The Nucleotide Excision Repair System of Borrelia burgdorferi Is the Sole Pathway Involved in Repair of DNA Damage by UV Light

Abstract

To survive and avoid accumulation of mutations caused by DNA damage, the genomes of prokaryotes encode a variety of DNA repair pathways most well characterized in Escherichia coli. Some of these are required for the infectivity of various pathogens. In this study, the importance of 25 DNA repair/recombination genes for Borrelia burgdorferi survival to UV-induced DNA damage was assessed. In contrast to E. coli, where 15 of these genes have an effect on survival of UV irradiation, disruption of recombinational repair, transcription-coupled repair, methyl-directed mismatch correction, and repair of arrested replication fork pathways did not decrease survival of B. burgdorferi exposed to UV light. However, the disruption of the B. burgdorferi nucleotide excision repair (NER) pathway (uvrA, uvrB, uvrC, and uvrD) resulted in a 10- to 1,000-fold increase in sensitivity to UV light. A functional NER pathway was also shown to be required for B. burgdorferi resistance to nitrosative damage. Finally, disruption of uvrA, uvrC, and uvrD had only a minor effect upon murine infection by increasing the time required for dissemination.

INTRODUCTION

All living organisms encounter DNA damage that, if left unrepaired, would cause mutation or lethality. Damage in the DNA can result from spontaneous base alteration, exposure to a DNA-damaging agent, or the environmental conditions (1, 2). The genomes of bacteria encode multiple pathways for detecting and repairing damaged bases. The main pathways, base excision repair (BER), methyl-directed mismatch correction (MMC), nucleotide excision repair (NER) (including transcription-coupled repair), and homologous recombination, have been best characterized in Escherichia coli (1, 3–10). In addition to naturally occurring DNA damage, pathogens also encounter stress from the host tissue environment and immune response. The capacity to detect and repair DNA damage is essential for a pathogen to successfully infect its host (11–13).

The Lyme disease spirochete Borrelia burgdorferi is an obligate parasite transmitted to a vertebrate host, including human and mouse, by the Ixodes tick (14–17). During murine infection, B. burgdorferi disseminates hematogenously to multiple organs, including joint, heart, skin, and bladder (18). Previous studies suggested that during its infectious cycle, B. burgdorferi could be exposed to DNA-damaging conditions and agents (19–27). However, little is known about the pathways used by B. burgdorferi to repair damaged DNA and if these systems are required for infection in mice.

Sequencing of B. burgdorferi revealed a relatively small but highly segmented 1.5-Mbp genome (28). Interestingly, no orthologs of the lexA, ruvC, mutH, sbcB, and recFOR DNA repair/recombination genes have been identified. However, sequence comparison revealed that all four genes of the NER pathway appeared to be present. The E. coli NER pathway is involved in detecting and removing various types of damaged bases, including those with damage caused by UV light (1) (see Fig. 1 for a schematic representation of the E. coli NER pathway). Briefly, E. coli UvrB is loaded at the damage site by UvrA and recruits the UvrC endonuclease, which incises the damaged DNA strand. The DNA helicase II (UvrD) displaces the damaged strand, and then polymerase I and DNA ligase fill the gap. Amino acid sequences of B. burgdorferi and E. coli UvrA, UvrB, UvrC, and UvrD are similar in length and have 53, 50, 29, and 30% identity, respectively, and 72, 68, 53, and 49% similarity, respectively (see Table S1 in the supplemental material). Moreover, microarray results show a constitutive level of expression similar to that of B. burgdorferi flgB (A. Salman-Dilgimen and G. Chaconas, unpublished data). However, DNA repair mechanisms in B. burgdorferi and their importance for infectivity remain poorly understood.

Fig 1.

Schematic of the bacterial NER pathway in response to DNA damage. (A) The damage site (red nucleotides) is recognized by the UvrA dimer, as part of the UvrA₂B₂ tetramer complex. UvrB is then loaded by UvrA to scan the DNA for damage. (B) Once an altered base is found, UvrB remains bound to the undamaged strand at the site of damage, forming the UvrB-DNA preincision complex, while UvrA₂ is released. (C) Following damage recognition, UvrB recruits the UvrC endonuclease, which first cleaves the phosphodiester backbone on the damaged strand 4 or 5 nucleotides 3′ from the damage and then 8 nucleotides 5′ from the damage site. (D) The damaged strand is then removed by the DNA helicase II (UvrD), leaving a gap of single-stranded DNA (E). (F) DNA polymerase I and DNA ligase are then recruited to fill the gap (G).

In the present study, we assessed the importance of 25 B. burgdorferi DNA recombination/repair genes for survival to DNA damage. A previously described fluorescence assay (29) was optimized for B. burgdorferi and used to expediently quantify survival of multiple B. burgdorferi mutants after exposure to DNA-damaging agents. Surprisingly, only disruption of the B. burgdorferi nucleotide excision repair pathway resulted in a significantly increased sensitivity to UV light. Moreover, all four NER genes were shown to be required for repairing nitrosative damage in B. burgdorferi. Finally, disruption of B. burgdorferi uvrA, uvrC, and uvrD only slightly decreased the dissemination time in mice.

MATERIALS AND METHODS

Strains and primers used and culture of B. burgdorferi.

All primers used are listed in Table S2 in the supplemental material. The plasmids and E. coli strains used in this study are listed in Table S3 in the supplemental material, and the B. burgdorferi strains used are listed in Table S4 in the supplemental material. B. burgdorferi was cultivated in BSK-II medium prepared in-house (30) and supplemented with 6% rabbit serum (Cedarlane Laboratories, Burlington, ON, Canada). Cultures were incubated at 35°C with 1.5% CO₂. For samples recovered from mice, 1× Borrelia antibiotic cocktail (20 μg/ml phosphomycin, 50 μg/ml rifampin, 2.5 μg/ml amphotericin B; Sigma-Aldrich, Oakville, ON, Canada) was added to the culture medium.

