The Fur Homologue in Borrelia burgdorferi

Abstract

Borrelia burgdorferi contains a gene that codes for a Fur homologue. The function of this Fur protein is unknown; however, spirochetes grown at 23 or 35°C expressed fur as determined by reverse transcriptase PCR. The fur gene (BB0647) was cloned and overexpressed as a His-Fur fusion protein in Escherichia coli. The fusion protein was purified by zinc-chelate chromatography, and the N-terminal His tag was removed to generate recombinant Fur for use in mobility shift studies. Fur bound DNA containing the E. coli Fur box sequence (GATAATGATAATCATTATC) or Bacillus subtilis Per box sequence (TTATAAT-ATTATAA) with an apparent K_d of ∼20 nM. Fur also bound the upstream sequences of three Borrelia genes: BB0646 (gene encoding a hydrolase of the α/β-fold family), BB0647 (fur), and BB0690 (napA). Addition of metal ions was not required. Binding activity was greatly decreased by either exposure to oxidizing agents (H₂O₂, t-butyl hydroperoxide, cumene hydroperoxide, or diamide) or by addition of Zn²⁺. B. burgdorferi NapA is a homologue of Dps. Dps functions in E. coli to protect DNA against damage during periods of redox stress. Fur may function in B. burgdorferi as a repressor and regulate oxidative stress genes. Additional genes (10 chromosomal and 15 plasmid) that may be Fur regulated were identified by in silico analysis.

Lyme disease is a tick-borne illness caused by the spirochete Borrelia burgdorferi (7, 12, 59). To infect a human (or other mammalian host), the spirochete must move from the gut of the tick (where it normally resides), through the gut wall, into the hemolymph, and to the salivary glands, where it gets deposited with the saliva at the site of tick feeding (8, 52, 67). From here it first moves through the skin, enters the bloodstream, and then disseminates to distant organs (9). Many virulence factors are required in order for the spirochete to complete this journey. Which genes are expressed in the tick versus the host has been a subject of much recent interest. How the spirochete orchestrates its response to its changing environment is essentially unknown (66).

Fur (ferric uptake regulation protein) is a transcriptional repressor that regulates expression of genes involved in iron uptake and iron storage (4). Under high-iron conditions, Fur binds with its corepressor Fe²⁺ to sites (Fur boxes) located within the promoter region of iron-regulated genes and in so doing blocks transcription. Under low-iron conditions, Fur dissociates from its corepressor and DNA, allowing transcription to proceed (5).

Fur can also act as a global regulator and control expression of genes unrelated to iron uptake. For example, Fur may regulate the expression of the chemotaxis/motility genes in Escherichia coli and perhaps also in Vibrio cholerae (42, 60). Some have suggested that a shift from a high-iron to a low-iron environment may signal the bacteria that it has entered a host (36, 38, 45, 63). By sensing iron, Fur can direct changes in gene expression in response to a change in the environment.

Fur-like proteins can also regulate other functions. For example, Zur regulates uptake of zinc in E. coli (44) and in B. subtilis (24). Unlike Fur, Zur seems not to function as a global regulator (28). PerR, CatR, FurA, and FurS are homologues of Fur that function to regulate the response to oxidative stress (29, 39). Most act as repressors (like Fur) but then lose this activity upon exposure to hydrogen peroxide or other oxidizing agents.

According to the genome sequence, B. burgdorferi contains a gene that codes for a Fur homologue (22). Because B. burgdorferi does not require iron for growth (47), it seems unlikely that this Fur acts to regulate iron uptake. It may still, however, act as a “global regulator” and control expression of genes in response to the level of iron sensed in the environment. Alternatively, it may function as Zur and regulate zinc uptake or as PerR and regulate the response to oxidative stress. Or it may function in some other way that is yet to be defined.

Recently, Boylan et al. (10) renamed the Fur protein in Borrelia “BosR” (for Borrelia oxidative stress regulator). They propose that BosR functions in Borrelia as a zinc-dependent transcriptional activator of oxidative stress genes. In their report, they show that the oxidizing agent t-butyl hydroperoxide induces the expression of napA (neutrophil-activating protein) in Borrelia. They identify by footprint analysis a BosR binding site upstream of the napA promoter, which they note is atypical of Fur binding sites owing to its location (180 nucleotides [nt] upstream of the transcriptional start site) and large size (50 nt). In gel shift assays, they show that optimal binding requires Zn²⁺ and dithiothreitol (DTT) and that exposure to t-butyl hydroperoxide enhances binding. In transcriptional fusion studies done in E. coli, they show that BosR activates transcription from the napA promoter and that exposure to t-butyl hydroperoxide increases this expression.

napA (a homologue of dps) is one of only a few genes identified thus far in Borrelia that may participate in an oxidative stress response (22). In E. coli, Dps functions to protect DNA against damage during redox stress by binding DNA nonspecifically and by sequestering iron to prevent generation of hydroxyl radicals via the Fenton reaction (4). How NapA might function in B. burgdorferi is currently unclear.

As a first step toward defining the function of Fur in Borrelia, we have prepared recombinant Fur and used it to carry out a series of mobility shift studies. We show that this Fur binds with high affinity to both Fur and Per box sequences. Furthermore, we show that this binding does not require Zn²⁺ and that exposure to peroxides decreases binding rather than enhancing it. These data support a role for Fur as a repressor in Borrelia. Whether the Borrelia Fur can function as both an activator and repressor is an area for future study.

MATERIALS AND METHODS

Bacteria and culture conditions.

B. burgdorferi B31-MI was available from an earlier study (40). Spirochetes were grown in BSK-H medium containing 6% rabbit serum (Sigma-Aldrich, St. Louis, Mo.) at 23 and 35°C. E. coli strains were grown at 37°C, with vigorous shaking, in Luria-Bertani broth supplemented with the appropriate antibiotics at the following concentrations: ampicillin (50 μg/ml), kanamycin (30 μg/ml), or tetracycline (15 μg/ml). Competent cells of E. coli DH5α were obtained from Invitrogen Corp. (Carlsbad, Calif.); competent cells of E. coli BL21(DE3) and E. coli NovaBlue were obtained from Novagen, Inc. (Madison, Wis.). The Institute of Genomic Research (TIGR) sequencing clones GBBEG25, GBBDA28, and GBBBM12 were obtained as E. coli SURE2 cultures from the American Type Culture Collection (Manassas, Va.).

DNA manipulations and sequencing.

Standard procedures were done as previously described (55). PCR amplifications were carried out with Taq DNA polymerase (Roche Molecular Systems, Summerville, N.J.) according to the manufacturer's recommendations and with optimal annealing temperatures as determined by the MacVector version 7 program. Plasmid DNA was isolated with Wizard Plus SV minipreps, and PCR products were purified on Wizard PCR preps (Promega Corp.). Plasmid DNA was dialyzed and concentrated on Microcon 100 units (Millipore Corp., Billerica, Mass.) in preparation for sequencing. All constructs were sequenced on both strands by using the BigDye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems, Foster City, Calif.) and appropriate primers according to the manufacturer's instructions. Sequences were determined on an ABI PRISM 3100 genetic analyzer (Applied Biosystems) operated by the DNA sequencing facility at Stony Brook University. Oligonucleotides were synthesized and purified by gel filtration chromatography at Midland Certified Reagent Company, Inc. (Midland, Tex.).

