An Inverted Repeat in the ospC Operator Is Required for Induction in Borrelia burgdorferi

Borrelia burgdorferi, the spirochete that causes Lyme disease, differentially regulates synthesis of the outer membrane lipoprotein OspC to infect its host. OspC is required to establish infection but then repressed in the mammal to avoid clearance by the adaptive immune response. Inverted repeats (IR) upstream of the promoter have been implicated as an operator to regulate ospC expression. We molecularly dissected the distal inverted repeat (dIR) of the ospC operator by site-directed mutagenesis at its endogenous location on the circular plasmid cp26. We found that disrupting the dIR but maintaining the proximal IR prevented induction of OspC synthesis by DNA supercoiling, temperature, and pH. Moreover, the base-pairing potential of the two halves of the dIR was more important than the nucleotide sequence in controlling OspC levels. These results describe a cis-acting element essential for the expression of the virulence factor OspC.


Introduction
Lyme disease is caused by the spirochete Borrelia burgdorferi, which is transmitted via an Ixodes tick [1][2][3]. B. burgdorferi is maintained in an enzootic cycle between its tick vector and a vertebrate host reservoir [4][5][6]. Naïve larvae acquire B. burgdorferi when feeding on an infected animal and can then transmit the bacterium to uninfected hosts as nymphs during feeding, completing the cycle. B. burgdorferi encounters disparate, and hostile, environments as it transitions through the enzootic cycle; the spirochete has evolved a repertoire of strategies, which involve the regulation of a variety of genes, to respond and adapt to these changes during its lifecycle [6][7][8].
The differential syntheses of outer surface lipoproteins (Osp), which are the interface between B. burgdorferi and its hosts, is paramount to the ability to infect, survive and replicate in both the tick and the vertebrate [6][7][8][9][10]. These regulated genes encode wide-ranging functions; for example, VlsE mediates evasion from the vertebrate immune system [11][12][13], while OspA binds to a tick midgut protein and protects the spirochetes from the incoming blood meal [14][15][16][17]. One of the best-studied lipoproteins is OspC [18], which is required for transmission and the early stages of mammalian infection [19][20][21][22]. The ospC gene is carried on the conserved 26-kb circular plasmid cp26 [23,24]. The precise function of OspC remains elusive, but the outer membrane lipoprotein appears to have several activities [25], including providing initial protection from the innate immune system [26] and facilitating dissemination [27]. Additionally, OspC has a ligand-binding domain essential for its function [28] and binds the tick salivary protein Salp15 [29] as well as mammalian plasminogen [30,31], which may assist in transmission and dissemination, respectively. OspC is highly immunogenic, so its synthesis must be repressed for the spirochete to persist in the mammal [32][33][34]. B. burgdorferi that continue to produce OspC during infection of immunocompetent mice are cleared [34] and ospC expressed in trans from a shuttle vector is selected against during mammalian infection [35]. While ospC is present in all B. burgdorferi strains examined, the sequence is variable, with only strains carrying certain alleles capable of disseminating and establishing infection in humans [13,[36][37][38][39].
Previous studies identified an ospC operator consisting of two overlapping inverted repeats (IRs) 42 bp upstream of the major transcriptional start site [23,[65][66][67]. The operator is highly conserved in B. burgdorferi sensu stricto [67], but there is some controversy about its role in ospC gene regulation. Eggers et al. [64] showed that the operator is required for full ospC expression in vitro, while Yang et al. [68] found the operator is dispensable for induction and repression. In addition, mutants lacking the operator were unable to persist in immunocompetent mice and were also cleared from SCID mice injected with transferred anti-OspC antibodies, suggesting that the operator is required to repress ospC expression during infection [69]. One caveat to these previous studies is that ospC was expressed in trans in an ospC null background strain. In the current study, we molecularly dissect the ospC operator in its native locus on cp26 and show that the distal IR is an important cis-acting element controlling OspC expression.

Bacterial Strains and Culture Conditions
Low-passage B. burgdorferi strain 297 (BbAH130) [49] and all mutant strains were maintained in Barbour-Stoenner-Kelly II (BSK-II) liquid medium, pH 7.6, containing 6% rabbit serum [70]. In temperature shift experiments, cultures were passaged twice at 23uC, before inoculating cultures at 1610 5 cells ml 21 and growing at 23uC to late log phase (5 to 9610 7 cells ml 21 ) or inoculating cultures at 1610 3 cells ml 21 and growing at 34uC to late log phase. For temperature downshift experiments, cultures were grown to log phase at 34uC before inoculating cultures at 1610 5 cells ml 21 and growing at 23uC to late log phase. In experiments examining the effect of pH, cultures at 34uC in BSK II at pH 8.0 were passaged into BSK II at pH 7.0. To determine the effect of DNA supercoiling, cultures were grown at 23uC (inoculated at 1610 5 cells ml 21 ) in the presence of 10 ng ml 21 coumermycin A 1 (50 mg ml 21 stock in DMSO) or DMSO (solvent control) until late log phase [42]. Cell density was determined using a Petroff-Hausser counting chamber [71].

