INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Moraxella (Branhamella) catarrhalis is a human pathogen that has only recently been recognized as a significant health problem (5). It is a commensal organism colonizing the respiratory tract in children and adults, with the highest incidence in young children and in adults of >60 years of age (10). It is the third most common cause of otitis media and sinusitis in children, after Streptococcus pneumoniae and Haemophilus influenzae, and is responsible for an estimated 15 to 20% of disease. In adults, M. catarrhalis infection can lead to exacerbation of chronic bronchitis or development of pneumonia in patients with pre-existing pulmonary disease. More rarely, it also causes bacteremia and meningitis (10, 17, 23).

Otitis media affects approximately 70% of all children by the age of three, with many children experiencing recurrent disease (2). Chronic otitis media can lead to hearing, speech, and cognitive impairment in children, since it tends to occur at a time when language is developing. The incidence of M. catarrhalis-induced otitis media is variable depending upon the population being studied, ranging from 7 to 20% in the United States, (21), 0 to 10% in Greenland (16), and about 1% in Spain (7). Antibiotic resistance, especially penicillin resistance due to the expression of -lactamase, is very common in clinical isolates of M. catarrhalis, reaching 80 to 85% in United States and European isolates (24). A vaccine against otitis media caused by M. catarrhalis is clearly needed.

Iron restriction is a general host defense mechanism against microbial pathogens, and in the human host, iron is sequestered by transferrin, lactoferrin, hemoglobin, and other complex molecules. A number of bacterial species, including Bordetella pertussis (22), Helicobacter pylori (9), M. catarrhalis (33), Neisseria gonorrhoeae (1), Neisseria meningitidis (29, 33), Prevotella nigrescens (8), and Treponema spp. (34), have been shown to express outer membrane proteins which specifically bind human lactoferrin. M. catarrhalis, N. gonorrhoeae, and N. meningitidis utilize both transferrin and lactoferrin binding complexes, and a single lactoferrin binding protein of ~105 kDa was originally identified in these organisms (33). The lbpA genes from N. gonorrhoeae and N. meningitidis have been cloned and sequenced (1, 27), but until recently there was no evidence for the existence of an lbpB gene (3, 13, 25, 28).

We report here the cloning and sequencing of the M. catarrhalis lactoferrin binding protein genes lbpA and lbpB. The recombinant proteins were expressed in Escherichia coli, and high-titer antibodies were raised. Anti-rLbpA antiserum was not bactericidal, but anti-rLbpB antisera were weakly bactericidal against the autologous and heterologous strains.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Recombinant DNA techniques. Restriction endonucleases were purchased from Boehringer Mannheim (Laval, Quebec, Canada), New England Biolabs, Bethesda Research Laboratories, or Pharmacia and were used according to the manufacturers' specifications. Oligonucleotides were synthesized on an Applied Biosystems, Inc. (ABI), model 380B DNA synthesizer and purified by chromatography (Oligonucleotide Purification Cartridge; Perkin-Elmer, Culver City, Calif.). Other recombinant DNA methods were performed according to Sambrook et al. (32).

Bacterial strains and media. M. catarrhalis otitis media clinical isolates 4223 and 3 were kindly provided by T. Murphy (State University of New York, Buffalo, N.Y.), strain Q8 was a gift from M. Bergeron (University of Laval, Laval, Quebec, Canada), strain VH19 was provided by V. Howie (University of Texas, Galveston, Tex.), strain H-04 was from G. D. Campbell (Louisiana State University, Shreveport, La.), and strain LES-1 was obtained from L. E. Stenfors (University of Tromso, Tromso, Norway). M. catarrhalis strains were maintained on Mueller-Hinton agar (Becton Dickinson, Cockeysville, Md.) or grown in brain heart infusion (BHI) medium (Difco, Detroit, Mich.) with or without the addition of ethylenediamine-di(o-hydroxyphenylacetic acid) (EDDA) (Sigma, St. Louis, Mo.), as described previously (15). E. coli strains were grown in YT medium supplemented with 50 µg of ampicillin ml¹ as required.

Purification of LbpA and protein sequence determination. Native LbpA was purified by affinity chromatography under high-stringency conditions with immobilized lactoferrin (3). The purified LbpA protein was digested overnight with cyanogen bromide; then, fragments were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and submitted for sequence analysis on an ABI model 477A protein sequencer. A 13-kDa protein fragment was found to have the N-terminal sequence MVQYTYRKGKENKAH.

