INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Haemophilus ducreyi is the etiologic agent of chancroid, a genital ulcer disease transmitted by sexual contact (reviewed in references 2 and 66). Prevalent in Africa, Asia, and certain developing nations, chancroid has recently gained importance as an independent risk factor for the heterosexual transmission of human immunodeficiency (67). There is accumulating evidence suggesting that the rapid spread of human immunodeficiency virus type 1 in eastern and southern Africa has been due, at least in part, to the presence of chancroid and other genital ulcer diseases (13, 51). At present no vaccine or practical field diagnostic test exists for H. ducreyi.

H. ducreyi is a fastidious slow-growing, gram-negative rod with an optimum growth temperature of 33°C. An important feature of H. ducreyi is its requirement for heme, which it is unable to synthesize. This obligate requirement for heme by H. ducreyi is in contrast with H. influenzae, which is able to use its enzyme ferrochelatase to ferrate protoporphyrin IX, bypassing the need for heme (24, 53). In H. ducreyi, heme is acquired from its only known host, humans, in several possible forms, including hemoglobin, hemoglobin complexed to haptoglobin, catalase, free hemin, or heme bound to carrier proteins such as albumin (3, 39, 64). Some or all of these heme sources could be released from host cells by the action of the H. ducreyi hemolysin (45, 46, 63) or cytotoxin (16, 37, 49).

Host heme compounds represent important sources of iron (44). Host iron is sequestered by several mechanisms, and invading bacteria must gain access to these iron sources to survive and initiate disease. Some bacteria utilize iron-scavenging siderophore systems; others directly bind host iron or heme/iron-containing compounds such as transferrin, lactoferrin, heme, or hemoglobin (19, 44). No siderophores have been found in H. ducreyi (39). Since H. ducreyi requires both heme and iron, differentiation between the requirement for both of these can be problematic.

Siderophore receptors and most receptors for host iron-binding compounds belong to a family of outer membrane receptors designated TonB dependent because they require the cytoplasmic membrane protein TonB for activity (7, 11, 57). TonB-dependent receptors are related structurally and functionally; where studied, they have been shown to form specific pores for the entry of low-molecular-weight molecules (12, 40). TonB-dependent receptors for host-iron complexes such as transferrin or lactoferrin have been shown to internalize the iron only (41), whereas receptors for siderophores internalize the iron-siderophore complex (43). TonB-dependent receptors are also related at the amino acid level by virtue of several regions of distinct homology, including an N-terminal sequence termed the TonB Box (18, 28, 35, 54). In Escherichia coli, the TonB protein has been shown to directly interact with TonB-dependent receptors FepA (5, 55). TonB and its accessory proteins ExbB and ExbD comprise a system for the transfer of energy from the cytoplasmic membrane to outer membrane receptors (10). Null mutations in either tonB, exbB, or exbD can variably affect the ability of TonB-dependent receptors to internalize or utilize their cognate ligand (48). The effect of these null mutations can vary from partial to total inhibition of uptake of ligands.

Previously, we identified and purified an outer membrane hemoglobin receptor from H. ducreyi termed HgbA (22) and showed that the HgbA protein is functionally and immunologically conserved between geographically diverse isolates. Further analysis was performed by using molecular techniques (23, 56). The hgbA gene was cloned and sequenced; its deduced amino acid sequence is similar to the TonB-dependent family of outer membrane receptors (23, 56). An hgbA isogenic mutant of H. ducreyi cannot bind or utilize hemoglobin as a source of heme but can utilize free hemin, implying that the utilization of hemin does not require HgbA. It has been shown by Stevens et al. (56) that an isogenic mutant of hgbA (hup) expressed reduced virulence in an animal model of H. ducreyi infection.