Determining the plasmid profile of B. burgdorferi.

The plasmid content for each clone was established either by PCR for each plasmid as described previously (31, 32) or by multiplex PCR with minor modifications to the method recently described (33). The multiplex PCR mixture was as follows, as described previously (33): 5 ng of B. burgdorferi genomic DNA, 1 unit Phusion high-fidelity DNA polymerase (New England BioLabs, Pickering, ON, Canada), 1× Phusion HF reaction buffer, 250 μM deoxynucleoside triphosphates, and 1× primer mix in a 20-μl reaction mixture. The circular and the linear plasmids were amplified in separate reactions. PCR products were analyzed by electrophoresis in a 3% Metaphor agarose gel (Lonza, Allendale, NJ). DNA was stained with GelRed nucleic acid gel stain (Biotium, Hayward, CA) and visualized under UV light.

Construction of mutants.

All gene targets except uvrA, uvrB, and uvrD were disrupted as described previously (34). Briefly, approximately 1.5 kb from each gene target was PCR amplified from B. burgdorferi B31 clone 5A4 (32), cloned into the pCR-BluntII-TOPO vector (Invitrogen, Burlington, ON, Canada) or into the pJET1.2/blunt vector (Fermentas, Burlington, ON, Canada), and used to transform E. coli DH5α competent cells. The middle 500 bp of each target was then removed by inverse PCR and replaced with PflgB-aacC1, a flgB-driven gentamicin resistance cassette amplified from pBSV2g (35). The resulting construct was used to transform B. burgdorferi B31 clone 5A4 by electroporation (36, 37). Disruption of uvrA was done essentially using the same strategy, but the complete uvrA gene flanked by an additional 500 bp of upstream sequence and 500 bp of downstream sequence was PCR amplified. The complete uvrA gene was replaced by a gentamicin resistance cassette. In order to ensure recovery of independent clones for uvrA, uvrB, and uvrD mutants, immediately following electroporation of B. burgdorferi, the culture was resuspended in 10 ml of culture medium and separated into five aliquots. After 20 to 24 h incubation at 35°C, each aliquot was diluted 1:12 into fresh culture medium containing 100 μg/ml gentamicin and plated on a 96-well plate (250 μl/well). For the uvrA, uvrB, and uvrD mutants, each independent clone used was recovered from a separate plate. To avoid polar effects from the insertion of the gentamicin resistance cassette, gene inactivation constructs with the gene cassette in both orientations were used. For all the inactivated uvr genes, at least two clones with transcriptional readthrough from the inserted gentamicin resistance cassette in the direction of the adjacent downstream gene were recovered.

Allelic exchange for the replacement of the middle of the gene by the gentamicin resistance cassette, the change in size of the targeted gene, and the insertion junctions of the gentamicin resistance cassette were confirmed by PCR, as previously described (34). Gene deletion was also confirmed by Southern blotting, as previously described (34). B. burgdorferi mutants for dnaK1, exoA, recD, uvrA, and uvrB contain all the B. burgdorferi plasmids tested. Both uvrC mutant strains are missing circular plasmid 9 (cp9), uvrD mutant clone 3 is missing linear plasmid 38 (lp38) and lp28-4, and uvrD clone 4 is missing only lp38.

Complementation of uvrA, uvrB, uvrC, and uvrD.

Complementation of uvrB, uvrC, and uvrD was achieved by allelic exchange (see Fig. S1A in the supplemental material). For all complementing clone constructions, the DNA sequence of the uvr gene from the transforming plasmids was verified. For uvrB, the open reading frame (ORF) was PCR amplified from B. burgdorferi B31 clone 5A4 and fused in order with a flgB-driven kanamycin resistance cassette (PflgB-kan) and the 500-bp sequence downstream of uvrB by overlap extension PCR. The PflgB-kan gene was first amplified from pBSV2 (35). The PCR product was then cloned into the pJET1.2/blunt vector (Fermentas, Burlington, ON, Canada) and used to transform the B. burgdorferi uvrB mutant. For uvrC and uvrD, the first 500 bp of DNA downstream from the target gene was PCR amplified from B. burgdorferi B31 clone 5A4 and inserted into the pJET1.2/blunt vector. PflgB-kan was then PCR amplified from pBSV2 (35) and cloned into the construct using BamHI and XhoI restriction sites. Finally, uvrC and uvrD ORFs were amplified from B. burgdorferi B31 clone 5A4 and cloned into their respective constructs using BamHI and XbaI restriction sites. Each construct was used to transform its respective B. burgdorferi mutant and cultivated in the presence of kanamycin (200 μg/ml) for selection. Furthermore, potential clones were tested for growth in the presence of gentamicin to confirm the replacement of the gentamicin resistance cassette by the wild-type sequence. Allelic exchange for PflgB-kan, the absence of the gentamicin resistance cassette, the presence of the sequence that was deleted in the knockout mutant, and the size of the gene target were confirmed by PCR (see Fig. S1B in the supplemental material). Since all our complementing constructs resulted in restoration of a wild-type phenotype in B. burgdorferi, DNA sequencing of the uvr genes in the B. burgdorferi constructs was not undertaken. For uvrA, reversion of the mutant to the wild type could not be accomplished. Therefore, we generated plasmids for complementation in trans. Two constructs similar to the plasmids previously described by Sambir et al. (38) were assembled. The first construct contained a PflaB-driven uvrA gene, and the second plasmid contained the uvrA ORF with a region upstream of the gene possibly encoding the native promoter. For the first construct, the PflaB promoter was amplified from B. burgdorferi B31 clone 5A4 using primers B2163 and B2164 and cloned into the shuttle plasmid pBSV2 using SacI and KpnI restriction sites. The uvrA ORF was then PCR amplified from B. burgdorferi B31 clone 5A4 using primers B2165 and B2166 and cloned into the construct using KpnI and PstI restriction sites, to obtain pPOH85-1. For the second construct, the uvrA ORF with 500 bp of DNA upstream was PCR amplified using primers B2167 and B2168 and cloned in pBSV2 using SacI and PstI restriction sites, respectively, to obtain pPOH83-2. The complementation constructs were used to transform the B. burgdorferi uvrA mutant. One clone complemented with each construct was used in all experiments testing for uvrA complementation. uvrA-complemented clone PflaB-uvrA is missing lp21, uvrC-complemented clone 2 is missing cp9, uvrC-complemented clone 3 is missing lp28-4 and cp9, uvrD-complemented clone 2 is missing lp28-4, lp36, and lp38, and uvrD-complemented clone 9 is missing lp36 and lp38.

