Haemophilus influenzae Surface Fibrils Contribute to Serum Resistance by Interacting with Vitronectin¹ Free

Haemophilus influenzae is a Gram-negative human pathogen responsible for a variety of diseases. Encapsulated H. influenzae strains belong to one of six serotypes (a–f), of which type b is the most virulent serotype (1, 2). The most serious and sometimes life-threatening conditions are invasive diseases (e.g., septicemia, epiglottitis, and meningitis) caused by encapsulated H. influenzae serotype b (Hib)³ (3). In contrast, nontypable H. influenzae accounts for the majority of local disease and upper and lower respiratory tract infections (e.g., bronchitis, sinusitis, and acute otitis media) and is after pneumococci the second most common pathogen isolated from children with acute otitis medium (1, 2).

A crucial factor in the pathogenesis of both encapsulated and nonencapsulated H. influenzae involves the initial adherence to the mucosa in the respiratory tract (4). If the bacteria manage to overcome the mucociliary escalator, they may colonize and cause damage to the epithelial cells and breakdown of tight junctions (5, 6). Consequently, the bacteria reach the basement membrane and the extracellular matrix (ECM), and may penetrate into deeper tissue layers and consequently into the circulation. Studies on the interaction of H. influenzae and tissue samples from the human respiratory tract show that H. influenzae has been associated with disrupted epithelial cells and exposed ECM proteins such as fibronectin, collagen, vitronectin, and laminin (7, 8, 9).

Both pilus and nonpilus adhesins of H. influenzae have displayed adherence to ECM proteins. H. influenzae pili, which hemagglutinate human erythrocytes and adhere to human oropharyngeal epithelial cells (10, 11, 12, 13, 14, 15, 16), exhibit adherence to fibronectin and heparin-binding ECM proteins. The nonpilus adhesin Haemophilus adhesion and penetration protein was reported as a binder of fibronectin, laminin, and collagen IV (8).

The major nonpilus adhesin in Hib is Haemophilus surface fibrils (Hsf) (17). The hsf gene is highly conserved among encapsulated H. influenzae strains and encodes a 2414-aa-long protein consisting of three repetitive domains with high sequence similarity. Hsf is found as short, thin surface fibrils at the bacterial surface and is associated with adherence to epithelial cells (16). In 25% of all unencapsulated strains, a homologue to the Hsf protein, H. influenzae adhesin (Hia), can be found (17, 18, 19). The hia gene, which is shorter than the hsf gene, encodes for a protein with a size of 1098 aa and harbors only one domain that corresponds to the three repetitive domains in Hsf. However, Southern blot analysis has revealed that hsf and hia are alleles of the same locus with 81% similarity and 72% identity (17).

The complement system is the first line of innate defense against pathogenic microorganisms, and activation of this system leads to a cascade of protein deposition on the bacterial surface, resulting in formation of the membrane attack complex (MAC) and opsonization of the pathogen, followed by phagocytosis. A regulatory component of MAC is the multifunctional glycoprotein vitronectin that is found both in plasma and in the ECM (20). It exists as a 75-kDa protein in the ECM and is found in plasma as two truncated forms: 75 and 65 kDa.

Both Hib and nontypable H. influenzae bind surface-associated vitronectin equally well, and it has been suggested that adhesins are involved in the binding because both fimbriated and nonfimbriated strains adhere to a similar degree (9). In this study, we demonstrate that Hsf is the major vitronectin-binding protein in Hib. A H. influenzae mutant devoid of Hsf displayed a decreased binding to both soluble and immobilized vitronectin. Furthermore, Hsf-dependent interaction with vitronectin was inhibited by heparin. Interestingly, H. influenzae wild type survived a significantly longer time as compared with the Hsf mutant counterpart when exposed to normal human serum (NHS). Finally, we show that two separate binding domains of Hsf are involved in the vitronectin binding.

The type b strain H. influenzae Eagan and the clinical capsule-deficient H. influenzae isolate RM804 have been described in detail (21, 22). Bacteria, wild type, and mutants were routinely cultured in brain-heart-infusion liquid broth supplemented with NAD and hemin (both at 10 μg/ml) or on chocolate agar plates at 37°C in a humid atmosphere containing 5% CO₂. The Hsf-deficient mutant was cultured in the presence of 15 μg/ml kanamycin (Merck). The Streptococcus pyogenes was a clinical isolate from our department and was grown in brain-heart infusion liquid broth. Escherichia coli BL21 (DE3) and DH5α were grown in Luria Bertani liquid broth, whereas Hsf transformants were cultured with 50 μg/ml ampicillin (Sigma-Aldrich).

