INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Gonorrhea, a sexually transmitted disease caused by the human pathogen Neisseria gonorrhoeae, primarily infects columnar epithelial mucosal surfaces including the urethra, endocervix, rectum, and oropharynx. It is a major cause of pelvic inflammatory disease in women and acts as a cofactor for the spread of human immunodeficiency virus (23, 41). Gonococci can grow within epithelial cells and are capable of invading the epithelial barrier and disseminating via the bloodstream to multiple sites, including the cerebrospinal fluid, skin, and joints (32). Therefore, N. gonorrhoeae must be able to obtain nutrients from a diverse range of environments.

Nearly all bacteria require iron as a nutrient for growth, and the ability of pathogens to obtain iron from their host is a major virulence determinant (33, 43). Gram-negative bacteria, including N. gonorrhoeae, have evolved elaborate iron assimilation systems involving outer membrane receptor proteins that are capable of recognizing and binding different sources of iron (19). These receptors are usually TonB dependent and are upregulated by low intracellular levels of iron through the repressor protein ferric uptake regulator (Fur) (2, 7, 53).

Many bacteria synthesize and secrete siderophores that chelate iron from the environment. Receptors bind iron-containing siderophores, which are transported through the periplasm and inner membrane by specific transporter systems (44, 47). Gonococci are unable to synthesize siderophores but take advantage of extracellular iron transport proteins made by their host (11). They utilize iron from human transferrin and lactoferrin through the TonB-dependent receptors transferrin binding receptor (TbpAB) and lactoferrin binding receptor (LbpAB), respectively (5, 11).

Although soluble transport proteins and siderophores provide a means of obtaining iron for pathogens, heme remains the largest potential source of iron from the host (38, 42). Heme is bound in hemoproteins to form cytochromes, catalase, hemoglobin (in erythrocytes), myeloperoxidase (in granulocytes), and myoglobin (in myocytes). Hemoproteins and heme also enter the extracellular space following increased cell turnover and breakdown, especially when inflammation and cell damage are present (38). In addition to providing a source of iron, heme is essential for several aerobic processes in bacteria (54). It acts as a cofactor for many of the cytochromes required for oxidative phosphorylation and, in the form of catalase, provides a defense against hydrogen peroxide, a by-product of this process (59). Accordingly, many pathogens, including gonococci, have retained the ability to make heme while developing means of acquiring heme from their host. Pathogenic neisseriae can utilize exogenous heme, hemoglobin, and hemoglobin-haptoglobin as the sole iron source in vitro (9, 30, 36, 50).

Gram-negative pathogens have adopted different strategies for obtaining heme. Some secrete hemophores, which are analogous to siderophores but specifically bind heme or heme-hemopexin (10, 18). Alternatively, heme or heme bound to carrier proteins can be utilized via heme receptors, as occurs with Yersinia spp., Plesiomonas shigelloides, Shigella dysenteriae, Escherichia coli O157:H7, and Vibrio cholerae (12, 21, 22, 39). Hemoglobin is a rich source of iron and heme, since there are four heme moieties contained within a single hemoglobin molecule. Some pathogens, including Haemophilus influenzae, H. ducreyi, N. meningitidis, and N. gonorrhoeae, have TonB-dependent receptors that recognize hemoglobin (9, 16, 49).

It is unclear whether pathogenic neisseriae internalize the heme ring intact (14, 50). One model has suggested that iron may be stripped from heme and transported to the inner membrane by the periplasmic ferric iron binding protein (FbpA) (14). FbpA is used to shuttle iron from transferrin and lactoferrin through the periplasmic space, where it is transported to the cytoplasm by an inner membrane permease/nucleotide binding protein complex (FbpBC) (1, 8, 37). However, a hitA (fbpA) nonpolar isogenic mutant of H. influenzae can still grow with heme as the sole source of iron (26). The fbpABC genes of N. gonorrhoeae and the analogous hitABC genes of H. influenzae are all located on a single operon (1, 26).

