MtrR Control of a Transcriptional Regulatory Pathway in Neisseria meningitidis That Influences Expression of a Gene (nadA) Encoding a Vaccine Candidate

The surface-exposed NadA adhesin produced by a subset of capsular serogroup B strains of Neisseria meningitidis is currently being considered as a vaccine candidate to prevent invasive disease caused by a hypervirulent lineage of meningococci. Levels of NadA are known to be controlled by both transcriptional regulatory factors and a component of human saliva, 4-hydroxyphenylacetic acid. Herein, we confirmed the capacity of a DNA-binding protein termed FarR to negatively control nadA expression. We also found that a known transcriptional regulator of farR in N. gonorrhoeae termed MtrR can have a negative regulatory impact on farR and nadA expression, especially when over-expressed. MtrR-mediated repression of nadA was found to be direct, and its binding to a target DNA sequence containing the nadA promoter influenced formation and/or stability of FarR::nadA complexes. The complexity of the multi-layered regulation of nadA uncovered during this investigation suggests that N. meningitidis modulates NadA adhesin protein levels for the purpose of interacting with host cells yet avoiding antibody directed against surface exposed epitopes.


Introduction
Neisseria meningitidis is a Gram-negative obligate human pathogen that colonizes the nasopharynx in 10-35% of adults [1]. For reasons not currently understood, commensal meningococcal (MC) colonization develops into an invasive disease causing septicemia and meningitis in 0.5 per 100,000 persons in the United States and up to 1,000 per 100,000 persons in sub-Saharan African epidemics [2]. The speed of disease progression results in up to 10-15% mortality even with antibiotic therapy [3], while often leaving survivors with permanent neurological complications [4]. Vaccines against the capsular polysaccharide of the most common disease-associated serotypes (A, C, W135, and Y) are available, leaving the hypervirulent and immune-evasive serotype B as a current focus for vaccine research [5].
Adhesion to the mucosal surface of the nasopharynx is the first step in successful colonization, mediated by a variety of factors, with type IV pili [6,7,8] and Opa and Opc proteins [9,10] produced in the greatest abundance. Recently, a non-fimbrial ''Oca'' family (Oligomeric coiled-coil adhesin) neisserial adhesin termed NadA was identified in 50% of hypervirulent MC capsular serogroup B lineages [11], but not in other capsular serogroup strains. Comprised of a leader peptide, globular ''head'' domain, a-helix intermediate region, and a C-terminal membrane anchor, NadA forms highly stable multimeric coiled-coil structures along the helical stalk, positioning the globular ''head'' for host cell interaction [12]. Importantly for consideration as a vaccine candidate, recombinant NadA lacking the C-terminal anchor elicits a bactericidal antibody response with epitopes accessible in encapsulated MC. Although nadA allele sequences differ between strains, varied antigen expression, not diversity, influences immune sera titer levels and protection [11]. Accordingly, the identification of factors influencing NadA levels at the gene expression level is critical for optimizing the efficacy of a NadA-targeted vaccine. Furthermore, understanding nadA expression may offer clues into the signals involved in converting a passive co-inhabitant of the human mucosal lining into an invasive and fatal septic infection.
MC uses a multi-tiered approach to control nadA expression. Maximum levels of the NadA protein are observed in stationaryphase in a growth-dependent manner [11], with expression of nadA varying widely among MC strains [13,14]. Upstream from the promoter are multiple tetranucleotide (TAAA) repeats whose number corresponds with varied nadA expression [13,15]. These repeats are phase variable, likely caused by slipped-strand mispairings during replication [16]. Several regulatory proteins bind to the nadA promoter (Figure 1), including integration host factor (IHF) and ferric uptake regulatory protein (Fur), though nadA expression is unchanged in a Fur null mutant [14]. Recently, a MarR-family transcriptional regulator, termed FarR and NadR in separate publications [14,17], was identified as a repressor of nadA, further expanding the list of nadA regulatory factors. This DNA-binding protein was first identified in the gonococcus (GC) and was shown to repress expression of the farAB-encoded efflux pump that is responsible for high levels of fatty acid resistance [18]. In contrast, MC FarR does not affect fatty acid resistance through FarAB, perhaps due to naturally high fatty acid resistance expressed by this pathogen [19]. Interestingly, however, MC FarR does bind to its farAB promoter region with relatively high affinity and represses farAB expression as shown by RT-PCR [20]. Because FarR regulates expression of farAB in both MC and GC, while nadA is present only in a subset of MC populations, we will continue to use the nomenclature of FarR for the repressor of nadA based on its more universal activity on farAB in both GC and MC.