In the course of these functional studies, we found that an E. coli hemA mutant expressing hgbA binds hemoglobin yet does not grow on hemoglobin as a porphyrin source (23). This result suggested that additional components are necessary for utilization of hemoglobin after the binding step. The objective of the present study was to identify the additional components necessary for removal of heme from hemoglobin and for the transport of heme across the outer membrane.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and media. Bacterial strains used in this study are described in Table 1. For routine growth, H. ducreyi was maintained on chocolate agar plates prepared by following gonococcal medium base (GCB) instructions (Difco). No fetal bovine serum was utilized in chocolate agar plates used for routine passage. Plates containing the various heme sources were prepared by using GCB and IsoVitaeleX supplements (Difco) (termed GCB-I). Human hemoglobin (H7379; Sigma Chemical, St. Louis, Mo.) was dissolved in phosphate-buffered saline at a concentration of 10 mg/ml and rocked at room temperature for 2 h or overnight at 4°C prior to filter sterilization. Heme (Sigma) was made 10 mg/ml in 0.1 N NaOH and was used without sterilization. To bind heme to either serum or serum proteins, various amounts of heme were added to the protein source and allowed to bind for at least 10 min at room temperature prior to incorporation into GCB-I plates. Human serum albumin (HSA; Sigma) was saturated to 50% by the addition of 1/2 mol of hemin per mol of albumin. We assumed that the albumin contained no hemin to start with and that it contained a single hemin binding site. Catalase was obtained from Sigma (C-100, lot no. 106H7055; C-6655, lot no. 23H70355). Five micrograms of hemin per milliliter corresponds to 7.6 µM hemin.

Growth conditions for testing phenotypes. For inoculation of plates for testing phenotypes, H. ducreyi strains were heme starved by growth anaerobically on GCB-I plates without a heme source (56). This method of inoculation reduced the possibility of heme carryover or differences in intracellular heme stores among the various strains. After H. ducreyi was grown anaerobically on GCB-I plates, inocula were standardized spectrophotometrically and 10-fold dilutions were made for each strain in a microtiter tray. A Steers replicator was used to inoculate agar medium containing dilutions of each heme source. Thus, the inoculum and heme sources were present in a checkerboard titration. Plates were incubated at 35°C in 5% CO₂.

Outer membrane isolation and analysis. Large-scale cultures of H. ducreyi were performed in Fernbach flasks with 1 liter of GCB-I broth containing 5% fetal calf serum (FCS) and heme at 2 µg/ml (heme limiting) or 50 µg/ml (heme replete) (22). It has been shown by several investigators that H. ducreyi and H. influenzae strains require less hemin when grown in liquid media compared to solid medium (3, 22, 60, 64). Therefore, these concentrations should not be compared with those used in solid medium used for testing phenotypes. Flasks were inoculated to a starting density of 2.4 × 10⁶ CFU/ml with growth harvested from chocolate agar plates and were incubated 24 h with shaking in a 5% CO₂ atmosphere at 35°C. The purity of each culture was routinely verified by streaking onto agar plates which did (chocolate agar) and did not support the growth of H. ducreyi but which supported the growth of most other bacteria (GCB-I, without hemin). This ensured that minor contamination could be detected and that the novel (previously undescribed) proteins detected (see Fig. 5A) were not the result of contamination. Outer membranes were harvested as previously described (22) except that lysozyme was added to the harvested cells prior to the French press step (200 µg/ml, 10 min, 4°C) and two solubilizations were performed with Sarkosyl rather than one. Protein concentrations were determined by using a bicinchoninic acid kit from Pierce (Rockford, Ill.). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting were performed as previously described (22) except for the antiserum to HgbA. Anti-HgbA was produced by immunizing mice with 25 µg of purified HgbA (22) three times at 2-week intervals. Complete Freund's adjuvant was used for the first immunization, and incomplete Freund's adjuvant was used for the last two immunizations.

Chemicals and reagents. All chemicals and reagents, unless otherwise noted, were from Sigma.

DNA manipulations. Standard recombinant DNA methods were used as described in reference 52 or following manufacturers' instructions.

