Two Distinct Mechanisms Govern RpoS-Mediated Repression of Tick-Phase Genes during Mammalian Host Adaptation by Borrelia burgdorferi, the Lyme Disease Spirochete

ABSTRACT The alternative sigma factor RpoS plays a key role modulating gene expression in Borrelia burgdorferi, the Lyme disease spirochete, by transcribing mammalian host-phase genes and repressing σ70-dependent genes required within the arthropod vector. To identify cis regulatory elements involved in RpoS-dependent repression, we analyzed green fluorescent protein (GFP) transcriptional reporters containing portions of the upstream regions of the prototypical tick-phase genes ospAB, the glp operon, and bba74. As RpoS-mediated repression occurs only following mammalian host adaptation, strains containing the reporters were grown in dialysis membrane chambers (DMCs) implanted into the peritoneal cavities of rats. Wild-type spirochetes harboring ospAB- and glp-gfp constructs containing only the minimal (−35/−10) σ70 promoter elements had significantly lower expression in DMCs relative to growth in vitro at 37°C; no reduction in expression occurred in a DMC-cultivated RpoS mutant harboring these constructs. In contrast, RpoS-mediated repression of bba74 required a stretch of DNA located between −165 and −82 relative to its transcriptional start site. Electrophoretic mobility shift assays employing extracts of DMC-cultivated B. burgdorferi produced a gel shift, whereas extracts from RpoS mutant spirochetes did not. Collectively, these data demonstrate that RpoS-mediated repression of tick-phase borrelial genes occurs by at least two distinct mechanisms. One (e.g., ospAB and the glp operon) involves primarily sequence elements near the core promoter, while the other (e.g., bba74) involves an RpoS-induced transacting repressor. Our results provide a genetic framework for further dissection of the essential “gatekeeper” role of RpoS throughout the B. burgdorferi enzootic cycle.

promoters, whereas modulation of bba74 expression involves a putative RpoSdependent repressor that binds upstream of the core promoter. Thus, Lyme disease spirochetes employ at least two different RpoS-dependent mechanisms to repress tick-phase genes within the mammal.
nothing is known regarding the mechanism(s) underlying repression by RpoS. This is, to a large extent, due to the fact that RpoS-mediated repression does not occur in vitro under experimental conditions (i.e., following temperature shift) in which RpoSdependent genes are known to be induced. Instead, repression by RpoS requires environmental stimuli that are unique to the mammalian host milieu (22,26). In the past, we have circumvented this limitation using the dialysis membrane chamber (DMC) peritoneal implant model to generate mammalian host-adapted B. burgdorferi (22,28,36,37). Among the cohort of tick-phase genes subject to RpoS-mediated repression are the glp operon (bb0240 to -243), ospAB (bba15/16), and bba74. To better understand the molecular mechanism(s) underlying RpoS-dependent repression, we used a series of green fluorescent protein (GFP) transcriptional reporter constructs, in conjunction with our DMC cultivation system, to explore the promoter elements of these three prototypical RpoS-repressed tick-phase loci. Our results suggest that B. burgdorferi employs at least two different mechanisms for RpoS-mediated repression within the mammal.

RESULTS
Comparison of ospAB, bba74, and glp operon promoter regions reveals few common regulatory motifs. We began by comparing the sequences upstream of ospAB, glpF, and bba74, three prototypical tick-phase genes, whose expression is known to be downregulated by RpoS in vivo. The transcriptional start sites (TSSs) for ospA and bba74 during growth in vitro have been previously reported (27,38). 5= rapid amplification of cDNA ends (RACE) was performed to definitively map the glpF TSS. Surprisingly, the TSS was located 195 bp upstream from the translation initiation site (Fig. 1). Consistent with these results, Adams et al. (20) recently reported the existence of a 195-bp untranslated leader sequence (UTR) in the glp operon by global 5=-end mapping and Northern blotting.
