INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Pathogenic bacteria must obtain essential nutrients in order to establish an infection; of these nutrients, iron plays a critical role (for a recent review, see reference 53). To obtain adequate iron for survival and replication, a successful pathogen must possess one or more efficient iron-scavenging systems capable of competing with or exploiting the iron transport and storage mechanisms of the host. During iron starvation, many gram-negative bacteria synthesize and secrete siderophores that bind iron with high affinity (44). Cognate, iron-repressed receptors facilitate the transport of the ferric siderophore complexes across the bacterial outer membrane in a process that is dependent on both the proton motive force and TonB (for reviews, see references 27, 32, 49, and 50). Other bacteria, including the pathogenic Neisseria species, express specific receptors that recognize host iron-binding proteins, in the apparent absence of siderophore secretion (2, 24, 40, 43, 62). Host iron-binding proteins that can be used as the sole source of iron by the pathogenic Neisseria spp. include human transferrin (Tf), human lactoferrin, and hemoglobin (39, 40).

The neisserial receptor involved in iron acquisition from human Tf is preferentially expressed under iron deprivation conditions (56). This receptor is composed of two iron-repressible proteins designated TbpA and TbpB (for Tf-binding protein A and B) (15). TbpA is similar to TonB-dependent, outer membrane receptors involved in ferric siderophore utilization and in vitamin B₁₂ acquisition (12, 34). TbpB is a surface-exposed lipoprotein (1, 34), which like TbpA specifically binds human Tf; however, unlike TbpA, TbpB is endowed with the capacity to discriminate between ferrated and apotransferrin (7, 14, 54). We have suggested that this characteristic contributes to the efficiency of the receptor and explains the observation that gonococcal mutants lacking TbpB are capable of Tf iron utilization, albeit at a lower efficiency (14, 15). Several lines of evidence suggest that TbpA and TbpB function together as the Tf receptor, the stoichiometry of which is the subject of some controversy. Published estimates of the TbpB/TbpA stoichiometry in the meningococcal receptor range from 1:2 (8, 9) to 9:1 to 15:1 (51). We have previously concluded that TbpB proteins on the gonococcal cell surface outnumbered TbpA proteins by as much as 5:1, on the basis of equilibrium-phase Tf-binding assays conducted in isogenic mutant strains (14).

Vaccine development efforts aimed at preventing gonorrhea have focused on the pilin (6, 59) or porin (63) proteins. Like gonococcal pilin, other potential outer membrane protein antigens are subject to high-frequency phase or antigenic variation (38), making them less-than-ideal vaccine targets. Because the components of the gonococcal Tf receptor are not subject to high-frequency variation and because expression of the Tf receptor (in the context of a lactoferrin receptor-deficient wild-type strain) is necessary to initiate signs and symptoms of urethritis in a human male urethral challenge model (13), TbpA and TbpB are considered viable vaccine candidates. Recent findings that link infection with a bacterial sexually transmitted disease pathogen and the transmissibility of human immunodeficiency virus (11, 37) have strengthened the resolve to rapidly identify protective gonococcal antigens for vaccine development.

The gene that encodes TbpB is located upstream of the gene that encodes TbpA (12, 34); the two genes are separated by an inverted repeat potentially capable of forming a stem-loop structure in mRNA. Insertional mutagenesis of tbpB with the  fragment had a polar effect on tbpA expression, suggesting that these genes are transcribed from a common promoter located upstream of tbpB (1). The presence of an inverted repeat between tbpA and tbpB is unique to this gene cluster; neither the genes that encode the components of the lactoferrin receptor (lbpBA) nor those that encode the hemoglobin receptor (hpuAB) are separated by similar sequence or spacing (4, 5, 35, 36, 46). The stoichiometry of multisubunit complexes, such as the photosynthesis reaction center (33) and the bacterial ATPase (16), are influenced by secondary structures located in the polycistronic operons from which they are encoded. Furthermore, the stem-loop structures in these systems influence the relative amounts of the subunit transcripts and thereby affect the final stoichiometry of the complex. Thus, by analogy, the inverted repeat between tbpB and tbpA may affect the relative expression of Tbp proteins, leading to differences in the structure and function of the Tbp complex in the outer membrane.

To understand the molecular mechanism that controls expression of the Tf-binding proteins, we first sought to directly demonstrate that tbpB and tbpA are cotranscribed as a single transcript using reverse transcriptase PCR (RT-PCR). We created tbpA'-lacZ and tbpB'-lacZ transcriptional fusions in the gonococcal chromosome and then measured -galactosidase activity as a measure of tbp expression under different environmental conditions. We also measured tbpB- and tbpA-specific mRNA levels detected in wild-type gonococci using competitive RT-PCR (cRT-PCR). In this report, we quantitated the influence of iron, external pH, and presence of ligand (Tf) on steady-state levels of tbp-specific mRNAs and measured the relative amounts of tbpB- and tbpA-specific transcripts under these environmental conditions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. Bacterial strains and plasmids used in this study are listed in Table 1. Neisseria gonorrhoeae strains were routinely maintained on GCB medium (Difco) with Kellogg's supplement I (28) and 12 µM Fe(NO₃)₃ at 37°C in a 5% CO₂ atmosphere. To generate iron-limited conditions, GCB agar or GCB broth were supplemented with 10 or 50 µM Desferal (Sigma), respectively. All glassware was washed in 3 N nitric acid to remove residual Fe and then rinsed extensively with deionized water. In experiments in which pH conditions were varied, gonococci were grown in GCB broth containing 30 mM Bis-Tris (Sigma) as a buffering agent. Escherichia coli strain DH5MCR (Bethesda Research Laboratories) was used as the host for plasmid manipulations and was maintained on Luria-Bertani medium (55) supplemented with antibiotics as necessary. The following antibiotics and concentrations were used: ampicillin, 100 µg/ml for E. coli; kanamycin, 100 µg/ml for E. coli; and erythromycin, 100 µg/ml for E. coli and 1 µg/ml for N. gonorrhoeae.