Detection of fur transcripts by RT-PCR.

RNA was extracted from a 50-ml culture of B. burgdorferi B31-MI by using TRI Reagent LS (Molecular Research Center, Inc., Cincinnati, Ohio) according to the manufacturer's instructions. The RNA was taken up in 100 μl of 10 mM Tris-HCl (pH 8.3) buffer containing 1.5 mM MgCl₂ and 50 mM KCl and treated with 10 U of RNase-free DNase I (Roche Molecular Systems) for 1 h at 37°C to degrade contaminating DNA, and the DNase I was then removed by passing the RNA through an RNeasy column (QIAGEN, Valencia, Calif.). cDNA was synthesized by using genome-directed primers prepared as previously described (51). The genome-directed primers were annealed to the RNA by heating at 90°C for 2 min and cooling to 42°C over a 20-min period. The cDNA was generated with avian myeloblastosis virus (AMV) reverse transcriptase (RT) (Roche Molecular Systems) by incubation for 2.5 h at 42°C. The reaction mixture contained RNA (1 μg), genome-directed primers (each at 250 ng/μl), AMV RT (30 U), deoxynucleoside triphosphates (dNTPs) (333 μM), 50 mM Tris-HCl (pH 8.5), 8 mM MgCl₂, 30 mM KCl, and 1 mM 1,4-dithiothreitol (DTT) in a volume of 25 μl. PCR amplification of the fur gene was done with forward primer BB0647FWD and reverse primer BB0647REV (Table 1) and with 1 μl of cDNA as a template under the following cycle conditions: 2 min at 94°C; 35 cycles of 30 s at 94°C, 30 s at 52°C, and 30 s at 72°C; and a terminal extension of 5 min at 72°C. A −RT control was done in parallel with cDNA prepared without added RT. The fur gene was PCR amplified from genomic DNA to establish the size of the PCR product. Genomic DNA was isolated from strain B31-MI by using TRI Reagent LS according to the manufacturer's instructions. Products were separated on a 1% agarose gel and visualized by staining with ethidium bromide.

TABLE 1.

Oligonucleotides used in this study

Oligonucleotide	Sequence (5′ to 3′)a	Purpose
BB0647FWD	ATGAACGACAACATAATAGACG	PCR primer
BB0647REV	TAAAGTGATTTCCTTGTTCTCATC	PCR primer
FLAF	GGTATAATCATATGAACGACAACATAATAGAC	PCR primer
FRR	ATGACTCGAGTTATATTCATAAAGTGAT	PCR primer
U-19 mer	GTTTTCCCAGTCACGACGT	PCR primerb
T7 promoter	TAATACGACTCACTATAGGG	PCR primer
GBBEG25F	CAAACTTGGAGAAAACTGG	PCR primer
GBBEG25R	CGAACATAAAGTGATTTCCTTG	PCR primer
GBBDA28F	TATGAAAATAAATAAATAATAAGTAG	PCR primer
GBBDA28R	CGAACATATGATTATACCTTTTTTG	PCR primer
GBBBM12F	TTATGCATTGTACTTAAATTGC	PCR primer
GBBBM12R	CGAACATAACTATCTCCTTTA	PCR primer
FurF	GGGGATAATGATAATCATTATCGGG	Fur box sequencec
FurR	CCCGATAATGATTATCATTATCCCC	Fur box sequence
PerRF2	GGGCTAAATTATAATTATTATAATTTAGGGG	Per box sequence
PerRR2	CCCCTAAATTATAATAATTATAATTTAGCCC	Per box sequence
LMFv2	GTATGCGTTGGTACCGCTGCTGGGGGC	Control sequence
LMRv2	GCCCCCAGCAGCGGTACCAACGCATAC	Control sequence

^aRestriction sites for subsequent cloning of the PCR products are underlined.

^bThe U-19 mer and T7 promoter primers were from Novagen.

^cThe oligonucleotides used for the Fur box sequence were from Gonzalez de Peredo et al. (25).

PCR cloning of the fur gene from B. burgdorferi.

The gene encoding Fur (TIGR designation BB0647) was amplified by PCR using DNA isolated from GBBEG25 as template and with forward primer FLAF and reverse primer FRR (Table 1) and the following cycle conditions: 2 min at 95°C, 30 cycles of 1 min at 95°C, 30 s at 48.5°C, and 1.5 min at 72°C, and a terminal extension of 15 min at 72°C. The product was cut with NdeI and XhoI and inserted into the NdeI-XhoI cloning site of pET28a(+) (Novagen, Inc.) to generate pLK1/5. pLK1/5 was introduced into E. coli DH5α for sequencing and into E. coli BL21(DE3) for expression of the recombinant protein.

Overproduction and purification of recombinant Fur.

E. coli BL21(DE3) cells containing pLK1/5 were grown under kanamycin selection to an optical density at 600 nm (OD₆₀₀) of 0.8 and stored overnight at 4°C. This starter culture was used to inoculate 200-ml cultures that were grown under kanamycin selection to an OD₆₀₀ of ∼1.0, at which point isopropyl-β-d-thiogalactopyranoside (IPTG) was added to 0.4 mM. These cultures were further incubated for 20 h, and the cells were harvested by centrifugation (7,000 × g for 15 min at 10°C) and stored at −20°C. Each 200-ml culture typically gave 1.0 g (wet weight) of cell pellet. Each cell pellet was thawed on ice and resuspended in 10 ml of lysis buffer (20 mM Tris-HCl [pH 8] buffer containing 0.5 mM DTT plus 1 tablet of complete, mini, EDTA-free protease inhibitor cocktail [Roche Molecular Systems] that contained aprotinin, leupeptin, and Pefabloc SC). Hen egg white lysozyme was added to 0.2 mg/ml, and the cells were disrupted on ice by pulsed sonication for 5 min. The lysate was clarified by centrifugation (40,000 × g for 20 min at 4°C) and passed through a 0.45-μm-pore-size filter. The filtered supernatant from 400 ml of culture was diluted with an equal volume of 2× buffer (0.1 M potassium phosphate [pH 7.8] buffer containing 1 M NaCl, 10 mM imidazole, 0.5 mM DTT) and applied to a column (1 by 2.5 cm) of IDA-agarose (Roche Molecular Systems) previously charged with zinc and equilibrated in 0.05 M potassium phosphate (pH 7.8) buffer containing 0.5 M NaCl. The column was washed with 0.05 M potassium phosphate (pH 7.8) buffer containing 0.5 M NaCl, 10 mM imidazole, and 0.5 mM DTT to remove unbound material, and the His-tagged protein was eluted with a gradient of increasing imidazole concentration (10 mM to 250 mM imidazole in 0.05 M potassium phosphate [pH 7.8] buffer with 0.5 M NaCl and 0.5 mM DTT). Fractions that contained the His-tagged Fur as determined by Tricine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting were stored at −80°C. To remove the His tag, the protein was thawed on ice and mixed with 100 U of biotinylated thrombin (Novagen, Inc.) and incubated overnight at 4°C with end-over-end mixing. Streptavidin beads (Novagen, Inc.) were added according to the company's recommendation (3.2 ml of a 50% slurry), and the incubation was extended for 1 h at 4°C. The beads were collected by spin filtration, and the resulting filtrate was clarified by centrifugation (40,000 × g for 20 min at 4°C) and stored with 10% glycerol at −80°C. The yield of Fur from 400 ml of culture was typically 6 to 10 mg. For use in gel shift assays, the Fur was thawed on ice, dialyzed at 4°C against 20 mM Bis-Tris-borate (pH 7.5) buffer containing 0.5 mM DTT, and then clarified by centrifugation (16,000 × g for 10 min at 4°C). In some instances, the Fur was stored at −80°C prior to clarification.

