Integrin αvβ1 promotes infection by human metapneumovirus

Abstract

Human metapneumovirus (hMPV) is a recently described paramyxovirus that causes lower respiratory infections in children and adults worldwide. The hMPV fusion (F) protein is a membrane-anchored glycoprotein and major protective antigen. All hMPV F protein sequences determined to date contain an Arg-Gly-Asp (RGD) sequence, suggesting that F engages RGD-binding integrins to mediate cell entry. The divalent cation chelator EDTA, which disrupts heterodimeric integrin interactions, inhibits infectivity of hMPV but not the closely related respiratory syncytial virus (RSV), which lacks an RGD motif. Function-blocking antibodies specific for αvβ1 integrin inhibit infectivity of hMPV but not RSV. Transfection of nonpermissive cells with αv or β1 cDNAs confers hMPV infectivity, whereas reduction of αv and β1 integrin expression by siRNA inhibits hMPV infection. Recombinant hMPV F protein binds to cells, whereas Arg-Gly-Glu (RGE)-mutant F protein does not. These data suggest that αvβ1 integrin is a functional receptor for hMPV.

Human metapneumovirus (hMPV) is a major cause of upper and lower respiratory infections in children and adults (1–3). hMPV is genetically most closely related to respiratory syncytial virus (RSV), which also is a significant respiratory pathogen of infancy and early childhood (4). hMPV is associated with significant morbidity in young infants and other high-risk populations with underlying medical conditions such as prematurity, asthma, cardiopulmonary disease, and immune compromise (5). Hospitalization rates for hMPV infection in previously healthy infants or high-risk groups are comparable to those caused by other common respiratory viruses such as RSV, parainfluenza virus, or influenza virus (2, 5). There currently is no licensed vaccine or antiviral therapy for this important human pathogen.

hMPV has a negative-sense, single-stranded RNA genome of ≈13 kb that encodes 9 proteins (6, 7). Two of these are membrane-anchored glycoproteins, the fusion (F) and attachment (G) proteins. Studies of mutant viruses and recombinant proteins show that F is the major protective antigen (8–10). The hMPV F protein is similar to other paramyxovirus fusion proteins and incorporates homologous regions that likely have homologous functions (6, 7). Paramyxovirus fusion proteins are synthesized as inactive precursors (F0) that are cleaved by host-cell proteases into the fusion-active F1 and F2 domains. hMPV F contains a conserved cleavage site, fusion peptide, and heptad-repeat domains. Concordantly, ectopic expression of hMPV F is sufficient to mediate cell-cell fusion (11). Synthetic peptides that correspond to the heptad-repeat regions are potent inhibitors of hMPV infection, suggesting that hMPV F is a class I fusion protein (12). Proteinaceous cellular receptors have been identified for some paramyxoviruses, including measles virus (13, 14) and Nipah virus (15, 16). Sialylated cell-surface proteins serve as receptors for other paramyxoviruses, such as parainfluenza virus (17, 18). For all paramyxoviruses identified to date, the G or H/HN protein rather than the F protein is the receptor-binding moiety. Receptors for hMPV have not been reported.

Integrins are heterodimeric cell-surface molecules that consist of α and β subunits (19). Integrins mediate cellular adhesion to the extracellular matrix, regulate cellular trafficking, and transduce both outside-in and inside-out signaling events (19). Integrins bind specific ligands in a divalent cation-dependent manner, and a subset of integrins bind to specific recognition sequences such as the amino acid Arg-Gly-Asp (RGD) motif (19). Several viruses use RGD-binding integrins as receptors or co-receptors, including adenovirus (20), echovirus (21), foot-and-mouth disease virus (22), hantavirus (23), human herpes virus-8 (24), human parechovirus (25), reovirus (26), rotavirus (27), and West Nile virus (28). A role for integrins in infection by any paramyxovirus has not been elucidated.

