SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We have characterized the interaction of the Neis-seria meningitidis TonB-dependent receptor HpuAB with haemoglobin (Hb). Protease accessibility assays indicated that HpuA and HpuB are surface exposed, HpuB interacts physically with HpuA, and TonB energization affects the conformation of HpuAB. Binding assays using [125I]-Hb revealed that the bipartite receptor has a single binding site for Hb (Kd 150 nM). Competitive binding assays using heterologous Hbs revealed that HpuAB Hb recognition was not species specific. The binding kinetics of Hb to HpuAB were dramatically altered in a TonB mutant and in wild-type meningococci treated with the protonophore carbonylcyanide m-chlorophenylhydrazone (CCCP), indicating that TonB and an intact proton motive force are required for normal Hb binding and release from HpuAB. Our results support a model in which both HpuA and HpuB are required to form a receptor complex in the outer membrane with a single binding site, whose structure and ligand interactions are significantly affected by the TonB-mediated energy state of the receptor.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Gram-negative diplococcus Neisseria meningitidis (NM) is currently the leading cause of bacterial meningitis in the United States (Schuchat et al., 1997; CDC, 2000) and is responsible for significant morbidity and mortality worldwide. NM is an obligate human pathogen, highly adapted to survive and acquire nutrients such as iron (Fe) needed to proliferate within the human host. However, Fe is not readily available in this environment. To cope with the insolubility and toxicity of free ferric iron (Fe3+) (Bagg and Neilands, 1987; Litwin and Calderwood, 1993), the majority of free Fe in the host is sequestered by carrier proteins such as lactoferrin (Lf) in secretions or transferrin (Tf), haemoglobin (Hb) and haemoglobin– haptoglobin (HbHp) in the bloodstream. The resulting reduction in free Fe to 1012-fold lower than is required to support microbial growth (Weinberg, 1978) provides a non-specific defence against infection, known as ‘nutritional immunity’ (Weinberg, 1975). To overcome nutritional immunity, NM has evolved efficient Fe-scavenging outer membrane receptors that bind to and internalize Fe directly from the host Fe carrier proteins Tf, Lf, Hb and HbHp (Dyer et al., 1987a; Schryvers and Morris, 1988a,b; Lee and Hill, 1992; Lewis and Dyer, 1995; Stojiljkovic et al., 1995). The importance of iron acquisition as a determinant of pathogenesis in meningococcal disease has been clearly established by Holbein et al. (Holbein et al., 1979; Holbein, 1980; 1981; Brener et al., 1981).

We have focused on studying mechanisms for Fe uptake from Hb, the most abundant haem-containing protein in the human body and the site of nearly two-thirds of the total Fe (Bridges and Seligman, 1995). Although most Hb is sequestered within erythrocytes where it is apparently unavailable to pathogens, spontaneous haemolysis results in low levels of Hb (80–800 nM) in normal serum (Sassa and Kappas, 1995). The availability of Hb is thought to increase in the later stages of meningococcal disease as a result of severe disseminated intravascular coagulation (DIC) and concomitant erythrocyte lysis (Wyngaarden and Smith, 1985; van Deuren et al., 2000). The dependence of NM on Hb as an Fe source is probably further enhanced by the host’s hypoferraemic response to infection, which lowers the amount of Tf-Fe by up to 70% (Weinberg, 1984). The in vivo importance of Hb as an Fe source during meningococcal disease was demonstrated by the ability of exogenously added Hb to increase NM lethality in mice (Schryvers and Gonzalez, 1989) and by the reduced virulence of NM mutants defective in Hb utilization (Stojiljkovic et al., 1995).

The serum glycoprotein haptoglobin (Hp) rapidly binds Hb released into the serum and directs its clearance through the liver (Dobryszycka, 1997). Thus, the most physiologically relevant haem-Fe source for the meningococcus is probably HbHp, not Hb (Muller-Eberhard et al., 1968; Hershko, 1975). Whereas microorganisms such as Escherichia coli are unable to use HbHp as an Fe source (Eaton et al., 1982), NM can acquire Fe from both Hb and HbHp (Dyer et al., 1987a; Lewis and Dyer, 1995). The meningococcus has multiple mechanisms for the acquisition of haem-Fe, including a poorly characterized TonB-independent haem uptake system (Stojiljkovic and Srinivasan, 1997) and two distinct TonB-dependent Hb receptors, HpuAB (Lewis and Dyer, 1995) and HmbR (Stojiljkovic et al., 1995). The single-component HmbR receptor is distinguished from HpuAB by the lack of an accessory lipoprotein and its inability to facilitate Fe acquisition from HbHp complexes (Stojiljkovic et al., 1995; Lewis et al., 1999). The redundancy of Hb receptors also argues that Hb utilization is important for NM pathogenesis, although the advantages of expressing two phase-variable Hb receptors are not known.

Neisserial Fe transporters exhibit sequence and structural similarities that reflect their shared dependence on TonB for the active internalization of Fe (Cornelissen et al., 1992; Biswas et al., 1997; Stojiljkovic and Srinivasan, 1997). In E. coli, TonB is anchored in the cytoplasmic membrane in association with ExbB and ExbD (Ahmer et al., 1995; Higgs et al., 1998) and trans- duces the energy of the proton motive force (PMF) to TonB-dependent outer membrane receptors (Postle, 1990; 1993; Klebba et al., 1993). TonB appears to interact directly with outer membrane receptors (Heller et al., 1988; Bell et al., 1990; Gunter and Braun, 1990; Tuckman and Osburne, 1992; Skare et al., 1993) to induce conformational changes that influence ligand binding and release, as well as Fe uptake (Cornelissen et al., 1997; Larsen et al., 1999). Homologues of the TonB–ExbB–ExbD operon have been characterized in N. meningitidis (Stojiljkovic and Srinivasan, 1997) and Neisseria gonorrhoeae (Biswas et al., 1997); as expected, neisserial TonB mutants have pleiotropic defects in Fe utilization from Tf, Lf, Hb and HbHp. The importance of TonB-dependent Fe transport in neisserial disease was demonstrated by the attenuation of a gonococcal Tf receptor mutant in human volunteers (Cornelissen et al., 1998). In addition, recent studies have shown that signature-tagged meningococcal mutants in tonB, exbB and exbD were unable to cause systemic infec- tion in an infant rat model (Sun et al., 2000). Studies of E. coli TonB-dependent transporters using the pro-tonophores dinitrophenol (DNP) and carbonylcyanide m-chlorophenylhydrazone (CCCP) have demonstrated that, in addition to TonB, an intact PMF gradient is required for ligand uptake (Reynolds et al., 1980; Bradbeer, 1993; Larsen et al., 1999). However, the direct involvement of the PMF in TonB-dependent Fe transport in Neisseria spp. has not been demonstrated.

HpuAB is a two-component, TonB-dependent receptor expressed by NM that binds to Hb, HbHp and apo- Hp (Lewis and Dyer, 1995; Lewis et al., 1997). We have shown that HpuAB internalizes the entire haem-Fe moiety from both Hb and HbHp complexes (Lewis et al., 1998a). The mechanisms of haem removal from Hb by HpuAB and subsequent internalization have not been elucidated. Based on its high degree of similarity to the TonB- dependent receptor family (Cornelissen et al., 1992; Lewis et al., 1997), HpuB probably forms a gated porin channel in the outer membrane. It is predicted to adopt a β-barrel structure with an N-terminal periplasmic ‘plug’ domain similar to the crystallographic structures of FhuA (Ferguson et al., 1998; Locher et al., 1998) and FepA (Buchanan et al., 1999). HpuA is a lipoprotein encoded directly upstream of HpuB, with both genes co-transcribed from a single Fe-repressible promoter (Lewis et al., 1997). There are currently no known homologues of HpuA in the public databases. In the structurally analogous Tf and Lf receptors, the lipoprotein is surface exposed and is not required for receptor function, playing an accessory role in ligand specificity and binding (Cornelissen and Sparling, 1996; Bonnah and Schryvers, 1998; Lewis et al., 1998b; Pettersson et al., 1998). In contrast, isogenic single mutants of meningococcal and gonococcal HpuAB suggested that both receptor components are required for the binding and utilization of Hb or HbHp (Chen et al., 1998; Lewis et al., 1999). The precise roles of HpuA and HpuB in this process have not yet been described in detail.

In this study, we begin to examine HpuAB receptor architecture, the interaction of HpuAB with ligand and the role of TonB energization in this receptor–ligand interaction. Of the three ligands known to interact with HpuAB, we have chosen first to focus our attention on Hb, the simplest ligand from which HpuAB extracts haem- Fe. Equilibrium binding experiments were performed to measure the binding kinetics of radiolabelled Hb to HpuAB and to each individual receptor component. Competitive binding assays examined the species specificity of the interaction of HpuAB with Hb. Protease accessibility studies to probe the structure of HpuAB confirmed the surface exposure of both receptor components and provided indirect evidence for the physical interaction of HpuA with HpuB. Finally, the integral role played by an intact PMF gradient and TonB in the conformation of this Hb receptor and in the binding and release of Hb from HpuAB was demonstrated. Fu-ture experiments will characterize the more complex interaction of HpuAB with HbHp complexes to determine how the presence of Hp alters the interaction between NM and Hb.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

HpuA is required for growth with Hb

We constructed a non-polar hpuA in frame deletion mutant to analyse what effect the loss of the HpuA lipoprotein had on HpuAB function. The HpuA deletion mutant, DNM69, expressed wild-type levels of HpuB (Fig. 1, lane 5) but was unable to use Fe from Hb or HbHp compared with controls (data not shown). These results are consistent with observations reported by lewis et al.Lewis et al., 1999) for the HpuA mutant strain DNM2A13, in which HpuB was poorly expressed as a result of polar effects of the kanamycin resistance cassette inserted in hpuA. Strains DNM143 (HpuA+BHmbR) and DNM68 (HpuAB HmbR) also were unable to use Hb or HbHp as Fe sources in a broth culture growth assay (data not shown). These results support the conclusion that both HpuA and HpuB are absolutely required for receptor function.

image

Figure 1. Expression of Hpu receptor components by mutant strains. Whole-cell lysates from Fe-depleted cultures were separated by 15% PAGE, transferred to nitrocellulose and probed with HpuA- and HpuB-specific antibodies. Arrows indicate the 85 kDa and 35 kDa bands corresponding to HpuB and HpuA respectively.