Target DNA.

Oligonucleotides containing the E. coli Fur box, the B. subtilis Per box, or an unrelated sequence (control) were prepared by mixing equal amounts of the synthesized single strands (Table 1) and then heating for 5 min at 95°C and cooling to 35°C over a 25-min period. Target DNA containing the Fur consensus sequence (Ecfur) or Per consensus sequence (BsperR) was prepared by inserting these sequences into the EcoRV site of pSTBlue-1 with Novagen's Perfectly Blunt Cloning kit (Novagen, Inc.) and then amplifying by PCR the insert plus flanking vector DNA with T7 promoter and U-19 mer primers (Table 1). Target DNA containing the upstream sequences of three Borrelia genes (BB0646, BB0647, and BB0690) was prepared in a similar way. Each upstream sequence was amplified by PCR with plasmid DNA from clone GBBEG25 (for BB0646), clone GBBDA28 (for BB0647), or clone GBBBM12 (for BB0690) as template and the appropriate forward and reverse primers (Table 1). Each product was then inserted into the EcoRV cloning site of pSTBlue-1 by using Novagen's AccepTor Vector kit, and the insert plus flanking DNA was amplified by PCR using the U-19 mer forward primer and the appropriate reverse primer (Table 1).

Mobility shift assay.

The mobility shift assay used was modeled after that of de Lorenzo et al. (18). Oligonucleotides were assayed on a 7% nondenaturing gel, while longer PCR products were assayed on a 6% gel. The gel was cast in 20 mM Bis-Tris-borate (pH 7.5) buffer and run in the same buffer with or without 0.5 mM DTT as indicated. Prior to loading the samples, the gel was prerun for 15 min at 100 V. In assays with oligonucleotides, the Fur was mixed with the oligonucleotide in a 10-μl volume in 15 mM Bis-Tris-borate (pH 7.5) buffer containing 0.375 mM DTT, 1 mM MgCl₂, 40 mM KCl, and 5% glycerol and incubated for 1 h at room temperature, and a 3-μl aliquot was then loaded onto a 7% gel (10 by 10.5 cm by 1 mm) and separated for 1 h at 100 V. In assays with PCR products, Fur was mixed with the DNA in a 100-μl volume in 20 mM Bis-Tris-borate (pH 7.5) buffer containing 1.7 mM DTT, 1.3 mM MgCl₂, 53 mM KCl, 3.3% glycerol, and 127-μg/ml bovine serum albumin and incubated for 30 min at room temperature, and a 5-μl aliquot was then loaded onto a 6% gel (10 by 10.5 cm by 1.5 mm) and separated for 1 h at 100 V. In competition assays, Fur was first mixed with the competitor oligonucleotide and then with the target DNA. Freshly prepared stocks were used in all assays containing metals (MnCl₂, ZnSO₄, ZnCl₂, or FeSO₄) or oxidizing agents (hydrogen peroxide, t-butyl hydroperoxide, cumene hydroperoxide, or diamide). To visualize DNA, the gels were immersed for 20 min in 50 ml of 20 mM Bis-Tris-borate (pH 7.5) buffer containing SYBR Green I (Molecular Probes, Inc., Eugene, Oreg.) at a dilution of 1:10,000 and then scanned on a Storm 860 phosphorimager (Amersham-Pharmacia Biotech, Inc., Piscataway, N.J.) set to blue fluorescence mode. The image was captured with ImageQuant version 1.2 (Amersham-Pharmacia Biotech, Inc.) and imported into Quantity One version 4.1.1 (Bio-Rad Laboratories, Hercules, Calif.) for densitometric analysis. K_d values were estimated as described by Carey and Smale (13), by using the formula K_d = [P][S]/[PS] where [P] is the concentration of unbound Fur, [S] is the concentration of unbound DNA, and [PS] is the concentration of bound Fur and bound DNA.

Tricine SDS-PAGE and Western blotting.

Tricine SDS-PAGE was carried out on 10% gels in a mini-format according to the method of Schagger and von Jagow (56). Gels were fixed in 50% methanol-12% acetic acid and stained with 0.1% Coomassie brilliant blue R-250 in the same fixative. Prestained protein standards (BRL, Invitrogen Corp., Carlsbad, Calif.) were used for estimation of molecular weight. Proteins were electrotransferred onto nitrocellulose (62). Blots were blocked in PBS containing 2% casein, incubated with a 1:1,000 dilution of anti-His monoclonal antibody (Sigma-Aldrich, St. Louis, Mo.) followed by a 1:1,000 dilution of goat anti-mouse immunoglobulin G conjugated to alkaline phosphatase (BRL), and then reacted with nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl-phosphate (Kirkegaard & Perry Laboratories, Inc., Gaithersburg, Md.). Both gels and blots were scanned on a model GS710 imaging densitometer (Bio-Rad Laboratories, Hercules, Calif.).

Protein analyses.

Protein was determined with the Coomassie Plus protein reagent (Pierce Biotechnology, Rockford, Ill.) following the manufacturer's instructions for the microplate version of the assay and with bovine serum albumin as the standard. N-terminal sequence analysis was carried out on bands electroblotted in CAPS (pH 11) transfer buffer onto Immobilon P^SQ polyvinylidene difluoride membrane (37) and visualized with Ponceau S according to the manufacturer's recommendations (Millipore Corp., Billerica, Mass.). The sequences were determined on a model 492 Procise Sequencer (Applied Biosystems, Foster City, Calif.) operated by the Proteomics Center at Stony Brook University.

Size-exclusion chromatography.

Recombinant Fur previously treated with thrombin to remove the N-terminal His tag was concentrated to ∼10 ml in a Centriprep-3 unit (Millipore Corp.) and separated on a Sephacryl S300HR 16/60 column (Amersham-Pharmacia Biotech, Inc., Piscataway, N.J.) previously equilibrated in 50 mM Tris-HCl (pH 8) buffer containing 0.5 M NaCl, 0.5 mM DTT, 10 μM ammonium sulfate, and 10 μM zinc acetate. A single peak appeared on the chromatogram, and fractions were pooled and concentrated to 5 mg/ml in a Centriprep-3 and Centricon-3 unit (Millipore Corp.). A 50-μl aliquot was separated on an analytical column of Shodex KW-802.5 on an Akta purifier (Amersham-Pharmacia Biotech, Inc.) at 0.5 ml/min at 4°C with a running buffer consisting of 100 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 0.025% NaN₃. The refractive index of the sample was obtained on an Optilab DSP interferometric refractometer (Wyatt Technology Corp., Santa Barbara, Calif.) at 35°C, and the static light scattering was measured by using a Dawn EOS enhanced optical system (Wyatt Technology Corp.).

Computer analyses.

