Neisseria meningitidis NhhA is a multifunctional trimeric autotransporter adhesin


*E-mail; Tel. (+39) 577 24 3414; Fax (+39) 577 27 8508.


NhhA, Neisseriahia/hsf homologue, or GNA0992, is an oligomeric outer membrane protein of Neisseria meningitidis, recently included in the family of trimeric autotransporter adhesins. In this study we present the structural and functional characterization of this protein. By expressing in Escherichia coli the full-length gene, deletion mutants and chimeric proteins of NhhA, we demonstrated that the last 72 C-terminal residues are able to allow trimerization and localization of the N-terminal protein domain to the bacterial surface. In addition, we investigated on the possible role of NhhA in bacterial–host interaction events. We assessed in vitro the ability of recombinant purified NhhA to bind human epithelial cells as well as laminin and heparan sulphate. Furthermore, we shown that E. coli strain expressing NhhA was able to adhere to epithelial cells, and observed a reduced adherence in a meningococcal isogenic MC58ΔNhhA mutant. We concluded that this protein is a multifunctional adhesin, able to promote the bacterial adhesion to host cells and extracellular matrix components. Collectively, our results underline a putative role of NhhA in meningococcal pathogenesis and ascertain its structural and functional belonging to the emerging group of bacterial autotransporter adhesins with trimeric architecture.


Neisseria meningitidis is a capsulated Gram-negative bacterium responsible of significant morbidity and mortality in humans, being the most common cause of pyogenic meningitis among children and young adults. Properties associated with pathogenesis of N. meningitidis include production of specialized surface structures, such as pili (Virji et al., 1991), lipooligosaccharide, and a number of outer membrane (OM) proteins, such as Opc and Opa opacity proteins (Virji et al., 1992; 1993), App and NadA (Virji et al., 1992; 1993; Comanducci et al., 2002; Serruto et al., 2003; Capecchi et al., 2005), mediating interaction of meningococcus to host cells.

In our efforts to discover novel surface-exposed proteins to be exploited in vaccine development, we have previously identified GNA0992, an OM protein able to induce a bactericidal antibody response (Pizza et al., 2000).

GNA0992, also termed NhhA (Neisseriahia/hsf homologue) (Peak et al., 2000) due to its sequence similarity to Hsf and Hia adhesins of Haemophilus influenzae (Barenkamp and St Geme, 1996; St Geme et al., 1996; 1998), is an oligomeric protein expressed in the vast majority of disease-associated strains (Peak et al., 2000; Pizza et al., 2000).

NhhA has been recently included in a new class of autotransporters (ATs) where a short translocation domain is able to trimerize so generating a complete β-barrel that directs the passenger secretion (Roggenkamp et al., 2003; Surana et al., 2004).

Such short translocation unit has been characterized in YadA adhesin of Yersinia spp. by Roggenkamp and coworkers (Roggenkamp et al., 2003). In their studies, the Authors showed that last 91 C-terminal residues of YadA formed the protein translocation unit. The extreme C-terminus of the protein folded into the β-domain, in which secondary structure predictions suggested the presence of four β-strands. The β-domain was preceded by a linking region having an helix secondary structure propensity. Mutagenesis experiments demonstrated that β-domain was responsible for the protein oligomerization and insertion into the OM, and that within linking region a proximal surface exposed (L2) and a transmembrane segment (L1) were recognizable, both required for the translocation of the passenger domain.

More recently, a shorter translocation unit consisting of 76 amino acids has been identified in Hia of H. influenzae (Surana et al., 2004).

These reported data led to the definition of a super family of trimeric ATs, which includes the YadA-like proteins as well as Hia/Hsf and NhhA, typified by a conserved short translocator domain whose trimerization is necessary to form a functional β-barrel. Members of the trimeric AT family are heterogeneous in their N-terminal domains,which are involved in adherence to a variety of substrates, including the extracellular matrix (ECM) components and host cell receptors, and participate in a more intimate step of attachment to the surface of epithelial cells. (Cotter et al., 2005).

In silico analysis of the NhhA amino acid sequence allowed so far to speculate about its possible role during meningococcal infection. The sequence similarity to Hia/Hsf, responsible of the formation of adhesive fibrils in H. influenzae, suggested the ability of NhhA to adhere to epithelial cells. Moreover, the presence, within the N-terminal domain, of a repeated amino acid motif common to eukaryotic adhesion molecules led to the prediction of putative binding sites to heparan sulphate (Peak et al., 2000; Scarselli et al., 2001).

In this study we explored the functional role of the different domains of NhhA. We defined the C-terminal translocation unit by analysing the ability of deletion mutants to export homologous and heterologous passenger domains. Our results indicate that last 72 residues of NhhA, corresponding solely to β and L1 segments, are able to trimerize and translocate the N-terminal portion of the protein to the bacterium surface.

Furthermore, we demonstrated the adhesive capabilities of NhhA, in particular its role in the interaction with epithelial cells and ECM components, such as laminin and heparan sulphate, providing additional insight into the mechanisms by which N. meningitidis colonizes the human respiratory tract.


Functional mapping of NhhA and design of mutants

In the 2996 strain of N. meningitis NhhA is encoded by a gene of 1797 nucleotides. The polypeptide precursor consists of 598 amino acids and contains all the three features characteristic of ATs (Fig. 1A). The first is a 51-amino acid long leader peptide (Henderson et al., 2004). The second is an N-terminal passenger domain, homologous to Hia/Hsf (residues 52–509), which is predicted to directly interact with the host environment.

Figure 1.