UV light survival assay.

To test the sensitivity of various B. burgdorferi mutants to UV light, cultures were grown to a density of 1 × 10⁷ to 5 × 10⁷ spirochetes/ml, harvested, and resuspended at 1 × 10⁸ bacteria/ml in phosphate-buffered saline (PBS). For each sample, 10⁷ cells were exposed on a 35-mm petri dish to 0, 1, 2, or 3 mJ/cm² of 254-nm UV light (Stratalinker UV cross-linker; Stratagene, La Jolla, CA). UV exposure conditions were empirically chosen to result in about 80% survival of the wild-type B. burgdorferi strain at the maximal dose, with our starting point based upon previously reported conditions (38–40). The UV exposure range was similar to that typically used in E. coli. Immediately following exposure to UV light, the samples were diluted 1:100 in fresh culture medium and incubated for 36 to 48 h. The growth of each clone was determined using a method adapted from a previously reported protocol (29) (Fig. 2A). The cell density of untreated wild-type B. burgdorferi was determined by counting with a dark-field microscope. When the culture reached 5 × 10⁷ to 1 × 10⁸ cells/ml, the same volume from each sample (equivalent to 10⁸ spirochetes for the counted culture) was centrifuged and spirochetes were washed twice with PBS. Cells were resuspended in 90 μl of water and incubated at 95°C for 5 min. The DNA was stained by adding 10 μl of 34 μM Syto-9 (Invitrogen, Burlington, ON, Canada) in a 96- or 384-well plate and incubated for 30 min at room temperature. The fluorescence was measured using a plate reader (Wallac 1420 VICTOR²; PerkinElmer, Waltham, MA) with a 0.5-s exposure time using a fluorescein filter. The percent relative fluorescence represents the fluorescence after UV light exposure compared to that after no exposure for a particular strain. Two independent clones were used for each mutated gene target. Experiments were done in triplicate.

Fig 2.

Method used to assess the sensitivity of B. burgdorferi to UV light. (A) Flowchart of the method used. Each B. burgdorferi clone was resuspended in PBS and exposed to the indicated dose of UV light. Each sample was then diluted 1:100 in BSK-II and cultured until the untreated control reached late exponential phase. Spirochetes were lysed, the DNA was stained with Syto-9, and the fluorescence of each sample was measured using a plate reader. The method was adapted from a previous study (29). (B) Correlation between the number of spirochetes per sample and the level of fluorescence observed in the assay described in the legend to panel A. Samples were analyzed in triplicate, and the mean and standard deviation were plotted.

ESS assay.

The endonuclease sensitivity site (ESS) assay was adapted from previous studies (41, 42) for B. burgdorferi (Fig. 3A). When a culture of B. burgdorferi lacking uvrC (GCB 537) reached 5 × 10⁷ spirochetes per ml, cells were resuspended in PBS at 1.5 × 10⁸ spirochetes per ml. Cells either were left untreated or were exposed to 15 mJ/cm² of 254-nm UV light (Stratalinker UV cross-linker; Stratagene, La Jolla, CA) in 100-μl aliquots in 35-mm cell culture dishes (BD Falcon, Franklin Lakes, NJ). Immediately following exposure, cells were lysed (100 mM Tris, pH 8.5, 10 mM EDTA, 30 mM NaCl, 0.5% SDS) and treated with RNase A (60 μg/ml) and then proteinase K (75 μg/ml). Genomic DNA was extracted with phenol-chloroform. To detect pyrimidine dimers, 1.2 μg of genomic DNA was either left undigested or incubated with either 10 or 15 units of T4 pyrimidine dimer glycosylase (PDG; New England BioLabs, Pickering, ON, Canada), following the supplier's instructions, for 30 min at 37°C. The reaction was then stopped by addition of an alkaline loading buffer (100 mM NaOH, 1 mM EDTA, 2.5% Ficoll, 0.05% bromocresol green) and loaded on a 10-cm, 0.5% alkaline agarose gel (42) run in alkaline buffer (300 mM NaOH, 10 mM EDTA) for 16 h at 15 V with buffer recirculation. Following electrophoresis, the gel was incubated for 2 h in neutralization buffer (1 M Tris, 1.5 M NaCl, pH 7.4) and the DNA was stained with SYBR gold (Invitrogen, Burlington, ON, Canada).

Fig 3.

