exposure. The mutant was not quite as UV sensitive as shown previously when uvrD303 was expressed from a plasmid (53). Table 2 shows that the uvrD303 mutation resulted in about a twofold decrease in conjugal recombination frequency compared to that of uvrD⁺ cells. Again, this recombination deficiency is slightly milder than that reported previously when uvrD303 was expressed from a plasmid. By use of a different recombination test, where the formation of Lac⁺ papillae is dependent on intrachromosomal recombination, the uvrD303 cells fail to form papillae (Fig. 2) (54). Lastly, Table 2 shows that the uvrD303 mutant has a twofold decrease in spontaneous mutation frequency compared to that of the wild-type strain. Except for the magnitudes of the UV survival and conjugational deficiencies, the uvrD303 allele imparted a similar phenotype whether expressed from a plasmid or the chromosome. It should be noted that in each of these experiments, a uvrD deletion strain has a phenotype different from that of the uvrD303 mutant (Fig. 1 and 2; Tables 2 and 3). It is much more UV sensitive; it has a hyperrecombination rather than a hyporecombination phenotype; and it gives rise to a mutator instead of an antimutator phenotype.

FIG. 1.

uvrD303 cells are defective in RecA-dependent repair (not NER) in response to UV treatment. (A) Circles, wild type (JC13509); squares, del(uvrD-yigE)::cat (SS3368); triangles, uvrD303 (SS5704). Survival was measured for untreated cells and at doses of 10, 20, and 60 J/m² UV light. (B) Filled squares, del(recA-srl)306::Tn10 (JC19328); filled triangles, del(uvrA)100::kan (SS5766); open triangles, uvrD303 (SS5704); open circles, del(uvrA)100::kan uvrD303 (SS5764); filled circles, del(uvrA)100::kan del(recA-srl)306::Tn10 (SS5769); open squares, del(recA-srl)306::Tn10 uvrD303 (SS5756). Survival was measured for untreated cells and at doses of 0.8, 2, and 4 J/m² UV light. Each data point represents the average for at least three independent experiments, and error bars represent standard deviations.

TABLE 2.

Conjugal recombination abilities and spontaneous mutation frequencies of some uvrD mutants used in this study

Strain	uvrD	Conjugal recombination		Spontaneous mutation
		Avg proficiency per 100 donors	Ratioa	Avg frequency	Ratioa
JC13509	+	3.58 ± 2.02	1.00	4.09 × 10−8 ± 2.92 × 10−8	1.00
SS5993	294	7.23 ± 2.17	2.02	NDb	ND
SS3368	cat	ND	ND	2.01 × 10−6 ± 8.27 × 10−7	49.20
SS5704	303	2.02 ± 0.62	0.56	2.78 × 10−8 ± 1.82 × 10−8	0.68
SS5450	701	4.48 ± 3.08	1.25	5.72 × 10−8 ± 6.23 × 10−8	1.40
SS5987	303, 701	2.68 ± 1.05	0.75	5.91 × 10−8 ± 1.51 × 10−8	1.44

^aThe ratio of the conjugal recombination proficiency or spontaneous mutation frequency of each strain to that of the wild type is shown.

^bND, not determined.

FIG. 2.

Qualitative analysis of the abilities of various mutants to perform intrachromosomal recombination. Cells were streaked onto LB plates containing lactose and the indicator dye 2,3,5-triphenyltetrazolium chloride (TTC) and were incubated at 37°C for 48 h. The ability to form lac⁺ (white) papillae depends on intrachromosomal recombination events (54). Strains are as follows: JC13509 (uvrD⁺), SS3368 [del(uvrD-yigE)::cat], SS5704 (uvrD303), SS5450 (uvrD701), SS5987 (uvrD303,701), and SS5987 [del(recA)200::frt].

TABLE 3.