In this study, we tested the hypothesis that integrins serve as functional receptors for hMPV. The divalent cation chelator EDTA, which reduces integrin-ligand binding, diminished hMPV infectivity but had no effect on RSV. Function-blocking αvβ1 integrin-specific antibodies inhibited infection by hMPV but not RSV. Transfection of αvβ1 cDNA into nonpermissive cells conferred hMPV infectivity. Reduction of αvβ1 expression by siRNA diminished hMPV infection. Finally, recombinant F protein bound specifically to permissive cells, and mutation of the F protein RGD motif abolished this binding. Collectively, these data provide strong evidence that the hMPV F protein engages αvβ1 integrin to mediate functional virus entry.

Results

hMPV F Contains a Conserved RGD Motif.

There are 4 distinct genetic lineages of hMPV, provisionally designated A1, A2, B1, and B2 (29). We compared full-length F gene sequences from 71 clinical isolates of hMPV collected worldwide over a 20-year period. An RGD motif at residues 329–331, along with 5 N- and 16 C-terminal flanking residues, is strictly conserved in all sequences studied, despite diversity in other regions of the F gene (Fig. S1). This motif, which is predicted to be in a solvent-exposed region of the protein (30) (Fig. 1), is unique to hMPV F among human paramyxovirus fusion proteins and absent in all published RSV F sequences (data not shown). Given the role of RGD sequences in integrin binding, this finding led us to hypothesize that integrins function as receptors for hMPV.

Fig. 1.

hMPV F has a conserved RGD sequence in the ectodomain. Schematic representation of putative domains of hMPV F protein. SS, signal sequence; FP, fusion peptide; HRA, heptad repeat A; RGD, RGD amino acid motif; HRB, heptad repeat B; TM, transmembrane domain; and CT, cytoplasmic tail. Arrowheads indicate N-linked glycosylation sites. (Scale bar, amino acids.)

EDTA and RGD-Specific Peptides Reduce hMPV Infectivity.

In many cases, integrin-ligand interactions are dependent on divalent cations, including Ca²⁺ and Mg²⁺ (19). Attachment of integrin-using viruses to RGD-binding integrins is diminished by the divalent cation chelator EDTA (19, 24, 31–33). We tested the effect of EDTA on hMPV binding and infection of cells using a quantitative immunostaining assay. The closely related paramyxovirus, RSV, which lacks an RGD motif in its analogous F protein, was used as a control. Rhesus monkey kidney LLC-MK2 cells, which are permissive for infection by both hMPV and RSV, were incubated with increasing concentrations of EDTA before adsorption with either hMPV or RSV (Fig. 2A). EDTA blocked hMPV infection in a dose-dependent manner but had no effect on RSV infectivity at any concentration tested. The cell sheet began to detach at EDTA concentrations greater than 2.5 mM; however, hMPV infection was not detected in the remaining adherent cells. RSV-infected cells were present at EDTA concentrations >2.5 mM in numbers that appeared similar to those following treatment with lower EDTA concentrations. We conclude that EDTA exerts a specific inhibitory effect on hMPV, supporting a role for integrins in hMPV infection.

Fig. 2.

EDTA and linear RGD peptides inhibit infection by hMPV but not RSV. LLC-MK2 cells were treated with increasing concentrations of (A) EDTA (0.3 mM, 0.6 mM, 1.25 mM, and 2.5 mM), (B) linear peptide GRGDSP (0.15 μM, 0.3 μM, and 0.6 μM), or (C) linear peptide GRGESP (0.15 μM, 0.3 μM, and 0.6 μM) before adsorption with either hMPV (black bars) or RSV (white bars). Infected cells were identified by indirect immunostaining for viral antigen. Black triangles represent increasing EDTA or peptide concentration. C indicates untreated controls. Results are expressed as the mean percentage of infected cells following inhibitor treatment compared with the mean percentage of infected control cells for at least three independent experiments performed in triplicate. Error bars represent the SEM. Only P values <0.05 (t test) are shown.

To determine whether the RGD tripeptide is required for hMPV infection, we incubated LLC-MK2 cells with the integrin-binding peptide, GRGDSP, and the control peptide, GRGESP, before viral adsorption. Treatment of cells with GRGDSP produced dose-dependent inhibition of hMPV infectivity (Fig. 2B), with ≈45% inhibition observed at a concentration of 0.6 μM, whereas the control GRGESP peptide had a minimal, nonsignificant effect (Fig. 2C). Neither peptide produced statistically significant inhibition of RSV infection (Fig. 2 B and C), suggesting a specific effect of the GRGDSP peptide on infection by hMPV. Although the inhibitory effect of the linear GRGDSP peptide is modest, the magnitude of the infectivity blockade approximates that achieved in studies of other viruses known to engage integrins (27, 33–36).