The pathways for heme synthesis in both nonplant eukaryotes and members of the beta and gamma subdivisions of the Proteobacteria (e.g., E. coli) are well characterized (3). The main difference occurs with the initial step of synthesizing -aminolevulinic acid (ALA). Mitochondria use succinyl coenzyme A and glycine as the substrates for the enzyme ALA synthetase to make ALA, whereas members of the beta- and gamma-Proteobacteria use -glutamyl tRNA as substrate for the enzyme -glutamyl tRNA reductase (HemA) to make glutamate-1-semialdehyde. Glutamate-1-semialdehyde transferase (HemL) converts glutamate-1-semialdehyde to ALA. The remaining steps of heme synthesis are similar, including the final step of inserting iron into the protoporphyrin IX ring by the enzyme ferrochelatase (HemH) to make heme.

Among E. coli K-12 strains, which lack a heme receptor, hemH isogenic mutants grow poorly because they are unable to utilize exogenous heme or heme precursors for aerobic growth. However, E. coli hemA isogenic mutants grow well when rescued by exogenous ALA, which bypasses the block in ALA synthesis. In fact, a E. coli hemA mutant has been used successfully as a model to demonstrate the function of heterologous heme and hemoglobin receptors, by recombinant expression, to restore growth aerobically in the presence of heme or hemoglobin (50).

For this study, we created mutants with mutations in the heme biosynthetic pathway (hemA and hemH) and in the periplasmic iron transporter (fbpA). Characterization of these mutants enabled us to determine whether exogenous heme can be utilized as a source of both iron and heme and to better elucidate the pathway for entry of heme. The heme mutants were also used to determine whether intracellular heme (from epithelial cells) could support the growth of gonococci.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains, plasmids, and culture conditions. The bacterial strains and plasmids used for this study are listed in Table 1. E. coli strains were grown at 37°C in Luria broth (LB) medium containing appropriate antibiotics for selection. Antibiotics were used at the following concentrations; ampicillin, 100 µg/ml, kanamycin, 60 µg/ml; and spectinomycin, 80 µg/ml. Chloramphenicol was used at 30 µg/ml for E. coli and at 1 and 10 µg/ml for gonococcal strains FA1090 and MS11, respectively.

PCR. The design of the hemH and hemA PCR primers was based on analysis of sequencing contiguities released from the University of Oklahoma Gonococcal Genome Sequencing Project Web site (46). Part of the hemH gene was amplified with primers HEMH1 (5'CGCAAACCGCATTACCTGAT3') and HEMH2 (5'ATTGGCTTTGGAACGATACG3'). The hemA gene was amplified with primers HEMA1 (5'AATCAAACCTGCCAATCCT3') and HEMA2 (5'TCTTCTTCCCCCGCCTTAT3'). For PCR amplification of fbpA, primers containing incorporated restriction sites for BamHI or EcoRI (underlined) were used: FBP1 (5'CGGGATCCTATGAAAACATCTATCCGATACGCACTG3') and FBP2 (5'GGAATTCAGGCAGGGTAAGCGGCAGGGCGATCAG3'). The template for PCR was chromosomal DNA extracted from strain FA1090 (genomic DNA kit; Qiagen) (13). The PCR conditions were as follows: 94°C at 3 min for 1 cycle; 94°C for 45 s, 58°C for 45 s, and 72°C for 3 min for 30 cycles; and 72°C for 3 min for 1 cycle. Genotypic confirmation of the plasmid constructs and isogenic mutants was obtained by analysis of PCR products from appropriate purified plasmid DNA (plasmid midi-prep kit; Qiagen) and phenol-chloroform extracted chromosomal DNA (see Fig. 1 and 2).