Generation of a probe for screening libraries. A degenerate oligonucleotide primer was prepared based upon the internal LbpA sequence QYTRKGENKA 5' CAA TAT ACI CGT/C AAA GGT/C GAA AAT/C AAA GC 3'

Construction and screening of genomic libraries. M. catarrhalis 4223 and Q8 EMBL3 libraries were prepared as described previously (20). Briefly, chromosomal DNA was partially digested with Sau3AI, and DNA fragments of 15 to 23 kb were purified. The DNA was cloned into BamHI-digested EMBL3 arms (Promega, Madison, Wis.) and packaged according to the manufacturer's instructions. The libraries in E. coli LE392 cells were plated, and plaques were lifted onto nitrocellulose membranes for hybridization with the labelled 2.2-kb lbpA PCR fragment. Several putative phage clones were obtained from each library, and phage DNA was prepared for further analysis. Restriction enzyme and Southern blot analyses indicated that at least a portion of lbpA was localized to a ~9-kb HindIII fragment from each phage clone. The Q8 HindIII fragment was subcloned into pBluescript, generating plasmid pLDW1, and the 4223 HindIII fragment was subcloned into pUC18, generating plasmid pLD1-8. Figure 1 illustrates the restriction map and gene placement within the M. catarrhalis lfr locus.

Sequencing of the lfr genes. Plasmid DNA was prepared from 50-ml overnight cultures by using the Qiagen Plasmid Midi kit (Qiagen Inc., Chatsworth, Calif.). DNA samples were sequenced on an ABI model 373A DNA sequencer using dye terminator chemistry. Oligonucleotide primers of 17 to 25 bases in length were used to sequence both strands of the DNA.

PCR amplification of the VH19 lbpB gene. Chromosomal DNA was prepared from M. catarrhalis VH19. Oligonucleotide primers were designed based upon the flanking sequence of the 4223 lbpB gene. The sense primer was 5' AAGCTTAGCATGATGGCATCGGCT 3', and the antisense primer was 5' TTAGCCCAAGGCAAATCTGGTGCA 3'. Two independent PCR amplifications were performed, as above, and specific 2.9-kb fragments were amplified and subcloned into pCR II (Invitrogen, Carlsbad, Calif.), generating plasmids pVH19pcr1 and pVH19pcr2 for sequence analysis. A third PCR amplification was performed without subcloning of the resultant DNA. PCR-amplified DNA was purified for direct sequencing with a Qiagen PCR purification kit.

Construction of clones expressing recombinant LbpA and LbpB. In order to express lbpA, 5' and 3' fragments of lbpA were generated by PCR amplification and were ligated to an internal 2.3-kb fragment to recreate a full-length gene. The primers used to amplify an ~200-bp 5' fragment to a BstEII site were MSKSIT 5' GGAATTCCAT ATG TCA AAA TCT ATC ACA AA 3'

Purification of recombinant proteins. The strain Q8 rLbpA protein was expressed at about 10% of total protein as inclusion bodies, but the strain 4223 rLbpA protein was expressed at substantially lower levels. E. coli cells from a 500-ml culture were resuspended in 40 ml of 50 mM Tris-HCl, pH 8.0, containing 5 mM 4-(2-aminoethyl)-benzenesulfonylfluoride protease inhibitor (Calbiochem, La Jolla, Calif.) and 0.1 M NaCl and disrupted by sonication (three times for 10 min each; 70% duty circle). The extract was centrifuged at 20,000 × g for 30 min, and the resultant supernatant, which contained >95% of the soluble proteins from E. coli, was discarded. The remaining pellet was further extracted in 40 ml of 50 mM Tris-HCl, pH 8.0, containing 0.5% Triton X-100 and 10 mM EDTA. The mixture was stirred at 4°C for at least 1 h and then centrifuged at 20,000 × g for 30 min, and the supernatant containing residual soluble proteins and the majority of the membrane proteins was discarded. The resultant pellet was further extracted in 40 ml of 50 mM Tris-HCl, pH 8.0, containing 1% octylglucoside. The mixture was stirred at 4°C for at least 1 h and then centrifuged at 20,000 × g for 30 min. The supernatant containing residual contaminating proteins was discarded. The resultant pellet obtained after the above extractions contained the inclusion bodies. The recombinant LbpA protein (rLbpA) was solubilized in 50 mM Tris-HCl, pH 8.0, containing 6 M guanidine and 5 mM dithiothreitol (DTT). After centrifugation, the resultant supernatant was further purified on a Superdex 200 gel filtration column equilibrated in 50 mM Tris-HCl, pH 8.0, containing 2 M guanidine and 5 mM DTT. The fractions were analyzed by SDS-PAGE, and those containing purified rLbpA were pooled. Triton X-100 was added to the pooled rLbpA fraction to a final concentration of 0.1%. The fraction was dialyzed overnight at 4°C against phosphate-buffered saline (PBS) and then centrifuged at 20,000 × g for 30 min. The purified rLbpA was stored at 20°C.