The small molecule 4-hydroxyphenylacetic acid (4HPA) was identified as an inducer or de-repressor of nadA by relieving the DNA-binding activity of FarR [14]. Being a colonizer of the oropharynx, MC is washed in saliva, in which 4HPA is a common metabolite [21], possibly leading to increased expression of nadA and subsequent invasive disease. Curiously, FarR-controlled targets in GC are directly and indirectly regulated by the TetR family regulator MtrR. Repression of farR by MtrR indirectly upregulates farAB [18], while the gene encoding glutamine synthetase (glnA) is directly regulated by both FarR and MtrR [22]. Therefore, we questioned whether MtrR similarly affects nadA expression in MC, adding to the growing list of regulatory factors targeting nadA. Here we confirm that FarR is the primary repressor of nadA, yet MtrR, when expressed at elevated levels, directly represses nadA as well. Furthermore, DNA-binding and DNase I protection assays suggest that MtrR influences FarR binding at the nadA promoter similar to the phenomenon seen in glnA expression in GC [22], suggesting a higher complexity to Neisserial regulatory schemes that is conserved across species.

Results and Discussion
Control of nadA, farR, and mtrR expression in MC strain M7 Co-regulation and competitive regulation between FarR and MtrR has been shown previously for multiple targets in GC [18,22] but not MC. Given previous observations that in MC FarR can regulate nadA, we asked if MC MtrR can control farR expression. We also tested if MC MtrR could control nadA directly. To investigate possible influences of MtrR on nadA and/or farR expression in MC, the expression profiles for these genes were determined using a promoterless lacZ-fusion expression system ( Figure 1) employed previously to monitor gene expression in GC [23]. Using translational lacZ fusions to each gene's promoter (ranging from 427 to 524 bp; Figure 1), mtrR expression was compared to farR and nadA across multiple growth phases of broth-grown cultures or from overnight, agar-grown cultures ( Figure 1) of MC strain M7, which is a capsule-deficient mutant of strain NMB that is used for biosafety purposes [24]. This was done because earlier work [25] reported that farR was maximally expressed between late-log and stationary phase, while nadA expression peaks at stationary phase [11]. Importantly, the lacZ fusions did not significantly impact growth rates in broth for the M7-derived strains (Figure 2A), suggesting that any expression profile differences are not growth rate-dependent. Aliquots from different growth phases (Figure 2A; boxed A, B, and C) were assessed for b-galactosidase activity and compared against the activity of overnight cultures grown on GCB agar plates ( Figure 2B). The results showed that agar-grown MC had higher levels of expression for all three genes compared to broth-grown strains, and this was especially true for nadA. With respect to agargrown cultures, we noted that nadA expression was considerably greater than farR or mtrR with the latter being the most poorly expressed gene ( Figure 2B). Based on these results, all subsequent gene expression studies were performed on cultures grown overnight on GCB agar plates.
