ABSTRACT
Iron is an essential nutrient for most living organisms. To acquire iron from their environment, Gram-negative bacteria use TonB-dependent transporters that bind host proteins at the bacterial surface and transport iron or heme to the periplasm via the Ton machinery. TonB-dependent transporters are barrel-shaped outer membrane proteins with 22 transmembrane domains, 11 surface-exposed loops, and a plug domain that occludes the pore. To identify key residues of TonB-dependent transporters involved in hemoglobin binding and heme transport and thereby locate putative protective epitopes, the hemoglobin receptor of Haemophilus ducreyi HgbA was used as a model of iron/heme acquisition from hemoglobin. Although all extracellular loops of HgbA are required by H. ducreyi to use hemoglobin as a source of iron/heme, we previously demonstrated that hemoglobin binding by HgbA only involves loops 5 and 7. Using deletion, substitution, and site-directed mutagenesis, we were able to differentiate hemoglobin binding and heme acquisition by HgbA. Deletion or substitution of the GYEAYNRQWWA region of loop 5 and alanine replacement of selected histidines affected hemoglobin binding by HgbA. Conversely, mutation of the phenylalanine in the loop 7 FRAP domain or substitution of the NRQWWA motif of loop 5 significantly abrogated utilization of heme from hemoglobin. Our findings show that hemoglobin binding and heme utilization by a bacterial hemoglobin receptor involve specific motifs of HgbA.
INTRODUCTION
Acquisition of iron by bacteria has been the focus of intense studies for many decades, not only for its importance in pathogenesis, but also to define key targets of vaccine development against pathogens. Although many bacterial receptors of iron and iron sources have been described over the years, the mechanism of heme acquisition in Gram-negative bacteria is still not thoroughly understood. In this article, we describe studies aimed at identifying residues and motifs involved in both hemoglobin (Hgb) binding and heme acquisition using the Hgb receptor HgbA of the Gram-negative bacterium Haemophilus ducreyi as a model system.
H. ducreyi is the etiological agent of the sexually transmitted genital ulcer disease chancroid, a risk factor for the heterosexual acquisition and transmission of HIV (1–4). H. ducreyi is strictly dependent on the availability of both iron and heme from its host since it is unable to synthesize heme de novo (5). H. ducreyi can obtain iron/heme from Hgb at the bacterial cell surface using its Hgb receptor, HgbA (6–8). HgbA is one of three TonB-dependent transporters (TBDT) present in the genome of prototypical strain 35000HP and the only TBDT absolutely necessary to establish infection in the human experimental model of chancroid (9, 10). Immunization with native HgbA (nHgbA) prevents infection in the experimental swine model of chancroid (11, 12). Furthermore, passive immunization with HgbA antisera protects against an experimental challenge, suggesting that vaccine protection is mediated by the humoral immune response elicited to nHgbA (13). Although it is not clear whether antibody protection is mediated by blocking Hgb binding or opsonization of the bacteria, polyclonal nHgbA antibodies partially abrogate binding of Hgb at the surface of H. ducreyi, suggesting that nutritional immunity may be a mechanism of protection of the HgbA vaccine.
HgbA is a member of a large family of outer membrane TBDT with a conserved tertiary structure comprised of a 22-stranded β-barrel, 11 extracellular loops, and a plug domain that occupies the pore inside the barrel (Fig. 1) (14). Although the basic structure of the HgbA β-barrel is known, it only accounts for less than 20% of the protein, and there is no accurate three-dimensional model of the surface and periplasmic portions of HgbA. We showed previously that all 11 extracellular loops of HgbA are necessary for utilization of iron/heme from Hgb (14). Conversely, substantial deletions in loop 5 or loop 7 abrogated Hgb binding to HgbA at the bacterial surface. The 16-residue deletion in loop 7 included the FRAP motif and histidine at position 706, but not NLYL (NXXL), conserved motifs found in many Hgb/heme receptors (15). Deletion of any of the other nine loops in HgbA does not impair Hgb binding (14).
Three-dimensional model of the Hgb receptor of H. ducreyi. Extracellular loops are identified by boxed numbers 1 to 11 and colored according to the following scheme: 1, magenta; 2, dark pink; 3, yellow; 4, salmon; 5, purple; 6, orange; 7, cyan; 8, pink; 9, yellow; 10, light pink; 11, light blue. Histidines are numbered according to their position in the predicted immature HgbA protein. The figure was created with PyMOL (version 1.2r3pre; Schrödinger, LLC.). Underlined and boldface histidines were targeted for further studies.