Lactoferrin binding and transferrin binding assays. Human lactoferrin (Sigma) was conjugated to horseradish peroxidase (HRP) by using an EZ-Link maleimide-activated HRP kit (Pierce, Rockford, Ill.) according to the manufacturer's instructions. Briefly, 1 mg of human lactoferrin, resuspended in 1 ml of PBS, was mixed with 20 µl of SATA solution (Pierce) to form SATA derivative. The solution was incubated for 30 min and then deacetylated for 2 h at room temperature. Separation of the deacetylated human lactoferrin derivative from hydroxylamine-HCl and by-products was achieved with a desalting column (1 by 10 cm). Fractions (500 µl) were collected; those containing deacetylated human lactoferrin were pooled, and the protein concentration (about 0.5 mg ml¹) was confirmed by measuring at A₂₈₀. The protein pool (1 ml) was added to 1 mg of EZ-Link maleimide-activated HRP and incubated for 1 h at room temperature. The resulting conjugate was used for the lactoferrin binding assay.

Immunization of animals and immunoassays. Groups of two guinea pigs (Hartley outbred; Charles River, LaSalle, Quebec) were immunized intramuscularly on day 1 with 5 µg of purified rLbpA or rLbpB protein emulsified in complete Freund's adjuvant. Animals were boosted on days 14 and 29 with the same doses of protein emulsified in incomplete Freund's adjuvant. Serum samples were collected on day 42 for determination of bactericidal activity. Anti-Lbp antibody titers in guinea pig immune sera were determined by antigen-specific enzyme-linked immunosorbent assays (ELISAs), as previously described (40). Microtiter wells (Nunc-MAXISORB; Nunc, Roskilde, Denmark) were coated with 50 µl of protein (0.5 µg ml¹). The reactive titer of an antiserum was defined as the reciprocal of the highest dilution consistently showing a twofold increase in absorbance at 450 nm over that obtained with the preimmune serum samples.

Whole-cell ELISAs. M. catarrhalis 4223 was grown in the presence of EDDA, as described above. Cell pellets were collected by centrifugation, washed with PBS, and resuspended in 50 mM carbonate-bicarbonate buffer, pH 9.6. The optical density of the suspension was adjusted to 0.5 at 490 nm, and 200 µl of a 1:100 dilution of whole bacteria was used to coat microtiter wells. The plates were air-dried at 37°C overnight and then blocked with PBS-0.1% bovine serum albumin (BSA) at 37°C for 1 h (250 µl per well). After three washes with PBS-0.1% Tween 20, 200 µl of antisera at an appropriate dilution (in PBS-0.1% gelatin) was added to the wells and further incubated at 37°C for 2 h. Affinity-purified F(ab')₂ fragment of donkey anti-guinea pig immunoglobulin G (H+L) antibodies conjugated to HRP (Jackson ImmunoResearch Laboratories) was used as reporter. The reactions were developed with tetramethylbenzidine-H₂O₂ (ADI), and absorbancies were measured at 450 nm (with 540 nm as a reference wavelength) in a Flow Multiskan MCC microplate reader (ICN Biomedicals).

Antigenic conservation of LbpA and LbpB in M. catarrhalis strains. To demonstrate the iron-dependent expression of the lbpA and lbpB genes, representative M. catarrhalis strains were grown in BHI with or without 25 µM EDDA. Whole-cell lysates were separated by SDS-PAGE and electrophoretically transferred to a nitrocellulose membrane. Guinea pig anti-Q8 rLbpA, anti-Q8 rLbpB, and anti-4223 rLbpB antisera were used as first antibodies, and HRP-conjugated protein G (Zymed Laboratories, San Francisco, Calif.) was used as reporter.

