Control of gdhR Expression in Neisseria gonorrhoeae via Autoregulation and a Master Repressor (MtrR) of a Drug Efflux Pump Operon

ABSTRACT The MtrCDE efflux pump of Neisseria gonorrhoeae contributes to gonococcal resistance to a number of antibiotics used previously or currently in treatment of gonorrhea, as well as to host-derived antimicrobials that participate in innate defense. Overexpression of the MtrCDE efflux pump increases gonococcal survival and fitness during experimental lower genital tract infection of female mice. Transcription of mtrCDE can be repressed by the DNA-binding protein MtrR, which also acts as a global regulator of genes involved in important metabolic, physiologic, or regulatory processes. Here, we investigated whether a gene downstream of mtrCDE, previously annotated gdhR in Neisseria meningitidis, is a target for regulation by MtrR. In meningococci, GdhR serves as a regulator of genes involved in glucose catabolism, amino acid transport, and biosynthesis, including gdhA, which encodes an l-glutamate dehydrogenase and is located next to gdhR but is transcriptionally divergent. We report here that in N. gonorrhoeae, expression of gdhR is subject to autoregulation by GdhR and direct repression by MtrR. Importantly, loss of GdhR significantly increased gonococcal fitness compared to a complemented mutant strain during experimental murine infection. Interestingly, loss of GdhR did not influence expression of gdhA, as reported for meningococci. This variance is most likely due to differences in promoter localization and utilization between gonococci and meningococci. We propose that transcriptional control of gonococcal genes through the action of MtrR and GdhR contributes to fitness of N. gonorrhoeae during infection.

the importance of bacterial species-specific studies for examining regulatory properties of a common DNA-binding protein.

RESULTS
The gdh locus in N. gonorrhoeae. Similar to N. meningitidis, the gdh locus in N. gonorrhoeae FA19 is positioned 954 bp downstream from the mtr locus ( Fig. 1) and contains two open reading frames, gdhR, which encodes a GntR-like DNA-binding protein, and gdhA, which encodes L-glutamate dehydrogenase. The gonococcal GdhR and GdhA proteins are 97% and 98% identical, respectively (data not shown) to the equivalent proteins described for meningococci (16). The end of gdhA is positioned 238 bp from the end of gdhR and is transcribed in the opposite direction; transcription of gdhA has been reported to be activated by GdhR in meningococci (16).
Bioinformatic analysis (http://www.ncbi.nlm.nih.gov) revealed that the 200 bp upstream of gdhR in five gonococcal strains (FA19, FA1090, MS11, FA6140, and F89) were identical, except for a C-to-T change in FA1090 29 bp upstream of the translation start codon but after the transcription start site (TSS) (see below). In these same gonococcal strains, 100% identity was noted for the 500-bp sequence upstream of gdhA (data not shown). When the same regions from N. meningitidis strain MC58 in the corresponding upstream regions of gdhR and gdhA were used in a BLAST search against wholegenome sequences, two other meningococcal isolates (LNP21362 and H44/76) were found to have identical sequences, while 10 others showed 99% identity (data not shown). Thus, our use of gonococcal strain FA19 for comparison to meningococcal strain MC58 is suitable for determining differences in regulation of the gdh locus in these pathogens. Although the DNA sequences of the gdh loci in gonococci and meningococci are very similar, important differences exist, especially in the location of promoters and potential cis-acting regulatory sequences (see Fig. 3 for the CE insertion in the meningococcal gdhR and see Fig. 6 for the gdhA promoters, respectively). We hypothesized that the differences in these sequences between gonococci and meningococci could impact GdhR-mediated regulation of gene expression in gonococci and influence gonococcal biology. Accordingly, we sought to identify a phenotype that is linked to GdhR production in gonococci and then to examine regulation of gdhR expression and the capacity of GdhR to control model genes.