Based on a multiple sequence alignment of Hgb binding proteins, we constructed multiple hgbA mutants using multiresidue deletions and single or multiple alanine substitutions in loops 5 and 7. By studying binding to Hgb agarose and growth in media containing Hgb as the only source of iron/heme, we identified amino acids in loop 5 and loop 7 of HgbA that are required for Hgb binding and others that are required for heme utilization. Identification of specific motifs and/or residues involved in Hgb binding and heme acquisition will aid in defining the mechanism of heme transport by a bacterial Hgb receptor. These studies will also guide vaccine studies by identifying putative epitopes involved in Hgb binding, which could be potential targets for nutritional immunity by preventing Hgb and heme acquisition by a bacterial Hgb receptor.
MATERIALS AND METHODS
Bacterial strains and culture conditions.Bacterial strains and plasmids used in this study are listed and described in Table 1. H. ducreyi strain 35000HP (16) is a human-passaged form of strain 35000 (17), the prototypical class I isolate (18). All plasmid constructs used in this study were expressed in strain FX547 (35000HP ΔhgbA), an isogenic hgbA deletion mutant in strain 35000HP that does not express an HgbA protein (14).
Bacterial strains and plasmids used in this study
H. ducreyi bacterial strains were routinely maintained by subculture at 34°C with 5% CO2 on chocolate agar plates (CAP) supplemented with 1× GGC (0.1% glucose, 0.001% glutamine, 0.026% cysteine) (19) and 5% FetalPlex (Gemini, West Sacramento, CA). For the whole-cell enzyme-linked immunosorbent assay (ELISA) and Hgb-agarose binding assays, H. ducreyi strain FX547 expressing mutant hgbA genes in trans from plasmids was grown overnight in low-heme broth (GC broth supplemented with 5% FetalPlex and 1× GGC). For growth in media with Hgb as the sole source or iron or heme, strains were grown on GC agar plates and GC broth, both supplemented with 1× GGC and 100 μg/ml bovine Hgb (catalogue no. H2500; Sigma-Aldrich, St. Louis, MO). Escherichia coli strain XL10-Gold (Agilent Technologies, Santa Clara, CA), used for mutagenesis experiments described below, was grown on LB plates. Streptomycin (100 μg/ml) was added to the media when appropriate.
Protein sequences used for alignment.The following sequences were used for the alignment shown in Fig. 1: gi 33152990 (accession no. NP_874343; H. ducreyi HgbA), gi 13936951 (accession no. AF359442; Actinobacillus actinomycetemcomitans HgpA), gi 33340606 (accession no. AAQ14873; Pasteurella multocida HgbA), gi 257441707 (accession no. EEV168491; Campylobacter gracilis Hgb and Hgb-haptoglobin binding protein B), and gi 1399345 (accession no. AAB36696; Haemophilus influenzae HhuA).
Construction of in-frame hgbA deletion mutants.Plasmids expressing mutated hgbA (Table 1) were constructed using the QuikChange II XL site-directed mutagenesis kit as per the manufacturer's instructions (Agilent Technologies, Santa Clara, CA). Forward and reverse primers containing the appropriate mutations (see Table S1 in the supplemental material) were designed using the QuikChange primer design program (Agilent Technologies, Santa Clara, CA). Ten nanograms of pUNCH1261 (14) was used as the template in each mutagenesis reaction, except for the reactions to obtain pUNCH1733 and pUNCH1734, which used pUNCH1710 as the template. The PCR parameters used to generate the mutated hgbA plasmids were as follows: 1 cycle at 95°C for 1 min, 18 cycles at 95°C for 50 s, 60°C for 50 s, and 68°C for 7 min, and 1 cycle at 68°C for 7 min. PCR products were treated with 1 μl of DpnI (10 U/μl) and incubated at 37°C for 1 h. Two microliters of DpnI-treated DNA was used to transform E. coli XL10-Gold competent cells, plated on LB plates containing streptomycin, and incubated overnight at 37°C. Transformants were grown in 5 ml of LB broth containing streptomycin, and plasmid DNA was purified using the Qiaprep spin miniprep kit (Qiagen, Valencia, CA). Plasmid DNA was sequenced at the Automated DNA Sequencing Facility (University of North Carolina at Chapel Hill, Chapel Hill, NC) to determine the presence of the appropriate mutation. If the correct mutation was present, plasmid DNA was electroporated in H. ducreyi 35000HP ΔhgbA (FX547) using the Bio-Rad Gene Pulser (BTX, Holliston, MA). Electroporated H. ducreyi FX547 cells were plated on streptomycin-containing CAP and incubated for 48 h. Whole-cell lysates were prepared from transformants and examined using SDS-PAGE and Coomassie blue staining for expression of a protein of the correct size. Plasmid DNA from clones that contained the correct mutations was sequenced along the entire hgbA locus to ensure the absence of mutations other than the intended one.