Construction of lacZ transcriptional fusions. Plasmids used in this study are listed in Table 1. To construct transcriptional fusions between the tbpA gene of N. gonorrhoeae and the lacZ-ermC' cassette, plasmid pAErmC'G (64) containing a promoterless lacZ gene with an ermC' marker was digested with BamHI. The 4.7-kb fragment containing lacZ-ermC' and a 10-bp gonococcal uptake sequence was released and ligated into the unique MluI site of tbpA in pUNCH405 (Table 1), generating plasmid pVCU108 (Table 1). Plasmid pVCU107 was identical to pVCU108 except that the lacZ-ermC' cassette was inserted in the orientation opposite that of tbpA expression (Table 1). This plasmid was constructed as a negative control. Plasmid pVCU109, carrying the tbpB'-lacZ transcriptional fusion, was constructed by ligating the lacZ-ermC' cassette into the PmlI site of pUNCH416 (Table 1) in the forward orientation relative to that of tbpB expression. Recombinants were selected in E. coli by their resistance to erythromycin, encoded by the ermC' cassette.

Generation of iron stress conditions over a 4-h time course. Gonococcal strains were grown overnight on GCB agar containing Kellogg's supplement I and 12 µM Fe(NO₃)₃ in a 5% CO₂ atmosphere at 37°C. A single colony was picked from these plates and inoculated in 20 ml of GCB broth with only Kellogg's supplement I added and grown at 35°C with 5% CO₂ and vigorous shaking. After one mass doubling, 50 µM Desferal was added and the cultures were allowed to grow for an additional 4 h. Samples of iron-stressed cultures were removed at 0.5-h intervals. All samples were analyzed for -galactosidase activity and standardized by culture density (see below).

Generation of iron stress and low-pH conditions for 2-h endpoint assays. For the iron-limited 2-h endpoint assays, gonococcal strains were grown as described above for the time course experiments. For comparison, iron-replete cultures were grown in GCB broth containing both supplement I and 12 µM Fe(NO₃)_3. When Desferal was added to the iron-depleted cultures, an additional 12 µM Fe(NO₃)₃ was added to the parallel iron-replete cultures. After 2 h, samples were analyzed for -galactosidase activity or harvested for RNA isolation (see below). To compare tbp expression during growth at low and neutral pHs, hydrochloric acid was added to decrease the pH of the growth medium from 7.2 to 5.8; this addition was made at the same time as the Desferal addition in the iron-depleted cultures. In parallel neutral-pH cultures, the growth medium was maintained at pH 7.2. After HCl addition, the cultures were incubated for an additional 2 h. Some samples were analyzed for -galactosidase activity, and other samples were harvested for RNA isolation (see below).

-Galactosidase assay. A culture sample of 0.5 ml was harvested and added to 0.5 ml of Z buffer (41). Cells were lysed with chloroform and sodium dodecyl sulfate, and -galactosidase assays were performed by the method of Miller (41). In endpoint assays, the results reported are the averages of at least three independent assays, each performed in triplicate on separate days. For time course assays, the data represent the averages of three independent assays conducted on different days.

RNA isolation. Total cellular RNA was isolated from N. gonorrhoeae strain FA19 grown under the conditions described above for iron-limiting, iron-replete, low-pH, and neutral-pH conditions. In some experiments, Tf (100% saturated with iron) was added to cultures grown in medium at pH 7.2 to a final concentration of 500 nM. This addition occurred at the same time as Desferal or HCl addition and was intended to assess the effect of the presence of ligand on tbp expression. RNA was isolated from these cultures using an RNeasy maxi kit as directed by the manufacturer (Qiagen). Purified RNA was treated twice with RQ1 RNase-free DNase (Promega) at 37°C for 1 h prior to storage at 80°C.

Detection of the bicistronic tbpBA operon by RT-PCR. For controls, portions of tbpA, tbpB, and porB1a were amplified by RT-PCR according to the following protocol. For cDNA synthesis, 1 µg of DNase I-treated total RNA was mixed with 50 pmol of each oligonucleotide primer (Table 2). Mixtures were incubated for 10 min at 70°C to allow template denaturation, which was followed by addition of a prewarmed mixture containing 10× PCR buffer, 50 mM MgCl₂, 0.1 M dithiothreitol, 40 U of RNaseOUT (Gibco) and 200 U of SuperscriptII RT (Gibco). The 25-µl reaction mixture was incubated for 1 h at 50°C, and the reaction was stopped by heating at 70°C for 15 min. RNase H was added to remove the RNA template. Following completion of the RT step, 5 µl of RT mixture was used as a PCR template. The amplification reaction mixtures contained 10× PCR buffer, 0.2 mM each deoxynucleoside triphosphate, 50 mM MgCl₂, 50 pmol (each) oligonucleotide primer (Table 2), and 5 U of Taq DNA polymerase (Gibco). After an initial denaturation step of 3 min at 94°C, DNA was amplified for 35 cycles, with 1 cycle consisting of 1 min at 94°C, 30 s at 54°C, and 1 min at 72°C, followed by a final extension step of 10 min at 72°C.