Download figure to PowerPoint

Binding kinetics of Hb to HpuAB

We developed a liquid phase equilibrium-binding assay to examine the Hb-binding defects of Hpu mutants and to measure the kinetics of the HpuAB–Hb interaction. The saturable binding of [125I]-Hb to DNM140 cells expressing wild-type HpuAB (Fig. 2A) was predicted to fit a one-site model, with an F ratio of 1.56 and a P-value of 0.218 (Motulsky, 1999). To validate these binding site predictions (one-site versus two-site), the binding kinetics of [125I]-Tf to TbpBA were examined (data not shown). Analysis of our results, virtually identical to those published by Cornelissen and Sparling (1996), correctly predicted that the TbpBA–Tf interaction conformed to a two-site binding model with an F ratio of 19.0 and a P-value of 0.009 (see Experimental procedures for an explanation of the F test). The apparent Kd of HpuAB Hb binding estimated from fitting the binding data to the one-site binding model was 149.2 nM with a receptor copy number of 4.05 × 1010 molecules μg–1 total cellular protein (assuming one α1β1 molecule of Hb bound per receptor). Copy numbers are expressed here as sites per μg of total protein rather than sites per cfu to describe most accurately the binding detected in this assay and to allow direct comparison of our data with analysis of the TbpBA receptor. As pointed out by Cornelissen and Sparling (1996), Hb bound to cell blebs or fragments that do not correspond to viable cfu may be retained upon filtration and thus contribute to total binding. We could estimate, however, that there were ≈ 8000 high-affinity Hb binding sites per cfu. Significant specific binding of [125I]-Hb to mutant strains lacking either HpuA (DNM69) or HpuB (DNM143) was not detected (Fig. 2A), consistent with both previous solid phase dot-blot binding data (Lewis et al., 1999) and a one-site model in which the two receptor components form a single Hb binding site. This precluded the determination of the copy number of HpuA and HpuB expressed individually. As expected, liquid phase binding assays showed that the HpuAB double mutant DNM68 did not bind Hb (Fig. 2A).

image

Figure 2. A. Hb binding kinetics of HpuAB. Liquid phase equilibrium binding analysis of [125I]-Hb to whole cells. The amount of [125I]-Hb specifically bound to Fe-starved DNM140 cells (HpuA+B+) is shown as a function of radioligand concentration (squares). The specific Hb binding to HpuAB DNM68 (circles), HpuAB+ DNM69 (triangles) and HpuA+B DNM143 (diamonds) is also shown. Error bars indicate standard error of the mean for each combined data set. (Error bars are included for DNM68 but are difficult to see because of the scale of this graph.)

B. Species specificity of HpuAB Hb binding. Competitive liquid phase binding assays using 100 nM 125I-labelled human Hb versus unlabelled heterologous Hbs (1nm–10 μM ). Representative competition curves are shown, with the inset table displaying the Ki value for each competitor assayed. The binding of each competitor was assessed in duplicate, and data points shown represent the means of two individual experiments. Error bars show standard error of the mean for each combined data set.

Download figure to PowerPoint

Species specificity of HpuAB Hb binding

In order to characterize the binding of heterologous Hbs to HpuAB quantitatively, we assessed the species specificity of HpuAB Hb binding using a competitive liquid phase binding assay. Human, baboon, porcine, bovine, rabbit and equine Hbs as well as human methaemoglobin (metHb) were all able to compete specifically with 125I-labelled human Hb for binding to HpuAB. These data, summarized in Fig. 2B, showed that HpuAB did not discriminate between human and non-human Hbs. The predicted Ki for each competitor (Fig. 2B, insert) suggested comparable binding affinities for each of the different Hbs to HpuAB. The slight difference between the Ki for un-labelled human Hb (Fig. 2B) and the Kd of radiolabelled human Hb (Fig. 2A) is probably the result of minor effects of iodination on the structure of Hb and its interaction with HpuAB. MetHb, the oxidized form of Hb unable to bind oxygen, and equine Hb had Ki values significantly different from that of human Hb (P = 0.0112 and P = 0.0061 respectively) with apparent affinities two- to threefold higher than human Hb. The ability of HpuAB to bind metHb with a higher affinity than Hb correlated with growth assays in which meningococci grew with a slightly higher rate and final extent when metHb, not Hb, was the sole Fe source (data not shown).

Previously, Stojiljkovic et al. (1996) proposed that the ability of HpuAB to use human and non-human Hbs may indicate that HpuAB recognizes the structurally conserved haem moiety of Hb. To test this hypothesis, we analysed the ability of unlabelled haemin to compete with [125I]-Hb for binding to HpuAB. We observed that, over a four-log concentration range of Hm, total specific binding of human Hb was reduced by only 25% (Fig. 2B). In control experiments to determine the specificity of this competitive inhibition, we assessed the ability of bovine haemin to compete with [125I]-Tf for binding to TbpBA. Total specific [125I]-Tf binding remained ≈ 100%, even in the presence of the highest Hm concentrations used (data not shown), suggesting that the ability of Hm to compete partially with [125I]-Hb appears to be specific to HpuAB. Thus, competitive binding assays between 125I-labelled Hb and haemin demonstrated that haem did not compete effectively with human Hb for HpuAB binding. These results suggest that the HpuAB–Hb interaction is largely mediated through binding to the globin moiety of Hb, although a small contribution of the porphyrin ring to binding may be possible.

Protease accessibility of HpuA and HpuB

To characterize the surface exposure and relative conformations of the individual Hpu components, Fe-stressed wild-type and mutant meningococcal strains were treated with exogenous trypsin and analysed by SDS–PAGE and immunoblotting (Fig. 3). In DNM140, the 35 kDa HpuA band is susceptible to degradation by trypsin (Fig. 3A), consistent with localization of the lipoprotein receptor component to the outer leaflet of the outer membrane. Control blots probed with antibody against the well- characterized neisserial periplasmic iron-binding protein FbpA (Chen et al., 1993) demonstrated that FbpA was not degraded by exogenous trypsin (data not shown), indicating that cells remained intact under these assay conditions. An HpuA-specific antibody generated against a C-terminal peptide (residues 325–341 of 341) was used to monitor HpuA cleavage by immunoblotting. Trypsin cleavage of 35 kDa HpuA in a wild-type receptor popu- lation yielded two distinct intermediates: (i) a transient ≈ 30 kDa band (HpuA*) seen at lower trypsin concentrations; and (ii) an ≈ 21 kDa species (HpuA**) resistant to further proteolysis at up to 250 μg ml–1 trypsin (Fig. 3A). In silico trypsin digests of HpuA suggested that the 30 kDa product resulted from cleavage at Lys-45 or Lys-51 yielding either a 31.5 kDa or a 30.7 kDa fragment, respectively, that would retain the Ab epitope. Similarly, the protease-resistant 21 kDa intermediate was predicted to result from cleavage at either Arg-125 (22.4 kDa) or Arg-131 (21.7 kDa). We suspected that cleavage of HpuA should release the remaining C-terminal portion of the lipoprotein from the N-terminal fatty acid tethering it in the outer membrane. However, the HpuA cleavage products remained associated with the cell pellets and were not detected in Western blot analysis of concentrated supernatants from trypsinized cells (data not shown). These data not only indicated that HpuA was surface exposed but also suggested that the C-terminal portion of HpuA may interact with some component of the outer membrane.

image

Figure 3. A–D. Protease accessibility of HpuA and HpuB in wild-type and mutant strains. Western blots of trypsin-treated whole-cell lysates of HpuA+B+ DNM140 (A), HpuAB+ DNM69 (B), HpuA+B DNM143 (C) and TonB DNM146 (D) probed with anti-HpuA and anti-HpuB antiserum. The trypsin concentrations used are shown below the blots. Bands corresponding to the receptor proteins and cleavage products are indicated on the left and by the asterisk in (B) (≈ 38 kDa HpuB cleavage product). Approximate positions of the molecular mass standards are included on the right.

Download figure to PowerPoint

The localization and accessibility of HpuB in the context of the wild-type receptor was also characterized by probing Western blots of trypsin-treated DNM140 cells with HpuB-specific antisera generated against an internal peptide of HpuB (residues 161–177 of 810; Lewis et al., 1999). The 85 kDa HpuB protein was accessible to cleavage by exogenous trypsin (Fig. 3A); however, no cleavage intermediates of HpuB were detected in trypsin assays of wild-type DNM140. The apparent absence of HpuB fragments may result from the lack of stable trypsin fragments; there are 94 predicted trypsin cleavage sites in this protein. Tryptic products of HpuB may also not have been detected because they did not contain the epitope recognized by the antibody or because of cleavage by trypsin within the epitope (Lys-166). (The anti-HpuB antisera did cross-react weakly with an ≈ 80 kDa band visible in some immunoblots that was unrelated to HpuB.) Consistent with its proposed function as a receptor, tryp- sin accessibility supported the hypothesis that HpuB is surface exposed. Similar protease accessibility profiles for HpuA and HpuB were observed when staphylococcal V8 protease was used in place of trypsin (data not shown).