DNA sequencing and analysis. DNA sequence analysis was performed at the University of North Carolina at Chapel Hill Automated Sequencing Facility, using Taq terminator chemistry. Both strands were completely sequenced and assembled by using AssemblyLIGN software (IBI). Amino acid alignments were done with the programs PILEUP and PRETTYBOX (25, 50). The default threshold parameters for PRETTYBOX were used: 1.5, identical (black boxes); 1.0, similar (dark grey boxes); 0.5, somewhat similar (light grey); 0.0, dissimilar (white boxes) (see Fig. 2). Table 2 was constructed by using data generated from the program OLDDISTANCES (25, 50).

Subcloning and construction of an H. ducreyi mutant. pUNCH 562 was subcloned by removal of two small HincII fragments by restriction and religation of the gel-purified 6.5-kb fragment to form pUNCH 563 (Fig. 1). pUNCH 563 was mutagenized by restriction with BbsI and MluI followed by Klenow enzyme treatment to create blunt ends (Fig. 1). A chloramphenicol-resistant (Cm^r) cassette was isolated from plasmid pNC40 (61) by digestion with BglII and end repair using Klenow enzyme. This cassette was ligated to digested pUNCH 563 to form pUNCH 568 (Fig. 1). The deleted DNA between the BbsI and MluI sites of pUNCH 563 contains sequences encoding the C terminus of ExbB, all of ExbD, and the N terminus of TonB. E. coli IR754(pUNCH 556) was electroporated with pUNCH 568, and transformants were selected on LB agar containing chloramphenicol (30 µg/ml), tetracycline (10 µg/ml), and ALA (25 µg/ml). To compare the phenotypes of E. coli IR754(pUNCH 556) containing pUNCH 563 and its mutagenized counterpart pUNCH 568, single colonies of each were streaked onto LB medium containing ampicillin, tetracycline, and either hemoglobin or ALA and on free hemin or 50% heme-saturated human albumin (Fig. 1 and 3).

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Restriction maps and phenotypes of relevant H. ducreyi exbB, exbD, and tonB clones. Not shown in pUNCH 568 are unique XbaI and XhoI vector sites used to release the insert prior to allelic exchange. E. coli IR754 (hemA tonB aroB) containing pUNCH 556 and the indicated plasmid was inoculated onto LB medium containing various sources of heme or ALA: hemoglobin (Hg), 100 µg/ml (6 µM hemin); heme, 50 µg/ml (77 µM) or 10 µg/ml (15 µM); heme-HSA (H-Alb), 3,300 µg/ml (50 µM), saturated at a 50% molar concentration with hemin, 16 µg/ml (25 µM); ALA, 25 µg/ml (4 µM). +, macroscopic growth at 48 h; , no macroscopic growth at 48 h.

Nucleotide sequence accession number. Relevant DNA sequence of the Ton system gene locus has been submitted to GenBank (accession no. AF001034).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning of the H. ducreyi exbB, exbD, and tonB loci. We found that cloned hgbA was able to confer upon E. coli EB53 (hemA aroB) (Table 1) the ability to bind human hemoglobin (23), but this strain was unable to utilize hemoglobin as a source of heme (data not shown). We reasoned that HgbA expressed in E. coli EB53 might not be energized by the E. coli Ton system, since E. coli and H. ducreyi are only distantly related (2). An experiment was designed to clone the Ton system of H. ducreyi which was similar to that used to clone the Ton system of meningococci (57). A partial Sau3AI library was transformed into E. coli IR754 (tonB hemA aroB) expressing hgbA from the compatible plasmid pUNCH 556 (Table 1) and plated on hemoglobin as the sole source of heme. If the hemoglobin receptor plus gene products from certain plasmids allowed the uptake of heme from hemoglobin, then the E. coli heme synthesis mutation would be bypassed and growth on hemoglobin plates would occur. Of approximately 5 × 10⁶ total transformants, 30 complementing clones were obtained on LB hemoglobin plates. To confirm that these clones required HgbA for growth on hemoglobin, plasmid DNA isolated from each was transformed into E. coli IR754 lacking pUNCH 556 (hgbA) and again plated onto hemoglobin agar plates. Thirteen of the 30 clones failed to grow without pUNCH 556, suggesting that the gene products expressed from those clones were insufficient to allow IR754 to utilize the heme from hemoglobin. This screen eliminated clones which allowed the heme from hemoglobin to leak in nonspecifically or directly complemented the hemA defect in IR754. Eleven of the 13 plasmids hybridized to the insert from pGJ300 containing the H. influenzae tonB operon (33, 34) (data not shown). The two plasmids which did not hybridize apparently suffered deletions during propagation since they were reduced to the size of the vector. Nine of 11 hybridizing plasmids contained a 6.5-kb hybridizing band; each conferred upon IR754 carrying pUNCH 556 (hgbA) the ability to grow on LB hemoglobin agar (data not shown). One plasmid, pUNCH 562, was chosen for further study (Fig. 1).