Relative RT-PCR. To compare the relative amounts of steady-state tbpA- or tbpB-specific transcripts, we employed relative RT-PCR using the porB1a and 16S rRNA genes as negative controls for regulated expression. For both tbpA and tbpB, a 30-cycle amplification protocol was used with 100 ng of total template RNA. Since the copy numbers of 16S rRNA and porB1a transcripts were likely higher than those of the tbp genes, we decreased the amount of input RNA in these experiments from 100 to 5 ng for 16S rRNA, used 50 ng of porB1a, and maintained the cycle number at 30. Otherwise, the RT-PCR procedures for tbpA, tbpB, and porB1a were identical to the procedure described above. For amplification of the 16S rRNA gene, the primers used are listed in Table 2 and the procedure was identical to that used for the other genes in this experiment.

Construction of native and competitor RNA species for cRT-PCR. We first constructed plasmids such that the native tbp gene could be distinguished from the competitor tbp gene by a 50-bp deletion in length but no difference in sequence. The oligonucleotides used to generate PCR products representing native tbpA, tbpA50, tbpB, and tbpB50 are shown in Table 2. Plasmids pUNCH411 and pUNCH416 (Table 1) were used as a PCR templates in the generation of the tbpA- and tbpB-specific constructs, respectively. The PCR conditions used were as described above for RT-PCR of the tbpA and tbpB genes. The resulting PCR products were inserted into plasmid pCRII-TOPO using the TA-Cloning System (Invitrogen, Carlsbad, Calif.). The tbpA and tbpB PCR products were oriented such that the 5' ends of the genes were near the T7 RNA polymerase promoter in pCRII-TOPO, generating plasmids pVCU115 to pVCU118 (Table 1).

Quantitation of tbp-specific transcripts with cRT-PCR. To quantitate the relative amounts of gonococcal tbp-specific messages, standard curves for tbpA- and tbpB-specific transcripts were generated in which various amounts of in vitro-synthesized, native-sized RNAs were mixed with 125 fg of competitor RNA (smaller by 50 bp) in each RT-PCR. For the experimental samples, 150 ng of total RNA from gonococcal strain FA19 grown under iron-replete conditions was mixed with 125 fg of competitor RNA. Only 15 ng of total RNA from iron-depleted cultures was mixed with 125 fg of competitor RNA. RT-PCR mixtures and conditions were as described above in the protocol for relative RT-PCR. To control for DNA contamination, reactions were conducted in the absence of RT and analyzed using the standard curve and experimental samples. Aliquots of all amplified products were electrophoretically separated on 2% agarose gels and imaged with a FluorChem imaging system (Alpha Innotech). Band intensities were quantitated, and a standard curve was generated by plotting the logarithm of the ratio of native to competitor band intensities against the logarithm of the amount of native RNA added to each RT-PCR mixture. Linear regression analysis allowed the transformation of the band intensity ratios to the amount of specific transcript present in unknown samples (60).

Confirmation of tbp ratios using real-time, quantitative RT-PCR. Total RNAs from gonococcal strain FA19 grown under various conditions were isolated as described above under "RNA isolation." TagMan real-time quantitative RT-PCR analysis was performed using the TagMan One Step RT-PCR Master Mix Reagents Kit (PE Applied Biosystems). Primers and probes for tbpA and tbpB were selected using Primer Express software (PE Applied Biosystems) and are shown in Table 2. The probe contained a fluorescent reporter (6-FAM) at the 5' end and a fluorescent quencher (TAMRA) at the 3'end. To construct the standard curves for tbpA- and tbpB-specific transcripts, a series of six 10-fold dilutions were made with starting concentrations of 1.2 amol for tbpA and 1.3 amol for tbpB. Both standard curves were generated with in vitro-synthesized transcript, produced as described above for cRT-PCR. For experimental samples, a 25-µl reaction mixture contained 20 ng of total RNA, 900 nM (each) primer, 250 nM (each) probe, 2× Master Mix without UNG, 40× MultiScribe, RNase Inhibitor Mix, and RNase-free water. Amplification and detection were performed with an ABI 7700 sequence detection system under the following conditions: 30 min at 48°C; 10 min at 95°C to activate AmpliTaq Gold DNA polymerase; and 40 cycles, with 1 cycle consisting of 15 s at 95°C and 1 min at 60°C.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of tbpBA cotranscription. We previously reported that the tbpA gene is located immediately downstream of the tbpB gene and that the genes are separated by an inverted repeat (12). Polar transposon insertions into the upstream tbpB gene resulted in mutants incapable of expression of either TbpB or TbpA, suggesting that the two genes are cotranscribed and coregulated from a common promoter upstream of tbpB (1). In order to investigate the cotranscription of tbpB and tbpA, RT-PCR analysis was performed on total RNA isolated from gonococcal strain FA19 under iron-replete and iron-limiting conditions. The locations of the primer pairs used for the detection of tbpA, tbpB, porB1a, and the intergenic region between tbpB and tbpA are shown in Fig. 1A, and the sequences of the primers are listed in Table 2. As shown in Fig. 1B, RT-PCR amplification generated products of the expected size for the tbpA, tbpB, and porB1a transcripts. Expression of the gonococcal porB1a gene was constitutive, i.e., not influenced by iron availability, and was thus used as a negative control (Fig. 1B). In contrast to porB1a, the tbpA and tbpB transcripts were expressed preferentially under iron-limiting conditions, indicating that transcription of both genes was influenced by iron concentration. To verify that the intergenic region separating the tbpB and tbpA genes was expressed as mRNA, we used RT-PCR to amplify the region between primers oVCU-63 and oVCU-64 (Fig. 1A). Indeed, the anticipated 336-bp fragment was amplified from mRNA templates expressed under iron-limited conditions, and to a lesser extent, the same-sized product was amplified from RNA isolated from iron-replete cultures (Fig. 1C). Without the addition of RT, no products were amplified from the RNA templates isolated from iron-replete or iron-limited cultures (Fig. 1B and C). From these results, we conclude that tbpB and tbpA were cotranscribed as a bicistronic mRNA and that their expression was controlled by iron availability.