Whole-cell binding ELISA.Two hundred microliters of a bacterial suspension (optical density at 600 nm [OD600] of 0.2) and 50 μl of primary antibody (anti-nHgbA pig antisera diluted in 0.25% Tween 20–GC broth [blocking buffer] at a final dilution of 1:2,000) were added to a MultiScreen HTS filter plate (Millipore, Bellerica, MA) and incubated with mixing at room temperature for 90 min. Wells were washed with 150 to 200 μl of 0.1% Tween 20–GC broth (wash buffer). One hundred microliters of a secondary antibody (horseradish peroxidase [HRP]-conjugated anti-pig IgG at 1:5,000 [Sigma-Aldrich, St. Louis, MO]) diluted in blocking buffer was added to the plate and incubated at room temperature for 60 min. Wells were subsequently washed four times, and 100 μl of chemiluminescence substrate (Fisher Thermo Scientific, Rockford, IL) was added. The plate was read using the Wallac 1420 Victor2 plate reader (PerkinElmer, Waltham, MA) under the luminescence setting.
SDS-PAGE.Total cellular proteins prepared from an H. ducreyi suspension (∼1 × 107 CFU) were subjected to a 4-to-12% gradient gel (150 V) and stained with Coomassie blue. For the Hgb binding assay (described below), three different samples were subjected to SDS-PAGE (4-to-12% gradient gel) followed by Coomassie staining: (i) starting material; (ii) unbound (material that did not bind to the Hgb-agarose beads), and (iii) boiled beads. Only the HgbA bands from the boiled-bead sample were subjected to densitometric analysis, which are shown in composite gels in the Results section.
Hgb-agarose binding assay.Bovine Hgb was coupled to Affigel-10 (Bio-Rad, Hercules, CA) as per the manufacturer's instructions. The Hgb-agarose was washed and mock eluted before use with elution buffer as previously described (11). Bacteria expressing mutant hgbA in FX547 grown in low-heme GC broth (3 ml at an OD600 of 1.0) were pelleted, suspended in 2% Zwittergent 3-14 (ZW) (Calbiochem, La Jolla, CA) in TEN (20 mM Tris-HCl, 5 mM EDTA, 150 mM NaCl [pH 8]), and incubated at 37°C for 1 h with mixing. Solubilized cells were centrifuged, and the concentration of ZW in supernatants was reduced to 0.5% with the addition of TEN buffer. Twenty microliters of Hgb-agarose was added to ZW-solubilized material and incubated at 4°C overnight with gentle shaking. The Hgb-agarose was washed several times with 0.5% n-octyl-β-d-glucopyranoside (OG) (Affymetrix, Maumee, OH) in TEN and then resuspended in 50 μl of 1× Laemmli loading buffer for analysis by SDS-PAGE and Coomassie staining.
Growth phenotype assay in Hgb-only media. (i) GC agar plates.H. ducreyi FX547 cells expressing mutant hgbA were initially grown on CAP overnight (from frozen stock) and then passaged 3 consecutive days onto GC plates containing bovine Hgb. The presence or absence of bacterial growth on day 3 was recorded (Table 2).
Summary of results from Hgb binding and growth assaysa
(ii) Hgb-only GC broth.H. ducreyi FX547 cells expressing mutant hgbA were initially grown overnight on CAP (from frozen stock) and then subsequently subcultured once onto GC plates containing bovine Hgb. A bacterial suspension was prepared from this subculture in GC broth (OD600 of 0.05 in 50 ml at time 0 h) containing 1× GGC and bovine Hgb. Every 2 h over a period of 12 h, the OD600 of the liquid culture was measured using the Victor2 plate reader. The OD600 of the culture was plotted against time using Prism software (version 5.0d; GraphPad Software, Inc., La Jolla, CA).