Loss of GdhR impacts in vivo fitness of gonococci independently of the mtr locus. Given the close location of gdhR to the mtr locus ( Fig. 1), we determined whether expression of GdhR influences transcription of mtrCDE and resistance of gonococci to antimicrobials recognized by the MtrCDE efflux pump (2,(6)(7)(8)(9)(10)(11)(12). For this purpose, we constructed a gdhR null mutant as well as a complemented strain. We found that the wild-type parent (FA19), the gdhR::kan mutant, and the complemented strain displayed identical levels of susceptibility to antimicrobials (erythromycin [Erm] MIC, 0.25 g/ml; penicillin MIC, 0.015 g/ml; Triton X-100 MIC, 100 g/ml), and these MICs varied according to the levels of the MtrCDE efflux pump (6,7,11,18). Moreover, results from quantitative reverse transcription-PCR (qRT-PCR) experiments indicated that expression of mtrC (the first gene in the mtrCDE operon [ Fig. 1]) was not impacted by loss of GdhR (data not shown).
We also assessed the fitness of the gdhR mutant relative to wild-type and complemented mutant bacteria during competitive infection of the lower genital tract of female BALB/c mice. Mice were inoculated vaginally with mixed bacterial suspensions containing similar numbers of the strains to be subjected to competition (total CFU per mouse, 10 6 ), and the number of each strain of bacteria recovered from vaginal swab suspensions on days 1, 3, and 5 postinoculation was expressed as a competitive index (CI), as described in Materials and Methods. Ninety percent of the mice were colonized throughout the 5-day study period, with 10 4 to 10 5 CFU/vaginal swab suspension recovered from the majority of mice on day 1 postinoculation and Ͼ10 4 CFU/ml on day 5 postinoculation (see Fig. S1 in the supplemental material). There was no difference in the relative fitness of the wild-type parental strain bacteria and the complemented gdhR::kan mutant strain (Fig. 2C). Interestingly, however, the gdhR::kan mutant strain Regulation of gdhR in Gonococci ® was significantly more fit than the complemented gdhR mutant strain on days 1, 3, and 5 postinoculation (geometric mean CIs, 18, 49, and 81, respectively) ( Fig. 2B) compared to the CIs for mice infected with the complemented mutant versus the wild-type strain (Fig. 2C). Competitive infections between the gdhR::kan mutant strain and the parental strain also showed elevated CIs (geometric mean CIs, 3.4, 30, and 21 on days 1, 3, and 5 postinoculation) ( Fig. 2A), although the differences were not statistically significantly different from those for the complemented mutant strain versus the wild-type strain. Consistent with the gdhR mutant strain outcompeting the GdhR-expressing bacteria in vivo, only mutant CFU were recovered from some mice inoculated with wild-type or complemented mutant bacteria mixed with the gdhR::kan mutant bacteria on days 3 and 5 ( Fig. 2A and B, open circles). These results indicated that GdhR production negatively influences the in vivo fitness of gonococci; importantly, loss of GdhR did not impact the growth rate of gonococci in GC broth (data not shown). This result circles and open triangles correspond to mice from which only mutant CFU or wild-type CFU were recovered, respectively. Bars represent the geometric mean CI values. Combined data from two experiments are shown, with data points from each experiment indicated in black or grey. Open circles indicate that only the mutant strain was recovered from the vaginal swabs at the indicated time point. Open triangles indicate that only the wild-type strain was recovered from the vaginal swabs at the indicated time point. The differences in the median CI between the FA19Str r versus FA19Str r gdhR::kan C3 and the FA19Str r gdhR::kan versus FA19Str r gdhR::kan C3 competitions were statistically significant on day 1 (P ϭ 0.02), day 3 (P ϭ 0.07), and day 5 (P ϭ 0.03) postinfection, based on the Kruskal-Wallis test with Dunn's multiple comparisons test (GraphPad Prism). Comparisons of the FA19Str r versus FA19Str r gdhR::kan competition with FA19Str r versus FA19Str r gdhR::kan C3 showed that the results approached but did not reach a statistically significant difference; P values for days 3 and 5 in this comparison were 0.055 and 0.056, respectively.