View larger version (29K): [in this window] [in a new window]   FIG. 1.   Detection of the tbpBA operon by RT-PCR. (A) Schematic representation of the tbpBA genes and the unlinked porB1a gene. The tbpA gene is flanked by inverted repeats represented by hairpin symbols. The arrows show the position and direction of the gene-specific primers used in RT-PCR. (B and C) RT-PCR analysis for detection of tbpA, tbpB, and porB1a transcripts and the intergenic (I.G.) region between tbpB and tbpA in gonococcal strain FA19. Lanes contain RT-PCR products amplified from the following templates: 1, FA19 chromosomal DNA (positive control); 2, RNA isolated from FA19 grown under iron-replete conditions; 3, RNA isolated from FA19 grown under iron-limiting conditions; RT, RT-PCR without RT using RNA from iron-replete or iron-limited FA19 cultures. Lanes labeled M contain molecular size standards. Images were acquired using the FluorChem imaging system (Alpha Innotech) and FluorChem version 2.0 software. Labeling was added using Adobe Photoshop version 4.0.

Time course of tbp expression during iron starvation. In order to quantitate the extent to which tbp transcription was influenced by iron, we inserted a promoterless lacZ gene into the tbpA and tbpB genes in the gonococcal chromosome. -Galactosidase activity expressed by these reporter constructs was used as a measure of tbpA- and tbpB-specific transcript levels. Fusion constructs used for this purpose are diagrammed in Fig. 2 and described in Table 1. For a negative control, we constructed an inverted lacZ fusion in which the orientation of lacZ was opposite that of the tbpA gene (Table 1). As a means of confirming the cotranscription of tbpB and tbpA, we inserted lacZ into the tbpA gene downstream of an existing, polar  insertion in tbpB (1). As shown in Fig. 3, tbpA-specific transcription was completely prevented by the upstream insertion of the polar  fragment, with -galactosidase activities indistinguishable from the negative control. At virtually every time point in the 4-h experiment, tbpB-specific expression was greater than tbpA-specific transcription, with the greatest differential of approximately twofold after 2 h of iron stress. After peak expression for both tbpA and tbpB, the transcription of both genes declined over the second half of the 4-h time course.

View larger version (3K): [in this window] [in a new window]   FIG. 2.   Schematic representation of the transcriptional lacZ fusions into either tbpB or tbpA. Plasmids and strains were constructed as detailed in Materials and Methods. The position of the lacZ-ermC' cassette () in tbpB (strain MCV111) or tbpA (strains MCV108 and MCV110) is shown; the inverted lacZ-ermC' cassette in tbpA (MCV107) is in the same position as in MCV108. The positions of the  marker in tbpB (MCV110) () and of the mTn3CAT marker in tbpA (MCV111) () are also shown. P, promoter.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Time course of tbpA- and tbpB-specific transcriptional activity. Gonococcal strains were grown in iron-depleted conditions for 4 h. Samples were removed at 0.5-h intervals and assayed for -galactosidase activity, which is expressed in Miller units (MU). Each point represents the mean ± standard deviation for three independent experiments. Symbols: , MCV111 (tbpB'-lacZ); , MCV108 (tbpA'-lacZ); , MCV110 (tbpB::  tbpA'-lacZ); , MCV107 (tbpA'-invlacZ).