DNA and deduced amino acid sequence of the H. ducreyi Ton system gene cluster. The location of the TonB gene cluster was revealed by subcloning and deletion analysis as shown in Fig. 1. The relevant regions were sequenced, and the order of the genes was exbB exbD tonB. Sequences similar to 35 and 10 consensus sequences were found 114 and 82 nucleotides (nt) upstream of the exbB start codon and were separated by 16 nt. Putative ribosome-binding sites were found at 9, 11, and 11 nt upstream of the respective exbB, exbD, and tonB ATG start codons. Just downstream of the tonB open reading frame there was an AT-rich region with dyad symmetry consistent with a transcription terminator. Notably absent were sequences corresponding to promoters upstream of the exbD and tonB structural genes.

View larger version (89K): [in this window] [in a new window]   View larger version (68K): [in this window] [in a new window]   FIG. 2.   Comparison of H. ducreyi ExbB (A), ExbD (B), and TonB (C) proteins with selected Ton system homologs from other gram-negative bacteria, generated with the programs PILEUP and PRETTYBOX. Black boxes, identical residues; grey boxes, related conserved substitutions; white boxes, nonconservative substitutions. In panel A, the N-terminal 74 aa are not shown for Pseudomonas putida ExbB. Hd, H. ducreyi; Ph, P. haemolytica; Hi, H. influenzae; Nm, Neisseria meningitidis; Pp, Pseudomonas putida; Ec, E. coli.

Phenotypic characterization of the H. ducreyi Ton system. Subcloning and deletion analyses were used to define regions of the original plasmid pUNCH 562 required to express a hemoglobin-utilizing phenotype. E. coli IR754(pUNCH 556) containing subclone pUNCH 563 also demonstrated a hemoglobin-utilizing phenotype (Fig. 1 and 3A), but not in the absence of pUNCH 556 (hgbA) (Fig. 1). pUNCH 563 was mutagenized by deleting the DNA between the unique BbsI and MluI sites and replacing it with a Cm^r (23) cassette to form pUNCH 568 (Fig. 1). This deletion included all of exbD and portions of exbB and tonB. IR754(pUNCH 556) containing the mutagenized Ton system plasmid pUNCH 568 was unable to grow on hemoglobin (Fig. 1 and 3B).

View larger version (67K): [in this window] [in a new window]   FIG. 3.   Growth of recombinant Ton system clones on LB hemoglobin agar or ALA (LB-ALA) E. coli IR754 (hemA tonB aroB)(pUNCH 556 [hgbA]) containing either pUNCH 563 (ExbB+ ExbD+ TonB+) (A) or pUNCH 568 (ExbB ExbD TonB) (B) was streaked onto LB agar containing hemoglobin (100 µg/ml [5 µM hemin]) or ALA (25 µg/ml [4 µM]) with antibiotic selection and incubated at 37°C for 48 h.

View larger version (51K): [in this window] [in a new window]   FIG. 4.   Growth of H. ducreyi Ton system isogenic mutant on hemoglobin agar. H. ducreyi parent strain 35000 (wild type [wt]), FX504 (hgbA), and FX514 (exbB exbD tonB) were streaked onto GCB-I medium containing human hemoglobin (200 µg/ml [12 µM hemin]) or chocolate agar and incubated at 35°C for 48 h with 5% CO2.