Densitometric and statistical analyses.Densitometry was performed on the HgbA band from total cellular proteins and Hgb binding assays using the AlphaView program (version 3.2.4.0; Protein Simple, Santa Clara, CA). Analysis of densitometry data was accomplished with the VassarStats website (http://vassarstats.net/) using a single-sample t test. To ensure that the relationship between the amount of HgbA present in the gels and the number of pixels obtained from analysis of the HgbA bands was linear, a standard curve of purified nHgbA subjected to SDS-PAGE and Coomassie blue staining was analyzed by densitometry. The results from this analysis, plotted in a graph using amounts of HgbA and pixel numbers as the x and y axes, respectively, were subjected to linear regression using Prism. The r2 of this curve was 0.9975, indicating a linear relationship between pixel number and amount of loaded HgbA. Results from the growth curves were first analyzed using Sigma Stat (version 3.5; Systat Software, San Jose, CA) to calculate the area under the curve (AUC) for each of the growth curves performed in triplicate. The mean ± standard deviation of the AUC for each hgbA mutant was compared to that of wild-type (WT) HgbA using a nonpaired, 2-tailed, equal variance t test (VassarStats). A P value of ≤0.05 was considered statistically significant.
RESULTS
The Hgb receptor of H. ducreyi HgbA contains conserved amino acids in loops 5 and 7.We previously demonstrated that loops 5 and 7 of the H. ducreyi Hgb receptor HgbA are required for binding Hgb, even though all 11 loops are necessary for utilization of Hgb as a source of iron/heme (14). To identify conserved motifs or amino acids within the loops that may control Hgb binding and/or heme transport and to define target epitopes for vaccine development, we aligned the predicted amino acid sequences of HgbA loops 5 and 7 with homologous loops of other Hgb binding proteins. We detected conserved amino acids at both N and C termini of loop 5, proposed to lie closest to the membrane in nHgbA (Fig. 1 and 2). We also identified areas of conservation in the N-terminal and central sections of loop 7 (Fig. 2), including the FRAP motif, which has previously been used to identify Hgb binding proteins, and a histidine residue shown to be important in Hgb binding by other Gram-negative bacteria (15, 20, 21).
Multiple-sequence alignment of a subset of Hgb-binding proteins from selected Gram-negative bacteria. Shown is an alignment of predicted amino acid sequences of loops 5 (top) and 7 (bottom) from selected Hgb receptors of Gram-negative bacteria, including H. ducreyi HgbA (14). Double-underlined regions are those that have been previously identified as critical for Hgb binding. Amino acids are color coded by functional group if they are identical in at least three of the five proteins in the alignment (http://bioinformatics.oxfordjournals.org/content/17/4/377.long). Boxes indicate residues that are identical among the loops and targeted for additional study. Hd, H. ducreyi; Aa, Actinobacillus actinomycetemcomitans; Pm, Pasteurella multocida; Cg, Campylobacter gracilis; Hi, Haemophilus influenzae.
To determine if any of the above-identified amino acids are essential for binding Hgb and/or acquisition of heme from Hgb, we deleted conserved regions of loop 5 (GYEAYNRQWWA) and loop 7 (FRAP). The loop 5 region was broken into two roughly equivalent regions (GYEAY and NRQWWA) to reduce structural perturbations in the mutant protein that might have resulted from large-scale alterations. For deletions that showed an effect on Hgb binding or heme acquisition, alanine substitution was performed on multiple residues to ensure that the phenotype exhibited by these mutant HgbA proteins is due to the absence of the motif and not to a change in the three-dimensional structure of HgbA induced by the deletion. Deletions and alanine substitutions of these motifs were made in the hgbA open reading frame (ORF) expressed from a plasmid in trans in the H. ducreyi isogenic hgbA mutant strain FX547 (35000HP ΔhgbA) (14). Binding to Hgb-agarose and growth in media with Hgb as the only source of iron/heme by mutant HgbA were compared to those of strain FX547 expressing an empty plasmid (pLSKS [negative control]) or a plasmid containing wild-type (WT) hgbA (positive control).