lacZ transcriptional fusions as a means of quantitating tbpA and tbpB expression. Because gonococci generate acid during growth with glucose as a carbon source (42), we tested the pH of the growth medium at the end of the 4-h time course and found it to be ~6.5, in contrast to the starting pH of 7.2. This suggested that the reason for the decline in -galactosidase activities during the last 2 h of iron stress could be that tbp expression was influenced by external pH. Using these lacZ fusion strains, we directly analyzed the effect of external growth pH on tbp expression in a 2-h endpoint assay (Fig. 4). Growth of the tbp'-lacZ fusion strains in media at pH 5.8 resulted in decreased -galactosidase expression relative to that from strains grown at pH 7.2, which suggested that external pH negatively influenced tbp expression. The induction ratios (iron-limiting/iron-replete) for tbpA and tbpB grown at pH 5.8 were 7 to 9, whereas the induction ratios for the tbp genes when strains were grown at pH 7.2 were 25 to 27 (Fig. 4). Thus, the activity corresponding to both genes was higher following 2 h of growth at pH 7.2 than at pH 5.8. Regardless of the external pH, tbpB expression was greater than that of tbpA in every experiment, with the iron-stressed tbpB/tbpA ratio in these endpoint assays being approximately 1.5:1. Taken together, these results indicate that the tbp genes were induced at the transcriptional level by low iron concentrations, that tbp expression was repressed at pH 5.8, and that tbpB-specific expression was always higher than tbpA-specific expression under iron stress conditions.

View larger version (21K): [in this window] [in a new window]   FIG. 4.   pH- and iron-dependent expression of tbpA and tbpB. -Galactosidase activity detected in strains MCV108 (tbpA'-lacZ) and MCV111 (tbpB'-lacZ) following 2 h of iron stress. Gonococcal strains were grown at pH 7.2 or 5.8 with a low (open bars) or high (filled bars) level of iron. -Galactosidase activity is expressed in Miller units (MU). Error bars represent standard deviations of the means of three experiments conducted on different days.

Evaluation of environmental influences on tbp expression by relative RT-PCR. The generation of single-copy lacZ fusions in the chromosome created tbp mutants (MCV108 and MCV111 [Table 1]), which are difficult to propagate in a coordinated fashion. Also, insertion of the large, nongonococcal lacZ gene might have had unanticipated effects on transcript stability. For these reasons, we sought to confirm the findings we had obtained with lacZ fusions using RT-PCR and native mRNA isolated from the wild-type strain FA19 (Table 1). The wild-type strain was grown under different environmental conditions (low iron, high iron, pH 7.2, pH 5.8, with and without ligand [saturated Tf]) for 2 h, after which total RNA was isolated so that it could be used as the template in RT-PCR experiments. The primers used in this assay are described in Table 2. As shown in Fig. 5, transcription of tbpB and tbpA was greatly increased under iron-limiting conditions, but tbp transcription during growth at pH 5.8 was affected only slightly, if at all. By adding ligand, we sought to address whether maximal tbp expression required both low iron conditions and the presence of saturated Tf. However, we detected no obvious effect by adding ligand under these experimental conditions (Fig. 5). As anticipated, expression of porB1a was not affected by iron availability, but we were surprised to find that expression of porB1a was influenced by external pH, with maximal expression resulting from growth at neutral pH. The expression of the 16S rRNA gene was used as a control in these experiments, as transcript levels were not altered under any experimental growth conditions tested (Fig. 5).

View larger version (29K): [in this window] [in a new window]   FIG. 5.   Relative RT-PCR analysis for detection of tbpA, tbpB, porB1a, and 16S rRNA transcripts. Lanes contain RT-PCR products amplified from mRNA templates isolated from cultures of strain FA19 grown under the following conditions: pH 5.8 or 7.2 or in the presence of 500 nM human Tf. Cultures from which template mRNA was isolated were additionally grown under iron-limiting (Fe) or iron-replete (Fe+) conditions as indicated below each blot. The RT-PCR products are those amplified from the tbpA, tbpB, porB1a, and 16S rRNA genes as indicated above each blot. Lanes: C, a representative negative control in which no RT was added; M, molecular size standards. The sizes of molecular standards are indicated to the left of each blot; the actual sizes of individual RT-PCR products are indicated to the right of each blot. Images were acquired using the FluorChem imaging system (Alpha Innotech) and FluorChem version 2.0 software. Labeling was added using Adobe Photoshop version 4.0.

Quantitation of iron-, pH-, and ligand-dependent expression using cRT-PCR. Because relative RT-PCR allows comparison of only the relative amounts of a single transcript isolated from cultures grown under different conditions, we used cRT-PCR to establish the ratio of tbpB- to tbpA-specific messages and to precisely quantitate absolute amounts of message detectable at steady state. In cRT-PCR, known amounts of competitor transcript are mixed with total RNA preparations such that the resulting amplified products are proportional to the relative amounts of native and competitor transcripts in the reaction. Typically, the competitor species is constructed with a small deletion so that products amplified from native versus competitor species can be discerned. The positions of the deletions in competitor tbpB and tbpA genes used in these experiments are shown in Fig. 6A. cRT-PCRs containing competitor- and native-sized transcripts and the primers diagrammed in Fig. 6A generated two PCR products that differed in length by 50 bp. The sequences of the primers used in these reactions are listed in Table 2. The agarose gels in Fig. 6B and D show the products that were generated in these cRT-PCRs as well as standard curves, which were generated by mixing various amounts of in vitro-synthesized, native-length transcript and a constant amount of competitor. After ethidium bromide staining, the intensity of each band was quantitated and the log ratio of native to competitor band intensities was plotted as a function of native transcript added to the reaction mixture (Fig. 6C and E). The amounts of the tbpA- and tbpB-specific RNA transcripts in experimental samples, determined from linear regression analysis of the standard curves, are shown in Table 3.