In silico analysis of NhhA protein.
A. Schematic illustration of the domain organization of NhhA in N. meningitidis strain 2996. Region 1–51 corresponds to the predicted autotransporter's leader peptide. Region 51–163 shows 81% of identity to Hsf but no significant similarity to Hia. The following region, corresponding to residues 164–380, shows 69,3% of identity to Hia and a variable similarity degree, from 53.5 to 66.2% of identity, to different portions of Hsf. As already pointed out, the tract 164–380 of NhhA contains two modules called R1 (residues 164–255) and R2 (residues 258–356), each holding a predicted binding site for heparan-sulphate (Scarselli et al., 2001). Within the region 381–509 the identity to Hia and Hsf falls to 23,8% and 30.1%, respectively, suggesting that this zone presumably contributes to the functional specificity of NhhA. The C-terminal region 510–598 corresponds to the conserved translocation domain. Positions of β (green), L1 (orange) and L2 (red) subdomains (see text for details) are indicated.
B. Multiple sequence alignment of the translocation unit. Identical residues are white on black background, conservative substitutions are enclosed in boxes. L2, L1 and β subdomains of NhhA (N. meningitidis, GenBank Accession number 7228548) are aligned to the regions of Hia (H. influenzae, GenBank Accession number 21536216), NadA (N. meningitidis, GenBank Accession number 21427183) Omp100 (Actinobacillus actinomycetemcomitans GenBank Accession number 19911154), YadA (Yersinia enterocolitica, GenBank Accession number 14579377), DsrA (Haemophilus ducreyi GenBank Accession number 7188589), UspA1 and UspA2 (Moraxella catarrhalis, GenBank Accession number 26284390 and 28866923 respectively). Upon the alignment, secondary structure element predicted with PHD algorithm (Rost and Sander, 1993) are represented as arrows (beta strands) and helices (alpha helix).

The third is the C-terminal domain (residues 510–598), putatively responsible for passenger secretion. Multiple sequence alignment shows indeed that the translocation unit proposed for YadA (Roggenkamp et al., 2003), consisting of β, L1 and L2 subdomains, is well conserved with NhhA and other members of the trimeric AT adhesin family (Fig. 1B).The secondary structure prediction of this region suggests the presence of four β-strands preceded by a α-helix. This architecture, typical of the trimeric AT proteins, suggests that the trimerization of this region could produce a 12-stranded beta barrel, in analogy to all the members of the trimeric AT family (Cotter et al., 2005).

Given the sequence similarity between NhhA and YadA C-terminal regions, we speculated that the partition of the linking region into L1 and L2 was applicable also to Nhha. Therefore, we split the 510–598 region of NhhA into β (545–598), L1 (527–544) and L2 (510–526) fragments (see Fig. 1). In order to identify the minimal C-terminal tract of NhhA that might efficiently export the N-terminal passenger domain, we expressed in Escherichia coli, under the control of T7 promoter, the full-length NhhA as well as different deletions of its C-terminus (namely NhhAΔL2, NhhAΔL1 and NhhAΔL2L1, see Fig. 2A) and investigated the secretion ability of the different proteins.

Figure 2.

Surface localization of NhhA passenger domain in NhhA and NhhA-deletions expressed in E. coli BL21(DE3).
A. Schematic representation of NhhA (2996 strain) and NhhA deletions constructed in this study. LP corresponds to N. meningitidis 2996 App leader peptide (residues 1–42).
B. FACS analysis of NhhA on whole cell bacteria performed with a mouse serum against purified 6× His-tagged NhhA (NhhA-His). E. coli BL21(DE3)/pET (empty vector) was used as negative control.
C. Western blot analysis of outer membrane proteins from recombinant E. coli strains using a mouse serum raised against NhhA-His. The band corresponding to NhhA trimers is indicated by a solid arrow. The heterogeneous population of dimeric forms is also indicated by the asterisk.

Preliminary attempts to express NhhA as full-length failed because its leader peptide was not functional in E. coli (data not shown). Therefore we decided to express NhhA, as well as its deletion mutants, using the leader peptide of the App protein, which was proven to be recognized by the secretion machinery of E. coli (Serruto et al., 2003).

The C-terminal L1β region of NhhA is able to trimerize and export the passenger domain(s)

We evaluated the export of the passenger domain by analysing the presence of the full-length and the different truncated forms of NhhA on the surface of E. coli by FACS analysis. In addition, we assessed the ability of these NhhA constructs to form stable oligomers on OM by immunoblotting, resolving OM protein preparations under standard denaturing conditions.

As shown in Fig. 2B, a polyclonal serum raised against purified recombinant NhhA (Pizza et al., 2000), revealed in FACS analysis that NhhA and NhhAΔL2 were surface-exposed, while NhhAΔL1 and NhhAΔL2L1 failed to translocate the passenger domain to the bacterial surface. Western blot experiments confirmed that NhhA and NhhAΔL2 were properly localized in the OM of E. coli and both were able to form stable trimeric oligomers, although traces of dimeric forms were also detectable (Fig. 2C). In the case of NhhAΔL2, the amount of surface exposed trimer was reduced compared with full-length NhhA. In contrast, NhhAΔL1 and NhhAΔL2L1 trimers were totally absent in the OM fraction, and their monomeric forms were observable in whole cell lysates (Fig. S1A).

In all OM preparation samples from bacteria expressing NhhA, faint bands around 60 kDa were observed likely corresponding to monomeric forms (Fig. 2C). They could be ascribed to contamination by other bacterial compartments.

To extend these results, we investigated the ability of the translocation unit of NhhA to export heterologous passenger domains. The C-terminal subdomains of NhhA (NhhAL2L1β, NhhAL1β, NhhAL2β and NhhAβ) were fused to the sequence coding for the leader peptide and the passenger domain of NadA (residues 1–315), an oligomeric adhesin/invasin of meningococcus (Comanducci et al., 2002; Capecchi et al., 2005), which possesses a similar C-terminal anchor (see Fig. 1B). The chimeric proteins (Fig. 3A) were expressed in E. coli and full-length NadA was used as a positive control. Immunoblot analysis on total lysates showed that the monomeric forms of all these proteins were expressed at comparable level, except for the chimera NadA-NhhAL2L1β, which resulted slightly less expressed (Fig. S1B).