We then analysed trypsin cleavage profiles of single mutants lacking either Hpu component to investigate the putative interaction between HpuA and HpuB. Trypsin accessibility assays of DNM143 (HpuA+B) indicated that HpuA was still outer membrane localized in the absence of HpuB, as shown by the disappearance of the full-length 35 kDa species with increasing trypsin concentrations (Fig. 3C). However, in the absence of HpuB, the trypsin-resistant 30 kDa and 21 kDa HpuA intermediates were absent or dramatically reduced (compare Fig. 3C with 3A). This suggested that the protease resistance of the 21 kDa HpuA** was dependent on the presence of HpuB. Likewise, the trypsin accessibility of HpuB in the hpuA deletion mutant DNM69 (Fig. 3B) indicated that HpuB localized properly to the outer surface, where it was degraded by exogenous proteases. However, the absence of HpuA significantly altered the cleavage profile of HpuB. Compared with DNM140, HpuB in DNM69 was more susceptible to proteolytic cleavage and yielded a transient ≈ 38 kDa intermediate (HpuB*) present at 5 μg ml–1 trypsin but degraded at higher trypsin con- centrations (Fig. 3B). This HpuB cleavage product was not detected in trypsin-treated DNM140 cells in which HpuA was also expressed. We were unable to deter- mine in silico the identity of the 38 kDa HpuB fragment because multiple internal trypsin fragments of HpuB could contain the epitope recognized by the antibody. The unique trypsin cleavage profiles of HpuA and HpuB when either was expressed alone indicated that the confor-mation and/or protease accessibility of each protein is dependent on the other, consistent with a model in which HpuA and HpuB interact physically to form a receptor complex.

Protease accessibility of HpuAB in the presence of specific ligand

To determine whether the exposure of trypsin cleavage sites on HpuAB was affected by ligand binding, cells were incubated with saturating concentrations of spe- cific ligands (Hb, HbHp) or a non-specific control ligand (Tf) before the addition of trypsin. As expected, DNM143 (HpuA+B) and DNM68 (HpuAB), which have previously been demonstrated to be defective in ligand binding, were not significantly protected by any of the ligands (data not shown). However, when DNM69 (HpuAB+) was incubated with Hb, but not other ligands, a very slight but reproducible protection of HpuB from proteolysis was observed (Fig. 4B, lanes 3 and 5; data not shown), seen as a faint 85 kDa band in 5 μg ml–1 and 20 μg ml–1 trypsin-treated samples (indicated by a star). This suggested that HpuB may be able to bind Hb weakly when expressed alone. In DNM140 (HpuA+B+), Hb and HbHp but not Tf protected both HpuA and HpuB from proteolysis (Fig. 4A). The protection of both receptor components indicates that the binding of specific ligands to HpuAB rendered surface-exposed trypsin sites on both proteins inaccessible either by inducing conformational changes in the receptor or by directly blocking access of the protease.

image

Figure 4. Protease accessibility of HpuAB in the presence of ligand. Cells were incubated with a saturating amount of ligand (1 μM ), as indicated below each image, before trypsin treatment and analysis as described in the legend to Fig. 3. For (A) and (B), cells were treated with the same trypsin concentrations ranging from 5 to 250 μg ml–1 as in Fig. 3; however, only representative trypsin treatments are shown as indicated.

A. DNM140 (HpuA+B+). Wild-type expression levels of receptor components by DNM140, without ligand or trypsin treatment, are shown in lane 1. DNM140 whole cells were incubated with the specific ligand Hb (lanes 3, 7 and 11) and HbHp (lanes 4, 8 and 12), the non-specific ligand Tf (lanes 5, 9 and 13) or no ligand (lanes 2, 6 and 10). Cells subsequently treated with 5 μg ml–1 (lanes 2–5), 50 μg ml–1 (lanes 6–9) and 250 μg ml–1 trypsin (lanes 10–13) are shown.

B. DNM69 (HpuAB+). Control cell lysates not treated with ligand or trypsin are shown in lane 1. No ligand controls treated with 5 μg ml–1 or 20 μg ml–1 trypsin are shown in lanes 2 and 4 respectively. Cells incubated with Hb before trypsin exposure are in lane 3 (5 μg ml–1) and lane 5 (20 μg ml–1). Intact HpuB and cleavage product HpuB* are indicated by arrows on the left. Ligand-protected HpuB in lanes 3 and 5 is indicated by stars.

Download figure to PowerPoint

Protease accessibility of HpuAB in de-energized cells

To assess the potential role of TonB in the structure and function of HpuAB, we first compared the protease accessibility of HpuAB in DNM140, DNM146 (TonB) and DNM140 cells treated with the protonophore CCCP. Neither TonB inactivation nor CCCP treatment affected HpuAB expression levels (Fig. 1, lane 3; data not shown). In de-energized receptors (DNM146), the kinetics of trypsin degradation of the 35 kDa HpuA band was comparable to that seen for the wild-type receptor in DNM140. However, a modest but reproducible decrease in the amount of detectable trypsin-resistant HpuA cleavage products was noted (Fig. 3D). Figure 3 shows that HpuB in the TonB strain DNM146 was more susceptible to trypsin degradation compared with HpuB in a wild-type strain (compare Fig. 3D with A). Whereas the 85 kDa HpuB band was still detectable in DNM140 cells treated with 20 μg ml–1 trypsin, DNM146 HpuB was barely visible in the 5 μg ml–1 trypsin sample and was completely degraded with 20 μg ml–1 protease. Treatment of DNM140 cells with CCCP before trypsin digestion also resulted in increased accessibility of HpuB to proteolysis (data not shown), mimicking the effect of a TonB mutation. Thus, it appeared that the energy state of the cell membrane influenced the accessibility of receptor components to proteases.

Protease accessibility of de-energized HpuAB in the presence of ligand

The ability of bound Hb to protect HpuAB from trypsin degradation was used to examine whether TonB- mediated receptor energization was required for ligand binding. In this assay, to analyse the kinetics of receptor degradation by exogenous protease, the duration of trypsin digestion was varied instead of the trypsin concentration, as shown previously in Fig. 3. Fe-stressed meningococci with energized (DNM140) or de-energized HpuAB receptors (DNM146) were incubated with saturating amounts of Hb before treatment with trypsin (100 and 250 μg ml–1) for 0–40 min. In the absence of ligand, we saw the enhanced proteolysis of HpuB in the TonB mutant strain, compared with HpuB in a wild-type receptor (compare Fig. 5B, lanes 1–6 with Fig. 5A, lanes 1–6). Whereas HpuB in wild-type cells treated with 100 μg ml–1 trypsin was still detectable after 30 min, the HpuB in de-energized cells was rapidly degraded and disappeared after only 5–10 min. This was consistent with the TonB-dependent protease sensitivity of HpuB seen in Fig. 3 when cells were treated with increasing trypsin concentrations. In the presence of Hb, both HpuA and HpuB were shielded from trypsin cleavage throughout the 40 min proteolysis of DNM140 (Fig. 5A, lanes 7–12) and DNM146 (Fig. 5B, lanes 7–12) with 100 μg ml–1 trypsin. However, when ligand-protected DNM140 and DNM146 cells were incubated with 250 μg ml–1 trypsin (Fig. 5C and D), HpuA and HpuB were degraded to near completion within 40 min in both strains. Western blot analysis of DNM140 cells treated with 10 μM CCCP before binding of Hb and trypsin digestion yielded cleavage profiles that mirrored those obtained from DNM146 (data not shown). Thus, although the energy state of HpuB affected its protease accessibility, ligand binding can occur in the absence of TonB or an intact PMF.

image

Figure 5. Kinetic protease accessibility assays of energized versus de-energized HpuAB. Cells from the indicated strains were incubated with HBSS as a no ligand control (lanes 1–6) or Hb (lanes 7–10) before exposure to 100 μg ml–1 (A and B) or 250 μg ml–1 trypsin (C and D). Trypsin digestion of whole cells proceeded for the times indicated below each blot before termination with TLCK. Whole-cell lysates were separated by SDS–PAGE and analysed by Western blot. HpuB** indicates a transient 83 kDa cleavage product of HpuB seen with treatment with 100 μg ml–1 trypsin for 5–10 min (distinct from the cross-reactive ≈ 80 kDa band visible even at maximum trypsin digest conditions).

Download figure to PowerPoint

Hb binding kinetics to de-energized mutants

To characterize quantitatively the importance of the energy state of HpuAB in its interaction with Hb, we measured binding of [125I]-Hb to a TonB mutant (DNM146) or to DNM140 whole cells treated with 10 μM CCCP (Fig. 6A). Comparing the affinity of Hb for DNM146 (Kd = 137.6 nM) and CCCP-treated DNM140 (Kd = 199.8 nM) with DNM140 (Kd = 149.2 nM) suggested that the affinity of the HpuAB–Hb interaction was largely insensitive to the receptor’s energy state. However, the Hb binding capacity (pg of Hb bound μg–1 total cellular protein) was significantly reduced in de-energized cells, yielding Bmax values (DNM146 Bmax = 502.9; CCCP/DNM140 Bmax = 220.9) eight- to 20-fold lower than energized receptors of DNM140 (Bmax = 4338). Receptor copy numbers in DNM146 (4.69 × 109 molecules μg–1 total protein) and CCCP-treated DNM140 cells (2.06 × 109 molecules μg–1 total protein) were estimated to be approximately 10-fold lower than in intact wild-type DNM140 (4.05 × 1010). The reduced binding capacity in CCCP-treated DNM140 cells was not due to CCCP toxicity because the viability of DNM140 was unaffected by CCCP treatment (data not shown). Thus, we found that HpuAB in de-energized cells exhibited dramatically different Hb binding kinetics than the energized receptor of DNM140 (Fig. 6A). Although the overall number of receptors competent to bind Hb was dramatically reduced in the absence of TonB or an intact PMF, the amount of HpuAB expressed or the affinity of HpuAB for Hb was not reduced.

image

Figure 6. A. Hb binding kinetics of de-energized HpuAB. Liquid phase equilibrium binding analysis of [125I]-Hb to whole cells of TonB strain DNM146 (triangles) or HpuA+B+ DNM140 treated with 10 μM CCCP (circles). The binding kinetics of Hb to untreated DNM140 (squares) are shown for comparison. The insert graph shows identical data from TonB mutant and CCCP-treated cells plotted on a smaller y-axis range to allow easier comparison with the much higher capacity wild-type binding curve. Error bars indicate the standard error of the mean for each combined data set.