The entire loop 5 GYEAYNRQWWA sequence, but not FRAP in loop 7, is involved in Hgb binding.Deletion of the amino acid sequence GYEAY in loop 5 significantly reduced Hgb binding compared to that of WT HgbA (Fig. 3A). Although this deletion also reduced HgbA surface expression, the reduction in Hgb binding was greater than the reduction in surface localization, suggesting that the portion of the mutant protein that localized to the bacterial surface was indeed defective in Hgb binding (Fig. 3A and Table 2; see Fig. S1A in the supplemental material). Furthermore, Hgb binding to a mutant HgbA containing polyalanine in place of the GYEAY sequence, which does not exhibit reduced surface expression, was also significantly reduced compared to that in WT HgbA (Fig. 3A; see Fig. S1A). Alanine scanning mutagenesis confirmed the role of the entire GYEAY sequence in Hgb binding since all four alanine replacements resulted in partial but significant reduction in Hgb-agarose binding (Fig. 3B).
Deletion or polyalanine substitution of the GYEAYNRQWWA motif abrogates Hgb binding by HgbA. Hgb-agarose binding was measured in mutant HgbA with deletions or polyalanine substitution in the GYEAY (A and B) and NRQWWA (A and C) motifs of HgbA. For each panel, a representative gel from at least 3 independent Hgb-agarose binding experiment with the indicated mutant HgbA proteins is shown on top; on the bottom, the mean (± standard error) densitometry of the HgbA band from all Hgb binding experiments is represented as % Hgb binding of WT HgbA, defined as 100% binding. *, P < 0.05. (See Table 2 for P values.)
We next tested the contribution of the adjacent loop 5 sequence NRQWWA in Hgb binding to HgbA. Deletion of these residues reduced Hgb binding by 86% (Fig. 3A). Substitution of NRQWWA with polyalanine also significantly reduced Hgb binding by 64% (Fig. 3A and Table 2). As with GYEAY, most of the NRQWWA sequence appears to be important in Hgb binding to HgbA since scanning alanine mutagenesis showed that four out of five substitutions significantly altered Hgb binding (Fig. 3C). Of note, the role of the asparagine in Hgb binding (N535A) may not be significant because this reduction may be fully accounted for by reduced expression at the cell surface (see Fig. S1B in the supplemental material).
We previously showed that deletion of 16 of the 36 amino acids in loop 7 significantly reduces Hgb binding (14). Others have identified two motifs in loop 7 required for interaction with Hgb (21). Deletion of one of these motifs, the FRAP domain, resulted in a 90% decrease in Hgb binding; however, polyalanine substitution of FRAP showed no significant change in Hgb binding (Fig. 3A and Table 2). Furthermore, single-alanine scanning of FRAP did not affect Hgb binding either (Table 2). These data indicate that deletion of FRAP, and not the FRAP motif itself, likely caused structural changes in HgbA, which affected its ability to interact with Hgb. Taken together, these data suggest that the entire GYEAYNRQWWA region in loop 5 of HgbA, but not the FRAP motif in loop 7, directly contributes to Hgb binding to the H. ducreyi Hgb receptor.
Polylanine substitution of NRQWWA and FRAP in HgbA caused reduced growth of H. ducreyi in Hgb-only media.We showed above that many residues of loop 5 are involved in Hgb binding by HgbA; however, Hgb binding is only the first step in the complex process of heme acquisition and utilization. Heme must first be extracted from Hgb and then transported into the cell for H. ducreyi survival, using Hgb as the sole source of heme and iron. To determine if heme acquisition by HgbA also involves the three motifs studied above, growth of FX547-expressing mutant hgbA was measured in liquid cultures in which Hgb is the sole source of iron/heme. Since deletion or polyalanine substitution of the GYEAY region in loop 5 significantly reduced Hgb binding (Fig. 3), we expected both mutations to also affect heme acquisition. Although the GYEAY deletion severely diminished growth of H. ducreyi in Hgb-only media, alanine substitution of the entire domain did not have any effect on growth (Fig. 4 and Table 2). By abrogating Hgb binding, we assumed the GYEAY deletion would also eliminate heme utilization, which is what was occurred (Fig. 4); however, residual binding of Hgb to the bacterial surface in the polyalanine GYEAY substitution, which also reduced Hgb binding under these conditions, was apparently sufficient to support H. ducreyi growth.
Deletion or polyalanine substitution of the NRQWWA and FRAP motifs reduces growth rates in Hgb-only media. Growth in Hgb-only media was measured in strain FX547 expressing mutant HgbA with deletions or polyalanine substitution in the GYEAY, NRQWWA, and FRAP motifs of HgbA. Shown are growth curves (mean ± standard error) of at least 3 independent experiments using the indicated mutant HgbA in media containing Hgb as the sole source of iron/heme. Residual growth of the hgbA mutant strain FX547 was subtracted from the data. *, P < 0.05. (See Table 2 for P values.)