Rabbits were immunized i.m. with 200 μg of rHsf^54–608 emulsified in CFA (Difco and BD Biosciences), and boosted on days 18 and 36 with the same dose of protein in IFA. Blood was drawn 3 wk later. To increase the specificity, the anti-Hsf antiserum was affinity purified with Sepharose-conjugated rHsf. To ensure that the polyclonal Ab (pAb) reacted with rHsf, the pAb was analyzed in ELISA. Hsf (100 ng/well) were immobilized in microtiter plates and incubated at increasing concentrations of the antiserum, followed by HRP-conjugated goat anti-rabbit pAb diluted 1/1000 (Dakopatts). The FITC-conjugated goat anti-human vitronectin and donkey anti-goat pAb were purchased from Serotec, and the mouse anti-human vitronectin and the HRP-conjugated anti-mouse pAb were from Invitrogen Life Technologies and Dakopatts, respectively.

The 5′ end of hsf (GenBank accession no. U41852) was amplified as two cassettes using DyNAzyme II DNA Polymerase (Finnzymes) introducing the restriction enzyme sites BamHI and EcoRI or EcoRI and XhoI in addition to specific uptake sequences in the two cassettes (23). Resulting PCR fragments (825 and 918 bp, respectively) were digested and cloned into pBluescript SK^+/−. A kanamycin resistance gene cassette from pUC4K was amplified by PCR, introducing the restriction enzyme site for EcoRI. After digestion, the PCR product was ligated into the truncated hsf gene fragment. H. influenzae strains Eagan and RM804 were transformed according to the M-IV method of Poje and Redfield (23). Resulting mutants were verified by PCR, and the Hsf expression was analyzed by Western blot and flow cytometry (Figs. 1 and 2 B).

All truncated Hsf constructs were manufactured using PCR-amplified fragments. The open reading frame of the hsf gene (U41852) from H. influenzae strain RM804 was used as a template. All Hsf constructs were amplified by PCR using DyNAzyme II DNA Polymerase (Finnzymes) with specific primers introducing the restriction enzyme sites BamHI and HindIII. The sequence encoding for the signal peptide was excluded. To express full-length Hsf, a NcoI restriction enzyme site was introduced. The PCR products were cloned into pET26⁺, except for the full-length Hsf, which was cloned into both pET26 and pET16. The resulting plasmids were transformed into the host E. coli DH5α, followed by transformation into the expressing host E. coli BL21(DE3) (Novagen). All constructs were sequenced using the BigDye Terminator Cycle Sequencing v. 3.1 Ready reaction kit (Applied Biosystems). To produce recombinant proteins, bacteria were grown to mid-log phase (OD₆₀₀ 0.5–1.0), followed by 1–3 h of induction with 1 mM isopropyl-1-thio-β-D-galactoside (Saveen Werner). Inclusion bodies were purified according to a standard protocol (Novagen). The resulting proteins were examined by SDS-PAGE, Western blots, and ELISA.

Bacteria grown to stationary phase were washed with 50 mM Tris-HCl buffer (pH 8.0). The pellet was resuspended in Tris-HCl buffer containing 3% Empigen (Calbiochem) and protease inhibitors (Complete; Roche) (24). OMPs were extracted by rotating the mixture at 37°C for 2 h. The bacterial cells, stripped of their outer membranes, were centrifuged, and the supernatants were collected. Thereafter, the supernatants were analyzed on SDS-PAGE and Western blots.

Recombinant proteins were subjected to SDS-PAGE (10%) (25) and stained with Coomassie brilliant blue R-250 (Bio-Rad). Electrophoretical transfer of protein bands from the gel to an Immobilon-P membrane (Millipore) was done at 35 V overnight to transfer the high m.w. complexes. After transfer, the Immobilon-P membrane was blocked in PBS with 0.1% Tween 20 (PBS-Tween) containing 5% milk powder. After several washings in PBS-Tween, the membrane was incubated with rabbit anti-Hsf antiserum diluted 1/100 in PBS-Tween, including 2% milk powder, for 1 h at room temperature. HRP-conjugated goat anti-rabbit antiserum diluted 1/1000 was added after washings in PBS-Tween. After incubation for 40 min at room temperature and additional washings in PBS-Tween, development was performed with ECL Western blotting detection reagents (Amersham Biosciences). To analyze the purity of vitronectin obtained from human plasma (Sigma-Aldrich), 2 μg was subjected to SDS-PAGE and Coomassie stained.