B. Dissociation kinetics of Hb from energized and de-energized HpuAB receptors. After cells were equilibrated with 100 nM [125I]-Hb, excess cold Hb (8 μM ) was added, and incubation continued for 2, 5,10, 20, 30 and 40 min. Dissociation kinetics are shown for Hb from DNM140 (squares), DNM146 (triangles) and CCCP-treated DNM140 (circles). Data presented are the mean of two independent experiments in which each sample was conducted in quintuplicate. Error bars indicate the standard error of the mean for each combined data set.

Download figure to PowerPoint

Hb dissociation kinetics from de-energized HpuAB

To investigate the potential energy requirement for ligand release from HpuAB, we examined the dissociation kinetics of [125I]-Hb from HpuAB in energized and de- energized cells. The results of these experiments, shown in Fig. 6B, are expressed as a percentage of [125I]-Hb remaining bound at various time points after the addition of excess unlabelled Hb. Haemoglobin dissociation from DNM140 proceeded from 100% down to ≈ 50% over the 40 min time course, whereas ligand dissociation was almost completely inhibited from de-energized receptors of DNM146 and CCCP-treated DNM140 cells (Fig. 6B). This suggested that TonB and an intact PMF gradient are necessary for ligand dissociation from HpuAB.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The ability to scavenge Fe from human carrier proteins is a virulence determinant believed to be important for bacteria to instigate and sustain an infection. Many successful pathogens such as NM have developed multiple redundant Fe acquisition mechanisms to satisfy this nutritional requirement within the Fe-limiting host environment. We have characterized the NM haem-Fe transporter HpuAB, and this report describes our investigation into the structure and energy-dependent interactions of this receptor with Hb, one of its three ligands. The genetic organization, proposed negative regulation of the hpuAB operon by Fur and Fe and the two-component structure consisting of a TonB-dependent transporter (HpuB) and a lipoprotein (HpuA) suggest that HpuAB is closely related to the neisserial Tf (TbpBA) and Lf (LbpBA) Fe transport systems (Lewis and Dyer, 1995; Lewis et al., 1997; Schryvers and Stojiljkovic, 1999). However, previous studies in our laboratory have revealed several unique features of HpuAB. Although HpuB is ≈ 25% identical (≈ 40% similar) to TbpA and LbpA, the lipoprotein HpuA is only half the size and has no homology with its counterparts TbpB and LbpB. In addition, unlike the gonococcal TbpB and the LbpB lipoproteins from NM and N. gonorrhoeae (NG), HpuA is absolutely required for receptor function (Lewis et al., 1999; this study). These unique characteristics of HpuAB are not surprising given that the receptor–ligand interactions and receptor function are dramatically different from the TbpBA-Tf system. Notably, HpuAB binds three structurally diverse ligands – Hb, Hp and HbHp – and extracts haem-Fe from both Hb and HbHp (Lewis and Dyer, 1995). Whereas Tf and Lf are homologous bilobed proteins that carry two Fe per molecule, the unrelated ligand Hb consists of four polypeptides (α2β2), each with a tightly bound Fe-centred porphy- rin ring. As discussed in more detail below, the iron- containing ligands encountered by HpuAB in vivo most probably include dissociated α1β1 Hb dimers, either free or complexed with Hp. Finally, the interaction of HpuAB with Hb ligands triggers the release and internalization of the intact haem-Fe moiety, in contrast to the transport of only the ferric ion from Tf or Lf. This study has uncovered additional properties of this receptor that distinguish it from other members of the family of two-component TonB-dependent Fe transporters.

To elucidate clearly the role of each Hpu receptor component in interactions with ligand, it was important to construct isogenic mutants lacking either HpuA or HpuB. As expected, studies of HpuA+B strain DNM2B6 (Lewis et al., 1999) and its HmbR derivative DNM143 (this study) demonstrated that the proposed haem- transporting channel protein HpuB is required for utilization of haem-Fe from Hb and HbHp. The inability of a non-polar HpuA deletion mutant to grow when Hb or HbHp was the sole Fe source indicated that the lipoprotein component of HpuAB is also indispensable to receptor function. This is in contrast to the non-essential accessory role played by the analogous lipoproteins of the Tf and Lf receptors. Binding of Hb to HpuAB was saturable and specific. With an estimated Kd of ≈ 150 nM, the affinity of HpuAB for Hb is approximately 10-fold lower than that reported for the binding of human Tf by the gonococcal receptor TbpBA (Cornelissen and Sparling, 1996). Assuming that the meningococcal Tf receptor has a similar affinity for Tf, this difference may be responsible for the ability of iron-loaded Tf to enhance mortality rates more efficiently in a mouse model of meningococcal disease (Schryvers and Gonzalez, 1989). Analysis of Hb binding to wild-type HpuAB indicated the presence of a single high-affinity binding site for Hb, consistent with the observation that mutants lacking either HpuA or HpuB did not significantly bind Hb. In contrast, Cornelissen and Sparling (1996) showed that TbpA and TbpB each bound Tf separately with high affinities (Kd = 2.3 nM and Kd = 7.4 nM respectively). They concluded that the two-site Tf binding kinetics (Kd1 = 0.8 nM, Kd2 = 16 nM) to wild-type gonococci was the result of the combined activity of TbpA and TbpB, whose conformation and ligand interaction change when both components are present (Cornelissen and Sparling, 1996). Thus, HpuAB is distinguished by an absolute requirement for both receptor components, which appear to interact very closely to form a single bipartite high-affinity binding site for Hb.

Previous studies revealed the high degree of species specificity exhibited by the N. meningitidis Tf and Lf receptors for the binding and use of human Tf and Lf respectively (Schryvers and Morris, 1988a,b; Schryvers and Gonzalez, 1989). Thus, it seemed reasonable to hypothesize that Hb utilization by NM would also be specific for human Hb. However, competitive binding assays demonstrated that HpuAB can bind both human and non-human Hbs. Compared with the Ki for human Hb, the binding affinity of these ligands to HpuAB was similar, within a two- to threefold range. This is in agreement with the observation that human and bovine Hb did not differ in the ability to support meningococcal growth in a mouse infection model (Schryvers and Gonzalez, 1989). Stojiljkovic et al. (1996) reported that strains expressing HmbR showed a nearly twofold preference for human Hb over various animal Hbs, whereas HpuAB did not discriminate between different sources of Hb in an assay measuring growth around Hb-saturated discs. They suggested that the different abilities of the two Hb receptors to discriminate between human and non-human Hbs might occur if HmbR recognizes the globin portion of Hb, whereas HpuAB binds to the conserved haem moiety of the ligand (Stojiljkovic et al., 1996). However, the broad specificity of HpuAB could also be explained by the high degree of homology between the Hbs tested. The β-chains of the non-human Hb used are 83.6–94.5% identical to human β-globin, whereas the level of identity of α-subunits ranges from 82.4% to 92.2% compared with human α-globin (data not shown). Further, we have shown that haem bound weakly to HpuAB and only competed poorly with Hb for binding. In native Hb, haem is bound within the globin polypeptides surrounded by non-polar side groups, with only limited exposure of the protruding propionate side groups of porphyrin (Hargrove et al., 1996). Our data suggest that haem is not a primary contact site of HpuAB with Hb. Although HpuAB must bind directly to haem at some stage in its transport, the initial ligand recognition and binding of Hb involves interactions of HpuAB with globin chains, with only minimal, low-affinity contacts with haem. It is unlikely that HpuAB extracts haem-Fe from Hb by direct competitive binding, as the affinities of apohaemoglobins for haem are very large, with equilibrium dissociation constants in the 10–12– 1015 M range (Hargrove et al., 1996). Thus, to allow the dissociation of haem from Hb before internalization by HpuAB, the interaction of the receptor with Hb must induce conformational changes in the ligand that reduce its affinity for the haem prosthetic group.

An unexpected result was the enhanced binding of the oxidized methaemoglobin (metHb) form versus reduced oxyHb, which suggested that metHb might be the preferred ligand. In erythrocytes, Hb is predominantly in the reduced form because of its high concentration (335 g l –1 or 5 mM ) and the presence of metHb reductases (Choury et al., 1981; Hargrove et al., 1997; Riggs, 1998). Upon cell lysis, Hb is readily oxidized and becomes diluted into the plasma, promoting the dissociation of α2β2 tetramers into α1β1 dimers (Bunn, 1987; Hargrove et al., 1997; Riggs, 1998) that form complexes with Hp (Nagel and Gibson, 1971; Langlois and Delanghe, 1996). In the oxidized form of Hb, metHb, the change in the haem-Fe valence state (Fe3+) and the subsequent loss of liganded oxygen results in dramatic allosteric changes in the tertiary and quaternary structure of the protein (Fung et al., 1976; 1977; Cordone et al., 1990; Marden et al., 1995). Thus, the Hb available to N. meningitidis as an Fe source is probably α1β1 dimers of deoxy-metHb, either free or complexed with Hp, which is structurally different from the oxyHb tetramers found in erythrocytes. The higher affinity of HpuAB for metHb may allow more efficient haem uptake, as haem release from Hb is enhanced upon haem oxidation and the dissociation of Hb into dimers (Benesch and Kwong, 1990; Gattoni et al., 1996; Hargrove et al., 1997). Notably, metHb concentrations were observed to be significantly increased in children suffering from septic shock (Kilbourn, 1997; Krafte-Jacobs et al., 1997), resulting from the reaction of Hb with nitric oxide (NO). NO elevation was correlated with increased severity of meningococcal disease (Kilbourn, 1997; Baines et al., 1999). Finally, bacterial endotoxin (LPS) binds to human Hb and mediates the oxidation of Hb to metHb (Weinberg, 1978; Kaca et al., 1994; 1995). Thus, it is likely that metHb becomes more abundant during meningococcal sepsis, and the HpuAB receptor may have evolved to take advantage of this.