Sequence alignment revealed that the extended promoter regions for ospAB, the glp operon, and bba74 differ at the primary sequence and secondary structure levels (Fig. 2). Multiple groups have examined the cis regulatory elements in the upstream region of ospA (26,39,40). Collectively, they suggested that a T-rich region is required for maximum expression in vitro. The region upstream of the bba74 TSS also contains a T-rich region, but no such element is discernible for the glp operon (Fig. 2). In addition, the ospAB upstream region contains a direct repeat and the bba74 extended promoter contains an inverted repeat. No obvious secondary structure motifs are discernible in the glp promoter upstream region (Fig. 2). The lack of any obvious shared upstream primary sequences or predicted secondary structure motifs raised the possibility that multiple mechanisms of RpoS-mediated repression are operative in B. burgdorferi.
Repression of ospAB and glp operon expression occurs by a mechanism different from that regulating bba74. In order to identify potential cis regulatory elements, GFP transcriptional reporter constructs containing various amounts of sequence upstream of the TSSs of ospAB, bba74, and the glp operon were transformed into B31 5A18 NP1 ( Fig. 3; see also Fig. 9). For each gene, we first examined GFP expression by the construct containing the greatest amount of upstream sequence during in vitro growth at 37°C. Pbba74(Ϫ275) elicited the highest average mean fluorescence intensity (MFI) [1.5-fold and 4.5-fold greater than the PospA(Ϫ102) and Pglp(Ϫ184), respectively]. The MFIs for both PospA(Ϫ102) and Pbba74(Ϫ275) were significantly greater than that of Pglp(Ϫ184) (Fig. 4).
We next compared GFP expression by full-length and 5= truncations for each upstream region following temperature shift in vitro and cultivation within DMCs. As shown in Fig. 5A, all of the Pglp reporters produced measurable expression of GFP during in vitro growth, with Pglp(Ϫ184) eliciting the highest MFI. When 42 bp was removed [Pglp(Ϫ142)], expression of GFP in vitro decreased significantly (P ϭ 1.31 ϫ 10 Ϫ6 ), suggesting that this region may contain an activator/enhancer site. Deletion of As shown in Fig. 5B, the full-length and truncated PospA-gfp reporters expressed measurable levels of GFP in vitro at 37°C. We observed a significant decrease in GFP expression for PospA(Ϫ47) relative to PospA(Ϫ102) and PospA(Ϫ86), suggesting that the region between Ϫ86 and Ϫ47 may contain a positive regulatory element (Fig. 5B). Deletion of these 39 bp removes the poly(T) tract ( Fig. 2) that Sohaskey et al. (39) suggested was responsible for enhanced expression of ospA in vitro. When B. burgdorferi strains harboring these PospA reporters were grown in DMCs, all three exhibited a significant decrease in GFP expression (Fig. 5B). Importantly, there were no significant differences in GFP expression between the PospA reporters, suggesting that, as with the glp promoter, sequences near the core Ϫ35/Ϫ10 sequence motif are sufficient for repression of ospAB.
GFP expression driven by Pbba74(Ϫ275) was also significantly repressed within DMCs (Fig. 5C). In contrast to the Pglp and PospA reporters, removal of a region between Ϫ165   (Fig. 5C). This region, therefore, may contain an effector site required for repression in vivo.
Repression of ospAB and the glp operon during mammalian host adaptation is RpoS dependent. In order to determine if repression of the Pglp and PospA reporters during mammalian host adaptation is RpoS dependent, we transformed the Pglp(Ϫ142) and PospA(Ϫ102) constructs into a B. burgdorferi strain 297 ΔrpoS mutant (25) and measured GFP in the transformants following cultivation in DMCs. Note that the RpoS mutant employed in these studies was on a strain 297 background, whereas the wild type was a B31 strain (B31 5A18). This was necessitated by the unavailability of a strain B31 ΔrpoS mutant at the time that these experiments were performed. Several factors suggest that this should not represent a problem. First, RpoS regulon expression levels have been shown to be similar between the B31 and 297 strains (22,28). In addition, we sequenced the regions upstream of the transcription start sites for glpF and bba74 in strain 297. For glpF, there is an A¡T change at position Ϫ3, and for bba74, there is a G¡A change at position Ϫ253 and an A¡G change at position Ϫ53. Comparison of the B31 and 297 sequences deposited in GenBank for the region upstream of ospA revealed a single G¡A change at position Ϫ26. None of these single nucleotide polymorphisms (SNPs) are likely to affect either expression or RpoS-mediated regulation. In accord with previous findings for the native genes (12,22,27), we saw substantially higher expression of GFP (i.e., derepression) for both reporters in the absence of RpoS (Fig. 6). Moreover, these results indicate that the Ϫ46/Ϫ47 to Ϫ1 regions of both genes are sufficient for RpoS-dependent repression.