Mutagenesis and gonococcal transformation. Plasmids containing gonococcal DNA insert were constructed as follows (Table 1; also see Fig. 1). The hemH and hemA PCR products were ligated into plasmid vector PCRII to generate pUNCH 1303 and pUNCH 1305, respectively (TA cloning kit; Invitrogen). To generate pUNCH1322, PCR-amplified DNA from fbpA and DNA from the gonococcal uptake sequence vector pUC9GCU were double digested with BamHI and EcoRI, gel purified, and directionally ligated (17). All the plasmids were transformed into competent E. coli DH5 MCR cells (Bethesda Research Laboratories), and positive clones were selected either by growth on LB ampicillin medium (pUNCH1322) or by the presence of a white phenotype on isopropyl--D-thiogalactopyranoside/5-bromo-4-chloro-3-indolyl--D-galactopy-ranoside (IPTG/X-Gal) LB medium containing ampicillin and kanamycin (pUNCH1303 and pUNCH1305).

Southern blot analysis and preparation of hybridization probes. Phenol-chloroform-extracted DNA from FA6973 (FA1090 hemH::CAT), FA6974 (FA1090 hemA::), and FA6978 (FA1090 fbpA::) was digested with ClaI-HincII, ClaI, and HincII, respectively. FA1090 DNA was digested with ClaI-HincII, ClaI, or HincII. Following electrophoresis, Southern blotting, alkali denaturation, and prehybridization, the restriction enzyme-digested gonococcal DNA was probed with an antibiotic cassette probe (CAT or ) or PCR-derived DNA with the HEMA1-HEMA2, HEMH1-HEMH2, or FBP1-FBP2 primers. The antibiotic cassette probe was labelled with the DIG High Prime kit (Boehringer Mannheim Biochemicals), and the PCR products were labelled with the PCR DIG labelling mix (Boehringer Mannheim).

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blots. Western blot analysis was performed on whole-cell lysates prepared from gonococci grown on GCB agar with 10 µM heme. To create iron-limiting conditions, Desferal was added to the medium. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was carried out in the discontinuous buffer system of Laemmli with a 4.5% stacking and a 12.5% resolving gel (27). Transfer and detection were performed as described previously (55). The anti-Fbp polyclonal antibody was a gift from T. Mietzner (University of Pittsburgh) and was used at a dilution of 1/5,000.

Intracellular growth assays. The intracellular growth assays were performed under similar conditions to those described previously, with minor modifications (32). Briefly, gonococcal strains MS11, FA6977 (MS11 hemH::CAT) and FA6976 (MS11 hemA::) were grown for 16 to 18 h on GCB plates containing 10 µM heme (35). Since phase variation may have affected the ability of gonococci to attach to and invade cells, all the strains were passed fewer than six times and checked for lipooligosaccharide and opacity protein (Opa) type. The lipooligosaccharide type was checked by analysis of silver-stained polyacrylamide gels, and the Opa type was checked by Western blotting of transparent colonies probed with anti-Opa monoclonal antibody 4B12 (6). A-431 cells (derived from a human endocervical columnar epithelial line) were seeded onto 24-well plates (Falcon) and grown to 80% confluency (5 × 10⁵ cells/well) (32). A 250-µl volume of Opa-negative, pilus-positive gonococci in Dulbecco's modified Eagle's medium (DMEM) with 10 µg of transferrin per ml (saturation, 30%) was added to each well at 2.5 × 10⁶ CFU/well (multiplicity of infection, 5), and the wells were incubated at 37°C under 5% CO₂.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The gonococcal hemH (encoding ferrochelatase) and hemA (encoding -glutamyl tRNA reductase) open reading frames were identified by TBLASTN searches of the N. gonorrhoeae FA1090 sequencing project with known HemH and HemA sequences from E. coli (40, 46, 57). The predicted amino acid sequence of FA1090 HemH had a size of 38 kDa with 41% identity and 62% similarity to HemH of E. coli. Similarly, the predicted amino acid sequence of FA1090 HemA had a size of 45 kDa with 48% identity and 65% similarity to HemA of E. coli (40, 57). Further analysis of the FA1090 sequencing database showed putative transcriptional terminator sequences located immediately downstream of both genes, with the gene downstream of hemH being in the opposite orientation. Furthermore, none of the downstream genes appear to be involved in heme biosynthesis. This is consistent with the arrangement of the corresponding hemH and hemA genes in E. coli (3). Thus, it is unlikely that interruption of hemH or hemA would lead to a polar effect on downstream genes.