Assessment of the proteolysis of HpuAB on intact cells by trypsin yielded significant insight into the topology, architecture and conformation of this bipartite Fe transporter. The accessibility of HpuA and HpuB to digestion by exogenous proteases is consistent with the proposed surface exposure of both receptor components. Although indicative of localization in the outer membrane, the rapid disappearance of the ≈ 85 kDa HpuB band without any detectable cleavage intermediates is most probably explained by cleavage of HpuB into fragments that we were unable to detect with the peptide-specific antibody available, because of either trypsin cleavage within the antibody epitope or generation of products lacking this epitope. The lipoprotein component HpuA is believed to be largely surface exposed, only peripherally associated with the outer leaflet of the outer membrane by its N- terminal cysteine-linked fatty acid (Lewis et al., 1997). This is supported by our observation that trypsin cleaved HpuA within 25–30 amino acids of the N-terminal membrane anchor to generate an ≈ 30 kDa cleavage intermediate. We also observed an ≈ 21 kDa HpuA trypsin product resistant to further digestion by the highest trypsin concentrations tested. The pronounced hydrophilicity of HpuA suggests that this protected domain does not result from an integral association of the C-terminal two-thirds of HpuA with the membrane. It is possible that the protease-resistant ≈ 210-amino-acid region is in a tightly folded conformation, such that trypsin accessibility is completely inhibited; however, this seems unlikely. Although its actual conformation is unknown, this C-terminal domain contains 23 trypsin sites, with many located in highly hydrophilic regions predicted to be solvent exposed. A more plausible explanation for the trypsin-resistant 21 kDa fragment is that this portion of HpuA is in direct physical interaction with HpuB. Consistent with this hypothesis, the HpuA trypsin products remained associated with the cell pellet after proteolytic release from N-terminal membrane anchors. In addition, the 21 kDa HpuA species was rapidly degraded in strain DNM143, in which HpuB is absent and thus unable to afford protection. The appearance of an ≈ 38 kDa HpuB trypsin-resistant fragment in HpuAB+ DNM69, which was not detected in analysis of the wild-type receptor, suggests that the presence of HpuA dramatically alters the conformation of HpuB. The formation of a single high-affinity Hb binding site by two separate proteins is presumably the result of this close structural relationship.

Both receptor components were protected from de- gradation in the presence of the specific ligands Hb and HbHp. Changes in protease sensitivity upon ligand binding could result from direct blocking of trypsin cleavage sites on the receptor proteins or ligand-induced conformational changes in HpuAB that render protease target sites inaccessible (Moeck et al., 1996; 1997). The fact that Hb and HbHp comparably protected both HpuA and HpuB reinforces our model in which both receptor proteins are involved in binding to both ligands. Trypsin assays of the HpuA deletion mutant DNM69 in the presence of ligand showed that Hb, but not HbHp, weakly protected HpuB from proteolysis. Thus, HpuB alone may bind free Hb, albeit with a very low affinity, as significant binding of [125I]-Hb was not detected in liquid phase binding assays. Overall, our findings are consistent with the results of Chen et al. (1998), who showed that an HpuAB+ strain of N. gonorrhoeae showed impaired, but detectable, Hb binding in a solid phase dot-blot assay. Nonetheless, although HpuB itself may interact with Hb, optimal binding of this ligand and acquisition of haem-Fe by HpuAB of both pathogenic Neisseria species require both receptor proteins acting in concert.

In its role as the energy source powering transport of Fe across the outer membrane, TonB is thought to interact directly with the periplasmic N-terminus of the transporter, transducing the potential energy of the PMF into conformational changes in the receptor protein(s) (Postle, 1990; Klebba et al., 1993; Braun, 1995). The TonB dependence of HpuAB protease susceptibility and of ligand interactions with extracellular receptor loops suggests that TonB-induced structural changes are transmitted through the membrane to surface-exposed regions of the transporter. When HpuAB was de-energized by the inactivation of TonB (DNM146) or the PMF (CCCP-treated DNM140), there was a pronounced increase in the protease susceptibility of HpuB in the absence of ligand. This suggested conformational changes in HpuB and/or HpuA that exposed additional trypsin cleavage sites. We also noted a modest, but reproducible, reduction in the amount of the protease-resistant ≈ 21kDa HpuA trypsin product detected. Thus, in the presence of TonB and an intact PMF, the conformation of HpuB, and perhaps its inter-action with HpuA, is altered. In contrast, Cornelissen et al. (1997) demonstrated that, in the gonococcal Tf receptor, the protease accessibility of the lipoprotein TbpB, but not the TonB-dependent protein TbpA, was increased in de-energized gonococci. Moeck et al. (1996) concluded that the binding of the siderophore ligand to FhuA altered the receptor conformation and triggered (or enhanced) its interaction with TonB, suggesting that the interaction of a TonB-dependent protein with TonB is ligand dependent. Although we cannot rule out a ligand-induced interaction of HpuB with TonB, our results clearly demonstrated TonB-mediated conformational changes in HpuB in the absence of ligand.

The protection of de-energized HpuAB from trypsin digestion by preincubation with Hb indicated that an energized state is not absolutely required for HpuAB ligand binding. However, TonB and the PMF profoundly influence the kinetics of the receptor–ligand interaction in two- component receptors such as TbpBA and HpuAB. Analysis of Tf binding to TbpBA revealed that Tf binds to deenergized receptors with an extremely high affinity in the subnanomolar range, compared with Kd values of 0.8 nM and 16 nM for the wild-type receptor (Cornelissen and Sparling, 1996). Although equally dramatic, the loss of TonB or the PMF had a quite different effect on HpuAB-mediated Hb binding. The affinity of Hb binding to HpuAB in a TonB mutant was indistinguishable from that of the energized receptor. However, the Hb binding capacity (Bmax) of the de-energized receptor was reduced almost 10-fold, indicating a decrease in either the overall receptor number or the number of ‘binding competent’ receptor complexes on the cell surface. Dissipation of the PMF by CCCP treatment mimicked these results, consistent with the central role of TonB as a transducer of PMF energy to HpuAB. Treatment of meningococci with CCCP did not affect cell viability, and Western blots indicated that the TonB mutant produced levels of HpuAB similar to those of wild-type meningococci. Several investigators have noted that strains with defects in the TonB–ExbBD system hyperexpress proteins from their iron acquisition systems (Guterman and Dann, 1973; Postle, 1990; Cornelissen et al., 1997). In our strain of N. meningitidis, neither the inactivation of TonB nor CCCP treatment appeared to affect the expression levels of HpuA or HpuB. Thus, in the absence of TonB or an intact PMF, only 10% of the HpuAB population was capable of binding Hb despite apparently normal levels of receptor expression. This subset of binding-competent HpuAB bound Hb with an affinity similar to that exhibited by energized HpuAB receptors.

Examination of the dissociation kinetics of [125I]-Hb revealed a profound impairment in Hb release from the Hpu receptor in a TonB mutant strain or from CCCP-treated wild-type meningococci. Cornelissen et al. (1997) also demonstrated that TonB was required for wild-type ligand dissociation rates from TbpBA. The rate of Hb dissociation from wild-type HpuAB was much slower than that observed for Tf from TbpBA, decreasing to 50% of total binding over 40 min compared with 18% for Tf during the same time. In protease accessibility studies of TbpBA in a TonB mutant, the irreversible binding of Tf to the majority of de-energized receptors correlated with the sustained protection of TbpA from trypsin cleavage during the 30 min experiment (Cornelissen et al., 1997). Un-expectedly, similar sustained protection of the majority of HpuAB by irreversibly bound Hb was not observed in kinetic trypsin assays. The protective effect of irreversibly bound Hb is counteracted by the increased sensitivity of HpuB to trypsin and the fact that only a fraction of HpuAB can bind Hb. Thus, the HpuAB that remains protected is assumed to represent the binding-competent fraction of the population. After treatment with high concentrations of trypsin for long time periods, the Hb ligand probably becomes degraded, exposing the underlying HpuAB to eventual degradation.

The ability of transporters such as HpuAB to release the apo-carrier protein after extraction and internalization of Fe or haem-Fe is essential for the receptor to recycle and acquire Fe from a second Fe carrier molecule. TonB may trigger Hb release directly through conformational changes in HpuAB that disrupt ligand-binding domains. Alternatively, TonB may affect Hb release indirectly by promoting haem extraction, leading to a reduced affin- ity of the receptor for apo-Hb and its subsequent dis- sociation. Figure 7 details our current model of TonB-dependent HpuAB interactions with Hb and contrasts it with the related but distinct mechanism of TbpBA- mediated Fe uptake from Tf. According to this model, HpuA and HpuB must both assume specific conformations and interact with each other to form a single high-affinity Hb binding site. TonB may serve to enhance the productive interaction of HpuB with HpuA, shifting the equilibrium of the receptor population to favour a binding-competent state capable of extracting haem-Fe from Hb. When TbpBA is de-energized, a similar decrease in binding capacity is not seen because both receptor components bind Tf independently and would not require TonB to form a binding-competent structure. The complex sequence of interactions and conformational changes involved in HpuAB-mediated haem uptake from Hb remains to be elucidated. A complete understanding of haem-Fe uptake by HpuAB will also require a detailed investigation of the interaction of HpuAB with apo-Hp and HbHp complexes, currently under way in our laboratory, as well as detailed characterization of the structure–function of the HpuAB receptor complex.

image

Figure 7. Comparison of models of HpuAB-Hb and TbpBA interaction – distinct mechanisms of Fe acquisition.

A. In de-energized cells (TonB mutants or CCCP-treated cells), the dramatic reduction in Hb binding capacity suggests that the majority of the HpuAB population is in a binding-incompetent state. The proposed dissociation of the HpuAB complex in this inactive ‘open’ conformation is consistent with the altered trypsin accessibility of receptor proteins in de-energized cells. The Hb binding site is thus disrupted, as the close association of both HpuA and HpuB is required to form a single high-affinity binding site. A small subset of the HpuAB receptor population binds Hb with approximately wild-type affinity, but haem-Fe extraction and ligand release are compromised.