Deletion or substitution of the NRQWWA motif of loop 5 from HgbA significantly reduced Hgb binding by HgbA and as expected impaired the growth of H. ducreyi strain 35000HP in Hgb-only media (Fig. 4 and Table 2). This deficiency in utilization of heme by the NRQWWA mutants is not entirely due to a reduction in Hgb binding since these mutants exhibited 20 to 40% of WT Hgb binding, but there was no growth in Hgb-only media (Fig. 3A and Fig. 4).
Although frequently used to identify TBDT likely to be Hgb binding proteins, the FRAP motif was not required for Hgb binding to HgbA (Fig. 3A). We therefore hypothesized that FRAP may be required for heme utilization. Consistent with this hypothesis, deletion or complete replacement of the FRAP domain by polyalanines showed no growth on Hgb alone, indicating that FRAP is required for heme utilization (Fig. 4). These data support a role for the NRQWWA and FRAP motifs in heme transport by HgbA.
The phenylalanine residue of FRAP is essential for heme uptake.Having shown that NRQWWA from loop 5 and FRAP from loop 7 are involved in heme acquisition by HgbA, we proceeded to determine the requirement for individual residues in those motifs. All mutated HgbA proteins with a single-alanine substitution in each residue of the NRQWWA motif grew as well as 35000HP ΔhgbA expressing WT HgbA (Fig. 5A), even though Hgb-agarose binding by several of these mutant proteins, especially the ones expressing mutations in the tryptophan residues, was significantly diminished (Fig. 3C). Taken together, these data suggest that multiple amino acids NRQWWA are involved in heme transport in HgbA and that no single substitution in this motif is enough to prevent HgbA-mediated heme utilization.
Single alanine substitution of the phenylalanine of loop 5 FRAP reduces growth rates in media with Hgb as the sole source of iron/heme. Shown are the growth curves of mutant HgbA proteins with single-alanine substitutions in the NRQWWA (A) and FRAP (B) motifs. Residual growth of the hgbA mutant strain FX547 was subtracted from the data. Shown are means ± standard errors from at least 3 independent growth experiments. (See Table 2 for P values.)
When the arginine and proline residues of the FRAP motif were mutated to alanine, we did not observe a growth defect on Hgb-only media (Fig. 5B). Conversely, the F692A mutant HgbA (HgbA-F692A) showed significantly reduced growth in Hgb-only media in the context of WT Hgb binding. Taken together, these data suggest that only the phenylalanine residue of the FRAP motif of loop 7 is involved in heme acquisition by H. ducreyi HgbA.
Select histidines in the plug domain and in loop 3 are important for Hgb binding to HgbA, but none is essential for heme acquisition.In other bacterial systems, histidines have been shown to be important residues in the interaction of bacterial Hgb receptors with heme (15). We therefore replaced selected histidines in loops 2, 3, and 7 and the plug region of HgbA with alanine. These residues were chosen because of their potential to interact either with Hgb at the surface of H. ducreyi or with heme during transit through the barrel of HgbA, as suggested by the location of these histidines in a three-dimensional model of HgbA (Fig. 1). Of the 2 histidines replaced in the plug domain of HgbA, only the one at position 44 (H44A) significantly reduced Hgb binding to HgbA (Fig. 6A and Table 2). Single-alanine substitution of the histidine residue in loop 2 (H227A) did not affect Hgb binding, but changing the one in loop 3 (H298A) significantly reduced Hgb binding by 41% (Fig. 6A). These findings indicate that histidines at positions 44 and 298 of the H. ducreyi HgbA protein interact with Hgb.
Alanine substitution of selected histidine residues of HgbA reduced Hgb binding, but none affect growth rates in media with Hgb as the sole source of iron/heme. Shown are Hgb binding assays (A) and growth curves (B) of HgbA encompassing substitutions in selected histidines. (A) Composite gel of one of at least 3 independent Hgb-agarose binding experiments. Below is a graph representing % Hgb binding (mean ± standard error) by the mutants compared to WT HgbA (100% binding). (B) Shown are means ± standard errors from at least 3 independent growth curves, from which the growth rate of the hgbA mutant strain FX547 expressing an empty vector was subtracted. *, P < 0.05 compared to WT HgbA. (See Table 2 for P values.)