Bactericidal antibody assay. The bactericidal antibody assay was performed as previously described (39). Briefly, the M. catarrhalis strains were grown to an optical density at 578 nm of 0.5 in BHI medium containing 25 µM EDDA. The bacteria were diluted so that 150 to 450 CFU were added to each reaction. Guinea pig anti-rLbpA or anti-rLbpB antisera and prebleed controls were heated to 56°C for 30 min to inactivate endogenous complement and were diluted with veronal buffer containing 0.1% BSA (VBS). Guinea pig complement (BioWhittaker, Walkersville, Md.) was diluted 1:10 in VBS. Twenty-five microliters each of diluted antiserum, bacteria, and complement was added to duplicate wells of a 96-well microtiter plate (Nunc). The plates were incubated at 37°C for 60 min with gentle shaking at 70 rpm on a rotary platform. Fifty microliters of each reaction mixture was plated onto Mueller-Hinton agar plates (Becton Dickinson) which were incubated at 37°C for 24 h, and then at room temperature for 24 h, before the bacteria were counted. Antisera were determined to be bactericidal if 50% of bacteria were killed compared with preimmune serum controls. Assays were performed at least twice, and two different guinea pig antisera were tested against the autologous strains.

Nucleotide sequence accession numbers. The sequence data in this report have been submitted to the GenBank database under accession no. AF043131 through AF043133.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning of the M. catarrhalis lfr genes. In order to clone the M. catarrhalis lactoferrin receptor genes, native LbpA protein was purified from strain 4223 by affinity chromatography under high-stringency conditions with immobilized lactoferrin (3) and submitted for N-terminal sequence analysis. The N terminus was found to be blocked, so the protein was digested with cyanogen bromide and the N-terminal sequence was obtained from an internal 13-kDa fragment. The sequence, MVQYTYRKGKENKAH, was not found in M. catarrhalis TbpA (manuscript submitted) or in other known TbpA or LbpA proteins. Based upon this unique LbpA sequence, a degenerate sense primer was designed for PCR amplification of part of the lbpA gene. An antisense PCR primer was designed based upon the sequence LEMKF, which has been found at the carboxyl terminus of all known LbpA and the related TbpA proteins. By using these primers, specific 2.2-kb fragments were PCR amplified from strain 4223 and Q8 chromosomal DNA. Partial sequence analysis determined that the fragments were not from the M. catarrhalis tbpA gene (manuscript submitted). The 2.2-kb fragment was used as a probe to screen M. catarrhalis 4223 and Q8 chromosomal libraries in EMBL3. Putative phage clones containing approximately 16-kb inserts were identified, and the lbpA gene was localized to a ~9-kb HindIII fragment by restriction enzyme and Southern blot analyses. The 9-kb HindIII fragments from the strain 4223 and Q8 libraries were subcloned, generating plasmids pLD1-8 and pLDW1, respectively.

Analysis of the nucleotide sequence of the lfr genes. The inserts from pLD1-8 and pLDW1 were sequenced, and three complete open reading frames (ORFs) and one partial ORF were identified. The gene arrangement was lbpB-lbpA-orf3, with a fourth partial ORF located downstream (Fig. 1). The coding sequence of the lbpB gene was approximately 2.7 kb, and putative promoter elements were identified upstream of lbpB (Fig. 2A). The potential 10 and ribosome binding site (RBS) sequences were more widely spaced than those found in E. coli consensus sequences. The separation of these two elements is greater in the Q8 lbpB sequence than in the 4223 lbpB sequence due to the presence of an extra 11 nucleotides (italicized in Fig. 2A). A putative Fur binding site was identified overlapping the 10 region of the lbpB promoter (Fig. 2B). The intergenic distance between lbpB and lbpA was 184 bp, and there was a second possible promoter region upstream of lbpA, which more closely resembles consensus E. coli promoters. The coding sequence of the lbpA gene was approximately 3.0 kb, and the intergenic distance between lbpA and the orf3 gene was only a single nucleotide. Possible consensus sequences for promoter elements were identified upstream of orf3, within the coding sequence of lbpA. The coding sequence of orf3 was 1.6 kb, and the intergenic sequence between orf3 and orf4 was 583 bp. Promoter elements were identified upstream of orf4.

View larger version (7K): [in this window] [in a new window]   FIG. 1.   Partial restriction map of the M. catarrhalis lactoferrin receptor locus. Restriction enzyme sites: Bg, BglI; Bs, BstEII; H, HindIII; R, EcoRI; S, SphI.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequences in the potential promoter regions of the strain Q8 lfr locus. (A) Potential promoter elements for the lbpB, lbpA, and orf3 genes. Potential 35, 10, and RBS sequences are underlined. The stop codon for lbpA is indicated by an asterisk. The italicized sequence in the promoter region of Q8 lbpB designates the 11 nucleotides missing from the 4223 lbpB sequence. (B) Comparison of Fur binding sequences in E. coli (14), N. meningitidis tbpB (19), H. influenzae tbpB (12), and M. catarrhalis lbpB. Dots indicate nucleotides identical to the E. coli consensus sequence. Underlining indicates dyad symmetry. The italicized sequence in the Fur binding region of Q8 lbpB is the 11 nucleotides missing from the 4223 lbpB sequence.