B. In the absence of TonB, the TbpBA Tf receptor also assumes an altered conformation and is unable to release bound ligand properly. However, the very high ligand-binding capacity of Tf receptors expressed in TonB mutants, distinguishing HpuAB from TbpBA, suggests that the majority of the de-energized Tbp receptor population is in a binding-competent state. The ability of TbpA and TbpB individually to bind Tf with high affinity probably contributes to the energy independence of binding capacity. This model also illustrates the ability of the TbpB accessory lipoprotein to bind Fe-loaded Tf preferentially. No accessory role in determining ligand specificity has been demonstrated for the analogous HpuA lipoprotein.

C. In energized cells, TonB- and PMF-induced structural changes in the receptor components favour the existence of the majority of the HpuAB population in an ‘active’ conformation. Both HpuA and HpuB interact productively to allow high-affinity Hb binding at a single site, removal and internalization of the haem-Fe complex and release of apo-Hb.

D. In energized cells, TbpBA binds the bilobed Tf molecule via two binding sites before transporting only the ferric ion into the periplasm. It should be noted that TbpB alone and intact TbpBA receptors also discriminate between apo-Tf and Fe-Tf in energized cells, as shown in (B). Finally, unlike HpuB, the TbpA gated porin is capable of internalizing Fe from its ligand in the absence of its partner lipoprotein (not shown here).

Download figure to PowerPoint

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and culture conditions

Neisseria meningitidis strains (Table 1) were routinely cultured on GC base agar (Difco Laboratories) supplemented with IsoVitaleX (Becton Dickinson) at 37°C with 5.2% CO2. Strains containing antibiotic resistance markers were grown on media supplemented with antibiotics (Sigma) as indicated. Strains DNM2B6 (HpuA+B), DNM2A13 (HpuAB+) and DNM2E4 (HpuAB) have been described previously (Lewis et al., 1999). Meningococcal strains constructed during this study are summarized in Table 1 and described in detail below. E. coli strains were cultured on Luria–Bertani (LB) medium containing antibiotics as specified.

Table 1. Bacterial strains and plasmids used in this study.
StrainDescriptionReference
DNM2 Neisseria meningitidis serogroup C, serotype 2a; Nalr Lewis and Dyer (1995)
DNM2A13DNM2 hpuA::kanr Lewis et al. (1999)
DNM140HpuA+B+hmbR::kanr derivative of DNM2This study
DNM64Spontaneous strepr isolate of DNM2This study
DNM65 hpuB::ermrstreps, DNM64 derivativeThis study
DNM2 B6 hpuB::ermr Lewis et al. (1999)
DNM143 hpuB::ermr, hmbR::kanrThis study
DNM66ΔhpuA, hpuB::ermrstrepsThis study
DNM67ΔhpuA, hpuB+ deletion mutantThis study
DNM68ΔhpuA, hpuB::ermCrpsL (ermrstreps), hmbR::kanrThis study
DNM69ΔhpuA, hpuB+, hmbR::kanrThis study
DNM2N1 tonB::kanrThis study
DNM146 tonB::kanr, hmbR::ermrThis study
IR2113 tonB::kanr Stojiljkovic et al. (1996)
Plasmids
pIRS525 hmbR::kanr in pUC19 Stojiljkovic et al. (1995)
pHpuA25 hpuB in pGEM-7Zf(+)This study
pA25.ES hpuB::ermCrpsL; ermrstreps, pHpuA25 derivativeThis study
pHSS8-HmbR661 bp hmbR PCR product in pHSS8 suicide vectorThis study

Reagents and enzymes

Unless otherwise noted, all chemicals were obtained from Fisher Scientific. DNA restriction enzymes were purchased from Promega, New England Biolabs or Gibco BRL. The high-fidelity DNA polymerase Pfu Turbo (Stratagene) was used for all cloning and mutagenesis polymerase chain reactions (PCRs). PCR screening of mutant constructs was performed with Taq DNA polymerase (Roche Molecular Biochemicals). PCR nucleotide mix used in all PCR reactions was obtained from Roche.

Primers

Oligonucleotide primers used for PCR and DNA sequencing (Table 2) were synthesized and column purified by Biosynthesis or Integrated DNA Technologies.

Table 2. Primers used in this study.
Primer namePrimer sequence (5′–3′)
Used for construction of hpuA deletion mutant DNM67
P17.85GCCTTGCCTTTGATGTAG
hA.del2AAAAAAGAACTTGAAAACAACACTG
hA.del6TTTTCAAGTTCTTTTTTGGCTTTGTATTTCATCGCATACT
hA.del7GCGGGGCATCATTTTCAACTTT
Used for hmbR mutant construction and screening
hmbr1CACGGCGCGGCTTTGTTTAC
hmbR4.2ACGTGTCAAGGTGGAAGAATC
hmbR4ACACCAACCTCTTTTACGAAT
hmbR5ACCGTGGTTGTAAGTGAAATA
Used to make hpuB::ermCrpsL mutant DNM65
M13F-ClaATATCGATTCGCCAGGGTTTTCCCAGTCACGAC
M13R-ClaTGAATCGATTCACACAGGAAACAGCTATGAC
Used to PCR hpuB::ermCrpsL from DNM65 to make DNM66
P17.85 (see above) 
P2.85GTACCACCCATCTCACAAA
Used to PCR screen DNM2N1 tonB mutant
TonB432CGAATTTTAACCCCCGCAGTC
TonB1568TCGGCGGTAGTTGCGGTAAAA

DNA isolation and transformation

Chromosomal DNA was isolated from N. meningitidis strains grown overnight on GC agar using the Puregene DNA isolation kit (Gentra Systems) as directed. Plasmid DNA was purified from E. coli using either a Wizard miniprep kit (Promega) or a QIAprep Spin miniprep kit (Qiagen). DNA fragments were purified by extraction from 1% agarose gels with Qiaex II resin or QIAquick kits (Qiagen).

Transformation of naturally competent N. meningitidis (with gel-purified PCR products or chromosomal DNA) was accomplished as described previously (Lewis et al., 1999) using a spot transformation protocol modified from the method of Gunn and Stein (1996). Briefly, a loopful of cells from a 16– 18 h CDM0 plate was placed onto a GC plate, and ≈ 0.5 μg of DNA (10 μl) was mixed with the cells using an inoculating loop. The transformation mixture was incubated at 37°C, 5.2% CO2 for 4–6 h before selecting for transformants by plating to appropriate selective media. Cells incubated as above with 10 μl of sterile water served as negative controls.

Fe-dependent growth

To assess the ability of meningococcal strains to use a specific Fe source to support growth, cells were Fe starved in Chelex-treated defined medium (CDM0) and then grown in CDM0 containing the Fe source of interest (Dyer et al., 1987b). Growth conditions for assaying Fe acquisition from Hb, Hm and HbHp have been described in detail (Lewis and Dyer, 1995). Briefly, meningococci from 18 h CDM0 agar plates were resuspended in CDM0 broth and a specific Fe source (see below) in acid-washed sidearm flasks and grown at 37°C with shaking. Spectrophotometric measurements at OD600 were taken to monitor growth kinetics.

Fe sources

Bovine haemin, human ferrous Hb (type A0) and human Hp (pooled mixture of 1–1, 2–1 and 2–2 phenotypes) were purchased from Sigma. Haem was dissolved as described previously (Dyer et al., 1987a) and added as an Fe source to CDM0 cultures at a final concentration of 5 μM. Lyophilized Hb was prepared as a 100 μM stock solution in 10 mM HEPES buffer, pH 7.4, and added to CDM0 cultures to give 1 μM Hb (4 μM Fe). Similarly, cultures were supplemented with HbHp complexes prepared as described by Dyer et al. (1987a) at a final concentration of 4 μM Fe where indicated. The Fe chelator Desferal (Df; Ciba-Geigy) was added (10 μM) to Hm-, Hb- and HbHp-containing cultures to remove any free Fe. Controls in which 10 μM Df was added to CDM0 cultures (without added Fe sources) were performed to ensure effective removal of all free Fe by Df, seen as complete growth suppression. In addition, 16 μM human serum albumin (SA) was added to Hb-containing cultures to distinguish use of intact Hb as an Fe source from growth with Hm (HpuAB independent) released into the media as a result of the breakdown of Hb over the course of the experiment.

Construction of HmbR derivatives of DNM2

HpuAB and HmbR have been shown to undergo phase variation by a slipped-strand mispairing mechanism resulting from alterations in the length of a poly(G) tract within the coding sequence (Lewis et al., 1999; Richardson and Stojiljkovic, 1999). To ensure that none of the strains examined in this study could express that alternative Hb receptor HmbR and confound the interpretation of results, the hmbR locus in DNM2 was inactivated essentially as described by Lewis et al. (1999). Briefly, plasmid pIRS525 containing the hmbR gene with a kanamycin cassette inserted in the NotI site (provided by Dr I. Stojiljkovic, Emory University, Atlanta, GA, USA) served as a template for PCR amplification with primers hmbR1 and hmbR4.2. The 2.2 kb product from pIRS525 encoding hmbR::kanr was gel purified (Qiaex II; Qiagen) and transformed into DNM2, creating DNM140. Transformants were selected on GC medium containing 100 μg ml–1 kanamycin and screened for insertional inactivation of hmbR with primers hmbR1 and hmbR4.2 as above.