Having established that the lacZ fusion technology could be employed to monitor gene expression in strain M7, we next asked if MtrR can regulate farR as it does in GC and if, in turn, it can modulate nadA expression. We first asked if loss of MtrR impacts MC farR expression levels and for this purpose constructed an M7 farR-lacZ fusion strain (427 bp promoter region; Figure 1) in an mtrR null mutant that carries the non-polar aphA-3 cassette within the mtrR coding sequence. With this fusion strain, we noted a small but significant (p,0.01; Figure 3) increase in farR-lacZ expression, which was reversed by complementation when the M7 mtrR allele was expressed ectopically under its own promoter or an IPTGinducible promoter. Having observed a significant, albeit modest, influence of MtrR on farR-lacZ expression, we next asked if nadA expression, known to be negatively controlled by FarR [17], would be impacted due to loss of MtrR using a nadA-lacZ fusion (514 bp promoter region; Figure 1). We first confirmed that loss of FarR significantly increased nadA-lacZ expression (strain JC5AZ; Figure 4A), and this was reversed by complementation with the wild-type farR gene from strain M7 (strain JC6AZ). Interestingly, loss of MtrR alone (strain JC2AZ) resulted in a consistent, albeit small, increase in nadA-lacZ expression, though not to significant levels, with restoration of nadA to wild-type levels when mtrR was expressed ectopically under its own promoter (strain JC3AZ; Figure 4B). Furthermore, complementation with the wild-type mtrR sequence under the control of an IPTG-inducible promoter significantly (p,0.01) reduced nadA-lacZ expression by 26% (strain JC4AZ; Figure 4B). Based on these results, we concluded that FarR-mediated repression of nadA-lacZ was greater than that mediated by MtrR, but that MtrR could impact nadA expression by a FarR-independent mechanism. In order to further test this, we compared nadA-lacZ expression in a DfarR/DmtrR double mutant (strain JC7AZ) against the single DfarR mutant (strain JC5AZ). Here, no significant impact on nadA-lacZ expression was observed, further supporting a dominant repression of nadA by FarR. However, IPTG-induction of mtrR in the absence of FarR (strain JC9AZ) did result in a significant reduction in nadA (p,0.01) compared to both the DfarR/DmtrR double mutant (strain JC7AZ) and DfarR mutant (strain JC5AZ; Figure 4C). Taken together, the data suggested that MtrR, when overexpressed, has an influence on nadA expression that is independent from FarR-modulated expression in MC strain M7. What effects MtrR may have on nadA expression in other MC strains that differ in either mtrR or nadA expression is not yet clear and the subject of future study.
Western blot analysis was performed to confirm the results observed with translational nadA promoter-lacZ fusions ( Figure 4D). As expected, FarR was absent in the farR null mutant strains (JC5AZ, JC7AZ, and JC9AZ), but was present in the complemented mutants that expressed farR ectopically (strains JC6AZ and JC8AZ); although FarR levels were reduced in the complemented strains, repression of nadA was still evident ( Figure 4A, C, D; arrow). NadA migrates in its multimeric form even under denaturing/reducing conditions [11], complicating discrimination of differences in NadA protein levels when analyzing the dominant band alone; therefore, minor NadA-dependent bands, observed elsewhere [20], were used for analysis of steady state levels of NadA ( Figure 4D; arrows). Despite repeated attempts, MtrR could not be detected by immunoblot in any strain except when overexpressed (strains JC4AZ and JC9AZ), suggesting that MtrR is maintained at low levels in N. meningitidis. This low level of MtrR may explain why enhanced expression of farR in the mtrR null mutant was modest ( Figure 4) and why its ability to repress nadA could only be observed when mtrR was over-expressed ectopically (strain JC4AZ; Figure 4B). In agreement with the data obtained with the nadA promoter translational fusions, steady-state levels of NadA exceeded wild-type in the absence of FarR ( Figure 4D; arrows point to minor bands used for comparison). Unfortunately, differences in NadA levels were not discernible when comparing +mtrR to DmtrR strains in the +farR (strains M7AZ, JC2AZ, JC3AZ, and JC4AZ) or DfarR (strains JC5AZ, JC7AZ and JC9AZ) backgrounds ( Figure 4D). Assessing differences in steady-state levels of NadA in other clinical MC strains following the loss of mtrR requires further study.