Analysis of the deduced amino acid sequences of the lactoferrin binding proteins. The M. catarrhalis Q8 and 4223 lbpA genes encode proteins of molecular mass 110.8 kDa that are 99% identical, with only seven different residues between them. Compared with known LbpA sequences from N. meningitidis (27, 28) and N. gonorrhoeae (1), there is about 32% sequence identity and 50% sequence similarity between the M. catarrhalis and the neisserial LbpA proteins (Fig. 3). The main differences between the M. catarrhalis and neisserial LbpA proteins are several small inserts, particularly in the N-terminal 80 amino acids and between residues 593 and 858. The deduced sequences of the LbpA proteins can be aligned with peptide sequences derived from purified native M. catarrhalis 141 LbpA (4). For the first peptide, 19 of 20 residues are identical, and for the second peptide, 16 of 16 residues are identical (Fig. 3).

View larger version (66K): [in this window] [in a new window]   FIG. 3.   Alignment of the amino acid sequences of LbpA from M. catarrhalis Q8 and 4223, N. meningitidis BNCV and H44/76 (27, 28), and N. gonorrhoeae FA19 (1). Dots indicate identical residues, and gaps have been introduced to maximize sequence alignments. The internal cyanogen bromide fragment used to design the oligonucleotide probe for cloning of the lbpA gene is underlined. The locations of the homologous peptide sequences obtained from M. catarrhalis 141 LbpA (4) are italicized and underlined.

View larger version (55K): [in this window] [in a new window]   FIG. 4.   Alignment of the amino acid sequences of LbpB from M. catarrhalis Q8, 4223, and VH19 and the partial carboxyl-terminal sequences of LbpB from N. meningitidis BNCV and H44/76 (27, 28) and N. gonorrhoeae FA19 (translated from Biswas and Sparling [1]). Dots indicate identical residues, and gaps have been introduced to maximize sequence alignments. The residues conserved with TbpB proteins (26) are underlined, and the RGD sequence is italicized.

View larger version (22K): [in this window] [in a new window]   FIG. 5.   Alignment of the amino acid sequences of the ORF3 protein from M. catarrhalis Q8 and 4223. Dots indicate identical residues, and gaps have been introduced to achieve maximum sequence alignment. The DGLG repeat sequence is underlined.

Expression of rLbpA and rLbpB from E. coli and protein purification. The lbpA and lbpB genes were cloned into plasmid pT7-7 (37), and the recombinant proteins were expressed in E. coli BL21(DE3) cells. Strain Q8 rLbpA was expressed at about 10% of total proteins as inclusion bodies, but the strain 4223 rLbpA protein was expressed at much lower yield and was not studied further. Two rLbpB proteins were expressed: one from the ATG start codon, which included the signal sequence, and the other as the mature protein starting from the cysteine residue. The mature LbpB proteins from strains Q8 and 4223 were made at 5 to 10% of total proteins as inclusion bodies, but the rLbpB proteins with the leader sequence were made in very low yield and were not studied further. The recombinant ORF3 protein could not be expressed in E. coli.

View larger version (41K): [in this window] [in a new window]   FIG. 6.   Purification of rLbpA and rLbpB proteins. (A) SDS-PAGE of the purification of Q8 rLbpA. Panels B and C show the purification of Q8 rLbpB and 4223 rLbpB, respectively. Lane 1, molecular weight markers; lane 2, whole-cell lysates; lane 3, inclusion bodies; lane 4, purified protein.

View larger version (41K): [in this window] [in a new window]   FIG. 7.   Lactoferrin binding of recombinant proteins. (A) SDS-PAGE of purified recombinant proteins. (B) Binding of recombinant proteins to human lactoferrin. (C) Binding of recombinant proteins to human transferrin. Lane 1, molecular weight markers; lane 2, Q8 rLbpA; lane 3, Q8 rLbpB; lane 4, 4223 rLbpB; lane 5, 4223 rTbpB.