The Hsf protein expression and the capacity for H. influenzae to bind vitronectin were analyzed by flow cytometry. The wild-type strains and the Hsf-deficient mutants were grown in broth overnight and washed once in PBS containing 2% BSA (PBS-BSA). Bacteria (10⁸) were incubated with rabbit anti-Hsf pAb. Bacteria were washed and incubated for 30 min on ice with FITC-conjugated goat anti-rabbit pAb (Dakopatts), diluted according to the manufacturers’ instructions. After three additional washes, the bacteria were analyzed in a flow cytometer (EPICS, XL-MCL; Corixa). To analyze H. influenzae binding to vitronectin, bacteria were incubated with 0.1–2.5 μg of vitronectin for 1 h in 37°C. After washings, bacteria were incubated with goat anti-human vitronectin pAb for 30 min at ice, before incubation with the FITC-conjugated donkey anti-goat pAb. After additional washes, bacteria were analyzed in the flow cytometer. All incubations were kept in PBS-BSA, and the washings were done with the same buffer. Secondary pAb were added separately as negative controls for each strain analyzed. Hsf-expressing E. coli was analyzed for vitronectin binding according to a standard protocol (26). In the competition assay, the H. influenzae wild type was preincubated with increasing concentrations of heparin (Heparin Leo, Lövens Kemiske Fabrik, or Sigma-Aldrich) or vitronectin-derived synthetic peptides, followed by 1 μg of vitronectin. Peptides spanning the heparin binding domain used in this study were vitronectin^348–361 (KKQRFRHRNRKGYR) and vitronectin^341–370 (APRPSLAKKQRFRHRNRKGYRSQRGHSRGR) (Innovagen).

Glass slides were coated with 2 μg of human plasma vitronectin, air dried at room temperature, and then washed twice with PBS. The slides were incubated with prechilled bacteria at late exponential phase (OD₆₀₀ = 0.9) for 2 h at room temperature, washed twice with PBS, and followed by Gram staining.

NHS was pooled from five healthy volunteers. Inactivated serum was used as a control. The H. influenzae wild type and corresponding Hsf mutant were diluted in DGVB²⁺ (2.5 mM veronal buffer (pH 7.3), containing 0.1% (w/v) gelatin, 1 mM MgCl₂, and 0.15 mM CaCl₂). Bacteria (10⁴ CFU) were incubated in 5% of NHS or heat-inactivated NHS in a final volume of 100 μl at 37°C. At different time points, 10-μl aliquots were removed and spread onto chocolate agar plates. After 18 h of incubation at 37°C, CFU were determined.

Microtiter plates (Nunc-Immuno Module) were coated with 40 μM purified rHsf fragments in 0.1 M Tris-HCl (pH 9.0) overnight at 4°C. Plates were washed with PBS-0.05% Tween 20 and blocked for 1 h at room temperature with PBS containing 2% BSA. After washings, the wells were incubated for 1 h at room temperature with vitronectin (5 μg/ml) in 2% BSA. Thereafter, the plates were washed and incubated with goat anti-human vitronectin pAb for 1 h. After additional washings, HRP-conjugated anti-goat pAb was added and incubated at room temperature for 40 min. Plates were developed and measured at an OD of 450 nm.

To determine the vitronectin concentration in NHS following incubation with wild-type or mutant H. influenzae, microtiter plates were coated with goat anti-human vitronectin pAb in 0.1 M Tris-HCl (pH 9.0) at 4°C overnight. Plates were washed, blocked, and incubated with NHS before (0 min) and after 20-min incubation with RM804 or the corresponding mutant. Thereafter, the plates were washed and incubated with mouse anti-human vitronectin mAb, followed by HRP-conjugated anti-mouse pAb.