FarR and MtrR binding to a target DNA sequence upstream of nadA The data obtained using translational fusion strains bearing an over-expressed wild-type mtrR allele in an mtrR null mutant background indicated that elevated levels of MtrR can exert negative regulatory influences on nadA expression in MC strain M7. Based on this hypothesis, we asked if MtrR can bind target DNA sequences upstream of nadA by employing electrophoretic mobility shift assays (EMSA) using FarR or MtrR fused to maltosebinding protein (MBP) at their C-terminus;maps describing the various DNA probes used are shown in Figure 1. As described previously [14], FarR was found to bind the target nadA sequence (457 bp) in a specific manner with at least three DNA-protein complexes observed ( Figure 5A, arrows). Furthermore, the results from binding specificity assays suggested that FarR has greater affinity for the nadA-promoter than to its own (443 bp) or farAB promoter-bearing sequences (497 bp and shown in Figure 5B); importantly, a probe lacking a FarR-binding site (rnpB; 609 bp) was unable to compete with the labeled nadA probe for complexing with FarR.
Having verified that the nadA probe could be recognized in a specific manner with an MC DNA-binding protein (FarR), we next asked if nadA could bind MtrR. First, we confirmed by EMSA the DNA-binding capacity of the MtrR-MBP fusion protein by evaluating its ability to bind a known target DNA sequence, namely the promoter-bearing region upstream of MC mtrCDE (811 bp). In GC, MtrR is a repressor of the mtrCDE-encoded antimicrobial efflux pump by virtue of its binding between the 210 and 235 promoter elements [26,27]. We found that as little as 1.0 mg of MC MtrR-MBP incubated with MC mtrCDE promoter-bearing probe resulted in virtually a complete shift ( Figure 6A; arrow) of the target sequence. Importantly, a similar shifting of the nadA probe by MtrR was observed ( Figure 6B; arrow). Specificity of MtrR-binding to the nadA probe was confirmed by use of competitive EMSA. Although the heterologous unlabeled rnpB promoter-bearing probe did to some extent compete for binding, especially at a high concentration (1006), the mtrCDE and nadA unlabeled probes proved to be more effective competitors at a lower concentration (256) with the mtrCDE probe appearing to be superior ( Figure 6C).
The ability of both FarR and MtrR to bind the nadA promoterbearing region in a specific manner was reminiscent of their ability to bind the DNA sequence upstream of glnA of GC (22). As the binding of either protein to the glnA target can impact binding of the other we asked if a similar situation might exist for the MCderived nadA target. To test this possibility, we pre-incubated the nadA probe with a fixed concentration of one protein and then introduced increasing amounts of the second protein. In the absence of a competing protein, both MtrR and FarR exhibited a distinct shift of the probe (Figure 7; arrowhead and arrows, respectively). As FarR-MBP concentrations increased following   In order to learn the mechanism by which MtrR could influence the formation of FarR::nadA complexes, we used DNase I protection assays to determine if their respective binding sites might be in close proximity. We confirmed FarR-MBP-binding to the three sites (data not shown) described by Metruccio et al. [14], which include the 210 promoter and TAAA phase-variable regions ( Figure 8B). In repeated DNase I protection assays, clear evidence for a sequence(s) capable of recognizing MtrR could not be obtained. However, in these experiments a DNase I hypersensitive site was identified positioned at the end of the phase-variable TAAA repeats ( Figure 8B; asterisk at nucleotide position 209). The presence of this suggests an interaction of MtrR with a sequence upstream of nadA that could influence binding of FarR.
While N. meningitidis colonizes up to 35% of humans [1], fewer than 1% of the population develops an invasive infection [2], suggesting that the bacterium focuses on a more commensal lifecycle. The ability to effectively transition between passive residence and active infection relies on tight transcriptional regulation involving an array of external and internal control systems. In GC, the transcriptional regulators FarR and MtrR have been well-characterized for their role in antimicrobial resistance, allowing for host persistence [28]. FarR represses expression of the fatty-acid efflux pump FarAB [29]; MtrR represses expression of the mtrCDE, which encodes an antimicrobial efflux pump [26,30]. Recently, MC FarR was shown to repress expression of nadA [17], whose gene product is a highly immunogenic adhesin and invasin associated with hypervirulent strains of serotype B MC [11,12]. Interestingly, GC MtrR has also been shown to repress GC FarR, thereby influencing transcription of farAB [18]. As this regulation of a regulator is not unique to GC [22], we explored whether MtrR likewise modulates farR in MC, thus affecting nadA expression.