®
suggested that GdhR may be a negative regulator of in vivo fitness; therefore, we sought to determine how gdhR is regulated in gonococci and if GdhR controls expression of model genes (gdhR and gdhA) previously studied in meningococci (16).
MtrR is a direct repressor of gdhR expression. MtrR exerts transcriptional repression on the mtrCDE operon by binding to a promoter located upstream of mtrC (8,19). Given the close proximity of gdhR to the mtr locus in both gonococci and meningococci, we asked if MtrR regulates gdhR expression in gonococci. Although gdhR was not previously assigned to be an MtrR-regulated gene in an earlier transcriptional profiling study that employed gonococcal RNA prepared from mid-logarithmic-phase cultures (12), we reexamined MtrR control of gdhR for two reasons. First, the presence of a putative MtrR-binding site upstream of the gdhR gene ( Fig. 3A) suggested such control is possible. This MtrR-binding site (boxed in Fig. 3A) was 60% homologous to the MtrR-binding site on the mtrC promoter region (Fig. 3B). Second, the work of Mercante et al. (20) showed that a different transcriptional factor (MpeR) expressed in gonococci displays growth phase-dependent regulons.
Results from qRT-PCR experiments indicated that deletion of mtrR increases gdhR transcription, supporting the hypothesis that MtrR controls gdhR expression in gonococci by functioning as a repressor of this gene in the late-logarithmic phase of growth (Fig. 3C). Using primer extension (PE) analysis (see Fig. S2), we identified three gdhR TSSs, positioned 73, 72, and 70 nucleotides upstream of the start of translation of gdhR. These TSSs allowed us to identify a promoter element (5=-TAGAAT-3= for the Ϫ10 hexamer and 5=-TTGACG-3= for the Ϫ35 hexamer) 81 bp upstream of the ATG translational start codon (Fig. 3A). Importantly, the putative MtrR-binding site overlapped the Ϫ10 hexamer sequence of the predicted gdhR promoter. Based on this promoter mapping and the identification of a predicted MtrR-binding site within the putative gdhR promoter, we tested if MtrR bound in a specific manner upstream of the gdhR coding sequence, and we used an electrophoretic mobility shift assay (EMSA) for this purpose. We found that 2 g of MBP-MtrR was sufficient to completely shift a 32 P-labeled probe, termed R3/R2, that consisted of 393 bp of sequence upstream of the gdhR translational start (Fig. 4, lane 2). In order to better localize the MtrR-binding site(s) within this region, we performed a competitive EMSA with nonradioactive fragments of the R3/R2 probe used in the aforementioned EMSA. Binding competition assays showed that a smaller probe encompassing the gdhR promoter and its downstream region (probe R4/R2 [ Fig. 3A]) competed with MtrR binding to the labeled R3/R2 probe  (Fig. 4, lanes 5 and 6). Accordingly, we propose that MtrR represses gdhR expression by binding within the promoter sequence that contains the predicted MtrR-binding site.
GdhR regulation of model genes in gonococci. We selected two GdhR genes for study: gdhR and gdhA, which constitute the gdh locus (Fig. 1). We chose these 2 genes to test if gdhR is subject to autoregulation by its gene product and because gdhA has been reported to be a GdhR-activated gene in meningococci (16) and is positioned near gdhR in both pathogens.