αvβ1 Integrin-Specific Antibodies Reduce hMPV Infectivity.

To identify integrins required for hMPV infection, we screened mAbs directed against the specific RGD-binding integrins expressed by LLC-MK2 cells (data not shown) for the capacity to inhibit hMPV infection. Cells were incubated with antibodies specific for RGD-binding integrins alone or in combination, adsorbed with either hMPV or RSV, and scored for infection by immunostaining (Fig. 3 A and B). Antibodies specific for αv and β1 integrins displayed the greatest inhibition of hMPV infectivity, with an αv-specific antibody demonstrating ≈55% inhibition and a β1-specific antibody demonstrating ≈40% inhibition. The combination of αv and β1 integrin-specific antibodies showed >80% inhibition. Antibodies specific for αvβ3 and α5β1 integrins exhibited 35% and 30% inhibition, respectively, whereas α5 and β3 integrin-specific antibodies had a minimal, nonsignificant effect. Thus, antibodies directed against either αv or β1 integrin significantly inhibit hMPV infection. Control antibodies specific for α2 integrin and junctional adhesion molecule-A (JAM-A) had no effect at any concentration tested (Fig. 3 A and C). Importantly, none of the integrin-specific or control antibodies had any significant effect on the infectivity of RSV (Fig. 3 B). The degree of inhibition of hMPV infection by αv and β1 integrin-specific antibodies was dose-dependent (Fig. 3C), providing further evidence that the inhibition of hMPV infection results from integrin blockade. These data suggest that cell-surface expression of αvβ1 integrins is required for infectivity of hMPV.

Fig. 3.

Integrin function-blocking antibodies inhibit hMPV infection. LLC-MK2 cells were incubated with the integrin function-blocking or control antibodies shown before adsorption with either hMPV (A, black bars) or RSV (B, white bars). Infected cells were identified by indirect immunostaining. C indicates untreated controls. (C) LLC-MK2 cells were preincubated with increasing concentrations of the antibodies shown, adsorbed with hMPV, and scored for infection by indirect immunostaining. Black triangle represents increasing antibody concentration. Results are expressed as the mean percentage of infected cells following antibody treatment compared with the mean percentage of infected control cells for at least three independent experiments performed in triplicate. Error bars represent the SEM. Only P values <0.05 (t test) are shown. Antibody concentrations in A and B: αv (7.5 μg/mL); β1 (4 μg/mL); αvβ3 (20 μg/mL); α5 (2 μg/mL); β3 (20 μg/mL); α2 (20 μg/mL); and JAM-A (20 μg/mL). Antibody concentrations in C: JAM-A (5, 10, and 20 μg/mL), α2 (5, 10, and 20 μg/mL), αv (3.75, 7.5, and 15 μg/mL), and β1 (1, 2, and 4 μg/mL).

Reduced Expression of αv and β1 Integrins Decreases hMPV Infectivity.

To determine whether diminished integrin expression limits hMPV infection, we transfected BEAS-2b human bronchial epithelial cells with siRNAs specific for αv, β1, or both integrins before adsorption with hMPV or RSV. Transfection with these siRNAs reduced cell-surface expression of αv and β1 integrins in the majority of the cells (Fig. S2), although a small subset of cells appeared to exhibit increased β1 integrin surface expression (Fig. S2 D and F). As controls, mock transfection or transfection with nonspecific siRNA did not diminish cell-surface expression of αv or β1 integrin (Fig. S2). The siRNA-transfected cells were adsorbed with hMPV or RSV and scored for infectivity by immunostaining. The αv and β1 integrin-specific siRNAs decreased hMPV infectivity by 26% and 35%, respectively (Fig. 4A). Transfection of cells with both αv and β1 integrin-specific siRNA exhibited the most potent inhibitory effect, reducing infection by ≈65% (Fig. 4A). In contrast, mock transfection or transfection of cells with nonspecific control siRNA did not affect hMPV infection. Consistent with results of experiments using integrin-specific antibodies, none of the transfected siRNAs (control or integrin-specific) inhibited RSV infectivity (Fig. S3). In fact, we detected an increase in RSV-infected cells following treatment with integrin-specific siRNAs. We speculate that this effect is attributable to altered cell morphology or cytoskeletal rearrangements that serve to promote RSV infection (37). Therefore, reducing cell-surface expression of αv and β1 integrins reduces the efficiency of hMPV infection.

