Abstract
We have identified a homologue of the adhesin AIDA-I of Escherichia coli in Neisseria meningitidis. This gene was designated nhhA (Neisseria hiahomologue), as analysis of the complete coding sequence revealed that it is more closely related to the adhesins Hia and Hsf of Haemophilus influenzae. The sequence of nhhA was determined from 10 strains, and found to be highly conserved. Studies of the localisation by Western immunoblot analysis of total cell proteins and outer membrane complex preparations and by immunogold electron microscopy revealed that NhhA is located in the outer membrane. A strain survey showed that nhhA is present in 85/85 strains of N. meningitidis representative of all the major disease-associated serogroups, based on Southern blot analysis. It is expressed in the majority of strains tested by Western immunoblot.
1 Introduction
Neisseria meningitidis is a Gram-negative bacterium which is the causative agent of meningococcal meningitis and septicaemia. Its only known host is the human, and it may be carried asymptomatically by approximately 10% of the population [1]. The factors influencing transition from asymptomatic carriage to invasive disease have not been completely elucidated, but are thought to include expression of adhesins. We previously reported the sequence of NhhA in N. meningitidis[2], a homologue of the adhesins AIDA-I of Escherichia coli[3] and Hia and Hsf of Haemophilus influenzae[4,5]. In this study we investigated its expression, cellular localisation and sequence variation.
2 Materials and methods
2.1 Bacterial strains
Strains were obtained from: World Health Organisation Collaborating Centre for Reference and Research on Meningococci, Oslo, Norway; The Public Health Laboratory Service Meningococcal Reference Laboratory, Manchester, UK; and Queensland Hospitals, except MC58, and ¢3 (an acapsulate mutant of MC58 [6]) and their respective nhhA mutants, MC58H:Kan and ¢3:2A (this study). E. coli strains were grown in liquid medium with shaking, or on solid L medium (1% agar w/v) at 37°C. Antibiotics (ampicillin or kanamycin) were added to liquid or solid media for selective growth where appropriate at a concentration of 100 µg ml−1
2.2 Nucleic acid manipulation
The expression construct pNhhAMBP was constructed by amplifying the nhhA gene from strain ¢3 using oligonucleotide primers NhhA-MBPA, 5′-GGTCGCGGATCCATGAACAAAATATACCGCAT-3′ and NhhA-MBPB, 5′-TCACCCAAGCTTAAGCCCTTACCACTGATAAC-3′ and cloning into BamHI/HindIII restriction-digested plasmid pMALC2 (NEB). Plasmid pNhhA:Kan (for inactivation of nhhA) was created by insertion of a BamHI KanR cassette from pUC4Kan into the unique BglII site of nhhA. N. meningitidis was transformed by incubating bacteria from overnight culture with plasmid pNhhA:Kan (linearised with EcoRI), or genomic DNA on solid media for 3–5 h before selection for transformants by growth on BHI plates containing 100 µg kanamycin ml−1. Strains MC58H:Kan, ¢3:2A, and H41H:Kan are nhhA::kan mutants of strains MC58, ¢3 and H41 respectively.
2.3 Production of rabbit anti-NhhA sera
Cell extracts from induced cultures containing pNhhAMBP were separated by SDS–PAGE, and the fusion protein was partially purified by elution using the Mini-Gel Electro-eluter (Bio-Rad). The purity of the NhhA–MBP fusion protein was determined by SDS–PAGE followed by Coomassie staining and Western immunoblotting using rabbit anti-MBP sera (NEB). Samples of eluted NhhA–MBP fusion protein were dialysed against sterile phosphate-buffered saline (PBS) pH 7.4, and mixed with adjuvant (MPL+TDM+CWS, Sigma), at a concentration of 50–150 µg protein ml−1 and inoculated at two weekly intervals into two New Zealand White rabbits. Serum was extracted and stored in aliquots at −80°C. For Western immunoblot, non-NhhA-specific antibodies were removed by absorbing sera against E. coli expressing MBP [7]. Bactericidal killing assays were performed essentially as previously described by Hoogerhout et al. [8].
2.4 Immunogold electron microscopy
Bacteria were grown overnight then subcultured for 5 h on solid media, resuspended in PBS, transferred to carbon-coated copper grids, and incubated at room temperature with 50 µl PBS containing fish skin gelatin, bovine serum albumin and glycine, before incubation in 10 µl PBS or absorbed rabbit sera diluted 1:5 in PBS. Grids were washed in PBS and incubated with gold-conjugated goat anti-rabbit antibody then washed in PBS before fixing in 1% glutaraldehyde (v/v). For each experiment, 100 diplococci were examined per grid, with operators blinded as to which grid was being examined. The number of gold particles adhering to diplococci was counted, and analysed by likelihood ratio χ2 test.