Figure 3.

Surface localization of NadA passenger domain in NadA-NhhA chimerae expressed in E. coli BL21(DE3).
A. Schematic representation of NadA (2996 strain) and NadA-NhhA chimerae constructed in this study. LP corresponds to NadA leader peptide (residues 1–23), regions 24–315 and 316–405 to the NadA putative passenger and translocation domains respectively.
B. FACS analysis of NadA on whole cell bacteria performed with a rabbit serum against purified NadAΔ351-405.
C. Western blot analysis of outer membrane proteins from E. coli recombinant strains using a rabbit serum against NadAΔ351-405.

Surface localization and oligomerization of all chimerae were investigated by using an antibody against NadA. FACS analysis on E. coli cells expressing NadA, NadA-NhhAL2L1β, NadA-NhhAL2β, NadA-NhhAL1β and NadA-NhhAβ detected the presence of NadA passenger domain on the bacterial surface for all the cases except NadA-NhhAL2β and NadA-NhhAβ (Fig. 3B).

Western blot analysis on OM preparations from the E. coli recombinant strains confirmed that all the chimerae except NadA-NhhAL2β and NadA-NhhAβ were present in the OM fraction, although at reduced levels compared with full-length NadA, and formed stable oligomers the size of which corresponded to trimers (Fig. 3C).

The correct folding of the surface-exposed NadA-NhhAL2L1β and NadA-NhhAL1 chimerae was confirmed by the ability of both these proteins to mediate adhesion of E. coli to epithelial Chang cells, as previously observed with E. coli strain expressing full-length NadA (Fig. S2).

Taken together, these results suggest that NhhAL1β is the protein domain necessary and sufficient to direct the translocation of own and NadA passenger domains, likely through the formation of a trimeric β-barrel pore.

NhhA plays a role in bacterial adherence to epithelial cells

Based on the homology of NhhA to Hia/Hsf adhesins, we hypothesized that also NhhA could have a role in adhesion to epithelial cells. To investigate this aspect we firstly examined by immunofluorescence microscopy whether NhhA was able to promote in vitro adherence by E. coli to epithelial cell lines Hec-1-B and Chang, both widely used to study in vitro adhesion of N. meningitidis (Virji et al., 1991; Nassif et al., 1994). Cells monolayers were infected with E. coli expressing the full-length NhhA (E. coli/NhhA) or with E. coli carrying the empty vector (E. coli/pET) and, after wash of non-adherent bacteria, cell-associated bacteria were detected using an anti-E. coli serum. As shown in Fig. 4A, E. coli/NhhA was clearly able to interact with both epithelial cell lines. In contrast, examination of adherence by E. coli/pET failed to reveal adherent bacteria.

Figure 4.

Adhesive properties of NhhA.
A. Adherence to Hec-1-B and Chang epithelial cells by E. coli BL21(DE3) expressing NhhA. Hec-1-B and Chang monolayers were infected with E. coli/NhhA or with E. coli/pET and adherent bacteria were detected using a primary anti-E. coli serum and secondary antibody Alexa Fluor 488. Samples were examined by immunofluorescence microscopy.
B. Dose-dependent binding of purified recombinant NhhA-His and GNA2132 to Hec-1-B and Chang cells. Binding was measured by FACS. The values are reported as net mean fluorescence intensity (MFI).

Binding of the recombinant NhhA-His protein (residues 52–598) to Hec-1-B and Chang cells was quantified by FACS analysis. As shown in Fig. 4B, NhhA binds to both cell lines in a dose-dependent manner, reaching a plateau at a concentration of ≈ 200 μg ml−1 in the case of Chang cells and ≈ 300 μg ml−1 in the case of Hec-1-B cells. GNA2132 (Pizza et al., 2000), an unrelated meningococcus protein used as negative control, failed to bind to both epithelial cell lines. These observations suggest a function of NhhA as an adhesin.

In order to investigate the contribution of NhhA in the adhesion of N. meningitidis we constructed an nhhA isogenic mutant in the capsulated, piliated MC58 strain, named MC58ΔNhhA. Lack of expression of NhhA in MC58ΔNhhA, as well as the presence of the main known meningococcal proteins involved in adhesion (PilC, Opa, Opc, NadA and App) was assessed by Western blot analysis on total protein extracts. As shown in Fig. 5A the protein band corresponding to oligomeric NhhA was clearly detectable in the wild-type strain (lane 1), while it resulted absent in the mutant MC58ΔNhhA (lane 2). All the known adherence factors tested were expressed at comparable level in both strains.

Figure 5.

Characterization of the isogenic mutant MC58ΔNhhA and the complementing strain MC58ΔNhhA-cNhhA.
A. Expression of NhhA and other known meningococcus adhesion factors (PilC, Opa, Opc, NadA and App) by Western blot analysis of total cell lysates of N. meningitidis strains: MC58 (lane 1), MC58ΔNhhA (lane 2) and MC58ΔΝhhA-cNhhA (lane 3). Specific antibodies against the different proteins were used.
B. Adherence to Chang cells by N. meningitidis strains. Adherence was expressed as number of adherent bacteria. Values correspond to the means of a representative experiment performed in triplicate ± standard deviation.

The ability of MC58ΔNhhA to adhere to Chang epithelial cells was investigated after 3 h of infection and compared with that of the wild-type MC58. The results, reported in Fig. 5B, showed that MC58ΔNhhA bacteria adhere to Chang cells approximately 10-fold less than the wild-type strain.