Repression of bba74 expression during growth in DMCs requires an RpoSdependent factor. Previously, we reported that repression of the native bba74 gene within DMCs is RpoS dependent (22,27). Above, we showed that repression of bba74 in vivo requires upstream sequences located between Ϫ165 and Ϫ82 (Fig. 5C). These results led to the prediction that this region contains the binding site for an RpoSdependent repressor. Despite multiple attempts, we were unable to obtain ΔrpoS transformants harboring Pbba74-gfp constructs to test this conjecture. As an alternative approach, we employed electrophoretic mobility shift assays (EMSAs) with a DNA fragment encompassing 275 bp upstream of the bba74 TSS (Fig. 7). Incubation of this fragment with a cell extract prepared from wild-type spirochetes cultivated in DMCs resulted in a mobility shift (Fig. 7, lanes 3 and 6), which was completely inhibited by the addition of an 80-fold excess of unlabeled target DNA (lane 4). Importantly, incubation of the target DNA with a cell extract prepared from DMC-cultivated ΔrpoS mutant did not produce a shifted product (lane 8). Incubation with a cell extract from wild-type B. burgdorferi grown in vitro at 37°C also failed to produce a shift (lanes 9 and 10). These findings not only confirm, as expected, that repression of bba74 expression is RpoS dependent (27), but they also indicate that repression appears to be mediated by a factor that is produced or functions only in response to mammalian host-specific signals.
Role of 5= UTR in regulation of glp operon expression. 5= UTRs have been shown to control gene expression transcriptionally and posttranscriptionally (41). To elucidate the possible contribution of the leader sequence to glp expression, we constructed a GFP reporter containing the core promoter plus the 5= UTR, Pglp(Ϫ46/ϩ195). The presence of the leader sequence resulted in significantly higher GFP expression in vitro compared with the minimal promoter (P ϭ 0.004) (Fig. 8). However, the two reporters expressed GFP at similarly low levels within DMCs (P ϭ 0.821) (Fig. 8). Thus, the 5= UTR does not appear to play a role in the repression of the glp operon in vivo. The enhanced expression in the presence of the 5= UTR in vitro may be the result of either stabilization of the longer transcript or enhancement of translation efficiency, but the precise mechanism remains to be elucidated.

DISCUSSION
The cytosolic concentrations of factors typically exceed those of RNAP (1,42). On this basis, repression of 70 -dependent gene expression by alternative factors has generally been thought to result from competition for limiting apo-RNAP (42)(43)(44)(45). Alternatively, RpoS may control the expression of one or more repressor proteins or regulatory RNAs that is/are not induced or not fully active until spirochetes are within the mammal (22,26,46). A third possibility is competition between 70 -RNAP and RpoS-RNAP holoenzymes for promoter binding. In a recent study, Levi-Meyrueis et al. (47) demonstrated that a Salmonella enterica serovar Typhimurium RpoS mutant defective in DNA binding (but capable of forming RNAP holoenzyme) had a global expression profile highly similar to that of an RpoS-deficient strain. Further, RpoSdependent repression of selected genes was not solely the result of factor compe-  tition for binding to apo-RNAP. The authors concluded that RpoS-mediated repression resulted from occlusion of certain 70 promoters by RpoS-RNAP holoenzyme, thereby reducing 70 -dependent transcription initiation (47). Our results suggest that RpoSdependent repression of tick-phase genes in B. burgdorferi can occur by at least two different mechanisms. Repression of ospAB and glp operon transcription most likely occurs due to increased competition between 70 -RNAP and RpoS-RNAP holoenzymes for binding to the respective 70 promoters. In contrast, repression of bba74 appears to result from RpoS-induced synthesis or activation of a repressor protein that binds upstream of the core promoter, blocking 70 -RNAP holoenzyme access to the promoter.