Pagliarulo et al. (16) suggested that the meningococcal gdhR transcript originates in a Correia element (CE) (21) located upstream of the gdhR gene. An examination of 23 publicly available gonococcal genome sequences revealed that this CE is absent in the gdhR promoter region (data not shown). However, a putative GntR-like binding site (5=-TGTCATTA-3=) was identified between the Ϫ10 and the Ϫ35 sites of the gonococcal gdhR promoter overlapping the predicted MtrR-binding site (Fig. 3A, underlined in green). In order to investigate autoregulation of gdhR, qRT-PCR analysis of total RNA prepared from mid-and late-log-phase cultures of strains FA19 and FA19 gdhR::kan was performed. The expression of gdhR was increased by 4-fold at mid-log phase and by a little more than 2-fold at late-log phase in the GdhR-negative mutant compared to the parental strain (Fig. 5A). We hypothesize that insertion of a CE upstream of gdhR in meningococci, but not in gonococci, results in the utilization of distinct gdhR promoters  by these two related pathogens. Therefore, competing mechanisms of gdhR regulation by MtrR and GdhR itself may not occur in meningococci. In this respect, it is important to note that most meningococci encode an MtrR protein that contains loss-of-function mutations in mtrR (22,23), while nearly 80% of gonococci encode a wild-type MtrR (reviewed in reference 2).
Previous work indicated that expression of gdhA in meningococci is directed by two promoters, only one of which is regulated by GdhR (16). The GdhR-activated promoter of gdhA in meningococci has a putative GdhR-binding site (5=-TGTCAACA-3=) upstream of the Ϫ35 hexamer, based on similarity to the known GntR-binding site (5=-TGTcaACA-3=; the lowercase letters refer to nucleotides that differ from consensus GntR-binding site) in other bacteria (14); this site is also located in the corresponding gonococcal sequence (underlined in Fig. 6). In order to investigate whether GdhR binds to this site in gonococci, as was shown previously in meningococci, we performed EMSA competition analysis using purified gonococcal His-tagged GdhR protein. The results showed that gonococcal GdhR binds specifically to a DNA fragment encompassing the GntRbinding site present upstream of gdhA in gonococci (Fig. 7).

®
In meningococci, the presence of another gdhA TSS was detected 207 bp upstream of the ATG translational start codon (represented in blue in Fig. 6). However, a GdhR-binding site was not identified within or near this distal promoter in meningococci. Interestingly, we did not detect a TSS 207 bp upstream of the ATG in gonococci. This could be due to the presence of a mutation which changes the Ϫ10 hexamer from 5=-TAATTA-3= in meningococcal strain MC58 to 5=-TAACTA-3= in gonococcal strain FA19. To determine if gonococci have an additional promoter(s) for gdhA transcription, we used PE analysis to identify transcription start sites. The results suggested the presence of two promoters (Fig. 6). We identified a TSS located 8 bp downstream from a Ϫ10 hexamer that constitutes the homolog of the above-mentioned meningococcal promoter (shown in red). We also identified three TSSs located upstream of a nonconsensus Ϫ10 hexamer (5=-ATTTGT-3=) that is spaced 17 nucleotides from a weak Ϫ35 hexamer (5=-ATATGG-3=) (represented in green in Fig. 6). Importantly, this putative promoter has the previously identified GdhR-binding site (underlined sequence in Fig. 6) between its Ϫ10 and Ϫ35 hexamer sequences. This second gonococcal promoter was not identified in meningococci. Based on the location of the two putative gdhA promoters in gonococci, GdhR could impact expression of gdhA from both promoters through interaction with the identified GdhR-binding site.
Taken together, our promoter mapping studies suggest that differences exist regarding gdhA transcription in gonococci and meningococci and that a GdhR-binding site influences gdhA transcription in gonococci. In order to assess promoter utilization in gonococci and any influence of GdhR on gdhA transcription, we performed qRT-PCR analysis on RNA prepared from strain FA19 and its isogenic gdhR::kan mutant at midand late-logarithmic phases of growth. Unlike its influence on gdhR expression, loss of GdhR did not impact gdhA expression in either mid-log-or late-log-phase cultures (Fig. 5B). One explanation for why we did not observe changes in gdhA expression is that GdhR binds upstream of the most proximal promoter (represented in red in Fig. 6) and inhibits the binding of the RNA polymerase to the second gdhR promoter (represented in green in Fig. 6), but it does not interfere with transcription from the proximal promoter, just as in meningococci. Consequently, when GdhR is present, the most proximal promoter (represented in red in Fig. 6) is the primary promoter used for transcription of gdhA. When GdhR is absent, the most distal promoter (green in Fig. 6) becomes the primary promoter for gdhR transcription.