Nucleic acid and protein sequences were analyzed by using the MacVector 7 and GCG Seqweb 2.1 programs (Accelrys, Inc., San Diego, Calif.). Paired alignments were done in Seqweb by using the Gap program with the Blosum62 amino acid substitution matrix (31). BlastP analyses were done at the National Center for Biotechnology Information using version 2.2.6 with the Blosum62 matrix and with no exclusion of regions of low complexity (3). Clustal W alignments were done at EMBnet (http:www.ch.embnet.org/software/ClustalW.html), using version 1.74 and the Blosum scoring matrix. The Protein Families (PFAM) database was accessed at http://www.sanger.ac.uk/Software/Pfam. Promoter sequences were predicted by using a neural network program accessed at http://www.fruitfly.org/seq_tools/promoter.html (50).

RESULTS AND DISCUSSION

The B. burgdorferi genome encodes for a Fur homologue.

B. burgdorferi contains a gene that codes for a Fur homologue (14, 22). This gene (528 nt) is located on the main chromosome and is designated BB0647 according to the TIGR nomenclature. The predicted product has a molecular weight of 20,205 and a pI of 8.19. A BlastP analysis of the GenBank database indicated that this Fur protein sequence has the highest homology to the sequence of a putative Fur (FN2045) in Fusobacterium nucleatum. A paired alignment of the two showed 35% identity (48% similarity) over the entire length of the sequence. The B. burgdorferi Fur belongs to the FUR family of prokaryotic transcription factors defined in Pf 01475 of the PFAM database. Currently, there are 243 proteins listed in this family, including members from both Bacteria and Archaea. Because the B. burgdorferi Fur shows little identity to any of the well-studied Fur proteins, the likely function of this Fur is unclear.

Zur is a Fur homologue that regulates uptake of zinc by the Znu ABC transporter system (43, 44). A Clustal W alignment of known Zur proteins and proteins putatively identified as Zur showed conservation of several residues (see Fig. S1 in the supplemental material). Alignment of the B. burgdorferi Fur with these proteins showed that some of these residues are also conserved in the B. burgdorferi Fur. However, one cysteine that is conserved among the Zur proteins is absent from the B. burgdorferi Fur. E. coli Zur has two metal binding sites that bind zinc in part through coordination with cysteines (41). Which cysteines are involved is not yet known; however, a cysteine uniquely conserved among the Zur proteins seems a likely candidate. Moreover, Zur proteins regulate zinc uptake through repression of the zinc ABC transporter genes (znuABC), which do not appear to be present in B. burgdorferi (14, 22, 53). For example, B. burgdorferi contains no homologue of the putative manganese/zinc ABC transporter (TroABCD) that is found in Treponema pallidum (30, 48). While the Fur in B. burgdorferi may regulate zinc uptake, in silico analyses do not support this function.

Fur homologues that regulate the response to oxidative stress have been variously termed PerR, CatR, FurA, or FurS (29, 39). B. burgdorferi Fur seems to be related to both CatR and PerR, according to BlastP analyses. Paired alignments showed that the B. burgdorferi Fur is 29% identical (46% similar) to the S. coelicolor CatR and 29% identical (42% similar) to the B. subtilis PerR. A Clustal W alignment of the PerR-like proteins showed that a number of residues are conserved (see Fig. S2 in the supplemental material). Four of these residues correspond to the structural zinc-binding site as defined in the PA-FUR crystal structure (46); however, only one of these residues (H111) is also conserved in the B. burgdorferi Fur. Thus, while the B. burgdorferi Fur shows some relatedness to the PerR-like proteins, it may or may not have the same function in Borrelia.

The B. burgdorferi Fur is expressed during growth in vitro.

To determine if B. burgdorferi fur is expressed, we grew Borrelia in culture, isolated the RNA, prepared cDNA with genome-directed primers, and then PCR amplified the fur with fur-specific primers. A fur PCR product was detected in Borrelia grown at either 23 or 35°C (Fig. 1). The −RT controls were negative for both, indicating that the PCR products were generated from cDNA and not from contaminating genomic DNA. These results agreed with those of Ojaimi et al. (40), who reported expression of BB0647 by B. burgdorferi grown at 23 and 35°C in a series of whole-genome-array studies. In their studies, expression of BB0647 was approximately the same at either temperature, suggesting that transcription of the B. burgdorferi fur is not regulated in response to temperature, at least not during growth in vitro.

FIG. 1.

Transcription of the fur gene in B. burgdorferi B31-MI. RNA was isolated from spirochetes grown at 23 or 35°C, and the cDNA was synthesized by genome-directed primers. fur was PCR amplified from the cDNA or genomic DNA by fur-specific primers. To control for possible contamination with genomic DNA, the cDNA was prepared with or without RT. DNA size standards (kilobases) appear on the left.

The B. burgdorferi Fur binds the E. coli Fur box and B. subtilis Per box in gel shift assays.

Because the B. burgdorferi Fur showed low homology to any of the well-studied Fur proteins, it was not possible to derive the Fur binding site by in silico analyses. We therefore adopted an empirical approach and looked at binding of the B. burgdorferi Fur to various target DNAs, using a gel shift assay. To do this, the fur gene (BB0647) was inserted into pET28a to generate a His-Fur fusion clone. The His-Fur fusion protein was overproduced in E. coli BL21(DE3) by overnight induction with IPTG, as higher yields of the fusion protein were obtained with the extended induction time (data not shown). Disruption of E. coli by sonication released >90% of the fusion protein into the soluble fraction, leaving <10% in insoluble inclusion bodies. The estimated size of the fusion protein as determined by Tricine SDS-PAGE was 22.6 kDa, in good agreement with the theoretical size of 22.4 kDa. The fusion protein present in the soluble fraction was purified by zinc-chelate chromatography so that DTT could be added to the buffers to maintain the cysteines within the Fur in a reduced state. The purified protein was stored at −80°C and later thawed for treatment with thrombin to remove the N-terminal His tag. Much of the protein precipitated during storage; however, after treatment with thrombin, the cleaved Fur was once again soluble. This allowed us to separate the Fur from uncleaved His-Fur by a simple centrifugation step and thereby avoid a second chromatography (Fig. 2). Thrombin cleaved >90% of the N-terminal His tag during an overnight incubation at 4°C. The size of the cleaved protein was estimated at 21.5 kDa (Fig. 2, supernatant), only slightly larger than the 20.5 kDa predicted for monomeric protein. N-terminal sequence analysis confirmed cleavage of the His tag at a sensitive arginine residue. The recombinant Fur (Bb-Fur) differed from native Fur by containing an extra three amino acids (Gly Ser His) at the N terminus as a result of the original His tag.

FIG. 2.

Purification of the recombinant Fur. His-Fur fusion protein was purified by zinc-chelate chromatography from the soluble fraction of a whole-cell lysate. The N-terminal His tag was removed by digestion with biotinylated thrombin, and then the thrombin was removed by adsorption onto streptavidin beads followed by spin filtration. Insoluble protein (including the uncleaved fusion protein) was removed by high-speed centrifugation to yield a supernatant fraction that contained the purified protein. The Tricine SDS-PAGE gel was stained with Coomassie blue; the Western blot was probed with anti-His monoclonal antibody. Protein size standards (kilodaltons) appear on the left.