To determine if exogenous heme is internalized intact and utilized for growth, we created gonococcal mutants with mutations in the two putative heme biosynthetic genes (hemH and hemA) and in a periplasmic iron transporter gene (fbpA) (Fig. 1). The general strategy was to clone the open reading frames of the genes by PCR, mutagenize them, and return the mutation to the gonococcal chromosome by allelic exchange. The insertions of the  and aphA-3 antibiotic cassettes in the fbp operon were designed to create polar and nonpolar mutations, respectively (34, 45). Since the hemH and hemA isogenic mutants were constructed to generate a heme auxotrophic phenotype, gonococcal transformants were recovered by antibiotic selection on heme-containing medium.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Construction of pUNCH1304, pUNCH1306, and pUNCH1324. A single line indicates a PCR product, and a box represents either the antibiotic cassette or plasmid vector. Only part of the plasmid vector is shown. The open reading frame (ORF) and direction of transcription for the relevant gonococcal gene, interrupted by insertion of the antibiotic cassette, are indicated in gray. The PCR fragment for hemH contains only a portion of the hemH gene. The restriction sites used to linearize the plasmid and insert the antibiotic cassette are shown in parentheses.

Confirmation that the hemH, hemA, and fbpA genes had been successfully cloned was obtained by PCR, restriction analysis, and Southern hybridization (Fig. 2 and data not shown). The PCR products obtained with either the hemH, hemA, or fbpA primers on template DNA from the isogenic mutants were similar in size to the PCR products from the plasmids that had transformed them, suggesting that allelic exchange had occurred (Fig. 1 and 2). Also, the difference in size between the PCR products from the parent strains and their isogenic mutants were compatible with a DNA fragment, of the length of the antibiotic cassette, being inserted into the open reading frames of hemH, hemA, and fbpA (Fig. 1 and 2). Full confirmation of allelic exchange was obtained by Southern blotting with digoxigenin-labelled DNA probes of the antibiotic cassettes and hemH, hemA, and fbpA PCR products (data not shown).

View larger version (39K): [in this window] [in a new window]   FIG. 2.   Ethidium bromide-stained 0.7% agarose gel showing the PCR products from plasmid and gonococcal chromosomal DNA following electrophoresis. Lanes: 1, FA1090; 2, pUNCH1304; 3, FA6973 (FA1090 hemH::CAT); 4, FA1090; 5, pUNCH1306; 6, F6974 (FA1090 hemA::); 7, FA1090; 8, pUNCH1324; 9, FA6978 (FA1090 fbpA::). Oligonucleotide primers for PCR were HEMH1 and HEMH2 for lanes 1 to 3, HEMA1 and HEMA2 for lanes 4 to 6, and FBP1 and FBP2 for lanes 7 to 9. DNA molecular size markers are shown in kilobases.

Growth phenotypes of the heme biosynthetic mutants. The heme biosynthetic mutants (containing the hemA and hemH mutations) grew on GCB agar containing 4 to 16 µM heme. Addition of Desferal, to create iron-limiting conditions, had no effect on growth. Interestingly, the hemH mutant turned reddish brown after prolonged incubation (36 to 48 h). Excessive accumulation of heme precursors, particularly protopophyrin IX, which could not be converted to heme in a ferrochelatase mutant may explain this phenomenon (3). Despite the color change, there was no obvious evidence of toxicity in the hemH mutant. However, when more than 16 µM heme was incorporated into the agar, neither the parent strain nor the heme mutants grew. Similarly, all strains showed a zone of inhibition around wells containing heme in a diffusion assay (Fig. 3). A possible explanation for the lack of growth around higher concentrations of heme was that sodium hydroxide (used to solubilize the heme) exerted an inhibitory effect. However, growth of the parent strains were unaffected around a control well containing 60 µl of 0.1 M NaOH on GCB agar (data not shown).