UvrD303 prevents complete SOS expression post-UV irradiation through a mechanism that requires its C terminus^a

Strain	uvrD	Before UV irradiation		After UV irradiation
		Avg RI	% Inductionb	Avg RI	% Induction
SS996	+	1.00	1.2 ± 1.0	20.8 ± 7.8	98.7 ± 1.7
SS5704	303	1.2 ± 0.6	0.7 ± 1.5	12.0 ± 4.9	79.8 ± 8.9
SS5450	701	2.3 ± 1.0	2.7 ± 5.2	24.2 ± 9.1	98.6 ± 1.0
SS5987	303, 701	2.9 ± 1.7	4.9 ± 10.2	31.6 ± 13.1	99.1 ± 1.9
SS3368	cat	3.9 ± 1.1	8.8 ± 5.7	48.4 ± 21.8	99.0 ± 0.5

^aCells were either left untreated or treated with 5 J/m² UV light. See Materials and Methods for further details. All values are given with standard deviations.

^bPercentage of cells with an RI sixfold or more greater than the average RI of the untreated wild-type strain.

uvrD303 mutants are defective in RecA-dependent repair (not NER) after UV treatment.

As mentioned above, UvrD functions in both homologous recombination and NER. Both processes contribute independently to survival after UV irradiation. An epistasis analysis was performed to test which of these two processes is affected in the uvrD303 mutant. It is known that mutations in the recA and uvrA genes confer sensitivity to UV light and that a recA uvrA double mutant displays an additive sensitivity because both repair pathways are affected (8, 21, 34). To test whether uvrD303 mutations are defective in RecA-dependent repair or in NER, the recA uvrD303 and uvrA uvrD303 double mutants were made. If uvrD303 is defective in RecA-dependent repair, the recA uvrD303 double mutant should be no more sensitive to UV light than the recA single mutant, but the uvrA uvrD303 double mutant should show an additive sensitivity compared to the uvrA single mutant. If uvrD303 is defective in NER, then the opposite scenarios should be true: the recA uvrD303 double mutant should show an additive phenotype, while the uvrA uvrD303 double mutant should show no more sensitivity to UV light than the uvrA single mutant. Figure 1B shows that the recA uvrD303 double mutant displayed no greater sensitivity to UV light than the recA single mutant. The uvrA uvrD303 double mutant, on the other hand, exhibited an additive phenotype in that it was more sensitive to UV light than either single mutant. These results suggest that uvrD303 is not defective in NER but rather displays negative effects on RecA-dependent repair of UV-induced damage.

uvrD303 mutants are defective in SOS induction after exposure to UV light.

While the results of the epistasis experiments (Fig. 1B) suggested that uvrD303 negatively affects the recA pathway of UV repair, other experiments showed that RecA-GFP focus formation in response to UV treatment was unperturbed in the uvrD303 mutant relative to uvrD⁺ cells (data not shown). This suggested that the UV^s phenotype of uvrD303 cells was not directly due to a decreased number of RecA filaments. Since RecA has multiple functions during the repair of UV-induced damage, including catalyzing homologous recombination, inducing the SOS response, and protecting ssDNA at stalled replication forks (10, 28), it is possible that while the RecA filament may still form, one or more of its functions may be decreased in the uvrD303 background. To test whether the ability of cells to turn on the SOS response to UV-induced damage was affected in the uvrD303 mutant, SOS expression was measured using fluorescence microscopy of cells expressing gfp under the regulation of the sulA promoter (31). Table 3 shows that 60 min after exposure to 5 J/m² of UV radiation, SOS expression was slightly reduced in the uvrD303 mutant from that in the wild-type strain. While the wild-type strain had an average relative fluorescence intensity (RI) of about 21, with 99% of cells induced (having an average RI sixfold or more higher than the average RI of the untreated wild-type strain), the uvrD303 mutant had an average RI of about 12, with a significantly (P < 0.03 by Student's t test) lower percentage (80%) of cells induced (Table 3). uvrD-deleted cells showed higher basal SOS expression and a robust increase after UV treatment to an average RI of 48, with 99% of cells induced after UV treatment. A broadening of the distribution of cells expressing gfp was also seen for the uvrD-deleted strain after UV treatment (see Fig. S1 in the supplemental material). The reason for this is currently unclear but may have to do with molecular noise associated with variations in gene expression (15, 35). Taken together, these results suggest that UvrD303 may prevent full SOS induction after UV treatment without disrupting RecA filament formation.