Immunogenicity and antigenic conservation of LbpA and LbpB. Both rLbpA and rLbpB were found to be immunogenic. Immunoblot analysis of M. catarrhalis isolates showed that eight of eight strains examined expressed an approximately 105-kDa protein recognized by anti-rLbpA antibody, and all of the approximately 90 strains tested expressed a protein recognized by anti-rLbpB antibodies. Representative immunoblots are shown in Fig. 8. The M. catarrhalis LbpB proteins were surprisingly homogenous in molecular mass, with about 57% of strains expressing an LbpB protein that comigrated with LbpA and the remainder of the strains expressing a slightly smaller protein (Fig. 8A). Both the LbpA and LbpB proteins appeared to be expressed constitutively in M. catarrhalis, although an increase in expression was observed with iron restriction (Fig. 8B and C). There was also weak recognition, by anti-LbpB antibody, of approximately 85- to 90-kDa protein bands in some strains grown under iron-reduced conditions (Fig. 8C). The anti-rLbpA and anti-rLbpB antibody titers were measured by ELISA, and the anti-rLbpB titers were found to be very high (Table 1).

View larger version (22K): [in this window] [in a new window]   FIG. 8.   Immunoblot of M. catarrhalis strains reacted with anti-rLbpA and anti-LbpB antibodies. (A) Whole-cell lysates probed with anti-Q8 rLbpA plus anti-Q8 rLbpB antisera. All cells were grown in the presence of EDDA. (B) Whole-cell lysates probed with anti-Q8 rLbpA antibody. (C) Whole-cell lysates probed with anti-Q8 rLbpB antibody. Lane 1, strain Q8; lane 2, strain 4223; lane 3, strain VH19; lane 4, strain LES-1; lane 5, strain H-04; lane 6, strain 3. +, with EDDA; , without EDDA.

Bactericidal antibody activity. Bactericidal antibody assays were performed with guinea pig antisera. Two guinea pig antisera were tested against the autologous strain, and in each case they were found to be equivalent. Neither of the two guinea pig anti-Q8 rLbpA antisera killed strain Q8, even at antibody dilutions of only 1:8. When bactericidal activity was defined as 50% killing, a 1:64 dilution of anti-4223 rLbpB antiserum and a 1:16 dilution of anti-Q8 rLbpB antiserum were bactericidal against their autologous strains. A 1:32 dilution of anti-4223 rLbpB antiserum also killed strain Q8, and a 1:16 dilution of anti-Q8 rLbpB killed strain 4223. Anti-4223 rLbpB antiserum at a dilution of 1:64 was used to screen for bactericidal activity against four additional heterologous strains, VH-19, LES-1, H-04, and 3, and was found to kill three of them (Table 2).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial transferrin and lactoferrin receptors are heterodimeric complexes of proteins, TbpA-TbpB and LbpA-LbpB, known to be functionally and genetically related (13). In order to clone the M. catarrhalis lfr genes and not the tfr genes, a specific lbpA probe was generated. PCR primers were designed based upon an internal cyanogen bromide fragment of affinity-purified M. catarrhalis LbpA and the conserved carboxyl-terminal sequence LEMKF, thus far identified in all TbpA and LbpA proteins. By this approach, specific 2.2-kb lbpA gene fragments were amplified from strains 4223 and Q8 and these were used to probe the gene libraries and clone the complete lfr loci. The sense primer used to PCR amplify the 4223 and Q8 lbpA fragments had two codons missing (Y at position 6 and K at position 10) due to an error in transcribing the N-terminal sequence analysis report. The fact that a fragment of the lbpA gene could still be cloned was probably due to the two-step process of first generating a PCR fragment and then using that as the probe for the libraries.

The N. meningitidis and H. influenzae tfr operons are comprised of the tbpB and tbpA genes arranged in tandem with a single promoter region upstream of tbpB. The intergenic distance between the tbpB and tbpA genes ranges from 13 to 87 bp. Pettersson et al. (28) proposed that the N. meningitidis lactoferrin binding proteins may also be encoded on an operon having the gene arrangement lbpB-lbpA, especially since the lbpB and lbpA genes overlap. In M. catarrhalis, the lfr genes are arranged as tandem genes, with lbpB followed by lbpA at an intergenic distance of 184 bp. The putative lbpB promoter sequences have an unusually large separation between the putative 10 and RBS sequences, especially in the Q8 locus, which contains an extra 11 nucleotides in this region. Promoter elements can be readily identified upstream of M. catarrhalis lbpA, suggesting that lbpB and lbpA may be independently transcribed. Of particular interest in the cloned lfr locus is the presence of a third gene immediately downstream of lbpA, which is apparently unique to M. catarrhalis. Since the orf3 gene was cloned from two independent libraries, it is unlikely to be an experimental artifact. Potential promoter elements for orf3 can be identified within the lbpA gene. What role ORF3 may have, if any, in the lactoferrin receptor protein complex is unknown.