In order to confirm that the impaired adhesive phenotype of MC58ΔNhhA was specifically imputable to NhhA, we complemented the nhhA deletion by inserting a functional copy of the gene into the chromosome of MC58ΔNhhA. The nhhA coding region was fused to a constitutive promoter (Ptac) and inserted between the convergent genes nmb1428 and nmb1429 along with a chloramphenicol resistance cassette. The expression of the gene in the resulting strain, named MC58ΔNhhA-cNhhA, was confirmed by Western blot. As shown in Fig. 5A, NhhA was produced in a trimeric form but at a lower extent compared with the wild-type strain (lane 1 and 3 respectively). On the contrary, the expression of the adherence factors analysed was not altered.

We then analysed the ability of the complementing strain to adhere to Chang epithelial cells. As shown in Fig. 5B, there is a partial restoration of the adhesive phenotype respect to the wild-type strain MC58 that could be imputable to the lower expression level of the NhhA protein detected in the complementing strain (Fig. 5A).

Taken together, the results strongly suggest that NhhA is involved in adhesion of capsulated meningococci to epithelial cells.

NhhA binds to laminin and heparan sulphate

In a previous study, we predicted possible heparan sulphate binding sites within the NhhA passenger domain (Scarselli et al., 2001). In order to investigate whether NhhA is able to interact with ECM, we evaluated by in vitro binding assays the interaction of purified recombinant NhhA to selected ECM components. Samples of purified NhhA-His at concentrations ranging from 1.17 nM to 1.75 μM were added to wells of 96-well plates coated with purified plasma laminin, heparan sulphate, fibronectin, collagen I, III and IV. Binding was detected with a polyclonal serum raised against NhhA-His and quantified by enzyme-linked immunoassorbent assay (ELISA). As shown in Fig. 6, dose-dependent binding of NhhA-His to laminin and heparan sulphate (HSPG – Heparan Sulphate ProteoGlycan) was observed; conversely, no binding was detected to fibronectin and to three different types of collagen. Bovine serum albumin (BSA) was used as negative control in all the assays.

Figure 6.

Dose-dependent binding of purified NhhA-His to selected ECM components. Shown are results of binding increasing concentrations of purified NhhA-His to laminin (A), HSPG (B), fibronectin (C), collagen I (D), collagen III (E), and collagen IV (F). In all the experiments BSA was used as negative control. Binding was quantified by ELISA at an absorbance of 405 nm. Points represent the means (error bars, standard errors of the means) of measurements made in triplicate.

We then analysed NhhA-mediated adhesion to ECM components by testing adherence of MC58 and MC58ΔNhhA strains to tissue culture plates coated with laminin or heparan sulphate. As shown in Fig. 7, there was approximately a 3.5-fold reduction of the association of the MC58ΔNhhA strain compared with the MC58 strain to laminin and a fivefold reduction in the case of heparan sulphate.

Figure 7.

Adherence to proteins of ECM by N. meningitidis strain MC58 and MC58ΔNhhA. Adherence to laminin (A) and heparan sulphate (B) was expressed as number of adherent bacteria. Values correspond to the means of a representative experiment performed in triplicate ± standard deviation.

These findings suggest that NhhA protein exposed on the bacterial surface is able to interact with laminin and heparan sulphate, promoting the adherence of meningococcus to the ECM components.


The ability of bacteria to colonize their hosts and cause infection is often linked to their ability to express several different adhesins with different receptor specificities.

Some potent adhesins of Gram-negative pathogens are anchored to the OM, being so able to promote an intimate contact to target cells as result of attachment to their receptor. A major class of these molecules is displayed onto the cell surface by virtue of the type V protein secretion system, whose members are known as ATs.

ATs are synthesized as large precursors proteins that reach the periplasm through a sec-dependent mechanism. Subsequently, their N-terminal passenger domains are exposed to the extracellular milieu by secretion through a β-barrel pore, formed by the C-terminal portion (Henderson et al., 2004).

Recently, a specialized class of trimeric AT adhesins has been described, in which the functional translocation unit has been proposed to result from trimerization of a short C-terminal domain (for a complete list, see Cotter et al., 2005). Members of this AT subfamily include also UspA (Hoiczyk et al., 2000; Lafontaine et al., 2000), DsrA (Cole et al., 2002), Hag (Holm et al., 2003), YadA (Roggenkamp et al., 2003), Omp100 (Asakawa et al., 2003), Hia/Hsf (Surana et al., 2004), BadA (Riess et al., 2004) and NadA (Comanducci et al., 2002; Capecchi et al., 2005), for which the ability to promote adherence to epithelial cells and/or ECM proteins has been demonstrated. In this study we have characterized NhhA of N. meningitidis and demonstrated that it is a functional member of such trimeric autotrasporter family.

NhhA is the orthologue of Hsf (St Geme et al. (1996), a protein involved in the formation of subtle fibrils on the surface of capsulated strains of H. influenzae, and of its allelic variant Hia, which is expressed by the majority of non-capsulate strains (Barenkamp and St Geme, 1996; St Geme et al., 1998). Hia and Hsf are modular proteins, whose passenger domains characterized by the presence of internal repeats, contain distinct cell binding sites (Laarmann et al., 2002; Cotter et al., 2005).

The amino acid sequence of NhhA shows a similar sequence organization of the passenger domain, although a reduced number of such repeats result in a remarkable difference in length respect to Hsf and Hia (Scarselli et al., 2001).

The NhhA translocation unit

The C-terminal 89 residues appear to be well conserved among NhhA, Hsf and Hia (Fig. 1B). Despite this similarity, there is not a complete concordance on the data reported so far about the length of the three translocation units.