Constitutive activation of the SOS response restores UV resistance to uvrD303 cells in a manner that is independent of recA expression.

To further test whether the uvrD303 strain's modest effect on SOS expression was the cause of its sensitivity to UV exposure, a lexA::Tn5 mutation was introduced. Cells defective in the LexA transcriptional repressor constitutively express the SOS genes (17). Therefore, if lack of full SOS expression is responsible for the UV^s phenotype of uvrD303 cells, introduction of the lexA::Tn5 allele should suppress the UV sensitivity of these cells. As seen in Fig. 3, the lexA::Tn5 uvrD303 double mutant became nearly as resistant to UV light as the lexA::Tn5 single mutant.

FIG. 3.

Constitutive activation of the SOS response, but not overexpression of RecA, restores UV resistance to uvrD303 cells. Filled squares, uvrD303 (SS5704); filled triangles, recAo281 (SS5962); filled circles, lexA::Tn5 (SS5959); open triangles, recAo281 uvrD303 (SS5963); open circles, lexA::Tn5 uvrD303 (SS5960). Survival was measured for untreated cells and at doses of 10, 20, and 60 J/m² UV light. Each data point represents the average for at least three independent experiments. Error bars represent standard deviations.

One possible explanation for this result was that UvrD303 was preferentially removing RecA from the DNA and that increasing the concentration of RecA with a lexA::Tn5 mutation would shift the equilibrium back to the bound state. To test this idea, a recAo281 uvrD303 double mutant was constructed. The recAo281 allele contains a mutation in the operator of recA that prevents binding of LexA, resulting in constitutive recA overexpression (48). It is seen in Fig. 3 that overexpression of recA via the recAo281 allele did not suppress the UV sensitivity of the uvrD303 strain.

These results suggest that the UV-sensitive phenotype of the uvrD303 strain stems from the inability to fully induce the SOS response, although this does not seem to be due to a limitation of available RecA molecules.

uvrD303 suppresses the constitutive SOS phenotype of recA730 but not that of recA4142 mutants.

recA730 (E38K) mutants constitutively express SOS genes in the absence of DNA damage (25, 51). Since uvrD303 was shown to have a negative effect on the cell's ability to induce SOS expression after UV treatment, we tested if uvrD303 would suppress the high levels of SOS expression conferred by the recA730 mutation. For this purpose, these two mutations were combined into a single strain containing the SOS reporter construct sulAp-gfp (31). Table 4 shows that 95% of recA730 uvrD⁺ cells were fully induced for SOS expression in a population of exponentially growing cells and that the addition of uvrD303 markedly decreased the average SOS expression across the population to near-wild-type levels (<1% cells induced >6-fold). Some specificity seemed to be associated with this suppression, however, because uvrD303 had no effect on the constitutive SOS phenotype of a different recA constitutive allele, recA4142 (F217Y) (45) (Table 4). Potential reasons for this specificity are discussed below.

TABLE 4.

uvrD303 suppresses the constitutive SOS phenotype of recA730 cells by a mechanism that is dependent on the UvrD303 C terminus

Strain	uvrD	recA	Avg RI ±SD	% Constitutivea ± SD
SS996	+	+	1.00	0.3 ± 0.5
SS5704	303	+	1.2 ± 0.6	0.7 ± 1.5
SS4629	+	730	39.2 ± 3.3	93.7 ± 11.0
SS5801	303	730	1.1 ± 0.1	0.6 ± 0.3
SS5990	701	730	43.8 ± 2.1	99.1 ± 0.8
SS5989	303, 701	730	43.6 ± 2.5	99.2 ± 0.5
SS4976	+	4142b	17.1 ± 2.4	99.3 ± 0.9
SS5809	303	4142b	17.5 ± 3.8	98.4 ± 1.0

^aPercentage of cells with an RI sixfold or more greater than the average RI of the wild-type strain (± standard deviation).