Although our promoter fusion experiments were designed to shed light specifically on the mechanism of RpoS-mediated repression, additional potential regulatory insights were revealed during the study. ospA activation has been examined by multiple groups (26,39,40). Consistent with the requirement for the T-rich region for full ospA expression in vitro (39,40), the ospA Ϫ86 GFP construct [PospA(Ϫ86)] containing the T-rich region had significantly higher expression in vitro than the ospAB core promoter construct [PospA(Ϫ47)] (Fig. 5B). Xu et al. reported that a large direct repeat is required for maximum expression in vitro (40). The Ϫ102 ospA reporter [PospA(Ϫ102)] contains the intact repeat, and this region is disrupted in the Ϫ86 ospA reporter construct [PospA(Ϫ86)] (Fig. 2). GFP expression did not significantly change in the shorter construct (Fig. 5B), suggesting that the direct repeat is not required for maximum in vitro expression. Importantly, GFP expression was not significantly different among the ospA reporters during growth in DMCs (Fig. 5B). Thus, the upstream cis elements required for full ospAB expression during cultivation in vitro cannot overcome the repression that occurs during mammalian host adaptation.
Recently, Li, Liang, and colleagues (48,49) identified two putative binding sites for BosR near the ospA promoter and proposed that BosR functions as an ospA repressor. One of the BosR binding sites is located downstream of the TSS and, therefore, is not included in our PospA reporters, whereas the second site overlaps the PospA Ϫ10 region. Thus, BosR may contribute to the reduced expression of our PospA reporters in DMCs (Fig. 5B). If, however, BosR is able to bind directly to the ospA promoter, why is RpoS also required for repression in vivo? One possible explanation is that BosR and RpoS act cooperatively to block transcription by 70 -RNAP holoenzyme at the ospA core promoter. While analysis of our ospA reporters in BosR-and RpoS-deficient backgrounds could be informative, these studies are complicated by the fact that ΔbosR mutants also lack RpoS (31,50). The only way to definitively establish whether BosR functions as a direct repressor of ospAB is to conduct studies of B. burgdorferi constructs in which expression of rpoS is independent of BosR under mammalian host conditions. Glycerol utilization has been shown to be a fitness requirement for B. burgdorferi during the tick phase (12,51) and is apparently dispensable during mammalian infection (12,28). The glp operon is reciprocally regulated by the RpoN-RpoS and Hk1-Rrp1 TCS (22,46,51,52). Rrp1 is a diguanylate cyclase that catalyzes production of cyclic di-GMP, whose effect is mediated through PlzA, the only cyclic di-GMP binding protein in B. burgdorferi (46,53). glp operon expression is severely reduced in Rrp1 and PlzA mutants (46,51,52). Recently, Rel Bbu has also been shown to be an activator of glp operon expression (54,55).
Comparison of GFP expression from Pglp(Ϫ184) and Pglp(Ϫ142) during in vitro cultivation at 37°C reveals that the 42-bp region between Ϫ184 and Ϫ142 is required for maximal expression. Further, removal of this 42-bp sequence also results in decreased expression during growth in DMCs (Fig. 5A). This suggests that the Ϫ184 to Ϫ142 region may represent an enhancer site. glp operon expression in the mammal appears to be a composite of two opposing regulatory mechanisms-induction that is partially dependent on an enhancer site (Ϫ184 to Ϫ142) and repression mediated by RpoS at the core promoter (Ϫ46 to Ϫ1) (Fig. 5A). It is unclear whether Hk1/Rrp1 (c-di-GMP) and/or Rel Bbu is involved in mediating interaction with this putative enhancer sequence; however, further elucidation of the mechanism for glp operon induction could potentially be assessed by measuring GFP expression driven by Pglp(Ϫ184) and Pglp(Ϫ142) in Rrp1, PlzA, or Rel Bbu mutant backgrounds.