Alternatively, to generate mutants with hmbR interrupted by an ermr marker, a 661 bp fragment of hmbR was amplified from DNM2 chromosomal DNA using primers hmbR4 and hmbR5 and cloned into pCR2.1-Topo (Invitrogen). After sequence confirmation, the 661 bp hmbR fragment was excised using EcoRI, gel purified (Qiaex II; Qiagen) and ligated into EcoRI-linearized pHSS8, a neisserial suicide vector. The cloned hmbR fragment was then mutagenized by shuttle mutagenesis as described previously (Seifert et al., 1990). Briefly, pHSS8-hmbR was transformed into RDP146+ pTCA, which constitutively expresses Tn3 transposase. This recipient strain was then mated with NS2114 bearing pOX38::mTn3 (donor) to allow transposition of mTn3::erm into hmbR. Selection for positive transconjugants was carried out on LB media containing erythromycin (500 μg ml–1), tetracycline (20 μg ml–1) and kanamycin (40 μg ml–1). A clone confirmed by PCR (primers hmbR4 and hmbR5) to contain an insertion of the ermr cassette in hmbR was designated DNM145. Mutations in hmbR did not alter the growth or expression of HpuAB (data not shown; Fig. 1, lanes 2, 3 and 5–7).

Mutagenesis of hpuB with ermCrpsL double selection cassette

To facilitate the construction of hpuAB mutant strains of N. meningitidis lacking antibiotic resistance markers, a two-gene ermCrpsL cassette (Johnston and Cannon, 1999) was introduced into the chromosomal hpuB locus. Repeated attempts to insert the ermCrpsL gene into hpuA were unsuccessful and appeared to be a consequence of the instability of this construct (data not shown). Thus, strategies for the introduction of targeted mutations in hpuA used the hpuB::ermrstreps strain DNM65 described below.

A full-length clone of hpuB was generated from two previously published overlapping partial clones (Lewis et al., 1997) as follows. A fragment corresponding to the 5′ end of hpuB was excised from pPDS85 using SacI and ClaI. Similarly, plasmid pCX1, containing a 1.7 kb insert of the 3′ end of hpuB in pGEM-7Zf(+), was linearized by digestion with SacI and ClaI. The SacI–ClaI fragment of pPDS85 was then ligated into the SacI–ClaI-cut pCX1 to yield pHpuA25, containing an intact hpuB gene cloned in pGEM-7Zf(+). Subsequently, a spontaneous streptomycin-resistant variant of DNM2 was isolated by passage on GC medium containing up to 1 mg ml–1 streptomycin and designated DNM64. The ermCrpsL cassette was amplified from pFLOB4300 (Johnston and Cannon, 1999) using primers M13F-C and M13R-C, which flanked the marker and introduced ClaI restriction sites. The resulting 1850 bp product was cloned into pCR2.1-Topo (Invitrogen). This fragment was then excised with ClaI, gel purified (Qiaex II; Qiagen) and ligated into the single internal ClaI site in hpuB in plasmid pHpuA25. The resulting hpuB::ermrstreps construct (pA25.ES) was linearized by digestion with ScaI to inactivate the Ampr gene, gel purified and transformed into DNM64; positive transformants in which the hpuB::ermrstreps construct had replaced the wild-type hpuB gene were selected for growth on GC agar containing 2 μg ml–1 erythromycin and screened for sensitivity to streptomycin (1 mg ml–1). The inactivation of hpuB in DNM65 was confirmed by PCR and Western blot analysis.

Construction of HpuA deletion mutant

Our laboratory has previously constructed hpuA mutant strain DNM2A13 by the insertion of a promoterless kanamycin resistance cassette (aphA-3) lacking transcription initiation or termination signals into hpuA (Lewis et al., 1997). Western blot analysis of DNM2A13 confirmed the lack of HpuA; however, the expression of HpuB was significantly reduced. To determine accurately the role of HpuA in Hb utilization and binding, we generated a non-polar mutation in hpuA that allowed the expression of wild-type levels of HpuB. Using overlap-extension PCR mutagenesis (Ho et al., 1989), an hpuAB construct lacking 95% of hpuA was made, in which the first five codons of hpuA (plus ≈ 800 bp upstream sequence) were fused in frame with the coding sequence for the last 17 C-terminal residues of HpuA and the N-terminus of the downstream hpuB gene. Briefly, fragment 1 (951 bp) was generated using primers hA.del6 and hA.del7 under standard conditions (3 min at 95°C, 30 cycles of 92°C for 1 min, 52°C for 1 min and 72°C for 3 min, followed by a final extension step of 72°C for 10 min) to amplify the sequence from ≈ 800 bp upstream of hpuA to the first five codons of the HpuA signal sequence. Fragment 2 (2092 bp), coding for the last 17 C-terminal residues of hpuA to the 3′ end of hpuB, was created with primers hA.del2 and P17.85. Primer hA.del2 contains a 25 bp extension complementary to hA.del6 to facilitate the fusion of the two PCR products. Gel-purified fragments 1 and 2 were fused together to yield an hpuAB construct (3043 bp) with an in frame deletion of 307 of 324 residues of HpuA by first denaturing the two fragments at 95°C for 3 min, followed by overlap-extension PCR carried out for 10 cycles (one cycle included denaturation at 92°C for 1 min, annealing at 40°C for 2 min and extension at 72°C for 5 min) with a final extension step at 72°C for 10 min. To amplify the 3043 bp full-length product, primers hA.del7 and P17.85 were added and cycles conducted as follows: template DNA was denatured at 95°C for 3 min followed by 30 cycles of denaturation at 92°C for 1 min, annealing at 50°C for 1 min and extension at 72°C for 4 min with a final extension step of 72°C for 10 min. This final PCR product was gel purified using Qiaex II resin (Qiagen) and introduced into DNM65 (hpuB::ermrstreps) by spot transformation as described above. Primary screening for clones that had replaced the ermrstreps marker with the hpuA-deleted PCR construct was performed by plating on GC media sup- plemented with 1 mg ml–1 streptomycin; reversion to erms was also confirmed. PCR screening with primers hA.del7 and HpuA8, which flank hpuA, sequencing and Western blot analysis served to confirm the construction of the HpuA deletion mutant, DNM67. An HmbR derivative of DNM67, named DNM69, was created by spot transformation of DNM67 with chromosomal DNA from strain DNM140 (HpuA+B+, hmbR::kanr) and confirmed by PCR using primers hmbR1 and hmbR4.2 as described above. Analysis of whole-cell lysates from Fe-depleted cultures of DNM67 clones (HpuAB+) and its HmbR derivative DNM69 by immunoblot demonstrated the loss of HpuA and expression of HpuB at levels comparable with strains DNM2 and DNM140 (data not shown; Fig. 1). Similarly, to generate an HpuA+B mutant in an HmbR background (named DNM143), the HpuA+B (hpuB::ermr) meningococcal strain DNM2B6 (Lewis, 1999) was transformed with DNM140 chromosomal DNA as above.

Construction of HpuAB double mutant

To construct an HpuAB derivative of DNM67, designated DNM66, primers P17.85 and P2.85 were used to amplify a fragment of hpuB containing the ermrstreps cassette from DNM65 chromosomal DNA using the following conditions: initial denaturation at 95°C for 3 min, followed by 30 cycles of denaturing at 92°C for 1 min, annealing at 55°C for 1 min, extension at 72°C for 4 min and a final extension cycle at 72°C for 10 min. The 3.9 kb PCR product was gel purified (Qiaex II; Qiagen) and transformed into DNM67 with positive transformants selected on GC plus erythromycin (2 μg ml–1) plates. Insertional inactivation of hpuB by the ermrstreps cassette was confirmed by PCR using primers P17.85 and P2.85 as described above. Loss of expression of HpuA and HpuB was assessed by Western blot as detailed previously. Finally, an HmbR derivative of DNM66, named DNM68, was generated as described above for the construction of strain DNM69. Both components of the Hpu receptor were absent in strain DNM68 (Fig. 1).

Construction of TonB mutant

To assess the role of TonB in the interactions of HpuAB with its ligands, a tonB knock-out mutant (DNM2N1) was generated in the DNM2 background by spot transformation of DNM2 with chromosomal DNA from meningococcal strain IR2113 (tonB::kanr) provided by Dr I. Stojiljkovic (Stojiljkovic and Srinivasan, 1997). The presence of the kanr cassette in the KpnI site of tonB was demonstrated by PCR using primers TonB432 and TonB1568, yielding the expected 2636 bp and 1136 bp products from DNM2N1 and DNM2 (control) respectively. Consistent with previously described NM tonB mutants (Stojiljkovic and Srinivasan, 1997), DNM2N1 in the DNM2 background was unable to grow using haemoglobin or transferrin as sole iron sources (data not shown). An hmbR derivative of the tonB knock-out, named DNM146, was constructed by spot transformation of DNM2N1 with chromosomal DNA from DNM145 (hmbR::ermr) with selection on GC-erythromycin (2 μg ml–1) media. Figure 1 (lane 3) shows that TonB and HmbR mutations did not affect the expression of HpuAB.

Protease accessibility assays

Meningococcal strains were grown in Fe-depleted CDM0 medium as described previously to an OD600 of 0.4–0.45. Cells were then diluted with CDM0 to equalize all cultures to the same optical density (the lowest optical density of cultures in particular assay). Cells were pelleted by centrifugation at 5000 r.p.m. for 5 min and resuspended in a 1:10 volume of cold Hanks’ balanced salts (HBSS) media prepared according to the manufacturer’s protocol (Sigma). Resuspended cells (100 μl) were treated with trypsin (Sigma) to final concentrations of 5, 20, 50, 100 and 250 μg ml–1 at 37°C for 30 min. Protease digestion reactions were stopped by the addition of 250 μM N-α-p-tosyl-L-lysine chloromethyl ketone hydrochloride (TLCK), an irreversible trypsin inhibitor (Sigma). Trypsinized cells were centrifuged, resuspended in 50 μl of HBSS, with 20 μl being solubilized and analysed by 15% SDS–PAGE and Western blot as described earlier.

For protease accessibility assays in the presence of Hb or HbHp, ligands were added to 100 μl of cells to a final concentration of 1 μM (saturating). After incubation with ligands for 15 min at room temperature, cells with bound ligand were centrifuged, and supernatants containing excess ligand were removed. Cells were resuspended in 100 μl of HBSS, treated with trypsin and processed as above. Control reactions included cells incubated with the non-specific ligand Tf (1μM) or with HBSS (no ligand control).