Fig. 4.

Expression of αv and β1 integrins correlates with infection of cells by hMPV. (A) BEAS-2b cells were mock-transfected, transfected with siRNAs specific for αv, β1, or αv and β1 integrins, or transfected with nonspecific control siRNA before adsorption with hMPV. Results are expressed as the mean percentage of infected cells following transfection of siRNA compared with the mean percentage of infected control cells for at least three independent experiments performed in triplicate. (B) CHO cells were transfected with expression plasmids encoding either human αv (CHOαv) or human β1 (CHOβ1) integrin, with and without additional incubation with function-blocking antibodies specific for αv (CHOαv+αv mAb) or β1 (CHOβ1+β1 mAb), before adsorption with hMPV. Surface expression of human αv or human β1 integrin was confirmed by flow cytometry (data not shown). Infected cells were identified by indirect immunostaining. Results are expressed as the mean percentage of infected cells following transfection of integrin cDNA compared with the mean percentage of infected control cells for at least three independent experiments performed in triplicate. Error bars represent the SEM. Only P values <0.05 (t test) are shown.

Transfection of αv and β1 Integrins Enhances hMPV Infectivity.

CHO cells are poorly permissive for hMPV infection. Following adsorption at equivalent multiplicity of infection, 80% fewer CHO cells are infected by hMPV compared with LLC-MK2 cells (Fig. 4B). To complement the loss-of-function experiments described thus far, we tested whether transfection of CHO cells with cDNAs encoding human αv or β1 integrin alters the infectivity of hMPV. Expression of human αv or β1 integrin in CHO cells substantially increased the efficiency of hMPV infection (Fig. 4B). Expression of αv integrin in CHO cells increased hMPV infection by 2-fold, and expression of β1 integrin increased infection by 3-fold. The increased efficiency of hMPV infection of human integrin-transfected CHO cells was inhibited by preincubation with αv and β1 integrin function-blocking antibodies (Fig. 4B). Thus, human αv and β1 integrin expression enhances the efficiency of hMPV infection of CHO cells, and blockade of these integrins reduces the magnitude of this effect.

Binding of hMPV F Protein to LLC-MK2 Cells Is Mediated by the RGD Motif.

To determine whether a soluble fragment of hMPV F protein is capable of binding to permissive cells, LLC-MK2 cells, which express abundant αvβ1 integrin (data not shown), were incubated with FΔTM protein, and binding was detected by flow cytometry. In contrast to LLC-MK2 cells incubated with primary and secondary antibodies alone (Fig. 5A), LLC-MK2 cells incubated with FΔTM protein and stained using hMPV-specific antibody showed a significant increase in fluorescence, indicating that FΔTM is capable of binding these cells (Fig. 5 A and F). As a critical specificity control, a mutant form of FΔTM protein in which the RGD motif was altered to Arg-Gly-Glu (RGE) demonstrated no specific binding (Fig. 5A). Furthermore, WT FΔTM protein showed minimal binding to CHO cells (Fig. 5A), suggesting that the limited infection of CHO cells by hMPV is attributable to diminished interactions of F with a cell-surface receptor. To quantify F protein binding to cells, we calculated the percentage of cells contained within identical gates of the flow cytometric profiles. Percent binding as determined by fluorescence intensity differed significantly between CHO cells incubated with FΔTM protein, LLC cells incubated with buffer alone or mutant F-RGEΔTM, and LLC cells incubated with FΔTM protein (Fig. 5B). Thus, soluble F protein binds to cells expressing αvβ1 integrin in a manner dependent on the RGD motif.

Fig. 5.