Expression of the tfr and lfr genes has been shown to be inducible under iron repression in vitro, a process thought to mimic the iron-restricted environment in the human host. From our data, there is a basal level of expression of the M. catarrhalis lfr genes observed in iron-sufficient medium, with an enhanced expression evident upon iron restriction. These data confirm the dot blot experiments of Schryvers and Lee, who showed that M. catarrhalis expressed low levels of transferrin and lactoferrin binding proteins under iron-sufficient growth conditions (33). The product of the ferric uptake regulation (fur) gene is thought to be responsible for this regulation of gene expression, and Fur binding sequences have been identified in the 10 region of the promoters for both the N. meningitidis and H. influenzae tbpB genes (12, 19). A potential Fur binding sequence was identified upstream of N. meningitidis lbpA; however, Pettersson et al. (28) were unable to demonstrate its functionality. In the case of N. meningitidis lfr, which is probably an operon, it seems likely that the Fur binding sequence is located upstream of lbpB, rather than lbpA, and will be identified once the complete N. meningitidis lbpB sequence is known. Compared with the consensus sequence for Fur binding sites (14), a homologous sequence can be identified in the 10 region of the M. catarrhalis strain Q8 lbpB promoter, but there is no obvious consensus sequence in the 10 region of the lbpA promoter. There are 11 nucleotides missing in the 4223 lbpB promoter region which are located within the putative Fur binding site of the Q8 lbpB promoter. The loss of these nucleotides in 4223 lbpB results in the loss of the Fur binding site and suggests another iron regulation mechanism for 4223 LbpB.

When the lactoferrin binding proteins from N. meningitidis, N. gonorrhoeae, and M. catarrhalis were first described, a single Lbp protein of an approximate molecular mass 105 kDa was identified (33). Subsequently, an 84-kDa protein isolated by low-stringency binding to lactoferrin was identified as LbpB (3). However, Bonnah et al. have recently demonstrated that the 84-kDa protein is CopB, and a 95-kDa protein has been identified as LbpB (4). Our data demonstrate clearly that, in some strains, the LbpA and LbpB proteins comigrate (Fig. 8A). The LbpA protein is quite homogeneous at about 105 kDa (Fig. 8B), and in 51 of the 90 strains examined, the LbpB protein comigrates with LbpA. In the remainder of the strains, the LbpB protein is apparently slightly smaller, but overall there is very little size heterogeneity for the M. catarrhalis LbpB proteins. This is in contrast to the TbpB proteins, which have been shown to be quite variable in size, ranging from about 68 to 88 kDa for N. meningitidis (30) and from about 60 to 90 kDa for H. influenzae (20), although less size heterogeneity was observed for the N. gonorrhoeae TbpB proteins, at 78 to 86 kDa (6).

The M. catarrhalis LbpA proteins were found to be 99% identical to each other. The N. meningitidis LbpA proteins from strains BNCV and H44/76 have been shown to be 95% identical to each other (27, 28) and 94% identical to the N. gonorrhoeae LbpA protein (1). Thus, as was previously found for the transferrin binding proteins, in which TbpA was highly conserved within a species, the LbpA proteins are also highly conserved. When compared with the neisserial LbpA proteins, there are several small inserts found in the M. catarrhalis LbpA proteins. Compared to the TbpA-LbpA topology model described by Gray-Owen and Schryvers (13), these inserts occur within the N-terminal periplasmic tail and extracellular loops 7, 9, 10, and 11.