Kinetic trypsin accessibility/ligand protection assays were performed essentially as described above. Briefly, cells were preincubated with 1 μM Hb or HBSS (control) for 15 min as before. Cells were then exposed to 100 μg ml–1 or 250 μg ml–1 trypsin for up to 40 min before being stopped with TLCK. Whole-cell lysates were prepared and analysed as described above.

SDS–PAGE and Western blot analysis

Whole-cell lysates were separated by SDS–PAGE using the discontinuous buffer system described by Laemmli (1970). Lysates were solubilized by boiling in SDS sample buffer (Maniatis et al., 1989) and electrophoresed on 15% acrylamide gels (10 mm × 12 mm) at 25 mA per gel for 2 h. The separated proteins were then transferred to nitrocellulose (Nitrobind; MSI, Fisher Scientific) at 0.98 A for 30 min. Blots were blocked for 1 h at room temperature with 0.5% skimmed milk (Difco Laboratories) in TBS buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.5) before being probed overnight at 4°C with antipeptide anti-HpuA or anti-HpuB antisera (Lewis et al., 1999) diluted 1:50 in blocking solution (see above). Blots were then washed three times (10 min each) with 1× TSB and incubated with goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP; Southern Biotechnology Associates) at room temperature for 1 h. Detection was performed using 4-chloro-1-naphthol HRP colour development reagent according to the manufacturer’s instructions (Bio-Rad). After we had established the tryptic fragments resulting from proteolysis of HpuA and HpuB independently, they were subsequently probed with anti-HpuA and anti-HpuB antibodies simultaneously as above.

Affinity purification of FbpA-specific antisera

Purified FbpA, a neisserial periplasmic Fe-binding protein, and polyclonal FbpA-specific antisera were kindly provided by Dr T. Mietzner (University of Pittsburg, Pittsburg, PA, USA). The FbpA antisera was affinity purified by adsorption to immobilized FbpA as described below. Briefly, 100 μg of purified FbpA was separated on two preparative 15% SDS–PAGE gels (50 μg per gel) and transferred to nitrocellulose. The band corresponding to FbpA, identified by amido black staining, was excised and blocked for 1 h with phosphate-buffered saline (PBS; pH 7.4) + 0.05% Tween 20 before incubation with polyclonal FbpA antisera overnight at room temperature. After two 10 min washes with PBS + 0.05% Tween 20 and one 10 min wash with PBS, the adsorbed antibody was eluted from the nitrocellulose by incubation with 3 ml of 100 mM glycine, pH 2.5, for 30 min. Finally, the affinity-purified FbpA-specific antibody was neutralized with 300 μl of 1M Tris (hydroxymethyl) amino-methane, pH 8.0, and dialysed against 4 l of PBS, pH 7.4, overnight at 4°C. To detect FbpA in Western blots, the affinity-purified FbpA-specific antibody was diluted 1:200 in blocking solution.

Preparation of radiolabelled ligands

Human ferrous haemoglobin A0 (Sigma) was resuspended as a 100 μM stock solution in 10 mM HEPES buffer, pH 7.4. Human Tf was prepared and Fe saturated (≈ 90% saturated) as described previously (Blanton et al., 1990). Hb and Tf were then radiolabelled with 125I (NaI; ICN Biomedicals) using the Iodobead reagent (Pierce) according to the manufacturer’s recommendations. Standard Hb iodination conditions (1 mg of Hb labelled with a single Iodobead and 1 mCi of 125I in a 250 μl volume for 5 min at room temperature) routinely yielded [125I]-Hb with a specific activity of ≈ 2.5 × 105 c.p.m. μg–1 Hb. Iodination of Tf (150 μg of Tf labelled with two Iodobeads and 1.5 mCi of 125I in a 250 μl volume for 15 min) yielded [125I]-Tf with a specific activity of 5.4 × 105 c.p.m. μg–1. Control mock iodinations, in which ligands were treated with just the Iodobead reagent or with Iodobead plus non-radioactive sodium iodide, were conducted to ensure that neither oxidative damage nor steric hindrance resulting from excess labelling had significant effects on ligand integ-rity. Excess free 125I was removed from the labelled Hb by passage over a PD-10 desalting column (Amersham- Pharmacia Biotech) and elution in 250 μl fractions with 0.1 M sodium phosphate buffer, pH 6.6 (Maniatis et al., 1989). The protein concentration of pooled fractions of [125I]-Hb thus purified was determined using the bicinchoninic acid (BCA) assay (Pierce). The integrity and purity of radiolabelled ligands was confirmed by native PAGE as described previously (Lewis and Dyer, 1995) and autoradiography (data not shown).

Liquid phase haemoglobin-binding assay

A liquid phase equilibrium-binding assay to measure the affinity and binding capacity of HpuAB for 125I-labelled Hb was developed based on the procedure described by Cornelissen and Sparling (1996). Meningococci grown overnight on CDM0 agar were inoculated into liquid CDM0 media and grown at 37°C with shaking for 6–7 h to an OD600 of 0.4–0.45. Approximately 5 × 107–1 × 108 cfu per well of cells in CDM0 medium were mixed with increasing concentrations of [125I]-Hb (2.5–500 nM) or [125I]-Tf (1–100 nM) in a total volume of 150 μl. Binding reactions were allowed to proceed for 20 min at room temperature in individual wells of 96-well Multiscreen microtitre filter plates (0.65 μm pore size, MADV N65; Millipore) that were preblocked with 2% bovine serum albumin (BSA; Sigma) for 2 h. Unbound ligand was removed by vacuum filtration and three CDM0 washes of 150 μl each. To determine the amount of cell-associated radioligand, filters were air dried, punched out using the Multiscreen filter punch apparatus and counted in a Packard Cobra II gamma counter. In addition, whole cells were washed with sterile water, and total cellular protein concentration was determined by BCA assay (Pierce). The total ligand (in pg) bound to whole cells was then normalized to μg of total cellular protein.

To determine the role of the proton motive force (PMF) in TonB-dependent HpuAB functions, Fe-starved meningococcal cells were incubated with the protonophore CCCP (Sigma) at 10 μM for 10 min before being analysed in binding or protease accessibility assays where indicated.

Analysis of binding kinetics data

For a single binding assay, binding at each ligand concentration was typically measured in triplicate. Each data point presented, unless otherwise noted, represents the mean of three independent experiments. In controls, non-specific binding was tested at each concentration by the addition of excess unlabelled Hb (12.4 μM ). To calculate specific Hb binding, non-specific binding counts were subtracted from total binding counts at each ligand concentration. Non-linear regression, curve-fitting analysis of binding data was conducted using GraphPad PRISM, version 2.0 software to estimate receptor copy numbers and Kd values. To apply this analysis to the kinetics of Hb binding to wild-type HpuAB, binding data were fitted to either a one-site or a two-site binding equation. The model that best describes the HpuAB–Hb interaction was determined by comparison of the goodness of fit (based on the lowest sum-of-squares) of the data to the two equations using the F test (Motulsky, 1999). Briefly, the curve generated by the more complex equation (two-site model) will almost always come closer to the data points because it has more inflection points. The F test determines whether the lower sum-of-squares is worth the ‘cost’ of additional variables (loss of degrees of freedom). The F ratio, then, compares the relative increase in sum-of-squares with the relative increase in degrees of freedom. If the one-site model is correct, an F ratio near 1.0 is expected with a high P-value. If the ratio is significantly >1.0 and the P-value is low, the two-site model of binding describes the interaction more accurately (Motulsky, 1999). Control experiments measuring the binding kinetics of Tf to TbpAB were analysed similarly and model predictions conducted using GraphPad PRISM.

Competitive liquid phase Hb-binding assay

A variation of the protocol described above was used to analyse the binding of non-human haemoglobins and methaemoglobin to HpuAB in a competitive binding assay. Non-human haemoglobins (baboon, porcine, bovine, rabbit and equine) and metHb (Sigma) were resuspended as described above for human Hb. Bovine haemin (Hm) was prepared as described previously (Dyer et al., 1987a). For this assay, the concentration of iodinated Hb was held constant at 100 nM, with the concentration of unlabelled competitors ranging from 1 nM to 10 μM in 10-fold intervals. Total binding controls (cells incubated with 100 nM [125I]-Hb alone) and non-specific binding controls (cells incubated with 100 nM [125I]-Hb and 125-fold excess cold human Hb) were each performed in quadruplicate and averaged. 100% specific binding was calculated by subtracting the average non-specific binding from the average total binding. Specific binding of human [125I]-Hb in the presence of heterologous competitor was then expressed as a percentage of total specific binding. The inhibition constants for each heterologous competitor Hb (Fig. 2B, insert) were compared with values obtained for human Hb using a two-tailed Student’s t-test.

Hb dissociation assay

Iron-stressed meningococci were mixed with 100 nM [125I]-Hb in a Millipore Multiscreen microtitre filter plate. Binding of Hb to cells was allowed to equilibrate for 20 min at room temperature. Excess cold competitor (8 μM ) was added to 125I-labelled Hb cells and incubated for up to 40 min. Unbound ligand was removed by filtration, and cell-associated radioactivity was detected using a gamma scintillation counter. Non-specific binding controls were performed in quadruplicate by adding excess cold Hb to cells coincident with 100 nM [125I]-Hb. Specific counts bound were determined by subtracting average non-specific counts from total counts at each time point. The percentage Hb bound was calculated by dividing specific counts at each time point by the specific counts bound at time 0 (no cold Hb competitor added).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank I. Stojiljkovich for providing HmbR mutant strains and plasmids used in this study. We also thank T. Mietzner for purified FbpA and FbpA-specific antisera. We appreciate the technical assistance of G. Caudle, K. Hartman and M. Gipson.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References