View larger version (22K): [in this window] [in a new window]   FIG. 6.   Quantitation of tbpA and tbpB expression by cRT-PCR. (A) Organization of the tbpBA locus, including the orientation and position of primers used in cRT-PCR. The lengths and positions of tbpB and tbpA deletions in competitor constructs are represented by the black bars. P, promoter. (B and D) Agarose gels containing tbpA-specific PCR products (B) and tbpB-specific PCR products (D). (B) The five lanes on the left contain the standard-curve samples generated from 0.04 to 28.18 amol of native-sized RNA coamplified with 1.06 amol of competitor RNA. Samples of total RNA isolated from strain FA19 grown under conditions listed were coamplified with 1.06 amol of competitor RNA. (D) The five lanes on the left contain the standard-curve samples generated from 0.04 to 27.06 amol of native-sized RNA coamplified with 1.02 amol of competitor RNA. Samples of total RNA isolated from strain FA19 grown under conditions listed were coamplified with 1.02 amol of competitor RNA. In both agarose gels (B and D), the top band represents the PCR product generated from the native-sized template, while the bottom band represents the PCR product generated from the competitor template. The lane labeled RT represents the negative control in which no RT was added. (C and E) The amount of tbpA- and tbpB-specific RNA transcripts was estimated by plotting the log ratio of native to competitor intensities versus the log amount of native transcript added to the reaction mixture. Images were acquired using the FluorChem imaging system (Alpha Innotech) and FluorChem version 2.0 software. Labeling was added using Adobe Photoshop version 4.0.

Significance and confirmation of tbpB/tbpA ratios. We used the Student t test to evaluate whether the differences detected between tbpA- and tbpB-specific mRNA levels were statistically significant. All tbpB-specific message levels were significantly different from tbpA-specific message levels under iron stress conditions at pH 7.2, as determined by cRT-PCR (P = 0.004) and by lacZ fusion analysis (P = 0.001). Using RNA isolated from iron-stressed cultures, we utilized real-time RT-PCR to confirm the tbpB/tbpA ratios detected by both lacZ fusion and competitive RT-PCR. As shown in Table 4, the results of real-time RT-PCR analysis indicated that regardless of pH or the presence of ligand, the tbpB-specific message was approximately twice as prevalent as the tbpA-specific message following iron stress.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

While much is known about the structure and function of the components of the gonococcal Tf receptor, relatively little is known about the molecular mechanism that controls their expression. We previously identified a putative promoter upstream of tbpB that contains a near-consensus Fur-binding site (1); however, no transcriptional start sites have been identified, and to date, no experimental evidence for the extent or function of the predicted promoter has been presented. Polar transposon insertions into the tbpB gene generated mutants incapable of expression of either TbpB or TbpA, suggesting that both genes are cotranscribed and coregulated from the putative promoter upstream of tbpB (1). Although the two genes are often referred to as the tbp operon, the cotranscription hypothesis had yet to be confirmed, until now. The RT-PCR analysis presented in this study, represents the first direct evidence that tbpB and tbpA are coexpressed on a single mRNA species. The components of the neisserial lactoferrin receptor, LbpA and LbpB, are encoded by linked genes that share a common upstream promoter (4, 36). Using similar techniques, Lewis et al. demonstrated that the meningococcal lbp genes were cotranscribed in a single bicistronic operon; however, there is no intergenic region separating these two genes (36). The genes that encode the components of the meningococcal hemoglobin-haptoglobin receptor are also linked without an intervening secondary structure (35). In contrast, the tbp genes are separated by an intergenic region of 86 bp, including a 35-bp inverted repeat capable of forming a secondary structure with a calculated free energy of dissociation of 23.6 kcal/mol (12). The fact that amplification of the intergenic region by RT required conditions distinct from those used for the flanking structural genes (higher temperature, temperature-resistant RT, and longer primers) suggests that the intergenic secondary structure did indeed form in the bicistronic message.

The presence of a unique inverted repeat separating the tbp genes implies that it might play an important role in modulating the relative expression of TbpA and TbpB, in contrast to the expression of the lactoferrin or hemoglobin receptor components. Thus, we determined the relative steady-state levels of tbpB- and tbpA-specific transcripts as measured by -galactosidase activity produced by lacZ transcriptional fusions. This approach has several drawbacks that compelled us to confirm the lacZ results with complementary approaches. Insertion of the lacZ-ermC' cassettes into the gonococcal chromosome resulted in tbp knockout mutants. Interruption of the native sequence with a large amount of foreign DNA could have unanticipated effects on mRNA transcription or transcript stability. Additionally, this approach necessitated quantitation of tbpB- and tbpA-specific transcripts expressed by two different strains, and it is difficult to synchronize precisely the growth of two strains.