Purified rHsf^608–1351 was labeled with 0.05 mol iodine (Amersham Biosciences) per molecule of protein, using the chloramine-T method (27). To define saturating conditions of ¹²⁵I-labeled Hsf^608–1351, vitronectin was incubated at increasing concentrations with ¹²⁵I-labeled Hsf^608–1351 in microtiter plates. The competition assays were essentially performed as described elsewhere (26). Briefly, microtiter plates were incubated with 0.065 μg of vitronectin overnight at 4°C in 75 mM NaCO₃ (pH 9.6). Thereafter, the wells were washed and blocked, as described above. After four washings, ¹²⁵I-labeled Hsf^608–1351 was added, together with various concentrations of unlabeled proteins diluted in blocking buffer, and followed by an overnight incubation at 4°C. After four washings, the radioactivity was measured in a gamma counter.

Two Hib strains (RM804 and Eagan) were mutated by introduction of a kanamycin resistance gene cassette in the gene encoding for Hsf (Fig. 1,A). Resulting mutants were confirmed by PCR, and the absence of Hsf expression was proven by analysis of OMPs in Western blots using a specific anti-Hsf antiserum (Fig. 1, B and C). The RM804Δhsf mutant was deficient in a high m.w. complex corresponding to Hsf (Fig. 1,C). The H. influenzae Hsf mutant was also analyzed by flow cytometry using anti-Hsf pAb (Fig. 2 B). Similar results were obtained with the H. influenzae Eagan wild type and the corresponding mutant (data not shown).

To determine whether Hsf interacted with vitronectin, soluble vitronectin (Fig. 2,A) at increasing concentrations was incubated with H. influenzae and the Hsf mutants, followed by flow cytometry analysis using polyclonal anti-vitronectin Abs (Fig. 2,C). The H. influenzae RM804 and Eagan isolates significantly bound vitronectin at a concentration of 0.5–2.5 μg. In contrast, a strongly decreased vitronectin binding was observed with RM804Δhsf and EaganΔhsf as compared with the wild-type counterpart (Fig. 2,D). To further show that Hsf interacts with soluble vitronectin, Hsf-expressing E. coli (Fig. 3,A) was included in our study. E. coli with Hsf at the surface bound vitronectin (1–5 μg) in a dose-dependent manner, whereas no binding was detected with the control bacteria (Fig. 3 B).

To investigate the attachment of bacteria to immobilized vitronectin, H. influenzae RM804 and its corresponding Hsf mutant were applied to vitronectin-coated glass slides. The H. influenzae RM804 wild type was found to strongly adhere to the vitronectin-coated glass slides (Fig. 4,A). This was in contrast to the H. influenzae RM804Δhsf mutant that barely bound the immobilized vitronectin (Fig. 4,B). Similar results were obtained with H. influenzae Eagan and the corresponding Hsf mutant (data not shown). To further prove that Hsf interacts with immobilized vitronectin, Hsf-expressing E. coli was tested. In parallel with H. influenzae, Hsf-expressing E. coli adhered to the vitronectin-coated glass slides (Fig. 4,C), whereas only a few bacteria were detected when the control E. coli wild type was analyzed (Fig. 4 D).

The vitronectin molecule harbors different functional groups, which are involved in, for example, cell attachment, collagen binding, and glycosaminoglycan binding. Three heparin binding domains exist in the N terminus and the C terminus of the vitronectin molecule (Fig. 5,A) (19, 28). To further investigate the nature of the interaction of vitronectin to H. influenzae, a series of blocking experiments with heparin was performed. The H. influenzae wild type was incubated with heparin at increasing concentrations, followed by addition of vitronectin. This commercially available heparin inhibited the binding of vitronectin to H. influenzae in a dose-dependent manner (Fig. 5,B). The binding was inhibited >80% when heparin at 10 μg/ml was added. BSA did not interfere with the binding (data not shown). Another commercial heparin preparation showed similar results. To exclude sterical hindrance of the heparin molecule, we also tested the capacity of heparin to block vitronectin binding to group A streptococci (Fig. 5 B). A significant vitronectin binding to streptococci was observed in analogy with previously published data (29), whereas any inhibitory effect of heparin on vitronectin binding to group A streptococci was not observed at heparin concentrations up to 500 μg/ml. To examine the vitronectin-Hsf interaction in detail, the peptides spanning the heparin binding site, vitronectin^348–361 and vitronectin^341–370, were preincubated with vitronectin, followed by addition of H. influenzae. These peptides did not block the vitronectin binding. Thus, the N-terminal part of the vitronectin molecule is most likely involved in the Hsf-vitronectin interaction.