Roggenkamp et al. (Roggenkamp et al., 2003), identified in the 91 C-terminal residues of YadA the shorter tract sufficient for the oligomerization and insertion into the OM. Surana and coworkers reported that the minimal tract necessary to translocate the passenger of Hia correspond rather to last 76 residues of the protein (Surana et al., 2004). In the same study, a longer region of NhhA (corresponding to residues 480–598 in 2996 strain) was demonstrated to promote the passenger secretion.

In order to precisely map the minimal tract of NhhA able to retain the functionality of the translocation unit, we designed deletion mutants at its C-terminus. Three subregions, namely β (residues 545–598), L1 (527–544) and L2 (510–526) were defined, according to the subdivision already proposed for YadA (Roggenkamp et al., 2003).

FACS analysis and Western blot data indicate that L1β (residues 527–598) is the translocation unit of NhhA, being able to localize into the OM, as well as to expose on the bacterium surface, NhhA and NadA passenger domains in their trimeric form.

The observation that NhhAΔL1 results in absence of detectable passenger domain on the bacterium surface indicates that the presence of L1 is necessary for the proper folding and/or functioning of the translocator domain, and that cannot be functionally substituted by L2.

This body of experimental evidences has been used, together with actual knowledge of crystal structures of bacterial OM proteins, to build the model of the NhhA translocation unit depicted in Fig. 8.

Figure 8.

Computer model of the putative 3D structure of NhhA translocation domain generated and refined by energy minimization with the DeepView/Swiss-Pdb viewer (Guex and Peitsch, 1997). The different subregions are coloured in green (beta), orange (L1) and red (L2).

We started with the assumption that the four beta strands predicted at the C-terminal end form the walls of the NhhA pore. Crystal structures of bacterial beta barrels solved so far have shown that in order to span the membrane bilayer, each strand has to be on average 10–11 residues long. A complete barrel is formed by an even number of antiparallel strands, each connected to the neighbour one in the chain, with the C-terminus always facing the periplasmic side of the OM (Schulz, 2000).

In order to reach a topology consistent with the extracellular localization of the passenger, the four beta strands have to be preceded by an additional structural element pointing outside. The secondary structure prediction suggests that the NhhA sequence form an alpha helix spanning from 523 to 540 residues. The presence of an alpha helical region immediately upstream the β-domain has been proposed to be a common signature of ATs (Henderson et al., 1998; Maurer et al., 1999; Oliver et al., 2003) and the crystal structure recently solved of the translocator domain of NalP shows that an alpha helix travels trough a 12-stranded beta barrel filling the cavity formed by beta strands (Oomen et al., 2004).

In our proposed model the β region forms the trimeric barrel embedded into the OM. Three L1 helices pass through the pore pointing towards the extracellular space, while the L2 fragments protrude outside, forming a neck between membrane-embedded anchor and the surface-exposed effector domains. Although a different topology could be also theoretically possible, in which the L1 helices cross the OM externally to the beta barrel, the model proposed in Fig. 8 seems preferable to describe the functional translocation unit for several reasons.

First, it is consistent with the secretion mechanism more widely accepted for AT, in which the passenger domain uses the cavity formed by beta strands as channel for secretion. Second, in this case the L1 helices can easily form an helix bundle which persists in L2 and seems fit to the intricately folded nature of the trimeric passenger domains. Third, the model proposed in Fig. 8 implies that L1 is more constrained than L2 to maintain small and/or hydrophobic residues as alanine, glycine and leucine in positions crucial for correct packing into the beta cavity.

Recently, a very similar model has been proposed for the YadA membrane anchor region (Wollmann et al., 2006); given the close relationship between the translocation units of NhhA and YadA, it is very plausible that the same architecture is shared by the two domains.

The model presented here is supported by the finding that NhhAL1β represents the translocation unit of NhhA, being able to export both homologous and heterologous passenger domains. The lower levels of NhhAΔL2 respect to the full-length protein observed in immunoblot analysis of the OM preparations could be due to the absence of a neck between the barrel edge and the globular passenger domain, which result in a minor stability of the trimer. However, we did not appreciate any difference in resistance to denaturation into monomers between NhhA and NhhAΔL2 after boiling of OM protein preparations for different time points (data not shown). The alternative hypothesis that NhhAΔL2 could be subjected to a less efficient translocation can not be therefore excluded. The failure of NhhAΔL1 to expose both NhhA and NadA passenger domains on the bacterial surface indicates in any case that L1 and L2 are not functionally exchangeable, and can be due to the difficulty to accommodate L2 within the β barrel due to its major steric hindrance.

The differences in translocation efficacy of the two passengers fused to NhhAL1β underlined that the nature of protein to be translocated has a great influence on the entire secretion process.

The biological function

N. meningitidis, has evolved a diverse array of surface structures to interact with host cells. Type IV pili are considered to be the prime attachment-promoting factor for capsulated meningococci to the nasopharyngeal mucosa (Virji et al., 1991). Recently, also the adhesion and penetration protein App was shown to contribute to adherence of capsulated meningococci to Chang epithelial cells (Serruto et al., 2003).

In a following intimate contact, pili are assumed to disappear and the synthesis of the polysaccharidic capsule is downregulated (Deghmane et al., 2002).

During this second step of adhesion a different repertoire of adhesins is involved, of which Opa and Opc are the most studied. Opa proteins have been shown to bind heparan-sulphate proteoglycans (HSPGs), as well as CEACAMs (Virji et al., 1996). Opc has been shown to bind α5-β1 integrin on the endothelial cells via interaction with fibronectin (Unkmeir et al., 2002). Differently, interaction of Opc with epithelial cells is mediated by binding to HSPGs (de Vries et al., 1998).