Fur is thought to bind DNA as a homodimer, much like DtxR (46). To determine if Bb-Fur could form dimers, we subjected a concentrated sample of Bb-Fur to size-exclusion chromatography and monitored the effluent for static light scattering and refractive index. Evaluation of the data by the two-detector approximation (61, 64, 65) gave an estimate of the molecular weight of the protein of 41,400 (data not shown). This is in agreement with the calculated molecular weight for a Fur dimer of 40,970, indicating that the Bb-Fur formed dimers under these conditions. Further examination of the data showed that the Bb-Fur existed as a monodisperse dimeric species with no measurable aggregate.

To determine if the recombinant Fur was capable of binding DNA, we looked at binding of the Bb-Fur to duplex oligonucleotides. Three sequences were tested: the E. coli Fur consensus sequence, the B. subtilis Per consensus sequence, and an unrelated control sequence (Fig. 3A). Bb-Fur was mixed with each duplex oligonucleotide in buffer containing DTT (but without added metal ions) and then separated on a 7% nondenaturing gel. Bb-Fur bound both the E. coli Fur box and the B. subtilis Per box with about the same affinity (data not shown). Although DTT was present in the initial incubation mixture, it was not present during the course of the electrophoresis. Without DTT, the Fur-DNA complexes seemed to dissociate while migrating through the gel. Bb-Fur has four cysteines, and one or more may be sensitive to oxidation.

FIG. 3.

Binding of Bb-Fur to Fur and Per box sequences. (A) Oligonucleotides were synthesized to contain consensus Fur or Per box sequences (boldface) as described in Materials and Methods. Terminal Gs and Cs were added to the single strands to help direct annealing. Inverted arrows indicate putative dimer binding sites (6, 34). (B, C, and D) Gel shift assays were carried out with 2.5 nM DNA and Bb-Fur atthe following concentrations: 0, 5, 11, 22, 54, 107, 215, and 322 nM. The target DNA consisted of the Fur box sequence plus flanking vector DNA (Ecfur), the Per box sequence plus flanking vector DNA (BsperR), or vector DNA alone (Vector). The arrows indicate shifted DNA that probably consisted of a single Fur dimer bound to a single DNA molecule (see text for discussion). The bottom band is due to nonspecific staining of a buffer component in this and other figures.

B. burgdorferi Fur binds both the E. coli Fur box and the B. subtilis Per box with an apparent K_d of ∼20 nM.

Although Bb-Fur formed complexes with the Fur and Per box sequences, the concentration of duplex oligonucleotide needed to generate a band that could be visualized with the SYBR Green I was too high to make K_d estimates possible. To circumvent this problem, we inserted these sequences into the cloning vector pSTBlue-1 at the EcoRV site and then PCR amplified the insert plus flanking vector DNA. This yielded two PCR products of 254 and 260 bp that contained the E. coli Fur box (Ecfur) and B. subtilis Per box (BsperR), respectively. To serve as a control, we also PCR amplified the vector alone, using the same two primers to yield a product of 229 bp that contained the two flanking sequences and no insert (vector). Each of these sequences was incubated with Bb-Fur in buffer containing DTT (and no added metal ions) and then loaded onto a 6% nondenaturing gel and separated in buffer supplemented with DTT to protect Bb-Fur against oxidation. The results showed that Bb-Fur bound both sequences—Ecfur and BsperR—about equally, as was suggested by the earlier data (Fig. 3B and C). At 22 nM Bb-Fur, approximately half of the DNA was present in the bound state and half in the unbound state for either construct. The major shifted band at 22 nM Bb-Fur probably consisted of a single Fur dimer bound to a single DNA molecule, if one assumes that Bb-Fur binds DNA as a homodimer, as has been suggested for other Fur homologues (46). In the case of the BsperR DNA, this would reflect binding of Bb-Fur to the single Per box (Fig. 3A). At higher concentrations of Bb-Fur (>50 nM), nonspecific binding seems to have contributed to the formation of Fur-DNA complexes, as evidenced by the presence of multiple shifted bands in the vector control (Fig. 3D).

The apparent K_d for binding of the Bb-Fur to either Ecfur or BsperR was ∼20 nM (expressed as monomer). This is the concentration at which half of the DNA was present in the bound state and half in the unbound state. The actual binding affinity, however, may be much higher if only a fraction of the purified protein retained activity. This apparent K_d value is within the range of values (1 to 60 nM) reported by others for recombinant E. coli Fur in assays with the E. coli Fur consensus sequence (2, 25, 34).

The region upstream of the fur gene contains potential Fur binding sites.

Some, but not all, Fur proteins exhibit autoregulation (1, 16, 19, 23, 27, 33, 35, 54, 57). Examination of the region upstream of BB0647 revealed several potential Fur binding sites. Five sites having 11 out of 15 or 12 out of 15 matches to the Per consensus sequence were identified in the 200 bp upstream of the putative translational start site (Fig. 4). In addition, a single site having 11 out of 15 matches to the Per consensus sequence was identified at the 3′ end of the fur coding region, 45 bp upstream of the putative translational start site for the downstream gene, BB0646 (Fig. 4). According to the genome sequence, BB0646 codes for a hypothetical protein of the α/β-fold hydrolase family (22). The deduced protein, however, shows low homology to several prokaryotic lipases and contains the GXSXG active site motif often referred to as the lipase box (data not shown).

FIG. 4.

Putative Per box sequences identified upstream of BB0647 (fur) and BB0646 (putative lipase gene). The translational start sites for BB0647 (nt 201) and BB0646 (nt 728) are boldface and underlined. The translational stop sites for BB0648 (nt 95) and BB0647 (nt 729) are boldface and overlined. Putative ribosomal binding sites are boldface. Putative −10 and −35 promoter sites are underlined. Arrows denote the location of an inverted repeat upstream of BB0647. Per box sequences with at least 11 out of 15 matches to the consensus sequence are in boxes: matches to the Per consensus sequence are boldface. The DNA used to generate target sequences for gel shift analyses is marked by brackets. The entire sequence corresponds to nt 685977 to 686707 (reverse complement) on the main chromosome of B. burgdorferi (GenBank accession no. AE000783).

The same region of DNA (Fig. 4) was also examined for matches to the Fur consensus sequence. Five sites having 12 out of 19 matches to the Fur consensus sequence were identified: one site was located in the region upstream of the ATG start site for BB0647, three sites were located in the region upstream of the ATG start site for BB0646, and one site was located in the middle of the BB0647 coding region. Although the exact locations of these Fur boxes differed from those of the Per boxes described above, the general pattern was the same, with the majority of the Fur boxes clustered in the two promoter regions (data not shown).

B. burgdorferi Fur binds to multiple sites within the putative promoter region of BB0646 and BB0647.

To determine if Bb-Fur binds to sites within the putative promoter region of BB0647 (fur) or BB0646 (putative lipase gene), we PCR amplified the two upstream sequences (Fig. 4, nt 87 to 203 and 602 to 730) and then inserted the products into the EcoRV cloning site of pSTBlue-1 and PCR amplified the inserts plus flanking DNA. This generated two target DNAs—BB0647up and BB0646up—that were 273 and 285 bp, respectively. Each target DNA was incubated with Bb-Fur in buffer containing DTT (and no added metal ions) and then separated on a 6% nondenaturing gel in buffer with DTT to protect the Fur against oxidation. Both BB0647up and BB0646up bound Bb-Fur, generating multiple shifted bands (Fig. 5). The ladder-like appearance of the bands suggested that each band consisted of DNA plus some number of Fur dimers. At 22 nM Bb-Fur, BB0647up gave five shifted bands. The slowest moving band, which was present in higher yield than the others, formed the sixth rung of the ladder and probably consisted of DNA plus five Fur dimers (Fig. 5B, arrow). In a direct comparison of the four target DNAs (Fig. 5C), 22 nM Bb-Fur gave 61% shifted DNA for BsperR, 54% for Ecfur, 35% for BB0646up, and 35% for BB0647up, as determined by densitometry.