Vitronectin plays a major role in the complement cascade by inhibiting the MAC of complement (19). To analyze the importance of Hsf in H. influenzae survival when exposed to NHS, the wild-type strains RM804 and Eagan, in addition to the corresponding mutants, were tested in a serum bactericidal assay. The wild-type strains were significantly more resistant to NHS as compared with the mutants devoid of Hsf (Fig. 6). Both the wild-type strain and the mutant were resistant to heat-inactivated NHS.

Vitronectin was bound to wild-type bacteria, but not to the mutant after incubation with NHS for 20 min. Quantification of residual vitronectin in serum was performed with ELISA after bacteria were spun down. Interestingly, the vitronectin serum concentration decreased with 13.3 ± 2.5% after exposure to H. influenzae, whereas no difference was seen with the Hsf-deficient mutants. Taken together, Hsf significantly contributed to H. influenzae serum resistance.

To further analyze the interactions of Hsf with vitronectin, Hsf^54–2414 was recombinantly produced in E. coli, coated on microtiter plates, and incubated with increasing concentrations of vitronectin. Bound vitronectin was detected by an anti-human vitronectin pAb, followed by incubation with an HRP-conjugated anti-goat pAb, as can be seen in Fig. 7. Hsf^54–2414 bound soluble vitronectin, and the interaction was dose dependent.

To define the vitronectin binding domain of Hsf, recombinant proteins spanning the entire molecule were manufactured. Vitronectin was incubated with immobilized Hsf fragments, and the interaction was quantified by ELISA. Interestingly, two major binding domains were found, i.e., Hsf^608–1351 and Hsf^1536–2414 (Fig. 8).

The interaction between vitronectin and the most efficient binding domain (Hsf^608–1351) was further confirmed using a competition assay after the saturated conditions of vitronectin and ¹²⁵I-labeled Hsf^608–1351 had been defined (Fig. 9,A). Vitronectin was incubated with ¹²⁵I-labeled Hsf^608–1351 in the presence of increasing Hsf^608–1351 concentrations. Unlabeled Hsf^608–1351 specifically inhibited the binding between ¹²⁵I-labeled Hsf^608–1351 and vitronectin (Fig. 9 B). A total of 95 nM Hsf^608–1351 was required to block the vitronectin/¹²⁵I-labeled Hsf^608–1351 interaction by 50% (IC₅₀).

In the present work, we demonstrate a novel interaction between the encapsulated respiratory pathogen Hib and the important complement inhibitor vitronectin. Complement resistance is crucial for bacterial virulence. Binding of complement inhibitors such as vitronectin, C4BP, or factor H is an efficient strategy used by serum-resistant pathogens (30, 31, 32). Several studies have indicated that complement proteins and regulators are present in the human respiratory tract (33, 34). Moreover, complement activity can be detected in the ECM during inflammation (34, 35).

Hsf is a major adhesin in Hib (17). It is a large, highly conserved autotransporter protein, which extrudes as thin fibrils from the bacterial surface. Flow cytometry analysis of the Hsf-deficient mutant revealed that Hsf is the major vitronectin-binding protein in encapsulated H. influenzae. Hsf-expressing H. influenzae and E. coli transformants bound soluble vitronectin at increasing concentrations (Figs. 2 and 3,B). This finding is in contrast to what has been shown in a previous study, in which no binding of H. influenzae to soluble vitronectin was found (9). In addition to H. influenzae, Staphylococcus aureus, E. coli, and β-hemolytic streptococci are efficient binders of soluble vitronectin (29, 36). Furthermore, we demonstrate that the Hsf-expressing H. influenzae and E. coli transformants both bound to immobilized vitronectin (Fig. 4, A and C). In contrast, when Hsf was deleted in H. influenzae, a significantly decreased binding was observed (Fig. 4,B). Three heparin binding domains of the vitronectin molecule (residues 82–137, 175–219, and 348–376) have been identified (Fig. 5,A) (20, 28). Heparin inhibited the binding between Hsf-expressing H. influenzae and vitronectin. Two different commercial preparations of heparin were tested with similar results. To further prove the specificity of the heparin blocking, S. pyogenes was included in our study. The binding of S. pyogenes to vitronectin was not inhibited by heparin (Fig. 5 B). This result suggests that the interaction between heparin and H. influenzae is specific. Peptides encompassing the C-terminal heparin binding domain, vitronectin^348–361 and vitronectin^341–370, were also used in blocking experiments. However, any blocking could not be detected with the two peptides, suggesting that the N-terminal heparin binding domains are involved in the interaction.