HMPV F protein binding to LLC cells is mediated by the RGD motif. (A) LLCMK2 cells were incubated with buffer alone (LLC), FΔTM protein (F), or RGE mutant FΔTM protein (F RGE) or CHO cells were incubated with FΔTM protein (CHO). Cells were stained with guinea pig anti-hMPV antibody and AlexaFluor-568 anti-guinea pig Ig and analyzed by flow cytometry. (B) The results are presented as the mean percentage of events within the gate region (M1) indicative of binding from four independent experiments conducted as in A. The M1 gate region was set at the first decade of fluorescence for all experiments. Error bars represent the SEM. Only P values <0.05 (t test) are shown.

Discussion

hMPV F protein contains an RGD motif that is absolutely conserved, along with 5 N- and 16 C-terminal flanking residues, among 71 clinical isolates collected over a 20-year period throughout the world. The observed sequence conservation suggests that this region serves an important biological function in hMPV infection, as F is the major antigenic target and subject to variation in other domains (7, 38, 39). The structure of hMPV F has not been reported to our knowledge, but the region containing the RGD sequence is predicted to be solvent-exposed (12, 30). Among the human paramyxoviruses, the RGD motif is unique to hMPV. Treatment of cells with EDTA, which inhibits integrin function, potently decreased hMPV infectivity while showing no effect on infection by RSV. Furthermore, RGD peptides inhibited hMPV at sub-micromolar concentrations, whereas RGE peptides had a minimal effect that was not statistically significant. The degree of inhibition was modest, but it is noteworthy that similar levels of inhibition for other viruses often require millimolar concentrations of peptide (27, 33–36).

Integrin-specific function-blocking mAbs exhibited potent inhibition of hMPV infection, with antibodies specific for αv and β1 integrin showing the greatest effect, both independently and in combination. α5+β1 and αvβ3 integrin-specific antibodies also inhibited hMPV, whereas either α5 or β3 integrin-specific antibodies alone had minimal effects. We interpret these data to indicate that the αvβ1 heterodimer is a functional hMPV receptor and that antibodies directed against either subunit can effectively inhibit infectivity. However, substantial amino acid sequence identity exists among the β1, β3, and β5 integrin subunits (19), raising the possibility that αv paired with other related β integrin subunits is capable of acting as an alternative hMPV receptor. Such is the case with foot-and-mouth-disease virus, which preferentially utilizes αvβ3 integrin as a receptor but is also capable of using αvβ1 and αvβ6 (22, 35, 40).

Further evidence for the utilization of αvβ1 integrin as a receptor for hMPV comes from complementary approaches to alter the expression of αv and β1 integrin subunits. Diminished surface expression of both molecules by RNA interference resulted in significant inhibition of hMPV infection. Importantly, control siRNA had no effect on either integrin expression or hMPV infectivity. Moreover, the integrin-specific siRNAs tested here did not inhibit RSV infectivity, suggesting that dampened hMPV growth is not attributable to a nonspecific effect of siRNA (41). A small number of siRNA-treated cells exhibited increased surface expression of β1 integrin, possibly via dysregulation of integrin homeostatic networks, altered β1 integrin mRNA transcription, or β1 integrin pairing with alternate α subunits (42, 43). Therefore, we may have underestimated the effect of siRNA-mediated knockdown of integrin expression on hMPV infection. In concordance with the siRNA experiments, transient transfection of CHO cells with human αv or β1 integrin allowed hMPV to infect these poorly permissive cells. Because hamster αv and β1 integrin are expressed on the CHO cell surface (44–47), it is possible that the hamster integrin subunits partner with the human counterparts to provide partial complementation of hMPV infectivity, as observed in our experiments. Nonetheless, it is possible that species-specific integrin expression is a host-range determinant for hMPV.

Recombinant hMPV F protein bound specifically to LLC-MK2 but not to CHO cells, an effect that was abrogated by mutation of the RGD motif to RGE. Although we have not demonstrated direct binding of virus or F protein to αvβ1, our data strongly suggest that the F protein interacts directly with αvβ1 integrin. To our knowledge, fusion protein-receptor binding has not been reported for other human paramyxoviruses. The related RSV F protein induces innate immune responses mediated by Toll-like receptor (TLR) 4, but direct interactions between RSV F and TLR4 have not been described (48, 49). RSV induces alterations in epithelial sodium transport, and purified F protein recapitulates this effect (50). However, a cell-surface ligand for RSV F has not been identified. Measles virus F protein interacts with DC-SIGN, although this interaction does not lead to functional cell entry (51).