View larger version (67K): [in this window] [in a new window]   FIG. 3.   Growth phenotypes of FA1090, FA6973 (FA1090 hemH::CAT), FA6974 (FA1090 hemA::), and FA6978 (FA1090 fbpA::) around 0.6-cm wells cut in GCB agar with 50 µM Desferral. The wells were filled with 60 µl of one of the following: heme, 1 mg/ml (Hm); hemoglobin, 10 mg/ml (Hg); transferrin, 25 mg/ml (Tf); iron citrate, 10 mM (FeCit); and iron chloride, 10 mM (FeCl3). All the plates were incubated at 37°C for 24 h in 5% CO2.

View larger version (47K): [in this window] [in a new window]   FIG. 4.   Growth phenotypes of FA6973 (FA1090 hemH::CAT) and FA6974 (FA1090 hemA::) around 0.6-cm wells cut in GCB agar both with and without 50 µM Desferral (DF). The wells contain 60 µl of either 10-mg/ml hemoglobin (Hg) or 10 mM ALA. The plates were incubated at 37°C for 24 h in 5% CO2.

View larger version (13K): [in this window] [in a new window]   FIG. 5.   Growth of MS11(), FA6976 (MS11 hemA::) (), and FA6977 (MS11 hemH::CAT) () in GCB broth with and without additional heme. (A) Strains grown in GCB broth without heme; (B) strains grown in GCB broth with 1 to 3 µM heme. The results are based on three independent experiments, and the vertical bars represent one standard deviation. Similar results were obtained with FA6973 (FA1090 hemH::CAT) and FA6974 (FA1090 hemA::), respectively (data not shown).

Growth with hemoproteins as a heme source. The heme biosynthetic mutants were tested for the ability to grow in the presence of various hemoproteins, including hemoglobin, myoglobin, catalase, and cytochrome c. No detectable growth occurred around wells containing these heme sources, with the exception of hemoglobin (data not shown). Growth of the heme biosynthetic mutants with hemoglobin (>0.3 µM) as the sole heme and iron source was comparable to growth of the parent strain, provided that the HpuAB receptor was expressed (Fig. 3). However, after 24 h on GCB plus hemoglobin, the colony size of the heme biosynthetic mutants was reduced compared with that of the mutants on the same medium containing Desferal (Fig. 4). The essential role of the HpuAB receptor in the hemoglobin-utilizing biosynthetic mutants was demonstrated by generating hemH and hemA mutants from HpuB-expressing and -nonexpressing strains of FA1090. These strains were tested for growth on GCB agar containing either 2.4 µM hemoglobin or 8 µM heme. Outer membranes of colonies picked from hemoglobin plates were probed with polyclonal antibody to HpuB on Western blots and showed the expected 85-kDa band (data not shown) (9). Furthermore, a double isogenic mutant containing antibiotic cassette insertions in hpuA and hemH, FA7008 (FA1090 hemH::CAT hpuA::), grew well on GCB agar with heme after 24 h but not on GCB agar with hemoglobin even after 48 h (9). Heme utilization from hemoglobin was also TonB dependent, since an HpuAB-expressing, TonB isogenic mutant, FA7007 (FA1090 hemH::CAT tonB::), grew with heme but not with hemoglobin (reference 4 and data not shown).

Characterization of the fbp mutants. Since heme can be internalized and used by gonococci as an iron source for growth, an attempt was made to construct two isogenic fbp mutants of FA1090 and rescue them on GCB agar containing heme or hemoglobin (1). The construction of the fbp mutants was designed to create both a nonpolar (FA1090 fbpA::aphA-3) mutation and a polar (FA1090 fbpA::) mutation in the fbp operon (fbpABC) (34, 45).

View larger version (50K): [in this window] [in a new window]   FIG. 6.   Western blot of gonococcal whole-cell lysates from FA1090, FA6978 (FA1090 fbpA::), and FA6979 (FA1090 fbpA::aphA-3) probed with polyclonal antibody against FbpA. Lanes: 1, FA1090 (iron replete); 2, FA1090 (iron limited); 3, FA6979 (iron replete); 4, FA6979 (iron limited); 5, FA6978 (iron replete); 6, FA6978 (iron limited).