The H. influenzae devoid of Hsf had a markedly reduced survival as compared with the wild type when exposed to NHS (Fig. 6). The ability to bind vitronectin suggests that Hsf uses the capacity of vitronectin to inhibit the complement-mediated attack. An important function of vitronectin is inhibition of the MAC of the complement cascade that is the first line of defense (20, 37). Vitronectin binds C5b-7, which is one of the end products in the complement cascade, and inhibits attachment to the cell membrane and induction of cell lysis of the bacteria (37, 38). It also binds the C5b-9 complex and blocks the tubular polymerization of C9, which is responsible for induction of the cell lysis. These two mechanisms make it impossible for the complement to attach to the surface of the bacteria, resulting in bacterial survival. Interestingly, it has been suggested that Yersinia enterocolitica OMP YadA acts in a similar fashion as vitronectin by sterically hindering the formation of the MAC (39). In addition to inhibiting the complement system, vitronectin is involved in attachment and spreading of endothelial cells to the ECM and in wound healing by binding the thrombin-antithrombin III complex (20, 38). Furthermore, vitronectin promotes the coagulation cascade by binding and activating the plasminogen activator inhibitor 1.

Localization of the vitronectin binding domains of Hsf is an important step in defining the function of Hsf. Recombinantly produced Hsf^54–2414 bound vitronectin in a dose-dependent manner (Fig. 7). In addition, Hsf^608–1351 and Hsf^1536–2414 from the clinical isolate H. influenzae RM804 contained two vitronectin binding domains. Hsf^608–1351 displayed the highest affinity for vitronectin and showed both a dose-dependent and specific binding to vitronectin (Fig. 9). These data were confirmed by dot-blot assay, in which nitrocellulose membranes were coated with vitronectin and incubated with ¹²⁵I-labeled Hsf fragments (data not shown). During preparation of this manuscript, St. Geme and coworkers (40) demonstrated that Hsf binds Chang epithelial cells via two acidic binding domains. These domains comprise aa 537–652 and 1904–2022. They also identified a third binding pocket (Hsf^1214–1338), which did not bind to Chang cells. Our strongest binding domains contain two of these three adhesive binding pockets. Hsf is anchored in the outer membrane by its C-terminal translocator domain, and one Hsf fiber may thus be able to bind two vitronectin molecules and stabilizing adherence despite the physical forces in the respiratory tract, which includes the mucociliary escalator, sneezing, and coughing (41).

Vitronectin is also a component of the ECM (20). Binding of H. influenzae to exposed ECM components may contribute to bacterial adherence, which is an essential step in the bacterial pathogenesis. One hypothesis is that these interactions contribute to the spread of bacteria through tissue barriers into secondary infection sites. Previous studies have shown that H. influenzae can interact with ECM and reconstituted basement membranes from cultured human epithelial cells (12). Binding to ECM proteins makes the bacteria able to reach deeper tissue layers of the mucosa. Hsf-mediated bacterial attachment to vitronectin may be an important factor in initial colonization and spread of the bacteria to new sites of infection. The ability to bind vitronectin is of great importance for several bacterial species. Neisseria gonorrhoeae probably uses vitronectin as a bridge for attachment and invasion of human cells (41). Binding of N. gonorrhoeae to specific integrins can trigger endocytosis of vitronectin and consequently engulfment of the bacteria. The interaction with the integrin receptor occurs by direct binding or through binding of vitronectin. In addition, vitronectin mediates attachment of Candida albicans to endothelial cells and of Pneumocystis carinii to bronchial epithelial cells in the lower respiratory tract (42, 43).

In conclusion, we have presented several lines of evidence on H. influenzae Hsf binding to vitronectin, a factor that inhibits the MAC formation in the complement system, preventing complement-induced cell lysis. This interaction may also contribute to bacterial colonization and spread of H. influenzae. Hsf is the major vitronectin-binding protein and consists of two separate binding domains. Hsf binding to vitronectin contributes to H. influenzae serum resistance and, consequently, virulence.

The authors have no financial conflict of interest.