Our results suggest that MC employs a dual-repressor approach to control nadA expression. Using lacZ translational fusions, EMSA, and DNase I protection assays, we confirmed earlier work [14,17,25] that FarR is a negative regulator of nadA due to its ability to bind target DNA upstream of the coding region ( Figure 7B) and affect subsequent expression ( Figure 4A). Complicating this regulatory scheme, our results indicate that MtrR can exert an influence on nadA directly by interacting with the upstream DNA sequence and indirectly through its ability to reduce farR expression (Figure 3). With respect to the first mode of MtrR regulation over nadA, our DNA-binding studies indicate that MtrR can bind upstream of this gene in a specific manner ( Figure 6C) and can impact the formation and/or stability of FarR::DNA complexes (Figure 7) when its level exceeded that of FarR. The stronger influence of FarR on nadA expression is likely due to its ability to recognize three target sites ( Figure 7B) while under the conditions employed in the DNase I protection assay. A possible site for MtrR binding could only be surmised by the presence of a DNase I hypersensitive site ( Figure 7A). Interestingly, this site is positioned within a tract of tetranucleotide repeats and a FarR-binding site ( Figure 7B).
Under what conditions might MtrR-mediated regulation of nadA have biologic relevance given the strong regulatory action of FarR? We propose several potential mechanisms: The develop- ment of mutations impacting FarR regulation of nadA would require alternative mechanisms of transcriptional regulation that could in part be fulfilled by MtrR. Thus, mutations in FarR that reduce its DNA-binding activity or mutations in FarR-binding sites could enhance nadA expression unless other regulatory processes are available. Alternatively, mutations that enhance MtrR levels or interactions with nadA-binding sites might repress nadA expression above that seen by FarR alone. Precedent for clinical isolates of Neisseria bearing regulatory mutations impacting gene expression exists in that gonococcal strains isolated from patients frequently contain mutations in mtrR and these can cause dysregulation of the mtrCDE-encoded efflux pump operon. Furthermore, cis-acting regulatory mutations can influence transcription of mtrR and/or mtrCDE or directly enhance expression of the mtrCDE-encoded efflux pump [28,31,32,33]. In M7, mtrR expression is typically low, yet overexpression of mtrR results in almost 30% repression of nadA ( Figure 4B and D). Accordingly, it will be important to evaluate MC clinical isolates to determine if they may develop mutations impacting nadA expression directly or indirectly; the latter being due to mutations in farR or mtrR. With the expression of nadA varying significantly between MC strains [13,14], mutations affecting farR and/or mtrR expression may have more profound effects on nadA expression than those observed here. Deletion of the Correia element or IHF binding site upstream from mtrR affects expression of mtrCDE, which is an MtrR target [32]; Enriquez et al. observed several MC isolates with Correia element deletions upstream from mtrR, including one serotype B, suggesting that these mutations are not an exception [34]. We propose that this multi-layered regulation of nadA, which now includes direct regulation by MtrR, reflects an effort by MC to balance levels of the NadA adhesion important for interacting with host cells yet avoiding potentially protective antibody responses.

Bacterial strains and growth conditions
All N. meningitidis strains listed in Table 1 are derivations of strain M7 constructed for this study; M7 is a stable capsule-negative variant of strain NMB and was used for biosafety purposes. MC were cultured on GCB agar (Difco Laboratories, Detroit, MI) with defined supplements I and II [35] at 37uC under 3.8% (vol/vol) CO 2 . For growth-phase analysis, MC were grown in a shaking incubator at 37uC in GCB broth with sodium bicarbonate and defined supplements I and II as previously described [35].