3 Results and discussion
3.1 Similarity to other bacterial proteins
The nhhA gene (Neisseria hiahomologue) was first identified because of its similarity to AIDA-I of E. coli[2]. NhhA exhibits 47% similarity to AIDA-I of E. coli[3], but is more similar to the Hia protein (67%) and its allelic variant Hsf (74%) from H. influenzae[4,5]. Although NhhA is similar to Hia and Hsf, their sizes are markedly different: NhhA is 589–599 aa long whereas Hia and Hsf are 1098 and 2353 aa long respectively (see Fig. 1 and [5]). The size difference between Hia and Hsf is partly explained by the presence in Hsf of three copies of a region of approximately 400 amino acids which is present only once in Hia (Fig. 1 and [5]). Hia and Hsf have a highly conserved carboxy-terminal domain of approximately 440 aa. NhhA shares the highly conserved amino-terminal 50 aa (predicted to be a signal peptide) and portions of the conserved repeat region (thought to be associated with adhesin function), and carboxy-terminal region (Fig. 1).
A schematic representation of similarity of NhhA, Hsf, and Hia. Signal sequences are shown by a black box, the conserved C-terminal regions by vertical shading, and the ‘Hsf repeat element’ by diagonal shading. An additional region of similarity between NhhA and Hsf is shown by hatching.
A schematic representation of similarity of NhhA, Hsf, and Hia. Signal sequences are shown by a black box, the conserved C-terminal regions by vertical shading, and the ‘Hsf repeat element’ by diagonal shading. An additional region of similarity between NhhA and Hsf is shown by hatching.
NhhA also exhibits similarity to several other bacterial outer membrane proteins, including HMW1, UspA1 (a high molecular mass protein of Moraxella catarrhalis), and SepA (involved in tissue invasion of Shigella flexneri) [9–11], all of which are members of the autotransporter family of outer membrane proteins. Autotransporters are a family of proteins which share a highly conserved signal sequence and conserved carboxy-domain which forms a transmembrane domain. The mature amino-domain exhibits little sequence homology between members of the family, and has diverse functions, many of which are involved in virulence (see recent reviews [12,13]). NhhA fulfills several of the criteria for inclusion in this family including an unusually long signal sequence, a paucity of cysteine residues, and a consensus sequence at the C-terminus terminating with an aromatic amino acid, usually phenylalanine or tryptophan [14].
The similarity between NhhA and these other autotransporters is restricted to the predicted signal sequence and consensus C-terminus. The degree of similarity over the entire length between NhhA and Hia and Hsf of H. influenzae suggests that there may be functional conservation and that NhhA represents a novel adhesin of N. meningitidis.
3.2 Sequence variation of nhhA between strains
The nhhA gene was sequenced from 10 strains to examine the degree of variation of this gene. These sequences and that of nhhAZ2491 (from the group A strain Z2491) were compared and showed 89.5–99.8% nucleotide sequence identity and 85.3–99.8% amino acid identity between strains. The nhhA open reading frame ranges from 1770 to 1800 bp in size between strains, encoding predicted proteins of 589–599 aa. Comparison of the predicted polypeptides of NhhA demonstrated that the majority of the sequence diversity was confined to the region aa 51–250 of the alignment, i.e. the first 200 aa of the mature protein. Outside this region, 93% of amino acids are conserved in all strains. Even within the region aa 51–250, there are conserved and variable stretches (see Fig. 2A). A comparison of the nucleotide sequence of nhhA between strains indicates that the variable region is mosaic in structure (Fig. 2B), which implies that horizontal exchange has occurred between alleles of this gene between strains. For example, nhhA from strains Z2491 and H41 are identical in the first 585 nucleotides (to the end of the C3 region), but only 86% identical to strain H15 over the same region. The H41 sequence is then 100% identical to H15 sequence to the end of the V4 region, whereas nhhA of Z2491 shares only 56% identity with nhhAH41 in the same region. Conversely, the P20 nhhA sequence differs significantly from nhhAZ2491 up to the end of the C3 region, and is then identical to nhhAZ2491. A further example, nhhAH38, is also shown in Fig. 2B.
Sequence conservation of NhhA. A: Schematic representation of NhhA, showing conserved (C1–4) and variable regions (V1–4) in the NH2-terminal 250 amino acids. The remaining approximately 350-amino acid C5 region is shown truncated. Numbers indicate percentage of fully conserved aa residues. B: Schematic representation of mosaic structure of nhhA of strains H15, H41, Z2491, P20 and H38. Shading indicates regions of common origin between strains (see text).