^bAll strains containing recA4142 also have the recAo1403 operator mutation, which increases the percentage of the population that displays constitutive SOS activity from about 15% to nearly all of the cells (27a).

The C terminus of UvrD303 is required for its antagonistic effects on SOS expression.

A C-terminal peptide of the S. cerevisiae homolog of UvrD, Srs2, binds tightly to Rad51 (24). Additionally, Srs2 is known to remove Rad51 from the DNA (47). For these reasons, it was of interest to see if the 40 C-terminal amino acids were also important for the ability of UvrD303 to remove RecA730 from the DNA. The 40 C-terminal amino acids had previously been shown to play an essential role in the dimerization of UvrD monomers (32). Although that study, using multiple-turnover kinetics, suggested that these residues were dispensable for the helicase activity of UvrD in vitro (32), a later study, using single-turnover kinetic experiments, contradicted this report and suggested that dimeric UvrD is required for active dsDNA unwinding (29). A third study proposed two possible mechanisms for dsDNA unwinding by UvrD and suggested that multiple monomers (rather than a dimer) may be required for the unwinding of certain substrates (26). Thus, while the role of dimerization in UvrD activity appears to be controversial, it is clear that dimerization depends on the C terminus. Additionally, while UvrD can prevent the formation of recombination intermediates and dissociate RecA from ssDNA in vitro, the highly homologous (40% identical) Rep protein cannot (19, 46). Intriguingly, UvrD is slightly larger than Rep, and alignments between the two show that the 40 C-terminal amino acids of UvrD are not present in Rep (19).

To test whether the 40 C-terminal amino acids of UvrD303 were required for its antagonistic effects on SOS expression, the uvrD del(40C) mutant (referred to below as the uvrD701 mutant) was constructed and transferred to the chromosome in place of the endogenous uvrD gene. Tables 2 and 3 and Fig. 2 show that uvrD701 mutants behave very similarly to the wild type. To test whether the C terminus of UvrD303 was important for its antagonistic effects on SOS, the uvrD303 allele was combined with the uvrD701 allele and characterized. Table 2 reveals that the uvrD303,701 allele is like the uvrD303 mutant in that it is slightly Rec⁻ as measured by conjugal recombination. It appeared more recombination deficient when the Lac⁺ papilla recombination test was applied (Fig. 2). The double mutant also showed a slight increase in mutability relative to the wild type (Table 2). However, when the uvrD303,701 allele was combined in a strain with the SOS reporter construct sulAp-gfp and SOS expression was monitored before and after UV treatment, uvrD303,701 cells, unlike those of the uvrD303 single mutant, had an induced SOS response 60 min after 5-J/m² UV treatment that was intermediate between those for the wild type and the uvrD deletion mutant (Table 3). The distribution of cells with binned RIs showed a broadening similar to that seen for a uvrD deletion mutant (see Fig. S1 in the supplemental material). The effects on the constitutive SOS expression of the recA730 strain were also examined in the absence of the C terminus of UvrD303. Table 4 shows that the addition of uvrD303,701 completely restored the constitutive SOS phenotype of recA730 cells. The uvrD701 mutation alone had no effect on the recA730 phenotype (Table 4).

These results suggest that the C terminus of UvrD303 is required for its antagonistic effects on SOS induction by recA⁺ cells after UV treatment and by recA730 cells under normal growth conditions.

DISCUSSION

With some small quantitative differences, a chromosomal copy of uvrD303 shows recombination-deficient, antimutator, and UV-sensitive phenotypes similar to those shown by uvrD303 when expressed from a multicopy plasmid (53). This report reveals that the UV sensitivity is likely due to the ability of UvrD303 to interfere with RecA activity and is not due to a defect in NER. This conclusion is further supported by the ability of uvrD303 to suppress the constitutive SOS phenotype of recA730 cells and to interfere with normal SOS induction after UV treatment. Although the UV sensitivity of uvrD303 cells can be suppressed by constitutively activating SOS genes with the lexA::Tn5 mutation, overexpression of at least one SOS-regulated gene, recA, is not sufficient to achieve this. This suggests that some other SOS-regulated gene(s) is necessary for suppression. Future work will be necessary to determine which SOS genes are involved. Finally, the two anti-SOS phenotypes are dependent on the presence of the C terminus of UvrD303.