Based upon the sequence variability of known TbpB proteins, it was expected that the LbpB proteins would show significant sequence variability; however, the three M. catarrhalis LbpB proteins showed surprising similarity, with strains Q8 and 4223 having 92% identical LbpB proteins. Strain Q8 was originally isolated from patient sputum in Montreal, Quebec, Canada; strain 4223 was isolated from middle ear fluid from a patient in Buffalo, N.Y.; and strain VH19 was isolated from middle ear fluid from a patient in Galveston, Tex. Strains Q8 and 4223 are also phenotypically distinct (39). The M. catarrhalis LbpB proteins show limited homology with the partial sequences of the putative N. meningitidis LbpB proteins. There is also very little homology with the known TbpB proteins, with the exception of short scattered sequences, the most notable being NRFVG at positions 390 to 394 and LEGGFYG at positions 430 to 436 (25). The conservation of these scattered residues (underlined) in bacterial TbpB proteins and in M. catarrhalis LbpB suggests that they may play a functional role in these iron-binding molecules. There is an unusually high content of Asp and Asn residues in LbpB, with a region of 50 residues that is ~54% Asp and another region of 26 residues that is ~42% Asn. The purpose of such concentrations of identical residues is unknown, but the fact that they are conserved among the three encoded LbpB proteins in this study suggests that they serve some function. Another unique feature of the M. catarrhalis LbpB proteins is the presence of a conserved RGD motif. This sequence is well established as a site for attachment of bacteria to eukaryotic cells (31), suggesting that the M. catarrhalis LbpB protein may act as an adhesin. An RGD motif has not been identified in any of the published TbpB sequences, and it will be interesting to see whether it is present in LbpB proteins from other species.

When the H. influenzae Rd genome was sequenced, it was found that there were several copies of transferrin or lactoferrin binding-like proteins (11). To demonstrate that we had indeed cloned the M. catarrhalis lfr genes and not a variant of the tfr genes, we tested the binding of the recombinant Lbp proteins to human lactoferrin. As demonstrated for transferrin binding proteins where only the TbpB protein binds to human transferrin after gel electrophoresis and electroblotting, only the LbpB, not the LbpA protein, specifically bound human lactoferrin. Since lactoferrin is known to be a sticky molecule, we also demonstrated that rTbpB did not bind to human lactoferrin under the same conditions. In addition, a corollary experiment was performed in which it was shown that only rTbpB, not rLbpA or rLbpB, bound human transferrin. Finally, the internal peptide sequences identified by Bonnah et al. (4) from M. catarrhalis 141 LbpA can be found in our sequences (Fig. 3). These data clearly demonstrate that we have cloned the lbpA and lbpB genes of M. catarrhalis.

The most unique finding in the M. catarrhalis lfr locus is the presence of the third gene, orf3. The putative ORF 3 protein has no homology to known proteins in the databases, and it contains an internal repeat of the tetramer DGLG. Such repeats sometimes represent phenotypic switches used to regulate virulence factors (35). We had hoped to be able to generate anti-rORF3 antibodies in order to determine whether ORF3 is expressed in M. catarrhalis, but we were unable to express the recombinant protein.

Guinea pigs immunized with purified rLbpA or rLbpB proteins elicited high-titer antibodies. There is no animal model for otitis media caused by M. catarrhalis, but a bactericidal antibody assay has been established (39). The clumping nature of M. catarrhalis strains makes this assay difficult to perform, so the data is only qualitative, not quantitative. The anti-Q8 rLbpA antibody was not bactericidal against its autologous strain. Since native LbpA protein is a transmembrane protein, it is possible that antibody raised to inclusion body-derived rLbpA protein would not recognize the native protein in intact organisms. However, in whole cell ELISAs, it was demonstrated that both anti-rLbpA and anti-rLbpB antisera recognized intact cells at titers ranging from 400 to 1,600 (data not shown). The anti-rLbpB antisera were weakly bactericidal, although the anti-4223 rLbpB antiserum appeared to be slightly more potent than the anti-Q8 rLbpB antiserum against their autologous strains. Heterologous strains were screened with a 1:64 dilution of anti-4223 rLbpB antiserum, and an arbitrary cutoff of 50% killing was defined as bactericidal activity. The heterologous strains that were tested were chosen based upon the diversity of their geographic origins and the inclusion of a mixture of molecular masses and anatomical sources. The data in Table 2 show that anti-4223 rLbpB antiserum was able to kill three of five heterologous strains by this stringent definition. There does not appear to be any correlation between the antibacterial activity of anti-4223 rLbpB and any other factor.

In this study, we have characterized the genes of the M. catarrhalis lfr locus and found that there are three closely spaced genes encoding conserved proteins. The lbpA and lbpB genes show some homology to other bacterial lbpA, tbpA, and tbpB genes, but the nature and function of the third gene is unknown. Recombinant LbpA and LbpB proteins were produced as inclusion bodies, and the purified proteins were used to generate high-titer antibodies. The anti-rLbpA antibody was not bactericidal, but the anti-rLbpB antibodies were bactericidal for autologous and heterologous strains of M. catarrhalis. Thus, rLbpB proteins may represent candidate vaccine antigens to protect against M. catarrhalis disease.