Phenotype of the H. ducreyi Ton mutant on sources of heme other than hemoglobin. In other bacteria, Ton system mutations are impaired in the ability to use a variety of heme/iron compounds transported by TonB-dependent receptors; therefore, the growth phenotype of FX514 was examined on other heme-containing media that support the growth of H. ducreyi. H. ducreyi was first heme starved (56) and then serially diluted for use as the inoculum. There was no difference for either mutant on hemin or catalase relative to the parent (Table 3). All three strains grew with hemin at 50 µg/ml or catalase at 250 µg/ml but not lower amounts. None of the three strains grew on GCB-I agar containing 33 µM HSA (2 mg/ml) saturated with 50% hemin. Taken together, these data indicated that utilization of hemoglobin was TonB dependent whereas utilization of free hemin and catalase appeared to be TonB independent.

Novel proteins expressed by H. ducreyi Ton system mutant FX514. Previous work in E. coli demonstrated that Ton system mutants are relatively iron starved based on the observation that genes normally regulated by Fur are derepressed (48). SDS-PAGE analysis of outer membrane proteins from H. ducreyi Ton mutant FX514 demonstrated several heme-regulated gene products present in increased amounts (Fig. 5A). The major protein present in increased amounts in FX514 was HgbA, confirming that the inability to utilize hemoglobin in this mutant was not due to the lack of synthesis of HgbA. Three other outer membrane proteins, designated 1, 2 (protein 2 refers to the lower band of the doublet), and 3 in Fig. 5A, were also present in increased amounts in certain of the mutants under heme-regulated conditions. Protein 1 was present only in FX514 grown under heme-limiting conditions. Proteins 2 and 3, although present in strain 35000, were more highly expressed in the two mutants than in strain 35000 under heme limitation. In experiments not presented here, Western blotting of H. ducreyi using specific antibodies to proteins 2 and 3 demonstrate they were clearly heme regulated, albeit at lower amounts than HgbA (data not shown). Outer membranes prepared from FX514 which had been repaired by chromosomal transformation regained the protein expression pattern of parent strain 35000 (data not shown).

View larger version (39K): [in this window] [in a new window]   FIG. 5.   SDS-PAGE and Western blotting of outer membranes from H. ducreyi. The indicated strains were grown under heme-replete (+) or heme-limiting () conditions, and their outer membranes were isolated and subjected to SDS-PAGE (7.5% gel) and Coomassie staining (A) or Western blotting using anti-HgbA sera (B). Strains were H. ducreyi parent strain 35000 (wild type), FX504 (hgbA), and FX514 (exbB exbD tonB). Positions of molecular mass markers are shown at the left. HgbA, hemoglobin receptor protein; protein bands 1, 2, and 3, previously undescribed heme-regulated proteins expressed by Ton mutant FX514 (protein 2 refers to the lower band of the doublet). The Western blot (B) was purposely overloaded in order to detect potential proteolytic breakdown products of HgbA as a source of proteins 2 and 3 and other, smaller proteins.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Utilization of hemoglobin is TonB dependent. Several TonB-dependent receptors for heme or hemoglobin have been cloned from gram-negative pathogens, utilizing a strategy based on the complementation of a heme synthesis defect (hemA) or iron acquisition (aroB) defect in E. coli (20, 29-31, 42, 57-59, 62). This method relies on the principle that E. coli K-12 strains are unable to transport hemin or hemoglobin across the outer membrane. However, once hemin traverses the outer membrane, E. coli K-12 strains are able transport it across the cytoplasmic membrane, where it can be used to fulfill its heme or iron requirements. Using this method, we have also attempted to clone a heme receptor from H. ducreyi in E. coli EB53 (hemA aroB), but initial experiments have been unsuccessful. This failure, coupled with our inability to demonstrate a TonB-dependent hemoglobin utilizing phenotype in E. coli EB53 with cloned hgbA, prompted us to search for an explanation. It was found that the cloned H. ducreyi hgbA receptor required its homologous Ton system in E. coli for growth on hemoglobin. Furthermore, the Ton system mutant FX514 was unable to utilize hemoglobin at wild-type levels, unambiguously demonstrating that HgbA was TonB dependent.