It had been proposed previously that UvrD is able to remove RecA under certain circumstances, as described above. The results presented here suggest that like UvrD, UvrD303 is able to remove RecA from DNA. This allele can do so, however, under a wider set of conditions and/or to a greater extent than the wild type. This ability seems to be specific, because UvrD303 can decrease the constitutive SOS phenotype of recA730 but has no effect on a different recA constitutive SOS allele, recA4142. It has been proposed elsewhere that RecA730 binds at replication forks and that RecA4142 binds at dsDNA breaks to induce the SOS response in log-phase cells (27a). Thus, the difference in the ability of UvrD303 to suppress high levels of SOS constitutive expression may be due to the environment in which the RecA resides. It should be pointed out that the ability of UvrD303 to interfere with RecA in some situations is not absolute; UvrD303 decreases RecA activity but does not remove it altogether. For instance, levels of UV resistance and UV-induced SOS expression are much higher in a uvrD303 strain than in a recA deletion strain (37). It is possible that other proteins, such as DinI and RecX, which affect the stability of RecA filaments, may modify the ability of UvrD303 to affect RecA activity in the cell. Future experiments may test this hypothesis.

Previous work has shown that the action of UvrD on RecA filaments occurs during situations where RecA has been loaded in a RecFORJQ-dependent manner, as at a replication fork (6, 16). Additional support for UvrD at replication forks comes from studies of yeast Srs2. This protein removes Rad51 after recruitment to replication forks by a SUMO-modified PCNA processivity clamp during the S phase of the cell cycle (36). While the evidence provided here suggests that the UvrD303 mutant does not completely remove RecA filaments at UV-damaged replication forks, it is likely recruited to these forks and prevents full RecA activity. Given its ability for generating anti-SOS phenotypes (Tables 3 and 4) and the fact that UV induction of the SOS response requires DNA replication (42), it is possible that the C terminus of UvrD303 (or that of wild-type UvrD) may physically interact with RecA or another protein to manifest its effect on RecA activity. Given the observation that the C terminus is also critical for UvrD to form a dimer in vitro (see above), it is possible that a dimeric form of UvrD303 is the functional entity required for these phenotypes in vivo (29).

It is possible that the restoration of high SOS expression observed in the uvrD303,701 mutant post-UV treatment and the recA730 phenotype (Tables 3 and 4) in the absence of UV irradiation are due to the instability of the double-mutant protein. This would suggest that the uvrD303,701 mutant should have the same phenotypes as the uvrD-null mutant. It is seen, however, that the mutability of the double mutant is like that of the wild type (Table 2) and that the uvrD303,701 mutant, like the uvrD303 mutant, is deficient in the Lac⁺ papilla test (Fig. 2). In both cases, the phenotypes are not similar to what one would predict for a partially stable UvrD protein or for the absence of the UvrD protein. Taken together, these results suggest that uvrD303,701 folds into an active protein and that this protein is stable enough to produce the phenotypes described above.

How UvrD (or UvrD303) removes RecA at replication forks is still unknown. While it is tempting to speculate that a specific interaction between the C terminus of UvrD and RecA may serve this function, there is no evidence to support this model. Additional experiments will be needed to reveal how the cell determines which RecA filaments are appropriate for UvrD-dependent removal and when this should occur.

This work reveals a new layer of how RecA participates in recombination and regulates the SOS response. While the details of how these processes are initiated (RecA loading) are well understood, less is known about how these processes are “turned off” or completed (dissociation of RecA from the DNA). Previous work has categorized UvrD as an enzyme with involvement in NER and MMR and with a role as an antirecombinase during replication fork arrest (see above). Here it is shown that UvrD303 can counteract the ability of certain RecA filaments to induce the SOS response and that its C terminus may be important for this role. It is intriguing to speculate that UvrD may also play an important role in ending recombination and turning off the SOS response by actively removing RecA filaments from the DNA. Further experiments will be required to test this idea.