In contrast to ospAB and the glp operon, a sequence between nucleotides Ϫ165 and Ϫ82 upstream of the bba74 TSS is required for RpoS-dependent repression; removal of this region resulted in derepression under mammalian host-adapted conditions (Fig. 5C). An EMSA using the 283 bp upstream of the bba74 TSS as a target with cell extracts isolated from B. burgdorferi propagated in DMCs suggests the presence of a protein that specifically binds to this DNA region. This putative DNA-binding protein is present or active only in extracts from host-adapted spirochetes, consistent with the notion that its synthesis is dependent on RpoS (Fig. 7). The precise binding site at which this putative repressor binds is under investigation; however, a careful scan of the bba74 upstream region reveals several unusual sequence features that may play a role in the putative repressor protein binding (Fig. 2). In particular, there is an inverse repeat at nucleotides Ϫ138 to Ϫ132 (5=..TTTTAAT..3=) and Ϫ107 to Ϫ101 (5=..ATTAAAA..3=), both of which are flanked by CCC sequences that may be used to stabilize a potential stem-loop structure formed by these repeats. In addition, positions Ϫ125 to Ϫ121 (5=..TCTAA..3=) and Ϫ114 to Ϫ110 (5=..AATCT..3=) are reverse repeats that ensure that no alternative stem can be formed in this region of the promoter. Although transcriptional repressors typically bind closer to their cognate promoters, global bioinformatic analyses of transcription factors in E. coli and Bacillus subtilis identified numerous instances of repressor binding at a greater distance upstream (56,57). The putative RpoSdependent bba74 repressor is likely encoded by a gene induced by RpoS during mammalian growth.

MATERIALS AND METHODS
Identification of glp operon transcriptional start site. B. burgdorferi B31 A3 (58) was cultivated initially in Barbour-Stoenner-Kelly medium II (BSK-II) at 25°C to a density of 1 ϫ 10 7 cells/ml. Cultures were diluted to 3,000 cells per ml into 10 ml of fresh BSK-II and grown to late logarithmic phase (~1 ϫ 10 8 cells/ml) at 37°C. Cells were centrifuged for 15 to 20 min at 8,000 ϫ g, washed three times in 1ϫ phosphate-buffered saline (PBS), and resuspended in 10 l of 1ϫ PBS. RNA was extracted using the ToTALLY RNA kit (Ambion, Foster City, CA) according to the manufacturer's protocol. The RNA pellet was resuspended in 30 l of nuclease-free water and treated twice with DNase using the Ambion DNA-free kit (Ambion). Isolated RNA was stored at Ϫ80°C in the presence of RNasin until further use.
The transcriptional start site for glpF, the first gene in the glp operon, was identified using the 5= RACE System for Rapid Amplification of cDNA Ends kit, version 2.0 (Invitrogen, Carlsbad, CA). First-strand cDNA synthesis was accomplished using the gene-specific primer BB0240_race2 (Table 1). A 2.5-pmol amount of BB0240_race2 primer was added to 5 g of RNA and diethyl pyrocarbonate (DEPC)-treated water in a final volume of 15.5 l, and synthesis of first-strand cDNA, terminal deoxynucleotidyl transferase (TdT) tailing, and second-strand synthesis by PCR using BB0240_race/Abridged Anchor Primer pair (Table 1) were performed according to the manufacturer's instructions. cDNA recovery was checked by PCR  (Table 1), and PCR products were analyzed on a 1% agarose gel to confirm amplification of the 5= RACE products. The purified amplicons were cloned into pGEM-T (Promega, Madison, WI) and transformed into E. coli DH5␣ followed by blue/white selection on LB agar plates containing 100 g/ml ampicillin. Selected clones were confirmed by PCR using BB0240_race/Abridged Anchor Primer. Inserts from confirmed clones were amplified by PCR using the pGEM-T