Genetic knockouts of each Halobacterium sp. NRC-1 uvr gene homolog were constructed to determine whether these genes are functionally homologous to the bacterial uvrA, uvrB, and uvrC genes that are required for NER in Escherichia coli and other bacteria 19. Knockouts of the Halobacterium sp. NRC-1uvr genes were constructed using an established ura3 counterselection strategy in which a ura3⁺ non-replicating Halobacterium shuttle vector (pMPK428) was engineered to carry a deletion construct composed of two 400–500 bp fragments that flank the region to be deleted 2021. These constructs were engineered for the uvrA, uvrB, and uvrC sequence homologs from Halobacterium sp. NRC-1 and the resulting plasmids were named pDCΔuvrA, pDCΔuvrB, and pDCΔuvrC. Each plasmid was transformed into DJC501, a ura3^- derivative of Halobacterium sp. NRC-1. ura3⁺ primary integrants were selected by plating on a medium deficient in uracil and, following PCR-based confirmation of plasmid integration, these primary integrants were cultured in a uracil-deficient medium and subsequently plated on rich medium containing 5-Fluoroorotic acid, which is toxic to ura3 prototrophs. Only cells which lost the deletion plasmid through a second recombination event and restored the ura3^- genotype could form colonies on 5-FOA plates. 5-FOA-resistant colonies were cultured and screened by PCR for the presence of deletion alleles.

Using the appropriate primer sets for PCR (Table 3), results such as those presented in Figure 1 were obtained. The control ura3^- strain, DJC501, contained wild-type alleles of uvrA and uvrC (Figure 1A, lanes 2 and 3). In contrast, the strains DJC519 and DJC502 contained the deletion alleles of uvrA and uvrC, respectively (Figure 1A, lanes 4 and 7). As expected, both strains possessed wild-type alleles of the non-targeted gene (Figure 1A, lanes 5 and 6). In the double mutant DJC509, both uvrA and uvrC deletion alleles were amplified (Figure 1A, lanes 8 and 9). Similar results were obtained for strain DJC520 (uvrB, data not shown). Because Halobacterium sp. NRC-1 is reputed to carry up to 30 copies of the genome (J. Soppa, personal communication), it was important to show not only the presence of the deletion allele but also the absence of any wild-type copies in the putative mutant strains. To do this, we carried out Southern hybridizations for each uvr gene. The Southerns confirmed that no uvr⁺ alleles were present in any of the mutant strains (Figure 1B and data not shown).

If the uvr genes are necessary for dark repair of UV-induced photoproducts in Halobacterium, we predicted that our deletion strains would be hypersensitive to UV when irradiated and incubated in non-photoreactivating conditions. We performed quantitative survival assays on DJC501 and the uvr deletion mutants we constructed and found a high degree of UV sensitivity, in the dark, in strains carrying mutant alleles of uvrA, uvrB or uvrC (Figure 2). In these mutant strains, approximately two logs of cell killing were observed at a dose (48 J/m²) resulting in well over 50% survival in the control DJC501 uvr⁺ strain.

Restoration of the ura3^- genotype in the counterselection knockout strategy allows for multiple deletions to be made in a single strain. If the uvr genes are performing a coordinated excision repair process similar to that of bacteria, then deletion of a second gene in this pathway should not further sensitize the cells to UV. To test this, DJC 502 (uvrC) was transformed with pDCΔuvrA to construct DJC 509 which carries deletions of both the uvrA and the uvrC genes (Figure 1). As can be observed in Figure 2, this double mutant was no more sensitive to UV than either the uvrA or uvrC single mutants, confirming that these genes operate in the same pathway in Halobacterium sp. NRC-1, presumably nucleotide excision repair.

Previous studies have shown that photoreactivation is highly efficient in the halophilic archaea, including Halobacterium sp. NRC-1, which encodes the CPD photolyase phr2 1622. In some bacteria and eukaryotes, molecular interactions between photoreactivation and excision repair have been suggested 23. To test the effect, if any, of the uvr genes on photoreactivation, we performed experiments in which identical plates inoculated with UV-treated uvr⁺ or uvr mutant cells were exposed to photoreactivating conditions (24 hours of fluorescent bulb irradiation) or kept wrapped in aluminum foil. For all strains, minimal loss of viability was detected on the unwrapped plates after 48 J/m² UV, indicating efficient photoreactivation that was not dependent on uvr genes (Figure 3 and data not shown).