Sequence analysis suggests that NhhA is the orthologue of Hsf/Hia. Recent studies on Hia demonstrated its adhesive properties and identified two binding sites, one of which with high affinity to an unidentified host cell receptor.

The results reported in this article clearly demonstrated the adhesive role of NhhA, which promotes adherence of recombinant E. coli to two different epithelial cell lines. The NhhA binding activity was confirmed in binding experiments carried out with recombinant NhhA. On the other hand, inactivation of the nhhA gene in the virulent N. meningitidis MC58 strain significantly reduced its adherence capability compared with the wild-type strain. In presence of the capsule the contribution of NhhA in N. meningitidis adhesion was comparable to that observed for App protein (Serruto et al., 2003). Therefore, our data suggest that NhhA may play a relevant role, together with pili and App, during adherence of capsulated meningococci to human epithelium.

The capability of N. meningitidis to cause disease is dependent both on its ability to attach to nasopharyngeal mucosa and to penetrate the submucosal tissue to reach the bloodstream. Adherence to the respiratory epithelium represents an essential early step in the colonization process, while attachment to ECM components and subsequent their degradation may be crucial during penetration from the blood to the cerebrospinal fluid.

In normal tissues, most ECMs are covered by epithelial or endothelial cells and hence are not available for binding. However, any type of trauma that damages host tissues may expose the ECM and allow microbial colonization and infection. Many different pathogenic bacteria such as H. influenzae, H. pylori, Streptococcus pyogenes and Staphylococcus aureus have been demonstrated to interact with different ECM proteins (Ljungh et al., 1996; Finlay and Caparon, 2004; Preissner and Chhatwal, 2004). In the case of N. meningitidis, Eberhard and coworkers have shown that it is able to adhere to subendothelial ECM as well as to immobilized components especially fibronectin and collagen types I, III and V (Eberhard et al., 1998), but the mechanism and factors involved in this interaction remain to be identified.

Our data revealed the ability of recombinant NhhA to bind laminin and heparan sulphate. In addition, we demonstrated a significant reduction in adhesion by meningococcus MC58ΔNhha mutant respect to the wild-type MC58 strain to both these ECM molecules. This strongly suggests that NhhA present on the meningococcus surface is also able to promote adherence of live bacteria to laminin and heparan sulphate. To our knowledge, while Opa (Virji et al., 1999) and Opc (Unkmeir et al., 2002) have been demonstrated to bind HSPGs in absence of capsule, NhhA is the first meningococcal adhesin whose ability to bind ECM components has been observed in capsulated strains.

In absence of an animal model of meningococcal colonization and disease, which would be instrumental to definitively prove the NhhA function in vivo, our results suggest a central role of the trimeric AT adhesin NhhA in meningococcal infection, particularly during colonization of the nasopharyngeal mucosa.

Experimental procedures

Computer analysis and molecular modelling

All sequence comparisons were performed using the FASTA algorithm included within the gcg software package.

Secondary structure prediction was carried out with PHD (Rost and Sander, 1993), available at the PredictProtein web server (

A model of the three-dimensional (3D) structure of 510–598 NhhA residues was built by the combined use of homology modelling and molecular mechanics.

Three monomeric four stranded beta sheet has been obtained by threading 545–598-rr residues of NhhA onto the crystal coordinates of the NalP beta-barrel.

The DeepView software (Guex and Peitsch, 1997) was used for construction of the homology model and for manipulation of torsion angles.

The stereochemical validity of both monomeric and trimeric final models was confirmed using procheck (Laskowski et al., 1993).

Bacterial strains, cells and culture conditions

MC58 strain (McGuinness et al., 1991), the isogenic MC58ΔNhhA and MC58ΔNhhA-cNhhA are the N. meningitidis strains considered in this study. E. coli strains included in this study are DH5α (Invitrogen) and BL21(DE3) (Novagen).

Neisseria meningitidis strains were grown on GC agar plates or in GC broth at 37°C in 5% CO2. E. coli strains were cultured in Luria–Bertani (LB) agar or LB broth at 37°C. Antibiotic concentrations used when required included 100 μg ml−1 ampicillin for E. coli and 5 μg ml−1 erythromycin or chloramphenicol for N. meningitidis.

Tissue culture cells used in this study are Chang epithelial cells (Wong-Kilbourne derivative, clone 1-5c-4, human conjunctiva) and HEC-1-B cells (human endometrium). Both cell lines were maintained in Dulbecco's Modified Eagle Medium (DMEM; Gibco) supplemented with 15 mM l-glutamine and antibiotics.

Construction of plasmids used in this study

In order to express in E. coli all the recombinant proteins considered in this study, the specific DNA fragments were amplified by polymerase chain reaction (PCR) from N. meningitidis 2996 strain and cloned into the pET-21b expression vector (Novagen).

The DNA sequence corresponding to the leader peptide of App (amino acids 1–42) was amplified using as primers the oligonucleotides App-FOR and App42-REV. The DNA regions corresponding to NhhA protein devoid of the leader peptide (amino acids 52–598) and to NhhA passenger domain (amino acids 52–509) were amplified using the oligonucleotides NhhA52-FOR and NhhA-REV, NhhA52-FOR and NhhA509-REV respectively. In the case of the NhhAΔL1 construct, the DNA sequence of passenger domain was amplified using NhhA52-FOR and NhhAL2-REV. To amplify the DNA sequence coding for NadA passenger domain (amino acids1–315) the oligonucleotides NadA-FOR and NadA315-REV were used.