All of the single-histidine mutants exhibited normal growth in Hgb-only media (Fig. 6B), even those with alanine replacements in histidines at positions 44 (plug) and 298 (loop 3), which had reduced Hgb binding (Fig. 6A). Simultaneous replacement of 2 HgbA histidines (H227 and H706A or H298A and H796A) significantly enhanced growth in medium containing Hgb as the sole source of iron/heme (Fig. 6B). Therefore, although some of the selected histidine residues are involved in Hgb binding, none appears to play a role in heme acquisition by H. ducreyi HgbA.
DISCUSSION
As iron is vital for most bacteria, many express several different receptors to acquire this essential nutrient from different molecules present in their milieu. Gram-negative bacteria frequently acquire iron in the form of heme from abundant hemoproteins, such as Hgb (22). The Gram-negative bacterium H. ducreyi is unable to synthesize heme de novo and thus acquires iron in the form of heme from the human host (23). Acquisition of heme from Hgb at the bacterial outer membrane and subsequent extraction and transport into the periplasm are accomplished by TBDT (24). Of the three TBDT expressed by H. ducreyi, only the Hgb receptor HgbA is required for survival in the human experimental model of chancroid (10). All 11 loops of HgbA, but not the plug domain, are required for utilization of Hgb by H. ducreyi as a source of heme; however, only 2 out of 11 putatively surface-exposed loops (5 and 7) are required for binding of Hgb at the surface of this bacterial pathogen (14). This is different from the homologous Neisseria meningitidis HmbR receptor: mutations in loops 2 and 3 affected Hgb binding, while loops 6 and 7 targeted Hgb utilization (20). Furthermore, the plug domain of HmbR is also involved in Hgb utilization, but not Hgb binding (20). Neither loop 5 nor loop 8 of HmbR is required for Hgb binding or heme transport.
There are at least four steps involved in using Hgb as a source of iron/heme by Gram-negative bacteria: (i) the initial interaction of Hgb with the receptor at the surface of the bacterium, (ii) extraction or release of heme from Hgb, (iii) binding of heme to the Hgb receptor, and (iv) transport of heme through the pore of the Hgb receptor to the periplasm (25). Presumably, Hgb binding and heme transport each require amino acids from different loops of the Hgb receptor to interact with Hgb and heme. In this work, we sought to identify conserved motifs and residues in loops 5 and 7 of the H. ducreyi Hgb receptor HgbA involved in Hgb binding at the bacterial surface and/or utilization of Hgb as a source of iron/heme. Our analysis revealed that loop 5 of HgbA possessed novel conserved motifs at the N-terminal region of the loop, an area closest to the bacterial membrane. These conserved motifs are also present in several Hgb receptors of other Gram-negative bacteria (Fig. 2). Loop 7 also exhibited conservation in several areas, including the N-terminal and middle sections of the loop (Fig. 2). Such conserved residues are also found in the well-studied Hgb receptors from N. meningitidis and Porphyromonas gingivalis, including the FRAP and NXXL domains, and a conserved histidine residue (20, 21).
A conserved region of loop 5 is involved in Hgb binding by H. ducreyi HgbA.Data presented here show that the conserved region GYEAYNRQWWA in loop 5 is important for Hgb binding to HgbA (Fig. 3A and B). Furthermore, single mutation in the glycine of the GYEAY motif or alanine replacement of the complete NRQWWA motif of loop 5, significantly impacted heme acquisition by HgbA, even though polyalanine substitution of GYEAY had no effect on Hgb binding to HgbA (Table 2). If the glycine from loop 5 was truly important for heme transport in H. ducreyi, alanine substitution of the complete GYEAY region would also affect growth in Hgb-only media, but that was not the case. These findings suggest that alanine replacement of this particular glycine changes the tertiary structure of HgbA, which is in turn indirectly affects heme acquisition by the Hgb receptor. We postulate that the effect of the alanine substitution on heme transport is due to reduced conformational flexibility of the mutated hgbA. Because glycine has no side chain, it has greater conformational flexibility in the backbone bonds than another amino acid would in that position. Replacement of it with an alanine, which has a methyl group as a side chain instead of a single hydrogen atom, may therefore limit the potential movement or specific conformation of the mutant HgbA protein and impact heme acquisition. Crystallographic studies using these mutants will be pursued to determine the role of glycine in the interaction of HgbA with heme.