Given the sensitivity of the uvr mutants to UV and their predicted function in excision repair of UV damage, we measured the ability of the deletion mutants to remove CPDs and 6-4PPs after a dose of 150 J/m² UV-C. The data show that the uvr⁺ cells repaired virtually all the 6-4PPs and around 80% of CPDs within one hour, and almost all damage in 3 hours, but that the uvrA, uvrB, and uvrC single mutants as well as the uvrA uvrC double mutant were unable to repair any CPDs or 6-4PPs during 3 hours post-irradiation incubation (Figure 4). This confirms that these genes are absolutely required for repair of UV lesions in the absence of photoreactivation, indicating that there is no other 'dark' repair mechanism in Halobacterium sp. NRC-1. This does not preclude the possibility that some of the eukaryotic repair homologs such as rad3 or rad25 are also involved, though NER proceeds without the need for these proteins in bacteria.

The uvrA, uvrB, and uvrC genes are found in the halophilic archaea and mesophilic methanogenic archaea but are absent from the genome sequences of all other archaea sequenced to date. Given this distribution, we examined the phylogenetic relationships between the core archaeal-encoded proteins (for all archaea known to contain them) and protein sequences found in a few diverse families of bacteria. Phylogenetic analysis of each of the Uvr sequences gave star topologies at the root, indicating that origins of the protein sequences cannot be uncovered. Haloarchaeal UvrA, UvrB, and UvrC always formed a monophyletic clade. Sequences from the mesophilic methanogenic archaea, however, were paraphyletic, with sequences from Methanosarcina acetivorans being quite different from the sequences encoded in the genomes of Methanothermobacter thermoautotrophicus and Methanosphaera stadtmanae. For UvrA, the haloarchaea were found to group with the UvrA sequence from the extremely radiation resistant bacterium Deinococcus radiodurans, while M. thermoautotrophicum and M. stadtmanae formed their own unique clade. M. acetivorans UvrA branched with the enterobacteria Camphylobacter jejuni and Helicobacter pylori (Figure 5A). For UvrB, the haloarchaea, M. thermoautotrophicum, and M. stadtmanae formed a major monophyletic clade together, while UvrB from M. acetivorans branched off on its own (Figure 5B). For UvrC, the haloarchaea formed a unique monophyletic clade, M. thermoautotrophicum and M. stadtmanae formed a clade with the cyanobacterium Synechocystis sp. PCC, and M. acetivorans claded with the spirochetes Borrelia burgdorferi and Treponema pallidum (Figure 5C).

All mutants were constructed using published techniques 2021. In brief, 400–500 base pair (bp) flanking regions of each uvr gene were amplified by PCR and cloned into pMPK428 (generous gift of M. Krebs and R. Peck), which carries the wild-type allele of ura3 and the β-lactamase (bla) gene for ampicillin selection in E. coli. PCR primers targeted to the upstream flanking region of each gene were engineered with KpnI and XbaI sites on the forward and reverse primers, respectively (see Table 3). Primers targeted to the downstream region of each gene were similarly engineered with XbaI and HindIII sites. Following amplification and purification, the PCR fragments were digested with the appropriate enzymes and triple-ligated with pMPK428 digested with KpnI and HindIII. The ligation mixtures were transformed into competent E. coli (JM109) cells and transformants were selected by plating on LB agar containing100 μg/ml ampicillin. Ampicillin-resistant colonies were picked, cultured, and plasmids were purified and digested with HindIII and KpnI to check for appropriate inserts. Selected plasmids with predicted restriction patterns were sequenced and named pDCΔuvrA, pDCΔuvrB, and pDCΔuvrC.

Halobacterium sp. NRC-1 ura3^- cells (a generous gift of Dr. M. P. Krebs, renamed DJC501) were grown to mid-log phase in modified rich CM+ media (per liter: 250 g NaCl, 20 g MgSO₄·7H₂O, 3 g Na citrate, 2 g KCl, 5 g Bacto-Tryptone, 3 g yeast extract, 1 g casamino acids pH 7.2, plus trace metals) and transformed with pDCΔuvrA, pDCΔuvrB, and pDCΔuvrC in separate reactions following established techniques 1440. Primary integrants (via homologous regions in deletion construct) were selected by plating transformation mixtures on HURA plates (per liter: 250 g NaCl, 20 g MgSO₄·7H₂O, 3 g Na citrate, 2 g KCl, 10 g of nitrogen base (Sigma-Aldrich co. Y0626), 1.92 g synthetic uracil dropout formula (Sigma-Aldrich co. Y1501), 20 g agar, pH 7.0 21 on which only ura3⁺ cells can grow. Primary integrant colonies were picked, grown to log phase in HURA broth, and DNA was prepared. PCR was performed using primers for the ura3 and bla genes to confirm integration of deletion plasmids (see Table 3). Log-phase cultures of ura⁺bla⁺ primary integrants were plated on modified CM+ plates (as above with 20 g agar/liter) + 0.25 mg/ml 5-fluoroorotic acid (5-FOA; Research Products International Corp., F10501) to select for loss of plasmid (and its ura3⁺ allele) via a second homologous recombination event. In approximately 50% of 5-FOA resistant colonies, the plasmid was lost by a recombination event that resulted in replacement of the targeted wild-type allele with the engineered deletion construct. 5-FOA resistant colonies were screened by PCR and genotypes confirmed by Southern blotting and hybridization.