Finally, the DNA regions corresponding to NhhAL2L1β (amino acids 510–598), NhhAL1β (amino acids 527–598) and NhhAβ (amino acids 546–598) were amplified using the oligonucleotides NhhA510-FOR and NhhA-REV, NhhA527-FOR and NhhA-REV, NhhA546-FOR and NhhA-REV respectively. In the case of the construct NadA-NhhAL2β the NadA passenger domain was fused to the L2 region of NhhA through to steps of PCR: in the first we used NadA-FOR and NadL2a-REV; in the second step NadA-FOR and NadL2b-REV as oligonucleotides and the product of the first step as template. The resulting product was then ligated to the DNA region corresponding to NhhAβ.

NhhA-His construct contains the nhhA gene deleted of the sequence for the predicted leader peptide (amino acids 1–51) and the stop codon. The gene fragment was obtained by PCR using the NhhA52-FOR and NhhAhis-REV primers. In all the oligonucleotides a sequence corresponding to a specific restriction site was inserted.

The amplified DNA fragments were digested with the appropriate restriction enzymes and cloned into pET-21b vector. Sequential cloning generated all the recombinant plasmids. The insertion of NheI, BamHI or EcoRI site results in Alanine-Serine, Glycine-Serine or Glutamic acid-Phenylalanine additional amino acids respectively. The ligation products were transformed into E. coli DH5α. E. coli BL21(DE3) strain was used as expression host.

The plasmid pBSnhhAERM was constructed to generate N. meningitidis nhhA isogenic mutant. This construct contains an nhhA truncated gene and the ermC gene (erythromycin resistance) for allelic exchange. Briefly, an upstream flanking region and a downstream flanking region were amplified from MC58 strain using the oligonucleotides UnhhA-FOR and UnhhA-REV, DnhhA-FOR and DnhhA-REV respectively. The fragments were cloned into pBluescript plasmid and the recombinant plasmid transformed into E. coli DH5α. The plasmid was digested with HincII in order to insert the ermC gene.

Complementation of NhhA was achieved by insertion of a copy of the nhhA gene under the control of the Ptac promoter in the non-coding region of the MC58ΔNhhA chromosome between the converging open reading frames NMB1428 and NMB1429.

The plasmid for complementation of the NhhA null mutant is a derivative of the pSLComCmr plasmid (Ieva et al., 2005) containing a copy of the nhhA gene under the control of the Ptac promoter. The nhhA gene was amplified from N. meningitidis strain MC58 with oligonucleotides nhhAcompFOR and nhhAcompREV. The resulting plasmid was named pCOM-NhhA.

All the oligonucleotides used as primers are listed in Table 1. DNA cloning and E. coli transformations were performed according to the standard protocols (Sambrook and Russel, 2000).

Table 1.  Oligonucleotide primers used in this study.
PrimerSequence (5′−3′)Restriction site (underlined)

Gene expression, cell fractionation and protein analysis

Escherichia coli BL21(DE3) recombinant strains were grown overnight at 37°C then bacteria were diluted and grown at 37°C until OD600 = 0.4–0.8. Gene expression was induced by addition of 1 mM isopropyl-1-thio-β-d-galactopyranoside (IPTG, Sigma) and growth was continued for 3 h. OM proteins were recovered on the basis of sarkosyl-insolubility following the rapid procedure as described by Carlone et al. (Carlone et al., 1986).

SDS-PAGE electrophoresis, using 7% NuPAGE gels (Invitrogen), and Western blot analysis were performed according to standard procedures. Western blots were performed with a mouse polyclonal serum raised against purified NhhA-His or with a rabbit polyclonal serum raised against recombinant NadAΔ351-405 (Comanducci et al., 2002). An anti-mouse or anti-rabbit IgG antiserum conjugated to horseradish peroxidase (DAKO) was used as the secondary antibody. Antibody reactivity was detected by a chemiluminescent substrate (Pierce).

FACS analysis of bacteria

Escherichia coli BL21(DE3) strains expressing the different deletions of NhhA or chimeric proteins were harvested and pellets were washed and re-suspended in PBS to an OD600 = 0.5. Mouse or rabbit polyclonal antibody against NhhA-His and NadAΔ351-405, used as primary antibody, was added directly to the cell suspension and incubated for 1 h at room temperature (RT). Following two washes in PBS, cells were incubated for 1 h at RT with R-phycoerythrin-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories). Bacteria cells were washed twice with PBS, and then analysed with FACSCalibur flow cytometer (Becton Dickinson).

Membrane permeability controls of recombinant E. coli strains were performed using a rabbit polyclonal antibody against the periplasmic protein DsbA (StressGen Biotechnologies) as primary antibody and an R-phycoerythrin-conjugated anti-rabbit IgG as secondary antibody (Jackson ImmunoResearch Laboratories). Results are reported in Fig. S3.

Purification of recombinant protein

NhhA-His was produced as insoluble inclusion bodies and was solubilized with urea and renatured after purification (Coligan et al., 1997). The fusion protein was purified by affinity chromatography on Ni2+ conjugated chelating fast flow Sepharose (Pharmacia). The purity was checked by SDS-PAGE electrophoresis stained with Coomassie blue (Fig. S4). The protein content was quantified by Bradford reagent (Bio-Rad).

Adhesion assay by immunofluorescence microscopy

Hec-1-B or Chang cells (105 cells) were seeded on Laboratory-Tek chamber slides (Nunc) and incubated for 24 h. Cultures of bacteria after IPTG induction, were washed twice in PBS and re-suspended in DMEM + 1% FBS to a concentration of 1 × 107 bacteria ml−1. Aliquots of 1 ml of each strain were added to monolayers culture of Hec-1-B cells (or Chang) and incubated for 3 h at 37°C in 5% CO2. Non-adherent bacteria were removed by washing three times with PBS. Cells were fixed in 3.7% paraformaldhehyde for 30 min, washed two times with PBS and permeabilized with 1% Triton X-100 in PBS for 10 min. Cells were incubated for 1 h with rabbit polyclonal anti-E. coli serum (DAKO). The cells were washed twice in 1% BSA in PBS and incubated for 30 min with Alexa Flour 488 goat anti-rabbit IgG (Molecular Probes). After two washes in 1% BSA in PBS and one wash in water, the labelled preparation was mounted with SlowFade Light antifade kit (Molecular Probes) and analysed with a Zeiss Axiophot immunofluorescence microscopy.