Construction of ospC Promoter Mutants
Mutations in the ospC operator were generated by allelic exchange on cp26 [72]. Constructs containing the operator mutants were made by overlap extension PCR [73]. The 59 portion of the upstream construct was amplified by PCR of genomic DNA using KOD polymerase (Novagen) with the primers ospC U866F and ospCp mutHup(297)R or ospCp mutDH(297)R2 ( Table 1). The 39 portion of the upstream construct, which also includes the ospC gene, was amplified by PCR using the primers ospC D697R+AatII+AgeI and ospCp mutHup(297)F (Table 1). PCR products were separated and purified in a 1% agarose gel. Paired 59 and 39 portions of the upstream construct were then combined and extended for six cycles in a thermal cycler. Next, the primers ospC U866F and ospC D697R+AatII+AgeI (Table 1) were added and the combined upstream construct was amplified by PCR. To generate the construct for the wild-type control strain with the antibiotic resistance cassette, PCR was done using primers ospC U866F and ospC D697R+AatII+AgeI with genomic DNA as a template. PCR products were separated in a 1% agarose gel, gel purified, polyadenylated, and cloned into pCRH2.1-TOPO. PCR of genomic DNA with the primers ospC D673F+AatII and ospC D1572R+AgeI (Table 1) was used to amplify the downstream construct, which was cloned into pCRH2.1-TOPO as described above. The accuracy of all DNA constructs was confirmed by sequencing. The downstream sequence was inserted into the upstream ospC operator mutation constructs using the synthetic AatII and AgeI restriction sites. Lastly, the kanamycin resistance cassette flgBp-aphI [74] was inserted downstream of the ospC gene into the engineered AatII site. The orientation of flgBp-aphI was determined by PCR using the primers kanR 488R and ospC D1572R+AgeI (primers a and b, respectively). The plasmid was linearized by digestion with AhdI and electroporated into competent B. burgdorferi [61,71,72]. Transformants were cloned in liquid BSK-II by diluting the electroporated cells to less than one cell per well of a 96-well plate in medium containing kanamycin (200 mg ml 21 ) at 34uC and a 1.5% CO 2 atmosphere [16]. Total genomic DNA was isolated from positive colonies and sequenced by the Murdock DNA Sequencing Facility at The University of Montana using the primer ospC U291F to confirm the site-directed ospC operator mutations.

RNA Isolation and qRT-PCR Analysis
RNA was isolated from 40-ml cultures of B. burgdorferi grown at 23uC containing 10 ng ml 21 coumermycin A 1 or DMSO as a solvent control using TRIzol (Invitrogen) as previously described [61,75]. Samples were treated with Turbo DNase (Ambion) to remove contaminating DNA. Samples were screened by PCR to ensure that contaminating DNA had been removed using the primers flaB 423F and flaB 542R. cDNA was synthesized using 1 mg RNA with SuperScript H III for qRT-PCR (Invitrogen). flaB and ospC primers and probes were designed using Primer Express H version 3.0 (Applied Biosystems). TaqMan absolute qRT-PCR was performed in 96-well plates using an Applied Biosystems 7300 Real-Time PCR System and standard curves for flaB and ospC were generated using a portion of the flaB ORF (nucleotides 278-551 of the ORF) cloned into pCRH2.1-TOPO and B. burgdorferi strain 297 (BbAH130) genomic DNA, respectively. Values represent the mean (6SE) from three independent experiments.

SDS-PAGE and Immunoblotting
B. burgdorferi cultures were grown to late log phase at 23uC or 34uC and total cell lysates collected as previously described [61,72]. Equal amounts of protein were separated on pre-cast Novex 4-20% Tris-Glycine polyacrylamide gels (Invitrogen). Proteins were transferred by electroblotting to PVDF Immobilon TM membranes (Millipore) and membranes blocked in Blocking Buffer (138 mM NaCl, 2.7 mM KCl, 8.1 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , 0.05% Tween 20, 4% dry milk, and 1% goat serum) overnight at 4uC. Membranes were incubated with rabbit anti-OspC antibodies (1:1000) [68,76] or anti-FlaB antibodies (1:1000) (a kind gift from Tom Schwan) in Blocking Buffer for 1 h at room temperature. Rabbit antibodies were detected by incubating membranes with goat anti-rabbit HRP-linked antibodies (Bio-Rad Laboratories) (1:20,000) in Blocking Buffer for 1 h at room temperature. HRP-linked antibodies were visualized by chemiluminescence using Amersham TM ECL Plus (GE Healthcare) and images were collected using an LAS-3000 Intelligent Dark Box (Fujifilm Medical Systems USA). Images were processed using ImageJ (US National Institutes of Health and available at http://rsbweb.nih.gov/ij/) and Pixelmator (Pixelmator Team, Ltd).