UV-induced DNA damage in B. burgdorferi. (A) Flowchart representation of the ESS assay used to detect DNA damage in B. burgdorferi DNA after exposure to UV light (41, 42). Briefly, B. burgdorferi was resuspended in PBS and exposed to UV light. Genomic DNA was then extracted and digested with T4 PDG before being run on an alkaline gel to determine the presence or absence of damage induced by UV light. (B) SYBR gold-stained agarose gel from an ESS assay to monitor B. burgdorferi DNA damage by UV light. Genomic DNA recovered from a B. burgdorferi uvrC mutant exposed to 0 or 15 mJ/cm² of UV light, as indicated, was either left undigested or digested with 10 (+; lanes 2 and 5) or 15 (++; lanes 3 and 6) units of T4 PDG.

Reactive nitrogen survival assay.

Sensitivity to reactive nitrogen species (RNS) damage was determined as previously described (43). Cultures were grown to a density of 1 × 10⁷ to 5 × 10⁷ cells ml⁻¹, and then spirochetes were pelleted and resuspended in culture medium at 5 × 10⁷ spirochetes per ml with 2.5 mM diethylamine (DEA) NONOate diethylammonium salt (DEA/NO; Sigma-Aldrich, Oakville, ON, Canada) or with culture medium only. Samples were then incubated for 4 h at 35°C, diluted 1:100 in fresh culture medium, and incubated at 35°C until the culture of treated wild-type B. burgdorferi B31 clone 5A4 reached 5 × 10⁷ spirochetes per ml. Spirochete densities were determined using a Petroff-Hausser counting chamber (VWR, Edmonton, AB, Canada) on a dark-field microscope. The percent survival for each mutant strain was determined as the number of surviving spirochetes divided by the number of surviving spirochetes in the treated wild-type strain. Experiments were done in triplicate using two clones for each mutated gene target.

Mouse infection studies.

All animal studies were carried out in accordance with the principles outlined in the most recent policies and the Guide to the Care and Use of Experimental Animals of the Canadian Council on Animal Care. Our animal protocol (AC12-0070) was approved by the Animal Care Committee of the University of Calgary. Each 3- to 4-week-old male C3H/HeNCrl mouse (Charles River, St-Constant, QC, Canada) was infected by both subcutaneous and intraperitoneal injection of 10³ spirochetes at each site. Three mice were infected with each clone, and two independent clones were used for each NER gene disrupted, resulting in a total of 6 mice infected per gene target. On day 7 postinfection, approximately 50 μl of blood was recovered from the saphenous vein and diluted into culture medium for the presence of B. burgdorferi. On days 14 and 21, two ear punch samples were taken, and on day 23, the heart, the ear, the bladder, and the knee joint were collected and transferred into culture medium for growth of spirochetes. Cultures were considered positive for B. burgdorferi when spirochetes could be observed by dark-field microscopy.

RESULTS

An expedient fluorescence assay for B. burgdorferi sensitivity to DNA damage.

Previous studies on testing DNA damage in B. burgdorferi were based on a protocol designed for E. coli, where treated samples were plated and cultured until colonies could be counted (39, 40, 44). However, solid plating requires incubation for 2 to 3 weeks before individual B. burgdorferi colonies can be observed. To more expediently monitor the DNA damage of two clones for each of 25 mutated target genes, a more efficient strategy was adapted from that of Kim and Surette (29), where the growth of swimmer and swarmer populations of Salmonella enterica serovar Typhimurium were followed and compared by measuring the total concentration of DNA using a fluorescent stain.

The adaptation of this strategy for B. burgdorferi (Fig. 2A) is described in Materials and Methods. As an easy, controllable, and reproducible method of inducing DNA damage, we chose exposure to UV light. Briefly, following UV light irradiation, cells were diluted 1:100 in culture medium and incubated for 48 h until they reached late exponential phase. Cells were then washed with PBS and lysed, and the DNA was stained by the addition of the fluorescent dye Syto-9. The level of fluorescence was measured in 100-μl samples using a plate reader. PBS washes prior to lysis were required because residual BSK-II medium, even if it lacks phenol red, generated a strong fluorescence background. It is important to note that the fluorescence assay quantifies only living cells that have grown after a 48-h incubation period following the UV irradiation. Dead cells do not grow, and even when all spirochetes are killed by irradiation, following the 1:100 dilution, their concentration is too low to contribute any background fluorescence. The cell density of a given UV light-treated strain versus that of the untreated strain after 48 h of outgrowth was used as a measure of the UV sensitivity of the strain. Before using this strategy to compare the growth of cultures treated with UV light, serial dilutions of a growing B. burgdorferi culture were used to confirm a correlation between the number of cells and the amount of fluorescence measured (Fig. 2B). When lysed samples contained between 10⁷ and 10⁸ spirochetes per 100 μl, the level of fluorescence detected reflected the number of cells lysed. This positive correlation established that the fluorescence assay can be used as an alternative strategy to rapidly estimate the growth of multiple liquid cultures of B. burgdorferi.

Most replication/repair genes do not affect UV sensitivity in B. burgdorferi.

A wide variety of E. coli genes are involved in resistance to UV radiation. However, only a few B. burgdorferi mutants have been tested for UV sensitivity (39, 40). Surprisingly, disruption of neither the recA recombinase nor the Holliday junction branch migrase subunit ruvA resulted in a significant effect upon B. burgdorferi survival. In the present study, sensitivity to UV light of 21 mutant genes not part of the nucleotide excision repair pathway in B. burgdorferi was assessed. For each mutated gene, two independent clones were exposed to 0, 1, 2, or 3 mJ/cm² UV light, and their capacity to grow was determined using the fluorescence assay. For each UV dose, the relative fluorescence of the mutant was compared to that of the wild-type strain to establish its level of sensitivity (see Fig. S2 in the supplemental material). The disruption of dnaK1, bbg32, exoA, hrpA, mag, mfd, mutL, mutS1, mutS2, nth, nucA, priA, recA, recD, recG, recJ, rep, ruvA, ruvB, sbcC, or sbcD did not result in a significant increase in B. burgdorferi sensitivity to UV light. Disruption of mutS1 and sbcC (but not sbcD) did result in a significant decrease in B. burgdorferi sensitivity to UV radiation only at the highest dose used (3 mJ/cm²). The magnitudes of the decrease in sensitivity were only 0.7- and 1.5-fold different from the sensitivity of the wild type for sbcC and mutS1, respectively.