Sequence conservation of NhhA. A: Schematic representation of NhhA, showing conserved (C1–4) and variable regions (V1–4) in the NH2-terminal 250 amino acids. The remaining approximately 350-amino acid C5 region is shown truncated. Numbers indicate percentage of fully conserved aa residues. B: Schematic representation of mosaic structure of nhhA of strains H15, H41, Z2491, P20 and H38. Shading indicates regions of common origin between strains (see text).
The strains used in this study included 45 serogroup B strains for which there are considerable phylogenetic data, based on multi-locus sequence type (MLST [15]) which demonstrated the considerable genetic diversity of these strains. Our phylogenetic analysis of the sequence variation of nhhA from 11 strains (not shown) indicates that the differences do not reflect the phylogeny based on housekeeping genes. For example, nhhA from strains BZ10 and H15 is 99.7% identical, yet these strains are distantly related genetically by MLST analysis, representing types 8 and 43 respectively. Conversely, nhhA is more dissimilar (92.9%) from the more closely related strains BZ10 and NGP20 (MLST types 8 and 11 respectively). This is further evidence for recombination of this allele between strains.
3.3 Presence of nhhA homologues in N. meningitidis
A total of 85 strains were examined by Southern blot analysis for the presence of nhhA. These included: four strains of serogroup A, 60 strains of serogroup B, 12 strains of serogroup C, two strains of serogroup W, two strains of serogroup X, three strains of serogroup Y, and one strain of serogroup Z. All strains produced a hybridisation signal under stringent conditions (typical results are shown in Fig. 3), suggesting that there is strong selection pressure to retain nhhA. In addition, hybridisation signals were detected at high stringency in N. lactamica (two strains), but not in N. animalis, N. cinerea, N. elongata, N. gonorrhoeae (two strains), N. pharyngis, N. polysaccharea, or N. subflava (not shown). Consistent with this, homologues of nhhA were not detected in data from the project to sequence N. gonorrhoeae.
Southern blot analysis of genomic DNA of N. meningitidis strains, digested with ClaI, and hybridised with a probe consisting of nhhA of strain ¢3. A: Serogroup B strains: lane 1 PMC28, lane 2 PMC27, lane 3 PMC25, lane 4 PMC24, lane 5 PMC16, lane 6 PMC13, lane 7 PMC12, lane 8 size standards, lane 9 2970, lane 10, 1000, lane 11 528, lane 12 SWZ107, lane 13 H41, lane 14 H38, lane 15 NGH36, lane 16, H15, lane 17, NGG40, lane 18 NgF26, lane 19 NHE30, lane 20 NGE28. B: Non-serogroup B strains: lane 1 PMC3 (serogroup A), lane 2 PMC17 (A), lane 3 PMC20 (A), lane 4 PMC23 (A), lane 5 PMC8 (C), lane 6 PMC9 (C), lane 7 PMC11 (C), lane 8 PMC14 (C), lane 9 PMC18 (C), lane 10 PMC21 (C), lane 11 PMC 29 (C), lane 12 size standards, lane 13 PMC19 (W), lane 14 PMC1 (X), lane 15 PMC6 (X), lane 16 PMC10 (Y), lane 17 PMC22 (Y), lane 18, PMC26 (Y), lane 19 PMC2 (Z/29E).
Southern blot analysis of genomic DNA of N. meningitidis strains, digested with ClaI, and hybridised with a probe consisting of nhhA of strain ¢3. A: Serogroup B strains: lane 1 PMC28, lane 2 PMC27, lane 3 PMC25, lane 4 PMC24, lane 5 PMC16, lane 6 PMC13, lane 7 PMC12, lane 8 size standards, lane 9 2970, lane 10, 1000, lane 11 528, lane 12 SWZ107, lane 13 H41, lane 14 H38, lane 15 NGH36, lane 16, H15, lane 17, NGG40, lane 18 NgF26, lane 19 NHE30, lane 20 NGE28. B: Non-serogroup B strains: lane 1 PMC3 (serogroup A), lane 2 PMC17 (A), lane 3 PMC20 (A), lane 4 PMC23 (A), lane 5 PMC8 (C), lane 6 PMC9 (C), lane 7 PMC11 (C), lane 8 PMC14 (C), lane 9 PMC18 (C), lane 10 PMC21 (C), lane 11 PMC 29 (C), lane 12 size standards, lane 13 PMC19 (W), lane 14 PMC1 (X), lane 15 PMC6 (X), lane 16 PMC10 (Y), lane 17 PMC22 (Y), lane 18, PMC26 (Y), lane 19 PMC2 (Z/29E).