FIG. 5.

Binding of Bb-Fur to sites upstream of BB0646 (putative lipase gene) and BB0647 (fur). (A and B) Target DNA was assayed at 2.5 nM with Bb-Fur at the following concentrations: 0, 5, 11, 22, 54, 107, 215, and 322 nM. The target DNA consisted of the upstream sequence of BB0646 plus flanking vector DNA (BB0646up) or the upstream sequence of BB0647 plus flanking vector DNA (BB0647up). The arrow indicates a complex between BB0647up and Bb-Fur that appears to consist of DNA plus five Fur dimers. (C) Four target DNAs (BsperR, Ecfur, BB0646up, and BB0647up) were assayed at 2.5 nM with Bb-Fur at 0 or 22 nM.

The B. burgdorferi Fur binds to site(s) within the putative promoter region of BB0690.

BB0690 encodes for NapA, a Dps homologue (22). The function of Dps in E. coli is to protect DNA against damage during oxidative stress (4). In B. subtilis and Staphylococcus aureus, expression of MrgA (also a Dps homologue) is under the control of PerR (11, 33). It is therefore of interest to determine if expression of NapA in B. burgdorferi is similarly regulated by Fur. Examination of the region upstream of the putative ATG start site for BB0690 revealed several sequences having 11 or more matches to the Per consensus sequence (Fig. 6). In particular, one sequence located immediately upstream of the putative ribosomal binding site (and 6 bp downstream of the ATG start site identified by TIGR) has 14 out 15 matches to the Per consensus sequence (Fig. 6, nt 206 to 220).

FIG. 6.

Putative Per box sequences identified upstream of BB0690 (napA). The ATG start site designated by TIGR (nt 201) appears boldface and underlined. A second start site that is 33 bp downstream from the first site (nt 234) is boldface and double underlined. A putative ribosomal binding site is boldface. Putative −10 and −35 promoter sites are underlined. Per box sequences with at least 11 out of 15 matches to the consensus sequence appear in boxes: matches to the Per consensus sequence are boldface. The sequence reported to bind BosR according to footprint analyses (10) is overlined. The DNA used to generate a target sequence for gel shift analysis is marked by brackets. The entire sequence corresponds to nt 730951 to 731186 on the main chromosome of B. burgdorferi (GenBank accession no. AE000783).

To determine if Bb-Fur binds to sites within the putative promoter region of BB0690, we PCR amplified the upstream sequence (Fig. 6, nt 82 to 236) and inserted the product into pSTBlue-1 and then PCR amplified the insert plus flanking vector DNA to generate the target sequence BB0690up. As in previous assays, BB0690up (311 bp) was incubated with Bb-Fur in buffer containing DTT (and no added metal ions) and then separated on a 6% gel in buffer with DTT to protect Fur against oxidation. BB0690up bound Bb-Fur, yielding one major shifted band and two minor shifted bands (Fig. 7A). The ladder-like appearance of the bands suggested that the major shifted band consisted of DNA plus three Fur dimers. In a direct comparison, 23 nM Bb-Fur gave 61% shifted DNA with the consensus Per box sequence and 36% shifted DNA with the napA promoter sequence (Fig. 7B).

FIG. 7.

Binding of Bb-Fur to sites upstream of BB0690 (napA). (A) Target DNA was assayed at 2.4 nM with Bb-Fur at the following concentrations: 0, 6, 12, 23, 59, 117, and 234 nM. The target DNA consisted of the upstream sequence of BB0690 plus flanking vector DNA (BB0690up). (B) Two target DNAs (BsperR and BB0690up) were assayed at 2.4 nM with Bb-Fur at 0 or 23 nM. (C and D) BB0690up DNA was assayed at 2.4 nM with Bb-Fur at 0 or 23 nM. Competitor oligonucleotide (Per box oligo or control oligo) was added to the incubations at 240 nM, as indicated. (E) BsperR DNA was assayed at 2.4 nM with Bb-Fur at 0 or 20 nM. Competitor oligonucleotide (Per box oligo or control oligo) was added to the incubations at 240 nM, as indicated. Bands that are not readily apparent are marked by asterisks.

To confirm that the Fur-DNA interactions were specific (and not the result of nonspecific binding), we set up a competition assay. The Per box oligonucleotide and a control oligonucleotide (Fig. 3A) were each tested for their ability to block binding of Bb-Fur to the BB0690up DNA. While the Per box oligonucleotide inhibited binding completely (Fig. 7C), the control oligonucleotide had no effect (Fig. 7D). Similarly, the Per box oligonucleotide blocked binding of Bb-Fur to the BsperR DNA, whereas the control oligonucleotide did not (Fig. 7E). These data indicate that Bb-Fur is binding to these sites specifically.

Recently, Boylan et al. (10) reported that BosR (their designation for the B. burgdorferi Fur) bound a site upstream of BB0690. By footprint analysis, they identified this site as being 50 nt in length and located 137 to 187 bp upstream of the transcriptional start site (Fig. 6, nt 12 to 61). The site(s) bound by Bb-Fur in our gel shift assays could not have been this site, as our BB0690up target DNA did not contain this region (Fig. 6). Their gel shift assays showed shifted DNA with 300 nM BosR but only in the presence of 10 μM Zn²⁺. We did not add Zn²⁺ to our incubations, yet observed shifted DNA with Bb-Fur at concentrations as low as 6 nM (Fig. 7A). Boylan et al. used recombinant protein isolated from inclusion bodies, which needed to be folded to obtain activity. Here, we used recombinant protein purified from the soluble fraction of lysed cells, which was already folded and perhaps also had metal ions incorporated (see below). Whether the Borrelia Fur binds in vivo both the 50-nt sequence identified by Boylan et al. (10) and the site(s) identified here is an area for future study.

The B. burgdorferi Fur does not require Zn²⁺ for binding to DNA.

Although we obtained high-affinity binding without the addition of metal ions, it was still possible that the recombinant Fur retained sufficient metal ions to support DNA binding. To determine if metal ions were required, we looked at the effects of two metal chelators: EDTA and 2,2′-dipyridyl (DP). EDTA chelates divalent metal ions, including Mg²⁺, and so for these studies Mg²⁺ was omitted from the binding buffer. Note, Mg²⁺ was found not to be required for binding to BsperR (Fig. 8A) or BB0690up (Fig. 8B). EDTA at 5 mM inhibited binding of Bb-Fur to BsperR and BB0690up but did not completely eliminate it (Fig. 8A and B). A higher concentration of EDTA (20 mM), however, gave identical results (data not shown). DP at 5 mM had no effect (Fig. 8A). DP chelates Fe²⁺, and therefore this result indicates that Bb-Fur does not contain bound Fe²⁺.