DISCUSSION
Our interest in gdhR was spurred by its close location to the mtr locus, which encodes the tripartite RND-type efflux pump MtrC-MtrD-MtrE and a transcriptional repressor (Fig. 1). We hypothesized that GdhR and MtrR might have cross-regulatory activities on the mtr and gdh loci, respectively. While we did not find evidence that GdhR regulates mtrCDE or antimicrobial resistance, we did find that its loss significantly increased fitness of gonococci during an experimental infection of the lower genital tract of female mice. This experimental model of infection has been used by us to show the importance of the MtrCDE pump for gonococcal survival in vivo and that gradients of fitness can be observed (11,24), depending on the presence of distinct cis-or trans-acting mutations that influence mtrCDE expression (2).
Although the observed increase in fitness in the gdhR-negative mutant of strain FA19 was not as dramatic as previous findings (11) with an mtrR-negative mutant (5-to 10-fold versus ca. 100-fold), the impact on fitness was reproducible and significant. Therefore, we sought to define regulatory systems that influence gdhR expression in gonococci, and we compared our results with those obtained by others working on regulation of meningococcal genes controlled by GdhR. We discovered two trans-acting regulatory systems that were not previously observed in meningococci: (i) MtrR, which can directly repress expression of gdhR (Fig. 3), and (ii) GdhR, which can repress itself (Fig. 4). In the first instance, MtrR regulation of gdhR emphasizes the global regulatory properties of this DNA-binding protein in gonococci. It is noteworthy that this mechanism does not extend to most meningococcal isolates, since they harbor loss-of-function mutations in mtrR (22,23). In the second instance, GdhR autoregulation of gdhR may be unique to gonococci, because the presence of a CE in this region in meningococci, but not gonococci, likely influences promoter utilization and GdhR binding.
GdhR has been previously studied in meningococci for its capacity to regulate genes involved in metabolism, but heretofore it has not been investigated for its regulatory properties in gonococci. Although GdhR has been reported to activate expression of gdhA and gltT, which encodes an L-glutamate transporter that appeared to be essential for full virulence in a rodent model of invasive meningococcal disease (25), its capacity to autoregulate its own gene or be controlled by trans-acting factors has not been elucidated in either pathogen. The work presented here illustrates that although two genetically related pathogens can encode the same transcription factor (e.g., GdhR) and have conserved target genes (e.g., gdhA), gene regulatory principles that have evolved for one pathogen may not necessarily apply to the related pathogen. Thus, although the GdhRs in meningococci and gonococci are identical, regulation of one target, gdhA, is distinct. While gdhA is a GdhR-activated gene in meningococci, based on quantitative analysis of levels of mRNA transcripts, our work failed to reveal differences in gdhA transcript levels in isogenic GdhR-positive and -negative gonococci. This does not mean that GdhR cannot activate gdhA in gonococci. We draw this conclusion because PE analysis suggested the presence of two promoters in gonococci that could direct transcription and be differentially impacted (activated or repressed) by GdhR, thereby giving the impression of lack of gdhA regulation. We propose that differences in the DNA sequence in the gdh locus in gonococci versus meningococci result in distinct promoter utilization and regulation.