3.4 Expression and localisation of NhhA
The expression of NhhA was compared in strains ¢3 (wild-type with respect to nhhA) and ¢3:2A (nhhA::kan) using NhhA-specific rabbit polyclonal antisera. In addition, to determine whether NhhA is an outer membrane protein (as suggested by its similarity to other outer membrane proteins), outer membrane complexes (omc) were prepared and analysed by SDS–PAGE (Fig. 4). Western immunoblotting revealed the presence of an immunoreactive protein in ¢3 that was absent from strain ¢3:2A (Fig. 4C, lanes 1 and 2), confirming that NhhA is expressed in this strain. Mature NhhA has a predicted molecular mass of approximately 56 600, but the immunoreactive protein of ¢3 is greater than 200 kDa. This high molecular mass immunoreactive band may represent multimers or other complexes that are stable with respect to heat, SDS and reduction with β-mercaptoethanol (it is noteworthy in this regard that NhhA contains no cysteine residues). The identity of the high molecular mass immunoreactive band as NhhA was confirmed by loss of this band in 5/5 strains in which the nhhA::kan mutation was made (two of these mutants are shown in Fig. 6). Anomalous migration has also been observed for at least three of the similar autotransporter proteins: the high molecular mass outer membrane proteins UspA1 and UspA2 of Moraxella catarrhalis have predicted molecular masses of 62 500 and 88 300 respectively [10] but migrate at an apparent size of between 350 000 and 720 000 as the UspA complex [16]; and Hia of H. influenzae has a predicted molecular mass of 116 000 but when expressed in E. coli, Hia migrates at greater than 200 000, and is present as high molecular mass multimers in H. influenzae strain 11 [5].
Expression and localisation of NhhA. Whole cells and omcs were separated electrophoretically on 4–15% SDS–PAGE gradient gel. A: Coomassie stain. B: Same gel silver-stained. C: Western immunoblot using NhhA-specific rabbit polyclonal serum. Lane 1, ¢3, 100 µg total protein (whole cell sonicates); lane 2, ¢3:2A, 100 µg total protein; lane 3, ¢3, 7.5 µg omc; lane 4, ¢3:2A, 7.5 µg omc; lane S, prestained size standards. Approximate sizes are indicated in kDa.
Expression and localisation of NhhA. Whole cells and omcs were separated electrophoretically on 4–15% SDS–PAGE gradient gel. A: Coomassie stain. B: Same gel silver-stained. C: Western immunoblot using NhhA-specific rabbit polyclonal serum. Lane 1, ¢3, 100 µg total protein (whole cell sonicates); lane 2, ¢3:2A, 100 µg total protein; lane 3, ¢3, 7.5 µg omc; lane 4, ¢3:2A, 7.5 µg omc; lane S, prestained size standards. Approximate sizes are indicated in kDa.
Expression of NhhA in a range of strains of N. meningitidis. Anti-NhhA rabbit polyclonal serum was used to probe membranes of N. meningitidis total cell protein separated by 10% SDS–PAGE. Lane 1 MC58, lane 2 MC58H:Kan, lane 3 H41, lane 4 H41H:Kan, lane 5 BZ198, lane 6 EG327, lane 7 EG329, lane 8 H15, lane 9 H38, lane 10 P20, lane 11 BZ10, lane 12 PMC21. Migration of 187-kDa standard protein is indicated.
Expression of NhhA in a range of strains of N. meningitidis. Anti-NhhA rabbit polyclonal serum was used to probe membranes of N. meningitidis total cell protein separated by 10% SDS–PAGE. Lane 1 MC58, lane 2 MC58H:Kan, lane 3 H41, lane 4 H41H:Kan, lane 5 BZ198, lane 6 EG327, lane 7 EG329, lane 8 H15, lane 9 H38, lane 10 P20, lane 11 BZ10, lane 12 PMC21. Migration of 187-kDa standard protein is indicated.
The Coomassie- and silver-stained gels (Fig. 4A,B) indicate that the omc preparation contained three major proteins, and a number of minor proteins, suggesting that this preparation was enriched for outer membrane proteins. NhhA could not be identified (at these loadings) by Coomassie or silver staining suggesting either that it is expressed at low levels, or that it is insensitive to staining by these methods. NhhA could be detected in Western immunoblots using NhhA-specific sera at similar intensity in both total protein and omc preparations indicating that NhhA is co-purified with other outer membrane proteins using this method. This is consistent with the proposed outer membrane localisation of this protein. Isolation of the sarcosyl-insoluble fraction is commonly used to enrich for outer membrane components, and this method has been shown to produce protein profiles closely resembling those obtained by 125I labelling of surface-exposed proteins [17].