Results
The ospC operator consists of a set of overlapping inverted repeats that are highly conserved in B. burgdorferi sensu stricto strains ( Fig. 1) [65][66][67]. We hypothesized that the dIR has a specific role in responding to changes in DNA supercoiling, which is known to regulate ospC expression [42]. Increased negative DNA supercoiling can alter DNA structure, including extrusion of cruciforms from IRs, which can have a regulatory effect on transcription [77][78][79]. We generated site-directed mutations in the native ospC operator on cp26 that specifically disrupted the dIR (dIR -) and that changed the dIR sequence but retained complementarity (dIR + ) ( Fig. 2A; sequence changes are in bold). The orientation of the kanamycin resistance cassette, flgBp-aphI, downstream of the ospC gene was determined by PCR using the primer sets a+b (Fig. 2B). Changes in the operator sequence of the mutant strains were confirmed by DNA sequencing; about onethird of the kanamycin-resistant clones contained the site-directed mutations.
The Role of the dIR in ospC Expression Induced by Relaxation of DNA Supercoiling We have previously shown that a decrease in negative DNA supercoiling causes an increase in ospC transcription [42]. The antibiotic coumermycin A 1 relaxes DNA supercoiling by inhibiting DNA gyrase [80][81][82][83]. Wild-type and mutant strains were grown at 23uC in the presence of coumermycin A 1 or DMSO (solvent only control) until late log phase to examine if the increase in ospC expression by the relaxation of supercoiling is mediated by the dIR of the operator. Transcript levels of ospC and flaB were measured by qRT-PCR. The fold increase in ospC transcript in the dIR + strain grown in coumermycin A 1 , compared to the solvent control, was significantly greater than that seen in the dIR 2 strain (about thirteenfold compared to less than twofold) (Fig. 3). The change in flaB transcript levels was about twofold or less for both the dIR + and dIR 2 strains (Fig. 3). These data suggest that the ability to form the dIR is an important regulator of coumermycin A 1mediated ospC expression.
To examine if the changes in ospC transcript levels in response to coumermycin A 1 treatment were reflected in OspC protein levels, total cell lysates were analyzed by Western blot using polyclonal antibodies to OspC to determine the levels of OspC synthesis. OspC levels increased in the wild-type 297 strain in response to relaxation of supercoiling when grown at 23uC in the presence of coumermycin A 1 compared to cultures grown in the DMSO control (Fig. 4). Very little OspC was produced when the dIR 2 strain was grown in coumermycin A 1 compared with the wild-type strain and the dIR + strain (Fig. 4). We have confirmed these results with a second independently constructed clone of each of the mutant strains (data not shown). Taken together, these results suggest that the ability to form the dIR is more important than the actual sequence, at least in regard to the nucleotides that we mutated. These data imply that the increase in OspC levels stimulated by relaxation of DNA supercoiling is mediated via the  dIR of the operator and that the secondary structure of the operator, rather than its sequence, plays a dominant role in regulating OspC levels.

The Role of the dIR in Temperature-regulated OspC Synthesis
A temperature shift from 23uC to 34uC is commonly used in vitro to mimic an environmental signal during tick feeding that induces virulence gene expression [8,40]. To assess the role of the dIR in temperature-regulated OspC expression, cultures were grown at 23uC, shifted to 34uC and grown to late log phase. Total cell extracts were analyzed by Western blot. The wild-type 297 strain increased OspC synthesis upon a temperature shift (Fig. 5A). Similar to the coumermycin A 1 treatment, OspC was not induced in response to temperature shift when the dIR was disrupted, while regenerating the complementarity of the dIR restored the ability to respond to increased temperature with increased OspC levels in the dIR + strain (Fig. 5A). A second clone of each mutant, from an independent transformation, showed the same pattern of OspC levels during temperature shift (data not shown).
We next assayed if the dIR was also involved in reducing OspC levels at 23uC. Cultures were grown to late log phase at 34uC and then passaged and grown at 23uC to late log phase. Again, changing the dIR sequence but maintaining complementarity in the dIR + strain allowed for the reduction in OspC levels similar to the wild-type strain shifted to 23uC (Fig. 5B).