Heme auxotrophs fail to grow inside epithelial cells. Most heme is found within the cytoplasm and mitochondria of cells in the form of hemoproteins, providing a potential source of heme and iron for growth. Since little is known about how gonococci can utilize iron or heme from within cells, the heme auxotrophs were tested for the ability to survive and grow within an endocervical epithelial cell line. MS11 was chosen as the parent strain, because it has been characterized in attachment, invasion, and intraepithelial cell growth assays (32, 35).

View larger version (16K): [in this window] [in a new window]   FIG. 7.   Intracellular growth and survival of MS11 (), FA6976 (MS11 hemA::) () and FA6977 (MS11 hemH::CAT) () inside epithelial cell line A-431. Gentamicin was added 12 h postinoculation to allow time for attachment and invasion. Time refers to the time after addition of gentamicin. The results are standardized to give a ratio by dividing the number of surviving colonies (CFUs) at various time points between 2 and 10 h after addition of the gentamicin by the number of initial colonies (CFUi) at 2 h after addition of the gentamicin. Thus, CFUs/CFUi at 2 h gives a ratio of 1 for all three strains. Results are based on three independent experiments, each with three wells per strain, and the vertical bars represent one standard deviation.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We constructed gonococcal hemA, hemH, and fbpA isogenic mutants and confirmed their genotype by PCR and Southern hybridization. Although a polar effect downstream of hemA and hemH cannot be completely excluded, analysis of DNA sequences surrounding the hemH and hemA open reading frames released from the University of Oklahoma gonococcal sequencing project suggests that neither of these genes form part of an operon for heme synthesis.

The hemH mutant was predicted to lack the enzyme ferrochelatase, necessary for inserting iron into the porphyrin ring of protoporphyrin IX to make heme, whereas the hemA mutant was predicted to lack -glutamyl tRNA reductase, necessary for synthesis of ALA, which is an early substrate in heme synthesis (3). The phenotypes of both mutants were consistent with these predictions. Concern that the hemH mutant would fail to grow, due to an inability to internalize heme, proved to be unfounded, since the mutant grew well when either exogenous heme or hemoglobin was present. The hemA mutant also grew in the presence of ALA, provided that an iron source was available, consistent with an inability to synthesize ALA.

Utilization of exogenous heme by gonococci is advantageous, since a large proportion of the iron requirement is likely to be directed to endogenous heme synthesis. Exogenous heme can reduce heme biosynthesis by feedback inhibition and serve as substrate for synthesis of hemoproteins (3, 56, 58). In addition to providing heme for synthesis of hemoproteins, exogenous heme fulfills the iron requirements in gonococci. How this occurs is unclear; eukaryotic cells contain heme oxygenase enzymes (HO1 and HO2) which mobilize iron from heme, but only one heme oxygenase homolog has been found in proteobacteria (HmuO in Corynebacterium diphtheriae) (48). Searches of the Oklahoma genomic sequencing and National Center for Biotechnology Information GeneBank databases with HO1, HO2, and HmuO failed to show any significant prokaryotic homology.

We demonstrated that in addition to serving as an iron source, exogenous heme could be utilized as a heme source by gonococci. However, excess heme inhibited growth, possibly due to lipid peroxidation of membranes (28). The heme biosynthetic mutants were able to utilize heme from hemoglobin in a TonB-dependent manner, provided that the hemoglobin receptor (HpuAB) was expressed. Growth at 24 h with hemoglobin as the sole heme and iron source was better than growth on hemoglobin with additional iron in the medium, suggesting that iron limitation enhances heme utilization from hemoglobin. The most likely explanation for this observation is that more hemoglobin receptor is expressed under iron-limiting conditions (9). In contrast, iron limitation had no observable effect on heme utilization from free heme, which, like iron utilization from heme, was TonB independent (4).