FIG. 8.

Effects of metals on Bb-Fur binding activity. (A) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 27 nM. Mg²⁺ was present at 0 or 1.3 mM; DP and EDTA were present at 0 or 5 mM. (B) BB0690up DNA was assayed at 2.5 nM with Bb-Fur at 0 or 40 nM. Mg²⁺ was present at 0 or 1.3 mM; Zn²⁺ was present at 0, 10, or 100 μM; EDTA was present at 0 or 5 mM. (C) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 22 nM. Zn²⁺ was present at 0, 10, 20, 50, or 100 μM. (D) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 22 nM. Zn²⁺ was added at 100 μM, and Mn²⁺ was added at 2 mM. Bb-Fur was allowed to incubate with the first metal for 15 min prior to addition of the second metal. DNA was then added to the incubations after an additional 20 min. (E) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 22 nM. Fe²⁺ was present at 0, 50, 100, 250, or 500 μM.

Boylan et al. (10) reported that BosR requires Zn²⁺ for binding, and therefore we looked at the effect of Zn²⁺ in our assay. Rather than increase binding, Zn²⁺ decreased binding. This was true for both BsperR (Fig. 8C) and BB0690up (Fig. 8B) and for both ZnSO₄ (data shown) and ZnCl₂ (data not shown). Boylan et al. (10) reported that Mn²⁺ blocks Zn²⁺. We found that Mn²⁺ (10 μM to 2 mM) had no effect on binding (data not shown). Mn²⁺ also did not block the effect of Zn²⁺ nor reverse it (Fig. 8D). Finally, Boylan et al. (10) reported that Fe²⁺ has no effect on binding. We found that Fe²⁺ had an effect similar to that of Zn²⁺, although the concentration of Fe²⁺ needed to prevent binding was approximately 10-fold higher (Fig. 8E). This may simply be owing to the fact that Fe²⁺ becomes quickly oxidized to Fe³⁺, and therefore the concentration of Fe²⁺ actually present is much lower.

Although it is still not clear if Bb-Fur contains bound metal, it is clear that metal ions (Zn²⁺ and Fe²⁺) affect DNA binding in vitro. If B. burgdorferi Fur functions as a repressor in vivo, then increased levels of Zn²⁺ (or Fe²⁺) could cause derepression and hence appear as activation. Iron-dependent activation has been studied in Helicobacter pylori (17). H. pylori Fur functions as an iron-dependent “activator” in regulating the pfr (non-heme iron-containing ferritin) gene (17).

Exposure to peroxides decreases the binding affinity of B. burgdorferi Fur for DNA.

We had observed that DTT was required to maintain Bb-Fur in a form that could bind DNA. When DTT was omitted from the binding buffer, the affinity of Fur for DNA was much decreased (Fig. 9A and C). Therefore, to study the effect of various peroxides on binding affinity, we first treated Bb-Fur with DTT to generate fully reduced protein and only then added oxidizing agent. The results showed that Bb-Fur oxidized by hydrogen peroxide, cumene hydroperoxide, or diamide no longer bound DNA (Fig. 9A and C). To determine if this oxidation was reversible, Bb-Fur was first reduced with DTT (1 mM), then oxidized with hydrogen peroxide (20 mM), and finally rereduced with DTT (100 mM). The results showed that all the binding activity lost to oxidation was regained after reduction (Fig. 9B). This suggests that one or more cysteines present in Bb-Fur are susceptible to oxidation. Hahn et al. (26) reported that CatR is susceptible to oxidation by both hydrogen peroxide and diamide. Herbig and Helmann (32) noted that PerR previously oxidized by hydrogen peroxide could be reduced by DTT to regain DNA-binding activity.

FIG. 9.

Effects of peroxides and diamide on Bb-Fur binding activity. (A) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 27 nM. The Bb-Fur was incubated with the DTT (1 mM) for 20 min prior to addition of the oxidizing agents. Hydrogen peroxide, diamide, t-butyl hydroperoxide, and cumene hydroperoxide were each added at 20 mM and allowed to act on the Bb-Fur for 30 min prior to addition of the DNA. (B) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 22 nM. The Bb-Fur was incubated in binding buffer containing 1 mM DTT to generate fully reduced protein prior to addition of the other agents. Hydrogen peroxide was added at 20 mM, and after 5 min, DTT was added at 100 mM. DNA was added last, after an additional 10 min of incubation. (C) BB0690up DNA was assayed at 2.5 nM with Bb-Fur at 0 or 40 nM. The Bb-Fur was incubated with the DTT (1 mM) for 20 min prior to addition of the oxidizing agents. Hydrogen peroxide, diamide, t-butyl hydroperoxide, and cumene hydroperoxide were each added at 20 mM and allowed to act on the Bb-Fur for 20 min prior to addition of the DNA. (D) BsperR DNA was assayed at 2.5 nM with Bb-Fur at 0 or 22 nM. The Bb-Fur was incubated in binding buffer containing 1 mM DTT and 10 μM ZnSO₄ (as indicated) to generate fully reduced protein with bound metal. DNA was then added and allowed to bind for 30 min prior to addition of the t-butyl hydroperoxide at 0, 1, 5, or 10 mM.

Boylan et al. (10) reported that exposure to t-butyl hydroperoxide (1 to 10 mM) enhances the binding activity of BosR in a mobility shift assay. They suggested that BosR binds to DNA in a reduced state and is then activated by oxidation. We found that exposure to t-butyl hydroperoxide decreased binding for the napA promoter sequence (Fig. 9C), but not for the Per box sequence (Fig. 9A). These assays, however, were carried out without added Zn²⁺. Boylan et al. (10) observed enhanced binding with t-butyl hydroperoxide, but only in the presence of DTT (1 mM) and Zn²⁺ (10 μM). To mimic these conditions, we incubated Bb-Fur with DTT (1 mM) and Zn²⁺ (10 μM), then added target DNA (BsperR), and finally t-butyl hydroperoxide (1, 5, or 10 mM). As expected, Bb-Fur showed slightly less binding activity in the presence of Zn²⁺ (Fig. 9D). Exposure to t-butyl hydroperoxide, however, did not enhance this binding at any of the concentrations tested (Fig. 9D). Our data therefore indicate that t-butyl hydroperoxide has no effect on Bb-Fur binding to the Per box sequence whether or not Zn²⁺ is present (Fig. 9A and D).

These data are consistent with a role for Fur as a repressor of oxidative stress genes in B. burgdorferi. It seems likely that metal ions affect this regulation; however, studies to show the relationship between oxidation and bound metal have not been done. For napA, increased expression under conditions of high iron and high oxidative stress would make sense. In B. subtilis, regulation of mrgA is affected by whether the metal bound to PerR is iron or manganese. If iron is bound, PerR is more susceptible to oxidation and hence mrgA is derepressed more readily (32). In Fig. 8E, the effect of iron on Fur binding may actually reflect an increased sensitivity to oxidation. However, if B. burgdorferi does not contain iron (47), this may have no relevance in vivo.

Other genes that have potential Fur-binding sites are located on the main chromosome and on both circular and linear plasmids.