The Role of the dIR in pH-regulated OspC Synthesis
OspC levels have also been shown to be regulated by pH [43,44], which is considered an environmental signal that changes during tick feeding: reducing the pH to 7.0 increases and raising the pH to 8.0 decreases OspC levels. To examine if pH-regulated OspC expression is mediated through the dIR, cultures were  grown at 34uC in medium at pH 7.0 and then passaged into medium at pH 8.0. Total cell extracts were collected at late log phase and OspC levels were analyzed by Western blot. The dIRstrain did not show increased OspC levels at pH 7.0 and the dIR + strain behaved like the wild-type strain in response to changing the pH (Fig. 6). Thus, all the signals examined, DNA supercoiling, temperature, and pH, control OspC levels through the dIR of the operator in vitro, suggesting a common mechanism, and the complementarity of the dIR may be more important than the specific sequence.

Discussion
Induction and repression of ospC transcription are crucial for B. burgdorferi to establish and maintain, respectively, an infection in mammals. A number of external factors, including temperature, pH, oxygen, carbon dioxide, acetate, and transition metals, regulate ospC expression [40][41][42][43][44][45][46][47][48]72,84]. Induction of ospC expression is generally accepted to be dependent on the RpoN-RpoS sigma factor cascade [49], which includes the regulatory proteins Rrp2 [51][52][53][54] and BosR [55][56][57][58]. Considerably less is known concerning the repression of ospC transcription, including the signals and accessory proteins involved. Here we show that the cisacting dIR of the ospC operator functions to control expression and our data indicate that the base-pairing potential of the two halves of the inverted repeat, rather than the specific sequence, is essential for induction, thus providing a level of ospC-specific regulation downstream of RpoS.
The large intergenic region upstream of the ospC gene contains the operator and is highly conserved among B. burgdorferi sensu stricto strains, much more so than the ospC gene itself, likely indicating selective pressure on the regulation of ospC expression [65,67]. Certain features of the operator, including the proximal IR (pIR), are broadly conserved throughout B. burgdorferi sensu lato genospecies, but, inexplicably, the dIR does not overlap the pIR in B. afzelii strains and is not even obviously present in B. garinii strains [67], suggesting alternate modes of gene regulation. Although we and others previously had shown that deleting the operator has little effect on the regulation of ospC transcription [68,69], our current results more closely agree with Eggers et al. [64], who showed that the entire operator was required for full ospC expression in response to a temperature shift. These data suggest that the dIR plays a role in ospC regulation. This discrepancy may be explained by the differences in the experimental approaches between the studies. In the present work, we have utilized a more precise method to dissect the operator: site-directed mutations were generated in cis in the endogenous operator on cp26, while the other studies utilized truncated operator mutants in trans on a shuttle vector in an ospC null background. Thus, ospC expression in trans from a plasmid much smaller than cp26 (7 kb compared to 26 kb), albeit still circular, with a strong promoter fused to a selectable marker, may affect operator function, especially when DNA topology is likely involved [42,68,85]. In fact, OspC levels expressed in trans were elevated at   23uC compared to wild type, even though the plasmid-borne ospC contained the entire operator region [68].
Mutations that disrupt the dIR, but maintain the pIR (dIR 2 strain) prevent an increase in OspC levels in response to temperature, pH or relaxation of supercoiling, suggesting that all these signals function through a similar mechanism. Thus, the dIR is required for an increase in the amount of OspC. The finding that the ospC induction by relaxation of supercoiling at 23uC with coumermycin A 1 (Fig. 3) depends on the dIR suggests that DNA topology has a regulatory role. These data imply that the regulatory element may be the DNA structure rather than the sequence, although we may not have mutated the nucleotides in the dIR + strain that are important in regulation. We propose that the inverted repeats mediate the effect of DNA supercoiling, possibly by extruding a cruciform, or bind a trans-acting factor that recognizes an alternative DNA secondary structure. This provides a molecular mechanism for our previous observation that relaxation of supercoiling induces ospC expression [42].
Finally, we are aware of only a handful of studies in which sitedirected mutations were introduced in cis into the genome of B.
burgdorferi [51,73,86,87] and this work provides an important caveat for interpreting genetic experiments involving introduction of DNA in trans, albeit the methodology is considerably more convenient. Our results add another level of complexity to ospC regulation suggesting that the DNA structure of the operator serves to mediate the external signals affecting expression.