Mutants for 12 genes known to increase E. coli sensitivity to UV light have not been tested in B. burgdorferi for various reasons. No mutants could be recovered for ssb, recB, recC, and ligA. dnaX encodes the τ and γ subunits of DNA polymerase III and so was expected to be essential, as it is in E. coli (45, 46), and disruption of pnp was not attempted because its impact on E. coli resistance to UV light has been shown only more recently (47). Finally, no orthologs of recX, ruvC, recF, recR, recO, and mutH could be found in the B. burgdorferi sequenced genome (28).

B. burgdorferi DNA is susceptible to UV damage.

The lack of an effect from the mutation of 21 replication/repair genes on UV sensitivity in B. burgdorferi suggested the possibility that B. burgdorferi was significantly resistant to UV damage at the DNA level. In order to detect UV-induced DNA damage, B. burgdorferi genomic DNA was extracted after exposure to UV light and analyzed using an endonuclease sensitive site (ESS) assay (Fig. 3A) (41, 42, 48). In this assay, the damaged DNA is incubated with T4 pyrimidine dimer glycosylase (PDG), an endonuclease that recognizes cis-syn cyclobutane pyrimidine dimers (CPD), the most common type of DNA damage generated by UV-C light (1, 49, 50). The PDG-treated DNA is then separated by alkaline gel electrophoresis to detect single-strand DNA breaks. In E. coli, the ESS assay allows determination of the number of breaks/damaged sites generated in the DNA. However, because of the segmented nature of its genome, this was not possible in B. burgdorferi. Nonetheless, exposing B. burgdorferi to both UV light and PDG (Fig. 3B, lanes 5 and 6) clearly increased the migration rate of the DNA compared to that for DNA from UV-treated B. burgdorferi that was not treated with PDG (lane 4) or DNA from non-UV-treated cells that was incubated with PDG (lanes 2 and 3). This confirms that UV light does generate pyrimidine dimers in B. burgdorferi DNA.

All four NER genes are required to repair UV-induced and nitrosative DNA damage.

Although many genes influence E. coli resistance to UV light, disruption of the NER excinuclease UvrABC has the most dramatic impact and disruption of uvrD results in a more intermediate sensitivity phenotype (51). To confirm if the NER pathway has a similar role in B. burgdorferi, mutants for the uvrA, uvrB, uvrC, and uvrD genes were tested for their resistance to UV light using the fluorescence assay. As described for the other replication/repair gene mutants, each clone was exposed to 0, 1, 2, and 3 mJ/cm² of UV light and assayed for its level of survival, as described in Fig. 2A. The level of fluorescence for all four mutants treated with UV light was similar to the background level, even at 1 mJ/cm², the lowest dose of UV light used (Fig. 4). The uvrA data confirm those from a previous report by Sambir et al. showing that this gene is required for B. burgdorferi survival after exposure to UV light (38). Complementation of each NER gene restored B. burgdorferi UV resistance to a level similar to that for the wild-type strain. We conclude that all four uvr genes are required for survival of B. burgdorferi after UV light exposure.

Fig 4.

Sensitivity of uvr mutants to UV light. The fluorescence assay was used to evaluate the sensitivity of uvrA, uvrB, uvrC, and uvrD mutants (gray lines) and of their respective complemented strains (black dotted lines) compared to that of the wild type (solid black lines). Two clones for each mutant and their respective complemented clones were exposed to 0, 1, 2, and 3 mJ/cm² of 254-nm UV light. For the complemented uvrA mutant, both complemented clones were used and the data from the two were averaged. The percent relative fluorescence represents the fluorescence after UV light exposure compared to that after no exposure for a particular strain. Error bars represent the standard deviation from at least three experiments done with two clones per mutant. ***, P value corresponding to <0.001 for the comparison of a given mutant and the wild-type control for a given UV dose.

Previous studies showed that an E. coli uvrD mutant is not as sensitive to UV light as the uvrA, uvrB, and uvrC mutants (51) (see Table 2), but such a difference was not observed in B. burgdorferi when the fluorescence assay was used to compare culture densities. However, this assay has a limited dynamic range. We therefore used direct dark-field counting to investigate growth differences greater than 10-fold in B. burgdorferi wild-type and uvr knockout strains. Spirochetes were exposed to either 0 or 1 mJ/cm² UV light and cultured as described for the fluorescence assay. When the untreated wild-type B. burgdorferi strain reached a cell density of between 5 × 10⁷ and 1 × 10⁸ spirochetes per ml, the density of each culture was determined by direct counting using a dark-field microscope. The level of survival of each clone treated with UV light was then compared to that of the untreated culture for a particular strain. Similar to the results for E. coli, disruption of B. burgdorferi uvrD resulted in a 10- to 20-fold increase in sensitivity to UV light, while exposure of uvrA, uvrB, and uvrC mutants resulted in a more than 1,000-fold increase in sensitivity (Fig. 5). The cell density of the treated uvrA, uvrB, and uvrC mutant cultures was below 5 × 10⁴ spirochetes per ml, the lowest density that can be determined by direct counting. The correspondence of UV sensitivity for the uvrA, uvrB, and uvrC versus uvrD mutants in both B. burgdorferi and E. coli underscores the functional similarity of the NER pathway in both organisms.

Table 2.