To identify genes that may be Fur regulated, we searched the Borrelia genome for Per box sequences, using the BlastN program available at the TIGR website (http://tigrblast.tigr.org/cmr-blast/index.cgi?database=GBB.seq). Candidate binding sites were identified and then screened manually according to the following set of criteria: (i) no more than 400 nt upstream of a putative translational start site, (ii) no deeper than 50 nt into an upstream coding region, (iii) a minimum of 13 matches to the Per consensus sequence, and (iv) associated with an open reading frame (ORF) of at least 200 bp. These criteria were modeled after those of Panina et al. (42) with some modification. Table 2 lists the sequences (and associated genes) that were found in searches with the Per box sequence as a single dimer binding site (TTATAAT-ATTATAA), two overlapping dimer binding sites (TTATAATTATTATAATTATAA), or two adjacent dimer binding sites (TTATAATTATTATAATTATTATAA). A total of 26 genes were found to have candidate sites within their upstream regions: 11 were on the main chromosome, 13 were on linear plasmids, and 2 were on circular plasmids. Only sequences containing 13 or 14 matches to the Per consensus sequence were included in Table 2; however, other sequences having 11 to 12 matches to the Per consensus sequence were typically found in the regions surrounding these sites.

TABLE 2.

Candidate Per boxes identified upstream of Borrelia ORFs^a

Gene	Putative identificationb	Putative Per box sequence(s)c	Position relative to ATG start site	No. of matches/total	Location in Borrelia genome	Comment (reference)
BB0083	Hypothetical protein	TTAAAATAATTATTA	−52	13/15	Chromosome
		TTTTAATTATTATAT	−94	13/15
BB0084	NifS protein (nifS)	ATATAATAATTAAAA	−49	13/15	Chromosome	Putative cysteine desulfurase (20)
		TAATAATTATTTTAA	−91	13/15
BB0166	4-α-Glucano transferase (malQ)	TGATATTTATTATAA	−66	13/15	Chromosome
BB0167	Outer membrane protein (tpn50)	TTATAATAAATATCA	−88	13/15	Chromosome
BB0565	Purine-binding chemotaxis protein (cheW-2)	TGATAATAATTATTA	−253	13/15	Chromosome	First gene in BB0565-BB0570 chemotaxis operon (15)
BB0657	Ribose 5-phosphate isomerase (rpi)	TACTAATAATTATAA	−116	13/15	Chromosome
BB0658	Phosphoglycerate mutase (gpmA)	TTATAATTATTAGTA	−47	13/15	Chromosome
BB0664	Hypothetical protein	TTATAATAATCATAT	−13	13/15	Chromosome
		TTAAAATCCTTATAA	−33	13/15
BB0665	Conserved hypothetical protein	TTATAAGGATTTTAA	−24	13/15	Chromosome
		ATATGATTATTATAA	−44	13/15
BB0689	Hypothetical proteind	ATATAATTATTATAA	−71	14/15	Chromosome
BB0690	Neutrophil-activating protein (NapA)	TTATAATAATTATAT	+6	14/15	Chromosome	Per box located −28 relative to second ATG start site
BBA69	Hypothetical protein	TTAATATAATTATAA	−71	13/15	Plasmid lp54
BBE22	Pyrazinamidase/nicotinamidase (pncA)	TATTAATAATTATAA	−255	13/15	Plasmid lp25	Gene required for mammalian infection (49)
		TTAATATTATTATAA	−374	13/15
BBG06	Conserved hypothetical protein	TTATAATTATAATAT	−145	13/15	Plasmid lp28-2	First gene in 4-gene plasmid maintenance locus (14)
		TTTTTATAATTATAA	−148	13/15
BBH40	Putative transposase-like protein	TTATAATTATAAAAA	−18	13/15	Plasmid lp28-3
BBI16	Hypothetical proteind	TTATAATAAATATAA	−55	14/15	Plasmid lp28-4	Virulent strain-associated repetitive antigen A (58)
		TTTAAATAATTATAA	−237	13/15
BBI28	Hypothetical proteind	TTATAATAAGTATAA	−47	14/15	Plasmid lp28-4	Homologue of BBI16
BBI41	Hypothetical protein	TTATAATTATAAAAA	−18	13/15	Plasmid lp28-4	Homologue of BBH40
		ATATTATAATTATAA	−21	13/15
BBI42	Putative outer membrane proteind	TTATAATTATAATAT	−317	13/15	Plasmid lp28-4
		TTTTTATAATTATAA	−320	13/15
BBJ19	Conserved hypothetical protein	TAATAATAATTATTA	−37	13/15	Plasmid lp38	First gene in 4-gene plasmid maintenance locus (14)
BBK01	Hypothetical proteind	TTATAATAATTATTC	−41	13/15	Plasmid lp36
BBK15	Putative P35 antigen	TAATATTTATTATAA	−191	13/15	Plasmid lp36
		TTATAATGATTACTA	−208	13/15
BBK47	Hypothetical proteind	TTATAATTATTATTA	−79	14/15	Plasmid lp36	B31 equivalent of N40 arp gene encoding arthritis-related protein (21)
BBK49	Hypothetical proteind	TTATAATTATTATTA	−79	14/15	Plasmid lp36	Homologue of BBK47
BBL36	Conserved hypothetical protein	TTATAAATATTATAG	−275	13/15	Plasmid cp32-8	Per box located within 180-bp inverted repeat of unknown function (14)
		TTTTATTAATTATAA	−332	13/15
BBM33	Conserved hypothetical protein	TAATTATAATTATAA	−52	13/15	Plasmid cp32-6	Last gene in 4-gene plasmid maintenance locus (14)
		TTAAAATAATTATAA	−58	14/15

^aCandidate Per boxes were identified by BlastN analysis of the B. burgdorferi genome employing an Expect value of 1,000 and a word length of 11. Sequences were excluded if they did not meet the following criteria: (i) no more than 400 nt upstream of a putative translational start site, (ii) no deeper than 50 nt into an upstream coding region, (iii) a minimum of 13 matches to the Per consensus sequence, and (iv) associated ORF at least 200 bp in size.

^bPutative identifications were from the TIGR website (http://www.tigr.org).

^cMatches to the Per consensus sequence (TTATAAT-ATTATAA) appear in boldface.

^dPutative lipoprotein.

Of the 11 chromosomal genes, 5 code for hypothetical proteins and 6 code for proteins identified by sequence homology (22). In some cases, the putative binding site was located in the region between two divergent genes and it may be that it is associated with only one or the other gene. Both genes, however, are listed in Table 2. In addition to the genes listed in Table 2, other genes were identified that had candidate sites within their promoter regions but did not meet the stated criteria. For example, upstream of the ospAB operon, a potential Fur binding site was identified (−104 nt relative to the ATG start site) that had 12 out of 15 matches to the Per consensus sequence and so was not included in Table 2.

In summary, fur was expressed by B. burgdorferi during growth in vitro. In gel shift studies, purified recombinant Fur bound DNA containing Fur boxes or Per boxes with an apparent K_d of ∼20 nM. Addition of Zn²⁺ (or Fe²⁺) decreased binding. Exposure to oxidizing agents (peroxides or diamide) also decreased binding. Bb-Fur bound the promoter regions of three Borrelia genes: BB0646 (gene encoding a putative lipase), BB0647 (fur), and BB0690 (napA). Other genes that may be Fur regulated were identified by in silico analyses. Fur may function in vivo as a repressor to regulate oxidative stress genes.