With these concerns in mind, we quantitated tbp-specific transcripts using a cRT-PCR approach in which we could detect steady-state levels of tbpA- and tbpB-specific mRNAs simultaneously from a single wild-type gonococcal strain. In spite of our concerns, the results from the lacZ fusion and cRT-PCR approaches were remarkably similar in some aspects. Both methods yielded apparent tbpB/tbpA ratios of approximately 2:1. Likewise, in confirmatory experiments using real-time RT-PCR, we detected twice as much tbpB-specific transcript as tbpA-specific transcript. Since tbpA was exclusively transcribed on a bicistronic message, as determined by the lacZ fusion analysis, a 2:1 tbpB/tbpA ratio would most likely result from either premature transcription termination at the intergenic region or preferential stability of a tbpB-only, monocistronic mRNA species. The functional significance of segmental differences in transcript accumulation in polycistronic operons has been reported for other multicomponent protein complexes. For example, the differences in stability of mRNA segments within the puf operon of Rhodobacter capsulatus contribute to the optimal stoichiometry of the light-harvesting and reaction center complexes (33). In a situation very similar to the one described in this study, the genes encoding the Zymomonas mobilis glycolytic enzymes glyceraldehyde-3-phosphate dehydrogenase (gap) and 3-phosphoglycerate kinase (pgk) are cotranscribed in a bicistronic operon but are subject to segmental differences in transcript stability, leading to fivefold-more gap transcript than pgk transcript (20). This differential in the steady-state levels of the two transcripts is sufficient to account for the observed differences in the activities of these enzymes in Z. mobilis (20). While there may or may not be additional levels of translational regulation that further influence the relative amounts of TbpA and TbpB proteins in the gonococcal outer membrane, we have previously suggested that there are more TbpB-specific Tf-binding sites than TbpA-specific binding sites by as much as a factor of 5 (14).

Recently, another neisserial operon was demonstrated to be subject to differential expression (29). Within the operon that encodes the ferric iron-specific, ABC transport system, the fbpA and fbpB genes are separated by an inverted repeat similar to that located in the tbp intergenic region. Khun et al. (29) found that fbpA transcript levels were greater than those of the polycistronic message (fbpABC) by a factor of 10 to 20. In contrast, the differential between tbp-specific transcripts reported here was approximately twofold. The mechanistic explanation for this discrepancy is unclear, since the calculated free energy of dissociation for the predicted secondary structure in the fbpAB intergenic region is slightly lower (20.4 kcal/mol; calculated by the method of Tinoco et al. [58]) than that of the putative structure separating the tbp genes (23.6 kcal/mol). Possible explanations for these differences might include methodologies utilized or perhaps sequestration of the ribosome binding site by the stem-loop structure located between fbpA and fbpB. The intergenic region separating fbpA and fbpB is shorter than that separating the tbp genes and includes the only possible ribosome binding site for translation initiation of fbpB. Secondary structure in the region of transcript that encodes the ribosome binding site would effectively decrease translatability of the downstream gene and leave it accessible to endo- and exonucleolytic attack, resulting in decreased message stability. Since the only ribosome binding site preceding tbpA is not obviously overlapped by the potential intergenic inverted repeat, the result would be efficient translation of tbpA, effective protection by ribosomes, and concomitant increased stability of the tbpA message in contrast to that of fbpB.

This report is the first to quantitate the effect of iron on tbpA and tbpB expression. The induction ratios (under iron-limiting and iron-replete conditions) at pH 7.2 for both tbp genes using the lacZ fusion approach were 25 to 27. In contrast, using cRT-PCR, the induction ratios for tbpA were approximately 50 and the induction ratios for tbpB were approximately 100, regardless of the external pH or presence of ligand. Moreover, the cRT-PCR analysis suggests that iron influenced the ratio of tbpB/tbpA transcripts detected at steady state. Iron stress increased the tbpB/tbpA ratio to its highest level (1.86 at pH 7.2), while under iron-replete conditions, the tbpB/tbpA ratios were essentially 1. Concomitantly, the induction ratios for tbpB rose to nearly twice those for tbpA. These observations suggest that some unidentified factor, which is responsive to iron, influenced the relative abundance of the tbp-specific messages. Because this phenomenon was not detected in the fusion strategy, the presence and proximity of the inverted repeats that flank tbpA could be targets of this iron-responsive regulatory factor. As noted earlier, these secondary structures would have been absent in the transcriptional fusion strains due to the introduction of foreign sequences with subsequent changes in the promoter-distal or 3' ends of mRNA transcripts. The concept of an iron-responsive regulatory factor that affects message stability is not a novel one. The eukaryotic IRE-BP (iron-responsive element-binding protein) is a demonstrated RNA-binding protein that functions to increase the stability of the eukaryotic Tf receptor mRNA and simultaneously decrease the translatability and stability of the ferritin message in an iron-dependent manner (31). Thus, in the gonococcus, a similar protein might play a role in dissociating tbpB expression from tbpA expression and increasing the relative abundance of tbpB. In order to isolate the most intact RNAs, we cultured gonococci in complex media and induced iron stress with Desferal, although culturing in chemically defined medium (CDM) with no exogenous iron supply induces more profound iron stress (data not shown). Thus, it is possible that growth in CDM and perhaps growth in vivo, would result in accentuated tbpB/tbpA ratios relative to those presented here.