Comparison of UV sensitivity in B. burgdorferi with that in E. coli^a

^a1, genes are listed in order of decreasing sensitivity in E. coli. ², ++, highly sensitive; +, slightly sensitive; −, nonsensitive; absent, no ortholog found in B. burgdorferi; ?, sensitivity is not known in B. burgdorferi. ³, based on the literature. The sensitivity of the exoA and mag mutant is based on Bacillus subtilis and Saccharomyces cerevisiae. The mutant for E. coli alkA, the mag ortholog, does not present increased sensitivity to UV light. ⁴, sensitivity based on the sbcCD mutant in E. coli. ssDNA, single-stranded DNA; dsDNA, double-stranded DNA.

Fig 5.

Survival of B. burgdorferi uvr mutants after exposure to UV light. To detect a range of effect greater than that possible with the fluorescence assay, the cell density of each culture was determined by direct counting using a dark-field microscope. Each clone was exposed to 0 and 1 mJ/cm² of UV light and cultured as described when the fluorescence assay was used to compare sensitivity to UV light. When the culture of untreated wild-type (wt) B. burgdorferi reached a cell density of between 5 × 10⁷ and 1 × 10⁸ spirochetes per ml, the cell density of every culture was determined and is shown on the graph. Cell density could not be determined if less than 5 × 10⁴ spirochetes per ml were present. Error bars represent standard deviations from three experiments done using two clones per mutated gene.

Although reactive nitrogen species (RNS) have been shown to kill B. burgdorferi primarily through S-nitrosylation of proteins, disruption of uvrB or uvrC (43), but not uvrA (38), has also been shown to significantly reduce B. burgdorferi survival to RNS-induced DNA damage. We assessed the importance of all four genes in the NER pathway in response to RNS stress by comparing the growth of B. burgdorferi lacking uvrA, uvrB, uvrC, or uvrD with that of a wild-type strain following exposure to the nitric oxide (NO) donor DEA/NO. For this experiment, spirochete numbers were determined by direct counting using a dark-field microscope. This allowed detection of a range of effect greater than that obtained by the fluorescence assay but limited the number of cultures that could be analyzed. All four B. burgdorferi NER mutants showed between 10- and 20-fold decreases in survival compared to that of the wild-type strain (Fig. 6). There was no significant difference between the wild-type strain and any of the complemented clones. In addition to B. burgdorferi NER mutants, the sensitivity of the mfd, sbcC, exoA, ruvB, recG, mutS, recA, nth, recJ, and priA mutants was also tested, but none of the mutants displayed an increased sensitivity to DEA/NO (data not shown). Our results point to a requirement for all genes of the NER pathway for B. burgdorferi to repair nitrosative DNA damage and once again suggest that as for UV damage, the NER system is the sole pathway for repair of nitrosative DNA damage. There was no effect of oxygen radicals on the survival of mutants in the NER pathway (data not shown), as expected on the basis of previous work that established that oxidative damage is limited to the outer membrane and does not damage DNA in B. burgdorferi (52). This is because of a lack of iron in the cell and a corresponding absence of Fenton chemistry (53).

Fig 6.

Sensitivity of uvr mutants to nitrosative damage. Survival of uvrA, uvrB, uvrC, and uvrD mutants compared with that of their respective complemented clones after B. burgdorferi exposure to 2.5 mM DEA NONOate. After treatment, the samples were diluted in fresh culture medium and cultivated until the treated wild-type culture reached 5 × 10⁷ cells/ml. Spirochete densities were determined using a counting chamber in a dark-field microscope. The percent survival compares each particular mutant strain to the wild-type B. burgdorferi strain after treatment. Error bars represent standard deviations from three experiments done using two clones per mutant. ***, P value corresponding to <0.001 for a given UV dose.

Disruption of the B. burgdorferi NER system had a minor effect on murine infection.

The nucleotide excision repair pathway is required for survival to reactive nitrogen stress, which is known to affect the infectivity of pathogens (54). Although previous studies have described the role of some NER genes for resistance to DNA damage in B. burgdorferi (38, 43), our current study is the first time that the importance of the NER pathway for infectivity has been tested. In order to detect a difference in infectivity between each mutant, mice were infected with 10³ spirochetes at each of two sites at a dose close to the minimum infectious dose required for infection of all mice and for dissemination to all target organs tested with strain B31 (55). For each uvr gene, two independent clones were used to infect three mice each, resulting in six mice per NER gene mutant. Infectivity was determined from a blood sample taken at 7 days postinfection, and dissemination was monitored by the collection of ear punch samples at 14 and 21 days postinfection. On day 23, the ear, the heart, the bladder, and the knee joint were recovered to determine if the NER genes were required for the invasion of specific organs.

At 1 week postinfection, spirochetes were cultured from the blood of all mice infected with the wild-type strain and with uvrA and uvrB mutant strains. The blood from one mouse inoculated with a uvrC mutant was culture negative, and only 50% of mice infected with the uvrD mutants contained spirochetes in the blood at day 7 (Table 1). These data indicate the full infectivity of spirochetes carrying mutations in uvrA or uvrB and perhaps a small decrease in infectivity for spirochetes carrying mutations in uvrC and uvrD.

Table 1.