universal forward and reverse primers and sequenced (Genewiz, South Plainfield, NJ). Generation of transcriptional reporters and strains. Figure 9 contains a schematic describing the generation of the gfp transcriptional fusions for the glp operon, ospAB, and bba74 upstream regions. Chromosomal DNA containing the region upstream of the relevant TSS was amplified from strain B31 5A18 NP1 (59) by PCR using primers listed in Table 1, cloned into pGEM-T Easy (Promega) according to the manufacturer's instructions, and transformed into E. coli DH5␣ followed by blue/white selection on LB-ampicillin plates. Clones were confirmed by sequencing (Genewiz, South Plainfield, NJ) using pGEM-T universal forward and reverse primers. Insert orientation was determined by PCR and/or restriction enzyme digest. For the transcriptional fusions, the promoter regions of interest were amplified from each clone using pGEM-T Easy universal forward and reverse primers. Purified amplicons were digested with SphI (Fermentas, Pittsburgh, PA) and PstI (Fermentas) and ligated upstream of the promoterless gfp cassette in pCE191 (21) using the Rapid Ligation kit according to the manufacturer's protocol (Denville, South Plainfield, NJ). Ligations were transformed into E. coli DH5␣ and selected on LB agar plates containing 100 g/ml ampicillin. Clones were confirmed by PCR amplification and DNA sequencing. Promoter-gfp fusion cassettes were amplified from pCE191 using promoter-specific forward and pGFPrev primers (Table 1), cloned into pGEM-T Easy as described above, and subcloned into the B. burgdorferi-E. coli shuttle vector pBSV2-G (60) using SphI and SacI. Transformants were selected on LB agar plates containing 8 to 12 g/ml of gentamicin. Clones were confirmed by PCR using the corresponding forward promoter primer and pGFPrev (Table 1).
Each transcriptional reporter construct in pBSV2-G was isolated and purified in a large-scale plasmid extraction by the alkaline lysis method (61). Thirty to 40 g of plasmid DNA was electroporated into competent B31 5A18 NP1 (62), following which transformants were cultivated in 96-well plates in BSK-II containing 40 g/ml of gentamicin and 100 g/ml of kanamycin. For some experiments, reporter constructs were transformed into a strain 297 ΔrpoS mutant, CE174 (25), and cultivated in BSK-II containing 40 g/ml of gentamicin and 0.06 g/ml of erythromycin. Desired transformants were serially diluted and screened by PCR using the corresponding forward promoter primer and pGFPrev ( Table 1).
Cultivation of spirochetes in dialysis membrane chambers. B. burgdorferi strains containing the transcriptional reporters were cultivated in DMCs (22,36,37). Cells were initially grown in BSK-II at 25°C to a density of 1 ϫ 10 7 cells/ml, following which they were transferred to fresh BSK-II at an initial density of 3,000 cells/ml and cultivated at 37°C to late logarithmic phase (~1 ϫ 10 8 cells/ml). Temperature-shifted organisms were used to inoculate 10 ml of BSK-II within a DMC (8,000-Da cutoff) at 3,000 cells/ml. Chambers were implanted in the peritoneal cavities of 160-to 200-g female Sprague-Dawley rats (Harlan, Chicago, IL). Two weeks postimplantation, DMCs were removed and their contents were transferred to a sterile 15-ml tube. Spirochete density was immediately determined by dark-field microscopy (63).
Mammalian host adaptation of DMC-cultivated spirochetes was assessed by real-time reverse transcription-quantitative PCR (qRT-PCR) measurement of transcripts for ospC and the endogenous copies of ospA, glpF, and bba74 as described previously (28). Subsequent analyses used only samples that exhibited induction of ospC and repression of ospA, glpF, and bba74.