Strain construction and verification
For construction of strain JC2, overlapping PCR products were generated to replace mtrR with aphA-3 [36] by allelic exchange at the native N. meningitidis locus, conferring kanamycin resistance. Specifically, primers mtrC_R_out_59 and mtrR_R_Kan_59ovhg generated product A; kan_F_mtrR_59ovhg and kan_R_mtr-R_39ovhg generated product B; and mtrR_F_Kan_39ovhg and NMB1718_FWD generated product C ( Table 2). Products A, B, and C were used as template with flanking primers mtrC_R_out_59 and NMB1718_FWD (Table 2) to generate the final PCR product used for transformation of wild-type M7 with selection on kanamycin at 50 mg ml 21 . The substitution of mtrR with aphA-3 was verified by PCR (data not shown) and Western blot (Figure 4). For construction of JC5, PCR products generated by FarR_prom_F and FarR _Sma_R ( Table 2) were subcloned into vector pCRH2.1 (Invitrogen) generating pJC5a with selection on ampicillin at 100 mg ml 21 . PCR products from FarR_ Sma_F & FarR_pmalC_Xba_R (Table 2) were purified, digested with SmaI, and ligated into pJC5a at the SmaI site, generating pJC5b, and sequenced for accuracy. Finally, pHP45V was digested with SmaI to liberate the spc cassette then ligated into the SmaI site of pJC5b generating pJC5 used to transform M7 with selection on spectinomycin at 60 mg ml 21 . The interruption of farR with spc was verified by PCR (data not shown) and Western blot (Figure 4). pLES94 pUC18K-derivative allowing a translational fusion to a promoterless lacZ fusion and insertion between neisserial proA and proB genes. [23] pGCC3 NCIS vector for insertion of genes directed under their native promoter between neisserial lctp and aspC genes. [37] pGCC4 As pGCC3 except inserted genes are directed under an IPTG-inducible promoter [37] pMal-c2 IPTG-inducible expression vector for fusion of proteins to maltose binding protein (MBP) and cytoplasmic expression Complementation of JC2 and JC5 was accomplished by delivering the parent gene ectopically using vector pGCC3 or pGCC4 [37]. Briefly, primers mtrR_F_Pac_GC3 and mtrR_R_Pme or mtrR_F_Pac_GC4 and mtrR_R_Pme ( Table 2) were used to amplify the mtrR allele then digested with appropriate restriction enzymes and ligated into pGCC3 and pGCC4, respectively. Similarly, farR_F_Pac_GC3 and farR_R_Pme (Table 2) were used to amplify the farR allele for subsequent digestion and ligation into pGCC3. All constructs were verified by sequencing prior to transformation. Transformants were selected on erythromycin at 1 mg ml 21 .
Construction of lacZ reporter fusions, b-galactosidase assay and immunoblot analysis All lacZ fusions used in this study were prepared in pLES94 and performed as described previously [18,23] using appropriate primers. For nadA-lacZ, primers nadA_F_Bam and nadA_lacZ_R were used; for farR-lacZ, primers farR_F_Bam and farR_lacZ_R were used (Table 2); and for mtrR-lacZ, primers mtrR_F_Bam and mtrR_lacZ_R were used ( Table 2) Constructs encoding a lacZ fusion were grown overnight on GCB agar with supplements, 1 mM IPTG, and 5 mM 4-hydroxyphenylacetic acid, when appropriate. Cells were harvested directly from plates or used to inoculate GCB broth with appropriate supplements and 1 mM IPTG then grown through stationary phase for growth-phase analysis of protein expression. Cells collected from overnight plates were resuspended in phosphate-buffered saline pH 7.2 (PBS), centrifuged for 2 min at 13,000 rpm, and stored overnight at 220uC after the supernatant was removed. From liquid cultures, 5 mL aliquots were removed at the indicated growth phases ( Figure 1) and centrifuged for 15 min at 5,000 rpm. After removing the supernatant, cell pellets were resuspended in PBS, centrifuged for 2 min at 13,000 rpm, and stored overnight at 220uC. To determine b-galactosidase specific