Results reported here do not define the precise step in hMPV cell entry mediated by αvβ1 integrin. As most of the experiments we performed measured viral protein expression in infected cells, an alternative possibility is that integrin expression increases viral transcription and translation rather than mediating attachment or cell entry. However, the flow cytometry data demonstrate that F binding to cells is RGD-dependent. Based on the established function of integrins as receptors for other viruses, we favor a role for integrins in initiating the process of infection. For the analogous HIV class I fusion protein, gp120/gp41, binding to primary receptor (i.e., CD4) and co-receptor (i.e., CCR5) triggers the conformational changes associated with virus-cell membrane fusion (52), an event that occurs rapidly (53). Influenza virus HA-mediated fusion occurs within endosomes and is triggered by low pH (54). Paramyxovirus fusion is thought to occur at the cell surface at neutral pH (55). We speculate that hMPV F protein binding to αvβ1 integrin initiates the conformational changes that lead to fusion at either of these sites.

Our results indicate that αvβ1 integrin is required for efficient hMPV infection. Blockade of cell-surface αvβ1 integrin inhibits hMPV infectivity, and expression of αvβ1 integrin allows nonpermissive cells to be infected. The hMPV F protein binds specifically to cells, and alteration of the RGD motif abolishes this interaction. Collectively, these data suggest that αvβ1 integrin is a functional hMPV receptor and a paramyxovirus fusion protein receptor. These findings extend the widening role of integrins as receptors for intracellular microbial pathogens and provide potential therapeutic targets for hMPV.

Methods

Cells, Viruses, Peptides, and Antibodies.

LLC-MK2 cells (ATCC CCL-7) were maintained in OptiMEM medium (Invitrogen) supplemented to contain 2% FCS, 2 mM L-glutamine, 50 μg/mL gentamicin, and 2.5 μg/mL amphotericin. CHO cells were maintained in DMEM/F12 medium (Invitrogen) supplemented to contain 10% FCS, glutamine, and antibiotics as for LLC-MK2 cells. The hMPV strain, TN/96–12, a genotype group A1 virus, was used for all experiments. Virus was propagated using LLC-MK2 cells (ATCC CCL-7) maintained in serum-free LLC-MK2 growth medium with 5 μg/mL trypsin (Invitrogen) as described (57). The virus stock had a titer of 1 × 10⁶ pfu/mL by plaque titration using LLC-MK2 cell monolayers (56). HMPV TN/96–12 was a clinical isolate passaged 8 times in LLC-MK2 cells before this study. RSV strain A2 was cultivated using LLC-MK2 cells according to the method described for growth of hMPV and achieved an equivalent titer.

Linear GRGDSP and GRGESP peptides were purchased from American Peptide. Polyclonal anti-hMPV antibody was generated by infecting guinea pigs twice intranasally with 10⁵ pfu of sucrose-gradient-purified hMPV. This anti-hMPV antibody has low nonspecific background, does not cross-react with other viruses by immunofluorescence or immunoblotting, and detects recombinant hMPV F protein with high sensitivity (8, 56). Function-blocking human integrin-specific mAbs MAB2021Z (αv; clone AV1), MAB1957Z (β3; clone 25E11), and MAB1976Z (αvβ3; clone LM609) were purchased from Chemicon. Function-blocking human integrin-specific mAbs AIIB2 (β1) and BIIG2 (α5) were obtained from the Developmental Hybridoma Studies Bank at the University of Iowa. Human JAM-A-specific mAb J10.4 was provided by Charles Parkos (Atlanta, GA) (57). Function-blocking human α2-specific mAb 6F1 was provided by Richard Bankert (Buffalo, NY) (58).

Immunostaining Assay of Viral Infection.