Role of HgbA in heme utilization. In an experiment to examine the role of HgbA in heme acquisition, we tested whether cloned HgbA with its homologous Ton system could confer upon E. coli hemA tonB aroB strains the ability to grow on free hemin or hemin complexed to HSA. No growth was observed on HSA-heme agar. Growth on hemin agar was observed only at high concentrations (50 µg/ml) and growth was TonB independent, suggesting that HgbA alone does not function as a typical TonB-dependent receptor for free hemin or hemin complexed to HSA. This finding does not rule out the possibility that HgbA is involved in internalization of free hemin in H. ducreyi but may indicate that additional components are required. Alternatively, HgbA and other (heme/iron) receptors may form a complex, and the absence of HgbA may disturb the function of this complex. Similar nonspecific defects in the uptake of iron compounds have been found for TonB-dependent receptor mutants in other bacteria (1, 6).

Utilization of heme or catalase in H. ducreyi is TonB independent. Although neither H. ducreyi FX504 nor FX514 grew on hemoglobin as a sole source of heme, both mutants exhibited no such growth defect on hemin relative to parent 35000 when prestarved for heme by growth anaerobically in the absence of heme. Further, all three strains required relatively high concentrations of free hemin (50 µg/ml) for growth on plates. To date, bacteria containing documented heme receptors require much less heme (less than 10 µg/ml) than H. ducreyi to fulfill their iron requirement. The requirement of the wild-type H. ducreyi strain for high concentrations of heme is reminiscent of certain mutants of H. influenzae. H. influenzae tonB (34) or hxuC (17) mutants require high levels of heme (50 µg/ml), whereas wild-type strains require only 10 µg/ml or less (0.1 µg/ml) (60). Most previous studies have either shown the TonB dependence or implied the TonB dependence of heme uptake (20, 29-31, 42, 57-59, 62). However, in contrast to these previous studies, in Neisseria spp., heme utilization has very recently been shown to be TonB independent (7, 57). No heme receptor has been found in Neisseria spp.

The H. ducreyi Ton system mutant expresses novel heme-regulated outer membrane proteins. Regulation of most iron-repressed genes in E. coli is under the control of the global negative repressor Fur (4). In E. coli, derepression of the Fur regulon in tonB mutants is believed to be due the inability to transport iron by means of the various TonB-dependent receptors, resulting in low intracellular iron levels (47). H. ducreyi contains a functional Fur protein (14); however, the H. ducreyi genes regulated by Fur have not yet been identified. In H. ducreyi, expression of hgbA and perhaps hemolytic activity (9, 23) are regulated by the levels of heme in the medium. Since heme and hemoglobin contain iron, it is possible that these heme sources also serve as a source of intracellular iron, thereby indirectly affecting expression of the H. ducreyi Fur regulon. H. ducreyi Ton system mutant FX514 demonstrated increased expression of HgbA as well as several previously undescribed outer membrane proteins, which appeared to be regulated by the levels of heme in the medium. The possibility that these novel proteins are involved in heme/iron acquisition is consistent with observations made in other pathogens. Further work is needed in this area to understand the repertoire of receptors for heme/iron compounds in H. ducreyi and their regulation.

Structure and arrangement of the H. ducreyi Ton system. The arrangement of genes in the H. ducreyi tonB cluster was similar to that in certain other gram-negative bacteria described elsewhere (8, 33, 34), with the order exbB exbD tonB. The arrangement, the DNA sequence, and the proximity of the structural genes suggested that only one promoter was responsible for expression of all three genes. We speculate that these genes may be transcribed as a multicistronic message; however, this remains to be proven experimentally.