Flow cytometry. B. burgdorferi cells containing GFP transcriptional reporters were temperature shifted from 23°C to 37°C in BSK-II as described above and grown to a density of 1 ϫ 10 7 to 10 ϫ 10 7 cells/ml. A 1.0-to 1.5-ml amount of the culture was transferred to a microcentrifuge tube and centrifuged at 8,000 ϫ g for 10 min. For DMC-cultivated B. burgdorferi, 5 to 10 ml of DMC contents (~5 ϫ 10 7 organisms) was centrifuged at 8,000 ϫ g for 10 min. Cell pellets were resuspended in 500 l of 2.5 M SYTO 59 (Invitrogen) in TN buffer (10 mM Tris-HCl [pH 8.0], 1 mM EDTA, 100 mM NaCl) and incubated in the dark for 20 to 30 min at room temperature. The stained samples were washed twice in TN buffer, resuspended in 300 l of 1% paraformaldehyde in 1ϫ PBS, and stored in the dark at 4°C prior to analysis. Samples were analyzed by multiparameter flow cytometry using a MACSQuant Analyzer and MACSQuantify software (Miltenyi Biotec). Compensation for spectral overlap in the fluorescein isothiocyanate (FITC) (GFP) and allophycocyanin (APC) (SYTO 59) channels was performed using a GFPexpressing (no SYTO 59) strain and a non-GFP-expressing SYTO 59-stained strain (64). Fifty thousand events per sample were collected. Threshold values for GFP-and SYTO 59-positive cells were determined using unstained, nonfluorescent B. burgdorferi and SYTO 59-stained and unstained, GFP-expressing spirochetes. B. burgdorferi was gated to include only cells that stained with SYTO 59. The GFP (FITC) MFI of GFP-expressing SYTO 59 ϩ cells was determined. For each experimental parameter, technical duplicates were prepared from two or three independent cultures, and the values for technical replicates were averaged. GFP MFIs of experimental groups were compared by a two-tailed, unpaired t test. Significance was defined as a P value of Ͻ0.05.
EMSA. Cell lysates (from 5 ϫ 10 7 to 7 ϫ 10 7 cells total) were obtained from either B. burgdorferi B31 5A18 NP1 or CE174 cultivated at 37°C or in DMCs as described above. Cells were pelleted by centrifugation, washed twice in PBS, and suspended in 0.3 ml of BugBuster protein extraction reagent (Novagen, Madison, WI), 10 l phenylmethylsulfonyl fluoride, and 0.4 mg/ml lysozyme. Following incubation for 40 min on ice, extracts were cleared by centrifugation at 22,000 ϫ g for 30 min, and protein concentration was determined using the Pierce Coomassie blue assay (Pierce Biotechnology, Rockford, IL). Cell lysates were stored at Ϫ80°C in aliquots of 0.5 g/l and used only once. A 283-bp biotin-labeled DNA target representing the region upstream of the bba74 promoter was generated by PCR using primers 5=Biot-A74pfullF and A74pfullR (Table 1). Unlabeled competitor DNA was prepared by PCR using the same primer sequences, except that the forward primer was not biotinylated. Each EMSA was performed in a total volume of 20 l containing 50 ng of poly(dI-dC), 50 fmol biotin-labeled target DNA, and 3 g of protein lysate. In target competition reactions, an additional 4 pmol of unlabeled target DNA was added per reaction mixture. Protocols for the order of addition of reagents and reaction conditions followed those recommended by the manufacturer of the LightShift chemiluminescent EMSA kit (Thermo Fisher Scientific, Suwanee, GA). Since the order of addition of cell extracts and biotin-labeled target DNA may affect the specificity of the DNA-protein complexes, in all reaction mixtures that did not contain unlabeled competitor target DNA, components were premixed prior to adding the biotin-labeled target DNA. In reaction mixtures that did include competitive target DNA, unlabeled target DNA was added to the premixed components and chilled on ice for 20 min prior to the addition of biotin-labeled target DNA. All reaction mixtures were incubated at room temperature for 20 min. EMSA reactions were resolved by electrophoresis in 8% native polyacrylamide gels in 0.5ϫ TBE (10 mM Tris-borate, 10 mM boric acid, 50 mM EDTA, pH 8.0) buffer at 200 V. Gels were blotted to charged nylon membranes (Hybond-N ϩ ; GE Healthcare, Buckinghamshire, United Kingdom) and cross-linked by UV light. Visualization of DNA bands was accomplished with the LightShift EMSA chemiluminescent kit.