We have also demonstrated that tbp and porB1a expression is influenced by the pH of the growth media. This observation is relevant because the pH of the natural environment of the gonococcus ranges widely from 4.8 to 8.4, depending on the gender of the host and the particular niche inhabited (26). While a dramatic decrease in gonococcal porin expression at pH 5.8 has not been previously reported, pH-dependent porin expression has been reported for E. coli. In this system, acidification of the medium results in a shift from OmpF expression to OmpC expression in a glucose-dependent manner (25, 57). Likewise, Pettit et al. demonstrated that expression of the gonococcal protein RmpM (PIII) was reduced during growth at low pH (47, 48).

The pH effects on tbp expression were more complicated. By lacZ fusion analysis, we detected an apparent decrease in expression at pH 5.8 on the order of three- to fourfold under iron stress conditions. This phenomenon might explain the decrease in -galactosidase activity in the latter half of the 4-h time course, when pH decreases as a function of glucose metabolism and production of acid end products. Furthermore, we postulated that the pH effect might be directly related to external pH or due to increased iron solubility at low pH, which would result in Fur-Fe-mediated repression of tbp expression. However, upon closer inspection, using the relative RT-PCR and cRT-PCR approaches, we found that tbp expression remained relatively constant or increased, respectively. The different results obtained in the fusion approach versus the RT-PCR approach, which utilize native mRNAs, might again validate our concern about using insertionally inactivated mutants (generating the transcriptional fusions) that could have unanticipated effects on message stability. Alternatively, the apparent decrease in -galactosidase activity in cultures grown at pH 5.8 might reflect the requirement for translation or for extended transcripts, neither of which would be required for detection of short mRNA species via RT-PCR. Note that the activity of the -galactosidase enzyme is not affected in the pH ranges tested in this study (19). In the cRT-PCR experiments, which require the smallest stretch of transcript for detection, we observed the highest steady-state levels of tbp-specific message, suggesting that growth at pH 5.8 might actually induce transcription from the tbp promoter but that there is concomitant degradation of long messages, resulting in decreased translatability. This hypothesis predicts that although tbp-specific transcription might increase at pH 5.8, tbp translation and accumulation of Tbp proteins would decrease. In fact, TbpA expression was dramatically decreased after growth at pH 5.8 overnight, as assessed by amino-terminal sequencing of pH-influenced proteins (D. C. Ruffner and A. E. Jerse, unpublished observations).

In some iron acquisition systems, expression of the outer membrane receptor is not only derepressed by iron starvation but is also induced by the presence of the specific substrate. This dual form of regulation has been described for ferric enterobactin (17, 18) and ferric pyochelin receptors (22, 23) in Pseudomonas aeruginosa, for the ferric dicitrate transport system of E. coli (52), and for the alcaligin synthesis and transport systems in Bordetella spp. (3, 10). Regulation of these iron acquisition systems falls into two general categories. In the case of dicitrate transport, induction occurs via a two-component regulatory system that senses the presence of ligand and subsequently induces expression of the components of the acquisition system, including the receptor (30, 45). In the case of ferric enterobactin uptake by P. aeruginosa, substrate-specific induction occurs via an AraC-like regulator (17, 18). In most systems, the presence of a functional receptor is required for signal transduction; therefore, we conducted similar ligand induction experiments in wild-type gonococci (TbpA⁺/TbpB⁺), not in the lacZ fusion strains. In the present study, we found no evidence of ligand-specific induction of tbp expression, regardless of whether the strain was cultured in high or low levels of iron. This observation is consistent with that of Kim et al., who identified a novel amino-terminal extension in FecA that was required for signal transduction from the citrate receptor to FecR (30). As TbpA has no amino-terminal extension 5' to the TonB box (12), this protein can be grouped with similar TonB-dependent receptors for which ligand-dependent induction has not been observed.

In summary, we have demonstrated that coexpression of the gonococcal tbp genes was controlled at the transcriptional level by iron concentration and was further influenced by external pH. The tbp genes constitute a bicistronic operon; however, the individual steady-state levels of tbpB- and tbpA-specific messages were distinct. These differences in transcript levels may or may not be reflected in differences in the relative amounts of TbpA and TbpB expressed and exported to the outer membrane, depending upon whether subsequent translational regulation occurs. We determined that the ratio of tbpB- to tbpA-specific message was approximately 2:1 under iron-limiting growth conditions. Moreover, this ratio decreased to 1:1 under iron-replete growth conditions, suggesting that the tbpB/tbpA ratio was influenced by iron. The relative amounts of the Tbp proteins expressed and the conditions for their expression have implications for ongoing gonococcal vaccine development efforts. If there are conditions that further disconnect tbpA expression from tbpB expression, hopes for the use of TbpA as a vaccine target, as recently suggested (61), may be misplaced. On the other hand, if TbpB proteins outnumber TbpA proteins with a significant population of uncomplexed protein TbpB, as has been suggested (14, 51), then blocking the function of TbpB, which is not critical for Tf iron uptake, with a vaccine-derived antibody response, could also be ineffectual. Thus, in our efforts to define the Tbp proteins as viable vaccine candidates, it is critical that we understand the factors that regulate their expression, together and separately, at both transcriptional and translational levels. The focus of ongoing studies is to determine the relative levels of Tbp proteins expressed, exported, and complexed and to assess their expression in in vivo model systems.