There is circumstantial evidence for a gonococcal heme receptor, based on growth experiments in broth, in which a monoclonal antibody that recognizes a 97-kDa heme affinity-purified total membrane protein appeared to inhibit iron utilization from heme (29). Characterization of a putative heme receptor, even with an isogenic mutant, may prove difficult. Heme is a relatively small (660-Da), poorly soluble, hydrophobic molecule, thereby presenting potential difficulties in differentiating receptor-mediated utilization from nonspecific utilization. It is quite possible that gonococci do not express a heme receptor and that heme is internalized by nonspecific mechanisms, such as entry through a porin protein channel. Nonspecific uptake of heme by gonococci would account for the apparent lack of a TonB phenotype and the lack of inhibition of heme-dependent growth by free iron in the medium (4, 50). The inability of gonococci to utilize iron from heme-hemopexin and heme-albumin is also consistent with absence of a heme receptor (15). The available evidence is insufficient to settle the question of whether gonococci express a heme receptor.

Complementation of a hemA E. coli mutant with a cosmid containing meningococcal tonB, exbB, and exbD genes and a plasmid for the meningococcal hemoglobin receptor gene (hmbR) demonstrated growth with hemoglobin as a porphyrin source (50). This suggests that heme is internalized in this system. Since the E. coli outer membrane is normally impervious to heme, it is unclear how heme may be internalized through the periplasm and inner membrane into the cytoplasm. E. coli is not known to have a heme-specific transporter system as occurs in Yersinia enterocolitica and Y. pestis (22, 31, 51).

Western blot analysis of the fbpA mutants showed that the full-length product (37 kDa) was not made, but a smaller band (about 20 kDa), which could represent a partial protein product, was seen. However, it is unlikely that this product could retain functional iron-binding activity, since the fbpA mutant was unable to utilize iron from transferrin, ferric citrate, and ferric chloride. Lactoferrin was not tested as an iron source for the fbpA mutant because the parent strain FA1090 does not have a complete lactoferrin receptor.

One possible route for transporting iron from heme is via FbpA. This may occur if iron can be stripped from the heme ring before entering the cytoplasm. Liberated iron could bind to FbpA in the periplasm and be transported into the cytoplasm by the inner membrane FbpBC complex. Based on one uptake study of ¹⁴C- and ⁵⁹F-radiolabelled heme, iron appeared to be stripped from heme and transported to the inner membrane by FbpA (14). However, iron reentering the periplasm from the cytoplasm or nonspecific binding may have contributed to this effect. We found that FbpA was not necessary for heme utilization in gonococci, because the fbpA isogenic mutants grew normally with heme or hemoglobin as the sole source of iron. We cannot exclude the possibility that FbpA plays a role in heme utilization, but it was clearly not necessary for growth of the fbpA mutants, as it was for iron uptake from transferrin and ferric citrate.

The gonococcal fbpA mutant phenotype was similar to that of an H. influenzae fbpA mutant, which was unable to utilize iron from transferrin but could still utilize iron from heme and hemoglobin (26). Following submission of this manuscript, an N. meningitidis fbpABC mutant which also showed a similar phenotype to the gonococcal fbpA mutant was described (25). This mutant grew with heme and hemoglobin as an iron source but was unable to utilize iron from transferrin, lactoferrin, and iron chelates.

The heme biosynthetic auxotrophs were able to attach to and invade A-431 cells but, unlike the parent strain, failed to survive inside the cells. No clear difference in intracellular survival was observed between the heme biosynthetic mutants. An inability to assimilate heme or heme precursors from within epithelial cells seems the most likely reason for this observation. One possible explanation is that gonococci are unable to utilize heme from hemoproteins, other than hemoglobin, which form the largest source of intracellular heme (15, 38). Also, many of the heme precursors and hemoproteins are synthesized and located within mitochondria, making it difficult for gonococci to obtain these substrates for growth. It remains unclear how gonococci acquire iron during growth within epithelial cells.

In summary, gonococci can utilize heme and hemoglobin as a heme source. Within epithelial cells, heme sources do not appear able to support the survival or growth of heme auxotrophic mutants. Internalization of heme probably occurs independently of FbpA. Further studies to investigate and characterize possible heme receptors and heme transport systems are under way.