Binding assay: FACS analysis

Binding assay was performed as previously described (Serruto et al., 2003). Briefly, human cells were non-enzymatically detached using cell dissociation solution (CDS, Sigma), harvested and suspended in RPMI medium supplemented with 1% FBS. Cells (1 × 105) were mixed with different concentrations of purified NhhA-His, purified GNA2132-His or medium alone for 60 min at 4°C. Cells were then incubated with mouse polyclonal serum against NhhA-His (or against GNA2132-His) for 1 h at 4°C, and with R-phycoerythrin (PE)-conjugated goat F(ab)2 antibody to mouse Ig (diluted 1:100; Jackson ImmunoResearch Laboratories) for 30 min at 4°C. Cells were analysed with a FACS-Scan flow cytometer (Beckton-Dickinson). Cell-bound fluorescence was analysed with FACSCalibur flow cytometer (Becton Dickinson) by using the CellQuest software program. The mean fluorescence intensity (MFI) for each population was calculated.

Binding assay: ELISAs

Microtitre plates were coated (overnight at 4°C) with 0.1 μg of ECM proteins: HSPG (from basement membrane), fibronectin (from human plasma), laminin (from basement membrane), human collagen type I, type III, type IV (Sigma). BSA (Sigma) was used as negative control. Wells were washed three times with PBST (PBS + 0.05% Tween 20) and then blocked with blocking solution (2% milk in PBS) for 2 h at 37°C. After three washing in PBST, 50 μl of purified NhhA-His protein (1.17 nM to 1.75 μM) diluted in PBS plus 0.1% blocking solution was added in triplicate wells and plates were incubated at 37°C for 2 h. After three washing in PBST, wells were incubated at 37°C for 1 h with an anti-NhhA-His serum diluted 1:1000 in 0.1% blocking solution. Then, after three washing in PBST, wells were incubated with a secondary anti-mouse antibody AP-conjugated diluted 1:1000 0.1% blocking solution at 37°C for 1 h. Wells were washed three times with PBST and binding by NhhA was detected incubated wells with 100 μl of pNPP substrate and measurement of absorbance at 405 nm.

Construction of N. meningitidis nhhA isogenic mutant and complementation

A knockout mutant in MC58 strain in which the nhhA gene was replaced with an antibiotic cassette, was constructed by transforming the parent strain with the plasmid pBSnhhAERM. Naturally competent N. meningitidis strain MC58 was transformed as described (Masignani et al., 2003). The presence of deleted nhhA gene was confirmed by PCR and the lack of NhhA expression was confirmed by Western blot analysis.

The plasmid pCOM-NhhA was used to transform MC58ΔNhhA strain. Transformants were selected for chloramphenicol resistance and correct insertion was verified by PCR.

All these meningococcus strains were characterized for the expression of PilC, Opa, Opc, NadA and App by Western blot analysis using polyclonal antisera against recombinant 6× histidine-tagged PilC1 and App and against NadAΔ351-405, or the monoclonal antibodies B303 (against Opa) and 4B12 (against Opc).

Adhesion assay with N. meningitidis

The interaction of parent MC58 strain, the isogenic mutant MC58ΔNhhA and the complementing strain MC58ΔNhhA-cNhhA with cultured Chang epithelial cells was studied using a variant of the method described by Pujol C et al. (Pujol et al., 1999). Briefly, the cells were seeded at 1.5 × 105 cells well−1 in 24-well tissue culture plates (Corning Costar) the day before. The culture medium was removed and the cells were washed three times with DMEM + 1% FBS and fresh medium added. Bacteria were added to monolayers at a multiplicity of infection of 100. Incubation was carried out for 3 h (at 37°C in 5% CO2) and the medium was replaced every hour to minimize monolayer re-infection. The inoculating dose of bacteria was confirmed by serial dilution and plating. To determine association, monolayers were washed three times with DMEM + 1% FBS to remove non-adherent bacteria. The remaining bacteria were released by the addition of 1% saponin and incubation at 37°C for 10 min. The number of associated bacteria were determined by serial dilution and plating.

Quantitative adherence assay to laminin and heparan sulphate

Ninety-six-well tissue culture plates were coated (overnight at 4°C) with 0.1 μg of laminin or heparan sulphate. Meningococcus bacterial strains, resuspended in PBS, were added to each wells (approximately 1 × 107 well−1) and incubated for 3 h at 37°C in 5% CO2. After incubation, wells were rinsed four times with PBS to remove non-adherent bacteria and were treated with trypsin-EDTA to release the bacteria. The number of adherent meningococcus were determined by serial dilution and plating.


We thank Vega Masignani for valuable discussion, Isabel Delany for supplying the plasmid used for complementation of nhhA in N. meningitidis, Tiziana Spadafina for technical help, Catherine Mallia for editing and Giorgio Corsi for artwork. We gratefully acknowledge G. Morelli and M. Achtman for providing the monoclonal antibody 4B12, and R. Moxon for providing the monoclonal antibody B303. This work was supported by a grant from European Commission 5th Framework Programme: Mucosal Immunization and Vaccine Development (MUCIMM), Contract no. QLK2CT 1999 00228.

Note added in proof

While this work was under revision, an article reporting the 3D structure of the Hia translocation unit has been released by G. Meng and colleagues (EMBO J 2006 May 11), which strongly supported our predictions on NhhA.