Infectivity of B. burgdorferi carrying mutant NER genes in C3H/HeN mice

Genotype	Strain	Blood, day 7	Ear		Organs, day 23
			Day 14	Day 21	No. of sites culture positive/no. of sites tested					%a
		No. of sites culture positive/no. of sites tested (%)	No. of sites culture positive/no. of sites tested (%)	No. of sites culture positive/no. of sites tested (%)	Ear	Bladder	Heart	Joint	Total sites
Wild type	920	6/6 (100)	4/6 (66.7)	6/6 (100)	6/6	5/6	6/6	6/6	23/24	95.83
uvrA	564	3/3 (100)	0/3 (0)	0/3 (50)	3/3	3/3	3/3	3/3	12/12	100
	565	3/3	0/3	3/3	3/3	3/3	3/3	3/3	12/12
uvrB	545	3/3 (100)	2/3 (66.7)	3/3 (100)	3/3	1/3	3/3	3/3	10/12	87.5
	546	3/3	2/3	3/3	3/3	3/3	2/3	3/3	11/12
uvrC	537	3/3 (83.3)	1/3 (50)	2/3 (83.3)	3/3	3/3	3/3	3/3	12/12	83.3
	538	2/3	2/3	3/3	2/3	2/3	2/3	2/3	8/12
uvrD	541	1/3 (50)	0/3 (16.7)	0/3 (16.7)	2/3	2/3	1/3	2/3	7/12	70.83
	542	2/3	1/3	1/3	3/3	3/3	1/3	3/3	10/12

^aAt day 23, 100% of the needle-inoculated mice were infected, as determined by positive spirochete cultures from at least 1 tissue specimen.

Cultures of 2-mm ear punch samples at week 2 are a sensitive indicator of dissemination efficiency, as the small amount of tissue cultured will be negative without a substantial spirochete density in the ear. Cultures of ear punch samples from mice infected with the wild-type and uvrB mutant strains were positive for 67% of the mice infected, while cultures of ear punch samples from mice infected with the uvrC, uvrD, and uvrA mutants were positive at day 14 for 50%, 17%, and 0% of the infected mice, respectively (Table 1). These data suggest a reduced dissemination efficiency for the uvrA and uvrD mutants. This reduced dissemination into the ear persisted at day 21 (50% for uvrA and 17% for uvrD; Table 1).

Organs collected for culture on day 23 postinfection provided a general picture of dissemination to a variety of sites but provided a greater sensitivity than 2-mm ear punch samples, as the entire organ was cultured. At day 23 (Table 1), 100% of the needle-inoculated mice were infected and mutants for all four NER genes had successfully invaded the majority of the organs recovered.

In summary, at a low infectious dose, mutations in the NER pathway resulted in a minor decrease in spirochete burden in the blood at day 7 and in the efficiency of dissemination to the ear. Nonetheless, by day 23, invasion of ear, bladder, heart, and joint had occurred for all the uvr mutants. The attenuation in infectivity observed in all cases remained minor and did not affect the overall dissemination of the spirochetes.

DISCUSSION

Sensitivity of B. burgdorferi nucleotide excision repair mutants to DNA damage.

A fluorescence assay originally described to compare the growth of Salmonella enterica serovar Typhimurium cultures (29) was adapted to expediently compare multiple B. burgdorferi liquid cultures. This assay was used to evaluate the sensitivity to DNA damage of two clones for each of the 29 B. burgdorferi constructs described in this study. Although solid plating and direct microscopic counting can detect variations over a wider range of cell densities, both techniques require significant effort, limiting the number of samples that can be analyzed. Plating also requires a 2-week turnaround for results, due to the slow growth of B. burgdorferi.

In this study, the importance of various B. burgdorferi DNA replication/repair genes for survival to DNA damage was assessed. While the disruption of the B. burgdorferi uvrA, uvrB, or uvrC gene resulted in a complete loss of measurable survival following UV-induced DNA damage, an intermediate sensitivity was observed in the uvrD mutant, similar to what has been observed in E. coli (Fig. 5). Although the sensitivity of a B. burgdorferi uvrA mutant was previously reported (38), this is the first time that the importance of all four genes of the B. burgdorferi NER pathway has been shown to be similar to that of the genes of the E. coli NER pathway for survival to UV light-induced DNA damage.

The importance of the NER pathway to repair DNA damage in B. burgdorferi was also demonstrated for survival to nitrosative stress. A previous study has reported that disruption of B. burgdorferi uvrB and uvrC results in decreased survival following 4 h exposure to the NO donor DEA/NO (43). However, in another study, using a different strain of B. burgdorferi exposed for 1 h to different sources of NO, disruption of uvrA did not result in significantly increased sensitivity (38). The difference between the results of these two studies might be explained by the use of different NO donors or different times of exposure to the damaging agent. In the work reported here, disruption of any of the four B burgdorferi NER genes resulted in a similar significant increase in sensitivity to NO. These results further demonstrate the importance of the NER pathway for B. burgdorferi survival to chemical DNA damage.

Lack of sensitivity of B. burgdorferi replication/repair mutants to DNA damage.

Surprisingly, disruption of 21 replication/repair genes not part of the NER pathway did not result in significant increases of sensitivity to DNA damage. This includes mutation of 11 genes previously reported to increase E. coli sensitivity to a dose of UV light similar to that used for B. burgdorferi in the present study (Table 2). The absence of an impact is not due to a lack of sensitivity of the method used since the fluorescence assay is sensitive enough to detect a change in cell density of less than 2-fold. A small but significant increase in resistance to UV light was observed in mutS1 and sbcC mutants (see Fig. S2 in the supplemental material), although the meaning and possible biological significance of the mutS1 and sbcC results are unknown.

An interesting difference between B. burgdorferi and E. coli is the apparent lack of an SOS response in B. burgdorferi. In the presence of DNA damage, the E. coli SOS response increases the expression of over 40 genes, including DNA recombination/repair genes uvrA, uvrB, uvrD, recA, ruvA, ruvB, and ruvC (1, 56–59). There is no ortholog for the SOS response regulator lexA in the B. burgdorferi genome (28). Moreover, the expression of recA, ruvB, and the uvr genes in B. burgdorferi (see Table S1 in the supplemental material) occurs, without induction (40, 60–63), at a level close to that of the constitutively transcribed flgB gene (A. Salman-Dilgimen, personal communication). The lack of a requirement for recA to promote SOS induction in B. burgdorferi partially explains the lack of sensitivity of B. burgdorferi recA mutants to DNA damage, since induction of the NER pathway is apparently not needed.

