Search

We have characterized the interaction of the Neisseria 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.

We have previously described HpuAB, a two-component receptor that mediates binding to haemoglobin (Hb), haemoglobin-haptoglobin (Hb-Hp) and apo-haptoglobin (Hp). In this communication, we constructed non-polar mutations in the hpuA and hpuB loci to examine the individual roles of HpuA and HpuB. Our results indicate that both HpuA and HpuB are required for the acquisition of Fe from Hb and Hb-Hp. We isolated Hb utilization-positive (Hb+) variants of our Hb utilization-negative (Hb-) hpu mutants at a frequency of 10(-3) and demonstrated that the Hb+ phenotype resulted from the expression of a second Hb receptor, HmbR. Expression of HmbR in DNM2 was found to be controlled by translational frameshifting involving a polyguanine (G) tract located within the hmbR locus. The hpuA locus also contains a poly(G) tract, which suggested that meningococci could phase vary each Hb receptor independently by slip-strand mispairing in the poly(G) tracts found in hpuA and hmbR. Thus, we isolated a naturally occurring Hb- variant of DNM2, designated DNM2 Hb-, which did not express either HpuAB or HmbR. Hb+ variants of DNM2Hb- were selected and examined for HpuAB and HmbR expression. In each instance, acquisition of HpuAB or HmbR expression was correlated with phase variation in the poly(G) tract of each Hb receptor.

The meningococcal hemA gene was cloned and used to construct a porphyrin biosynthesis mutant. An analysis of the hemA mutant indicated that meningococci can transport intact porphyrin from heme (Hm), hemoglobin (Hb), and Hb-haptoglobin (Hp). By constructing a HemA- HpuAB- double mutant, we demonstrated that HpuAB is required for the transport of porphyrin from Hb and Hb-Hp.

We identified lbpB, encoding the lipoprotein component of the meningococcal lactoferrin receptor. An LbpB mutant was unable to acquire Fe from lactoferrin and exhibits decreased surface binding to lactoferrin. Primer extension and reverse transcription-PCR analysis indicate that lbpB and lbpA are cotranscribed on a polycistronic Fe-repressible mRNA.

We previously identified HpuB, an 85 kDa Fe-repressible protein required for utilization of Fe from, and binding to, haemoglobin and the haemoglobin-haptoglobin complex. The gene for hpuB was cloned from Neisseria meningitidis strain DNM2 and the predicted amino acid sequence indicates that HpuB is an outer membrane receptor belonging to the TonB family of high-affinity transport proteins. A second open reading frame, predicted to encode a 34.8 kDa lipoprotein, was discovered 5' to hpuB, and was designated hpuA. HpuA was identified in a total-membrane-protein preparation by construction of a mutant lacking HpuA. Acylation of HpuA was confirmed by [3H]-palmitic acid labelling of meningococci. Consensus promoter sequences were not apparent 5' to hpuB. The hpuA insertion mutation exerted a polar effect, abolishing expression of hpuB, suggesting that hpuA and hpuB are co-transcribed. The 3.5 kb polycistronic hpuAB mRNA was identified and shown to be transcriptionally repressed by iron. The transcriptional start site was identified 33 nucleotides 5' to the hpuA translational start site, appropriately positioned around consensus promoter and ferric uptake regulator (Fur)-box sequences. The structure of this operon suggests that HpuA-HpuB is a two-component receptor analogous to the bipartite transferrin receptor TbpB-TbpA.

The majority of in vitro-grown Neisseria gonorrhoeae strains were unable to use hemoglobin as the sole source of iron for growth (Hgb-), but a minor population was able to do so (Hgb+). The ability of Hgb+ gonococci to utilize hemoglobin as the iron source was associated with the expression of an iron-repressible 89-kDa hemoglobin-binding protein in the outer membrane. The N-terminal amino acid sequence of this protein revealed amino acids, from positions 2 to 16, identical to those of HpuB, an 85 kDa iron-regulated hemoglobin-haptoglobin utilization outer membrane protein of Neisseria meningitidis. Isogenic mutants constructed by allelic replacement with a meningococcal hpu::mini-Tn3erm construct no longer expressed the 89-kDa protein. Mutants could not utilize hemoglobin to support growth but still grew on heme. Thus, the gonococcal HpuB homolog is a functional hemoglobin receptor and is essential for growth with hemoglobin.

Hemoglobin complexed to the plasma protein haptoglobin can be used by Neisseria meningitidis as a source of iron to support growth in vitro. An N meningitidis mutant, DNM2E4, was generated by insertion of the mini-Tn3erm transposon into the gene coding for an 85-kDa iron-regulated outer membrane protein. Membrane proteins prepared from DNM2E4 were identical to those of the wild-type strain except that the 85-kDa protein was not produced. This mutant was unable to use hemoglobin-haptoglobin complexes as an iron source to support growth and was also impaired in the utilization of free hemoglobin. The mutant failed to bind free hemoglobin, hemoglobin-haptoglobin complexes, or apo-haptoglobin in a solid-phase dot blot assay. The 85-kDa protein was affinity purified when hemoglobin-haptoglobin complexes were used as a ligand but was not purified when free hemoglobin was used. We hypothesize that the 85-kDa iron-regulated protein is the hemoglobin-haptoglobin receptor and designate this protein Hpu (for hemoglobin-haptoglobin utilization).