activities, cell pellets  were resuspended in 50/50 PBS and Z-buffer (60 mM Na 2 HPO4, 40 mM NaH 2 PO4, 10 mM KCl, 1 mM MgCl 2 , 50 mM b-Mercaptoethanol; pH 7.0). After addition of 20 mL 0.1% SDS and 40 mL chloroform, samples were vortexed and incubated at room temperature for 5 min. Protein concentrations were quantified by Nanodrop1000 (Nanodrop Technologies,Wilmington, DE). To 200 mL of Z-buffer, 30 mL of each cell lysate and 70 mL of ONPG (2-Nitrophenyl-b-D-galactopyranoside; Sigma, St. Louis, MO) at a concentration of 4 mg ml 21 in Z-buffer was added. Following a color change, the reaction was stopped with 500 mL of 1 M Na 2 CO 3 solution. The reactions were centrifuged at room temperature for 5 min at 13,000 rpm to remove cell debris and 200 mL of supernatant was transferred to a 96-well microtiter plate and analyzed at 420 nM by a PerkinElmer Victor X3 microplate reader. For data analysis, specific activity was calculated using the formula: {[(OD 420 * v)/(4500 nL nmoles 21 cm 21 61 cm)]/t/mg protein} with v being the volume used and t being the reaction time. All reactions were performed in triplicate and repeated at least 3 times. Statistical analysis was performed using multivariate ANOVA followed by Tukey HSD post-hoc pairwise comparison using SAS 9.2 software (The SAS Institute, Cary, NC). Verification of protein absence, overexpression, and comparison between strains were assessed by immunoblot. Total protein was quantified by Nanodrop1000 (Nanodrop Technologies) prior to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) separation [38] and Western immunoblotting [39]. Rabbit polyclonal antibodies were used at the following dilutions: anti-NadA, 1:2,000; anti-FarR, 1:5,000; and anti-MtrR, 1:1,000. Antirabbit alkaline phosphatase-conjugated secondary antibody (Bio-Rad Laboratories, Hercules, CA) was used at 1:7,500.

Purification of MtrR and FarR
Fusion and purification of N. meningitidis M7 MtrR and FarR to maltose-binding protein (MBP) were performed per manufacturer's guidelines (New England Biolabs, Beverly, MA) and as described previously [27] with some exceptions. Primers mtrR_pmalC_F & mtrR_pmalC_Xba_R and primers farR_p-malC_F & farR_pmalC_Xba_R were used to PCR amplify the mtrR and farR alleles, respectively, before XbaI digestion and bluntend ligation into pMal-c2 (New England Biolabs). Both constructs were sequenced for accuracy. Factor Xa digestion was not performed due to protein stability issues; therefore, all DNAbinding investigations utilizing these proteins maintained an intact MBP fusion, which has been successful in prior investigations [27]. Analysis of eluted fractions by SDS-PAGE revealed a 65-kDa band consistent with the expected size an MtrR-MBP fusion (data not shown). For consistency, N. meningitidis M7 FarR was likewise fused to MBP and purified, in which the purified protein band migrated to about 60 kDa (data not shown).
DNase I protection assays were performed as previously described [22,27,40] with slight modifications. Target DNA probes were generated with PCR pimers nadA_prom_F and nadA_prom_R ( Table 2). The 59 end was labeled with T4 polynucleotide kinase as described above for EMSA probes. Purified MtrR-MBP was incubated with target DNA for 15 min at 37uC. Dnase I in loading buffer [10 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 4 mM CaCl 2 , 1 mM dithiothreitol, 40% glycerol] was added to each reaction for 1 min at 37uC. The reactions were stopped with DNase I stop buffer (95% ethanol and 7.5 mM ammonium acetate), plunged in an ethanol dry-ice bath for 15 min, and precipitated overnight at 280uC. Pellets were washed in 70% (vol/vol) ethanol, dried, and resuspended in loading buffer (Epicentre). Resuspended reaction mixtures were loaded on 6% denaturing polyacrylamide gel and resolved by autoradiography.