5-FOA resistant colonies were cultured at 42°C in a C76 water bath shaker (New Brunswick Scientific, Edison, N.J.) at 200 rpm in modified rich CM+ liquid media and genomic DNA was prepared as described 40. PCR was performed using primer sets shown in the bottom section of Table 3. For uvrC amplifications, the uvrC1 forward and uvrC2 reverse primers were used.

Genomic DNA was digested with PstI (uvrA and uvrB), or KpnI and PvuII (uvrC) for determining genotype by Southern blot hybridization. Samples were subjected to electrophoresis overnight in alkaline gels containing 0.8% agarose. DNA was transferred to Hybond N+ membranes by Southern blotting and hybridized with AlkPhos-Direct with ECF-labelled (GE Healthcare Life Sciences, RPN3692) PCR fragments from either the upstream or downstream flanking region of each gene (Table 3). Chemifluorescence signal was detected using a GE Healthcare Storm Imager.

Halobacterium cultures were grown to mid-log phase in modified CM+ liquid media, centrifuged, and washed twice with CM salts (per liter: 250 g NaCl, 20 g MgSO₄·7H2O, 3 g Na citrate, 2 g KCl, pH 7.2). Working in subdued yellow light to prevent photoreactivation, washed cells were irradiated with a 254 nm germicidal lamp at a dose rate of 0.8 J/m²/sec with mild agitation. Samples were diluted and 20 microliters of each dilution were spotted on modified CM+ plates. Plates were wrapped in foil and incubated 4–5 days at 42°C. To observe the effects of photoreactivation on survival, two identically spotted plates were exposed to Sylvania Gro-Lux fluorescent light (Sylvania F40/GRO/AQ/RP) for 24 hours prior to incubation at 42°C. One plate was wrapped in foil as a control.

The repair experiments and dot-blot immunoassays for UV photoproducts were carried out as described previously 12152141. Log-phase cells were harvested and irradiated with 150 J/m² UV-C at a dose rate of 1 J/m²/sec and the irradiated cells were incubated aerobically at 37°C to allow repair to proceed. We have previously shown that, after this dose of UV, there is no detectable DNA replication during a 3-hour post-UV incubation 15. All irradiation and post-UV incubation was carried out either under yellow light illumination or in the dark. Fifty-ml samples were harvested at timed intervals and genomic DNA extracted using Promega Wizard genomic DNA kits. DNA samples from the various time points were equalised by measuring fluorescence of ethidium bromide-stained DNA in agarose gels, adjusting DNA concentrations and repeating this analysis as many times as necessary until all samples were of equal concentration.

Two identical dot blots were prepared on nitrocellulose filters, each containing a set of dilutions of each DNA sample in 1 M ammonium acetate. One blot was used to measure total damage (CPDs and 6-4PPs); the other was exposed to CPD photolyase and visible light to eliminate CPDs and allow for detection of 6-4PPs alone. The blots were then exposed to rabbit polyclonal antiserum containing antibodies to 6-4PPs and CPDs, then to biotinylated anti-rabbit antibody followed by alkaline phosphatase-conjugated Extravidin (Sigma) and finally to Nitro Blue tetrazolium (NBT) and 5-bromo-4-chloro-indolyl phosphate (BCIP) substrate. The intensity of blue color was proportional to the amount of DNA damage in the samples and was measured using a scanning densitometer (BioRad GS-670) and compared to a set of standards included on each blot.

Phylogenetic analysis of Uvr protein sequences in archaeal genomes

Protein sequences for the core Uvr system components (UvrA, UvrB, and UvrC) from representative bacteria, Halobacterium sp. NRC-1 and Methanothermobacter thermoautotrophicus, were downloaded from COGs 0178, 0556 and 0322, respectively. Sequences for Haloarcula marismortui, Natronomonas pharaonis, and Methanosphaera stadtmanae were downloaded from NCBI (see Additional file 1). Amino acid sequences were aligned using CLUSTAL_X1.83 42. Alignments were manually inspected and edited if necessary. TREEPUZZLE5.2 was used for quartet puzzling consensus maximum likelihood phylogenetic reconstruction using the JTT amino acid substitution matrix 43.