Attempts were made to resolve the high molecular mass immunoreactive protein to its predicted subunit molecular mass of approximately 56 000 using formic acid treatment of omc samples (used to resolve the UspA complex of M. catarrhalis[16]). We were not able to demonstrate the presence of NhhA migrating at its predicted molecular mass by this method in strain ¢3 (not shown).
To confirm that the NhhA-specific sera recognised surface-exposed epitopes, the labelling of bacteria was examined by immunogold electron microscopy (Fig. 5). We examined ¢3 and its nhhA mutant ¢3:2A in two independent experiments and statistical analysis using the likelihood ratio χ2 test indicated in both experiments that ¢3 bound significantly more gold particles than its nhhA mutant ¢3:2A (P>0.001). These results indicate that NhhA is accessible to antibodies, which is consistent with it being surface-exposed. In addition, bactericidal assays revealed that the rabbit antisera raised against the recombinant NhhA–MBP fusion contained NhhA-specific bactericidal antibodies, as killing of NhhA wild-type was four-fold higher than NhhA mutant strains. After submission of this manuscript, another study was reported showing that mouse polyclonal sera raised against recombinant NhhA demonstrated bactericidal activity [18]. In addition, anti-NhhA antibodies can be detected in sera from patients convalescing from meningococcal disease [19], indicating that NhhA is immunogenic in humans.
Immunogold electron microscopy results. Grids were incubated with NhhA-specific absorbed rabbit sera followed by goat anti-rabbit-conjugated gold beads. Strains ¢3 (wild-type with respect to NhhA expression) and ¢3:2A (nhhA::kan) were examined. Each point plotted represents the number of gold particles on a single diplococcus. Controls included bacteria incubated with second antibody alone, and all these bound less than two gold particles/diplococcus (not shown).
Immunogold electron microscopy results. Grids were incubated with NhhA-specific absorbed rabbit sera followed by goat anti-rabbit-conjugated gold beads. Strains ¢3 (wild-type with respect to NhhA expression) and ¢3:2A (nhhA::kan) were examined. Each point plotted represents the number of gold particles on a single diplococcus. Controls included bacteria incubated with second antibody alone, and all these bound less than two gold particles/diplococcus (not shown).
3.5 Expression of NhhA in other strains
Having established that nhhA is present in all strains, we investigated expression of its gene product, NhhA, in a subset of these strains from which the nhhA gene had been partially or fully sequenced. Total cellular protein was Western-immunoblotted using the anti-NhhA rabbit sera. We detected NhhA in 15/19 strains examined by Western immunoblot (typical result shown in Fig. 6). Further sequence analysis of two serogroup B strains (NG6/88 and NGF26, which are closely related strains of MLST 13 and 14 respectively), in which NhhA could not be detected using the rabbit anti-NhhA antisera, revealed that they contain a 14-bp deletion in the predicted signal peptide region (not shown), leading to a premature stop codon, and presumably lack of expression. In other reports, 31/31 and 7/7 strains of N. meningitidis were shown to express NhhA [18,19].
We conclude that nhhA encodes a novel outer membrane protein based on (i) sequence similarity to a number of bacterial outer membrane proteins, (ii) enrichment in omcs and (iii) localisation by electron microscopy. The nhhA gene is present in 85/85 of N. meningitidis strains surveyed, and the NhhA protein is expressed in the majority of strains examined. In addition, this work and other studies show that recombinant NhhA is immunogenic and elicits bactericidal antibodies, and that convalescent sera recognise recombinant NhhA.
Acknowledgements
Work at the University of Queensland is supported by grants from the World Health Organisation (V23/181/115) and NHMRC (99/38007). E.R.M. is supported by a Program Grant from the Medical Research Council. We thank Dr Rick Webb for excellent electron microscopy studies, and Joan Hendrikz for statistical analysis. NhhA sequence of strain Z2491: data produced by the Neisseria meningitidis Sequencing Group at the Sanger Centre which can be obtained from ftp://ftp.sanger.ac.uk/pub/pathogens/nm/ and N. gonorrhoeae sequence data were obtained from Gonococcal Genome Sequencing Project, and are available at http://www.genome.ou.edu/gono.html.