The data presented above indicate that NRQWWA from loop 5 is required for heme transport in HgbA (Fig. 4). The involvement of the NRQWWA motif in acquisition of heme by HgbA may be supported by homology studies. Indeed, upon BLAST search, this motif was found in catalases/peroxidases of several species of bacteria, including Bacillus, Treponema, and Streptomyces, to name a few. Since catalases are known to use heme as a cofactor, this motif may be involved in the interaction of those proteins with heme.
The phenylalanine residue of loop 7 is involved in heme acquisition by HgbA.Our results show that the phenylalanine residue of the FRAP motif had a role in heme transport by HgbA (Fig. 5). The role of phenylalanine in heme acquisition was supported by three observations: (i) alanine substitution of the entire FRAP domain completely abrogated growth in Hgb-only media (Fig. 4), (ii) replacement of all residues except the phenylalanine supported normal growth (Fig. 5), and (iii) single-alanine exchange of the phenylalanine (F692A) without replacing other residues in the motif also significantly diminished growth to the level of the empty vector (Fig. 5). None of the other amino acids of the FRAP motif was involved in either Hgb binding or heme transport by HgbA (Fig. 3A and 5B). These results are divergent from those obtained with the HmuR receptor in P. gingivalis. Alanine substitution of the conserved YRAP domain to YAAA affected heme transport by this Hgb receptor, suggesting that the arginine and/or proline residues of this motif are required for heme acquisition by HmuR (21).
Select histidines of the H. ducreyi Hgb receptor HgbA are involved in Hgb binding, but none is required for heme acquisition.There are multiple studies demonstrating the importance of histidines as ligands of Hgb or heme in bacterial Hgb receptors (15, 21, 26). Selected histidines of HgbA were therefore targeted for alanine substitution since histidines in similar positions in HemR, HasR, and HmuR have been shown to be important in heme transport and Hgb binding. Furthermore, previous findings and our homology model of HgbA (Fig. 1) suggest that some histidines may work in concert to bind Hgb or transport heme. Surprisingly, our data show that the conserved histidine in loop 7 (H706A) was involved in neither Hgb binding nor heme transport by HgbA (Fig. 6). The corresponding histidines in other Hgb receptors have been involved in either Hgb binding or heme acquisition, but not both. In the Yersinia enterocolitica HemR and the S. dysenteriae ShuA receptors, the loop 7 histidine residues at positions 461 and 420, respectively, are only required for heme acquisition (15, 26). On the other hand, the histidine at position 434 of the P. gingivalis HmuR protein mediated Hgb binding, but was not involved in heme transport (21). Taken together, these data suggest that the conserved histidine from loop 7 may not always be involved in Hgb binding or heme transport in all Gram-negative Hgb receptors.
Our findings demonstrate that only one of the two histidines found in the plug domain (H44) was involved in Hgb binding by HgbA (Table 2). Furthermore, out of two other histidines targeted for mutations in HgbA, only the one in loop 3 (H298A) impacted Hgb binding by the H. ducreyi Hgb receptor. Even though H44 and H298 were shown to be involved in Hgb binding, none of the five histidines selected for site-directed mutagenesis had a role in heme transport. Moreover, alanine replacement of two histidines at one time even increased growth of bacteria in medium with Hgb as the sole source of Hgb (Fig. 6B), suggesting that these substitutions enhanced heme transport in H. ducreyi expressing these mutated HgbA proteins. These findings suggest that contrary to other Hgb receptors, either other histidines of HgbA not targeted in this study are involved in heme acquisition, or other amino acids such as tyrosines play this important role in HgbA.
In conclusion, our studies propose the critical involvement of novel motifs and amino acids in Hgb binding and heme acquisition by a bacterial Hgb receptor. These results therefore identify regions that differentiate Hgb binding and heme acquisition in Hgb receptors. Our findings also show that the histidine from loop 7 of Hgb receptors is not always necessary for heme transport in Gram-negative bacteria. Determination of the crystal structure of HgbA will help confirm these results and determine the mode of action of the H. ducreyi Hgb receptor, furthering our knowledge of heme transport in Gram-negative bacteria.
ACKNOWLEDGMENTS
This work was supported by public service grant AI088081 to E. Collins, funded by the National Institutes of Health.
We are grateful to Marcia Hobbs for help with statistical analyses and careful review of the manuscript.
FOOTNOTES
- Received 19 February 2013.
- Accepted 3 May 2013.
- Accepted manuscript posted online 10 May 2013.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00199-13.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