In contrast to our results, disruption of recA, ruvABC, recG, recBCD, priA, and mfd in Neisseria gonorrhoeae, also an obligate parasite missing the SOS response (64), results in a significant increase in sensitivity to UV light (64–69). These results suggest that, even in the presence of a functional NER system, recombinational repair (recA, ruvABC, recG, and recBCD), transcription-coupled repair (mfd), and repair of arrested replication forks (recG, priA) play a significant role in the repair of UV damage in N. gonorrhoeae, but not in B. burgdorferi. No orthologs of sbcCD (70) are found in the sequenced genome of N. gonorrhoeae, but their disruption did not affect B. burgdorferi survival to UV damage. Other recent studies have also reported that B. burgdorferi recA and ruvA are dispensable for repairing UV-induced DNA damage (39, 40). Recombinational repair of UV and nitrosative damage may well be occurring at low levels in B. burgdorferi; however, experiments to detect this would require recA uvr double mutants.

Our results also show that disruption of the methyl-directed mismatch correction (MMC) pathway (mutS and mutL) does not affect survival of B. burgdorferi after exposure to UV light but does affect survival of E. coli. Interestingly, no potential ortholog of the MutH endonuclease has been identified in the B. burgdorferi genome (28). N. gonorrhoeae is also lacking a mutH ortholog, and similar to the findings for B. burgdorferi, the MMC system is dispensable for DNA damage repair (71). However, disruption of N. gonorrhoeae MMC genes results in an increase of the spontaneous mutation rate due to base-pairing errors, suggesting a functional MMC pathway, even in the absence of mutH (71). In B. burgdorferi, disruption of mutS, but not mutL, resulted in a small but significant decrease in sensitivity to UV light. Whether this small difference has any biological significance remains unknown. A similar phenotype is observed for recD and recJ mutants. As opposed to N. gonorrhoeae (72, 73), disruption of E. coli (74–76) and B. burgdorferi recD and recJ does not result in an increased sensitivity to UV light. The possible role of dnaK (77), pnp (47), rep (78), and hbb (related to the HU/IHF family) (79, 80) in DNA damage repair also remains unclear. Sensitivity to UV light was also tested for B. burgdorferi exoA, mag, hrpA, nth, bb0098, nucA, and bbg32 mutants because of their predicted role in DNA replication, recombination, and repair (28). However, none of these mutants had an increased sensitivity to UV light compared to that of wild-type B. burgdorferi.

Effect of B. burgdorferi NER mutations on murine infection.

The nucleotide excision pathway is required for full infectivity of various pathogens. Garbom et al. showed that disruption of the Yersinia pseudotuberculosis virB gene, an ortholog of uvrA, results in a significant increase of the 50% lethal dose for mouse infection (81). In Mycobacterium tuberculosis, transcription of uvrA is significantly increased after the organism is internalized by human macrophages (82). Also, M. tuberculosis uvrB mutants are attenuated for infection in wild-type and inducible nitric oxide synthase-knockout (iNOS^−/−) mice but not in iNOS/phox-deficient mice, suggesting that the nucleotide excision repair pathway might be important for M. tuberculosis survival to phox-mediated stress (54). In B. burgdorferi, our data demonstrate that all four NER pathway genes are involved in survival to nitrosative stress.

B. burgdorferi has been shown to induce the release of NO from macrophages (25) and recruitment of macrophages and neutrophils to organs where it disseminates (22, 24, 83). However, previous studies showed that inhibition of iNOS using N^G-l-monomethyl arginine (LMMA) does not significantly affect the bacterial burden in the heart and joint of C3H/HeJ or BALB/c mice infected with B. burgdorferi (84) and has only a minor effect on survival of the bacterium in the presence of macrophages (25). Although B. burgdorferi appears to be resistant to nitrosative DNA damage, Bourret et al. showed that disruption of either uvrB or uvrC results in a significant increase in sensitivity to DNA damage (43). Our study demonstrates that a complete NER pathway is required for B. burgdorferi survival to nitrosative damage. This suggests that B. burgdorferi NER mutants could be impaired in mouse infection. However, our data show that disruption of uvrA and uvrD had only a minor effect upon murine infection. B. burgdorferi uvrA and uvrD mutants appeared to be attenuated, as determined by culture of ear punch samples, but did disseminate to all the organs tested by day 23 postinfection. This suggests that during mouse infection, B. burgdorferi would not be exposed to levels of nitrosative stress sufficient to prevent infectivity and dissemination of a sensitive strain.

Disruption of uvrD had the most noticeable effect, with decreased proportions of blood and ear biopsy specimens being culture positive on days 7 to 21. This could reflect a role for the UvrD helicase in other repair pathways, as described in E. coli. Previous studies showed that UvrD is also involved in MMC in E. coli (6, 85), but it was shown that disruption of B. burgdorferi mutS and mutL does not affect B. burgdorferi infectivity and dissemination to the ear (34). However, the mice in that study were infected with a higher number of spirochetes, and in the presence of a functional NER system, disruption of MMC might not have generated a discernible phenotype.

In conclusion, this study demonstrated that B. burgdorferi requires a functional NER system for survival of nitrosative and UV light-induced DNA damage and that the pathway appears to be conserved between E. coli and B. burgdorferi. Moreover, disruption of uvrA and uvrD had only a minor effect upon the infectivity of B. burgdorferi in mice. Further studies will be required to investigate the role of the NER pathway during the tick/mammal interphase or in the tick.