Iron-limited growth conditions under anaerobiosis were established for Actinobacillus actinomycetemcomitans strains Y4, JP2, and 75 by use of the ferrous ion chelator 2,2'-dipyridyl. Growth inhibition was reversible with both ferrous and ferric iron sources. Sarcosyl-insoluble membrane fractions of iron-stressed anaerobic A. actinomycetemcomitans cultures revealed a similar iron-repressible protein of approximately 70 kDa in A. actinomycetemcomitans strains Y4, JP2, and 75. This 70-kDa protein was recognized by serum from localized juvenile periodontitis patients and a periodontally healthy subject. This suggested that the 70-kDa iron-repressible protein may be expressed in vivo. When A. actinomycetemcomitans was grown under aerobic conditions, the ferric iron chelator, ethylenediamine di(o-hydroxyphenylacetic acid) (EDDA) was utilized for growth limitation. EDDA inhibition was reversible in strain Y4 with ferrous and ferric iron sources. An iron-repressible protein of approximately 70 kDa was also noted in iron-stressed aerobic cultures. The 70-kDa protein may be involved in iron transport by A. actinomycetemcomitans. Preliminary experiments were performed to examine potential iron transport systems for A. actinomycetemcomitans. Production of the two most common chemical types of siderophore was not detected in A. actinomycetemcomitans culture supernatants. Iron-starved A. actinomycetemcomitans cells did not bind transferrin or lactoferrin in a dot blot assay.

The human immunoglobulin G (IgG) immune response against Porphyromonas (Bacteroides) gingivalis A7A1-28 iron-repressible membrane proteins (IRMPs) and other membrane proteins was examined by immunoblot analysis. Thirty sera from patients with adult periodontitis and 30 sera from periodontally healthy subjects were included. Iron limitation of P. gingivalis was achieved by growing bacteria in brain heart infusion broth supplemented with protoporphyrin IX and 250 microM alpha, alpha'-dypyridyl, a ferrous iron chelator. Iron-sufficient growth was achieved by growing bacteria in the same medium without alpha, alpha'-dypyridyl. Human sera, in particular those from patients with periodontitis who exhibited high levels of IgG against whole cells of P. gingivalis A7A1-28 in serum in an enzyme-linked immunosorbent assay (ELISA), commonly reacted with five membrane proteins with apparent molecular masses of 80, 67.5, 51, 40.5, and 28 kDa and four IRMPs of 46, 43, 37.5, and 22 kDa. More than 80% of the sera from patients with periodontitis and high levels of IgG against strain A7A1-28 in serum by ELISA reacted with the 46-, 43-, and 37.5-kDa IRMPs, and 40% of these subjects expressed immunoreactivity against the 22-kDa IRMP. Sera from patients with periodontitis and low levels of IgG against strain A7A1-28 in serum by ELISA and sera from periodontally healthy subjects exhibited less immunoreactivity against IRMPs and the five membrane proteins of P. gingivalis. The present study indicates that P. gingivalis IRMPs are immunogenic and that these proteins are expressed in vivo.

Neisseria meningitidis grown under iron-limiting conditions in vitro expresses additional iron-repressible outer membrane proteins (FeRPs). To see which FeRPs were expressed and immunogenic in human infection, we examined purified membranes from four meningococcal disease isolates with Western blotting of patient sera. Convalescent serum from each patient contained immunoglobulin G (IgG) and IgM antibodies against the homologous 70-kilodalton (kDa) FeRP and IgG antibody to the homologous 94-kDa FeRPs. Three other immunoreactive FeRPs were identified in two or more strains. Neither acute-phase sera nor pooled normal human sera contained appreciable levels of these antibodies. Antigenic cross-reactivity among FeRPs was suggested by the observation that the convalescent sera of two patients contained IgG antibodies reactive with the 70- and 94-kDa FeRPs and IgM antibodies reactive with the 70-kDa FeRPs from all four strains. Additionally, rabbit antiserum against the 70-kDa FeRP from one of these disease isolates contained IgG and IgM antibodies that reacted in Western blots with the 70-kDa FeRPs in all four strains. These results demonstrate that meningococcal FeRPs are expressed and immunogenic in vivo and that certain of these proteins are immunologically cross-reactive.

We isolated an iron-uptake mutant of Neisseria meningitidis M986-NCV-1 that was severely limited in the ability to use several sources of iron in the form of Fe3+. This mutant, FAM11, grew poorly or not at all with human transferrin (TF) or lactoferrin (LF) as the sole iron source in a defined medium but grew as well as wild-type meningococci with hemin or hemoglobin. Uptake of 55Fe bound to TF, LF, dicitrate complexes, aerobactin, or nitrilotriacetate was reduced to 0 to 4% of the wild-type level. FAM11 did not produce an iron-repressible outer membrane protein (FeRP) of 70 kilodaltons (kDa) found in membranes of iron-stressed M986-NCV-1. Western blot (immunoblot) analysis using rabbit antiserum against this protein revealed that at least 17 of 18 meningococcal and 10 of 14 gonococcal strains produced an FeRP of ca. 70 kDa. The 70-kDa FeRP was shown to be surface exposed by radioimmunoprecipitation with human immune sera. These data suggest that the 70-kDa FeRP is somehow involved in Fe uptake from TF and LF. However, we were unable to transform the iron-uptake phenotype from FAM11 into wild-type meningococci to confirm this. Revertants of FAM11 that grew with TF and LF did not regain the ability to make the 70-kDa FeRP but also did not completely regain the Fe-uptake phenotype of M986-NCV-1.

In gonococci, the nonspecific antimicrobial resistance locus penB is known to be closely linked to loci designated nmp that alter the Mr and antigenicity of the outer membrane porin protein I (P.I). We report that after selection for the linked donor penB locus, occasional recombinants expressed P.I with some epitopes from each parent. These hybrid P.I antigens were stable on subculture and were transformed at a locus closely linked to penB. The hybrid P.I antigens were detected with monoclonal antibodies in both coagglutination and Western blot assays. The alterations of P.I antigenicity may have resulted from recombination between structural genes for P.I that are closely linked to penB.