Additional studies are needed to define the GdhR regulon in gonococci in order to understand the role of this DNA-binding protein in controlling genes important for metabolism and in vivo fitness of N. gonorrhoeae. In this respect, the increased fitness of the gdhR mutant observed on days 3 and 5 (Fig. 2) corresponds to the time inflammation is detected in the mouse model (26). With the protocol we use, proinflammatory cytokines and chemokines begin to increase on day 3 and peak on day 5, along with a peak polymorphonuclear leukocyte influx on day 5; expression of antimicrobial peptides also peaks on day 5 (A. E. Jerse et al., unpublished data). Thus, it is possible that gdhR may downregulate genes important in the invasion of innate defenses. Depression or induction of genes that are important in growth and metabolism could also contribute to the increased fitness observed with the gdhR mutant. With these possibilities in mind, which form the basis for future studies, our results emphasize that pathogen-specific regulatory actions of a common DNA-binding protein likely exist even between closely related bacteria (e.g., gonococci versus meningococci) and that differences in gene control, which could be influenced by cis-regulatory elements, may have consequences for the overall biology of members in same genus.

MATERIALS AND METHODS
Gonococcal strains, growth conditions, and determination of susceptibility to antimicrobial agents. Strains used in this study are presented in Table 1. Gonococcal strains were grown overnight at 37°C under 5% (vol/vol) CO 2 on GC agar containing defined supplements I and II (27). Determination of susceptibility of test strains to antibiotics was performed by the agar dilution method, and results were reported as the MIC (18). Antibiotics were purchased from Sigma Chemical Co. (St. Louis, MO). Escherichia coli strains were grown overnight at 37°C on LB agar.

Mapping transcriptional start sites by primer extension analysis.
Total RNA from strain FA19 was prepared at the late-logarithmic phase of growth in GC broth as described above, using the method of Baker and Yanofsky (31). Primer extension experiments were performed as described previously (7) with 6 g of total RNA with primers PEgntR (5=-CCAGTTTCATCACTCCTCCT-3=) or PEgdhA (5=-TTTGAGGTTGG CAAACAGGG-3=). The AMV reverse transcriptase primer extension system from Promega (Madison, WI) was used as described by the manufacturer. The TSSs were determined via electrophoresis of the extension products on a 6% (wt/vol) DNA sequencing acrylamide gel adjacent to reference sequencing reaction mixtures.
For protein expression, pET15bgdhR was transformed into E. coli BL21(DE3) cells. Cultures (5 ml) of BL21(DE3)-pET15bgdhR cells were grown overnight at 30°C and added to 500 ml of LB broth the next morning. The culture was grown at 30°C until mid-log phase and then induced with 0.3 mM isopropyl-␤-D-thiogalactopyranoside and grown overnight at 30°C. Cells were harvested and resuspended in 20 ml of 10 mM Tris (pH 7.5), 200 mM NaCl, and then EDTA-free protease inhibitor was added to the bacterial suspension. The cells were lysed by use of a French press cell as described elsewhere (32), membranes and unbroken cells were removed by centrifugation at 100,000 ϫ g, and the supernatant was collected and filtered. GdhR-His was purified over a 2-ml nickel-nitrilotriacetic acid (Ni ϩ2 -NTA) column. After flowing the supernatant over the Ni ϩ2 -NTA column, the resin was washed successively with buffer containing 20 mM and 50 mM imidazole to remove contaminants and weakly bound proteins, and GdhR-His was eluted successively with buffer containing 100 and 200 mM imidazole. The fractions containing GdhR-His were concentrated and the imidazole-containing buffer was removed by dialysis into storage buffer (10 mM Tris-HCl [pH 7.5], 200 mM NaCl, and 1 mM EDTA). Dithiothreitol and glycerol were added to final concentrations of 1 mM and 10%, respectively. To verify the stability and purity of the GdhR and MtrR fusion proteins, we subjected 1 g of purified proteins to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a 12% (wt/vol) polyacrylamide gel (33), and then stained the resolved proteins with Coomassie brilliant blue (CBB). Each protein preparation contained a single CBB-staining band; the respective proteins migrated in the SDS-PAGE gel with a molecular mass consistent with their fusion protein status (32.0 kDa for GdhR-His and 65 kDa for MtrR-MBP [data not shown]).