Confluent monolayers of cells grown in 48-well plates (Costar) were washed twice with Dulbecco PBS solution. Dilutions of EDTA, peptides, or antibodies were prepared in serum-free OptiMEM medium, and 75 μL of each dilution or medium alone was added to wells in triplicate. Cells were incubated at 37 °C for 90 min, followed by incubation at 4 °C for 30 min. Cells were adsorbed with hMPV at multiplicity of infection of ≈60 pfu per well and incubated at RT for 60 min (adsorption) and 37 °C for 60 min (internalization), with occasional rocking. Cells were washed twice with Dulbecco PBS solution and incubated in 0.5 mL cell growth medium at 37 °C for 24 h. Cells were fixed in formalin at RT for 1 h, washed with diH₂O, and incubated with blocking buffer (5% milk, 0.05% Tween in PBS solution) at 37 °C for 30 min. Cells were incubated with polyclonal guinea pig anti-hMPV antibody in blocking buffer at 37 °C for 2 h, washed, and incubated with an HRP-conjugated goat anti-guinea pig IgG antibody (Southern Biotech) at 37 °C for 2 h. Cells were washed, incubated with TrueBlue peroxidase substrate (Kirkegaard and Perry Laboratories) at room temperature for 10 min, washed, and air-dried. Infected cells were visualized by using a dissecting microscope.

Subcloning of Integrin Constructs Used for Transfection.

A cDNA encoding human αv integrin (59) cloned into the BamHI and XbaI site of pcDNA-1neo (Invitrogen) was excised by digestion and sub-cloned into pcDNA3.1+ (Invitrogen) using complementary restriction enzyme sites. A cDNA encoding human β1 integrin cloned into the EcoRI and XhoI sites of pBluescript SK (ATCC) was excised by digestion and sub-cloned into complementary restriction enzyme sites of pcDNA3.1+. A cDNA encoding human integrin β3 (60) cloned into the XbaI site of pcDNA-1neo was excised by digestion and sub-cloned into the XbaI site of pcDNA3.1. A cDNA encoding β5 (61) cloned into the EcoRI site of pcDNA-1neo was excised by digestion and sub-cloned into the EcoRI site of pcDNA3.1+. Integrin expression plasmids were sequenced to confirm fidelity of cloning.

Transfection of Human Integrin cDNA.

Cells grown in 24-well plates (Costar) were transfected with either empty vector or plasmids encoding human integrin constructs using TransFectin reagent (Bio-Rad). Cells were incubated for 24 to 48 h to allow receptor expression before adsorption with hMPV for infectivity studies.

Knockdown of Integrin Expression with siRNA.

BEAS-2b cells grown in 24-well plates to achieve 90% confluence on the day of siRNA transfection were mock-transfected, or transfected with integrin-specific or control siRNAs (Qiagen AllStars Negative Control siRNA) using HiPerFect reagent (Qiagen). After 48 h incubation, cells were either harvested for determination of cell-surface integrin expression by flow cytometry or adsorbed with hMPV or RSV for infectivity studies.

HMPV F Protein Binding Assay.

The hMPV F protein RGD motif was changed to RGE by site-directed mutagenesis using Quikchange II (Stratagene). WT and RGE mutant hMPV F ectodomain (FΔTM) constructs were purified as described (8). LLC-MK2 cells and CHO cells were detached from plates and pelleted by centrifugation. Cells (2–3 × 10⁵/mL) were resuspended in PBS/1% FCS (FACS buffer), pelleted, and washed with FACS buffer. Cells were resuspended in 0.2 mL FACS buffer with 10 μg of purified hMPV F ectodomain and incubated at 4 °C for 1 h with rocking. Cells were centrifuged, the supernatant was removed, and cells were incubated with anti-hMPV polyclonal guinea pig serum in 0.2 mL FACS buffer at 4 °C for 30 min. Cells were pelleted, washed twice, and stained with AlexaFluor 568-conjugated goat anti-guinea pig IgG secondary antibodies (Molecular Probes) in 0.2 mL FACS buffer at 4 °C for 30 min. Cells were resuspended in FACS buffer and analyzed by flow cytometry using a LSRII flow cytometer (BD Biosciences). F protein binding to cells was quantified by setting the M1 gate region at the first decade of fluorescence in all experiments and enumerating the percentage of cells contained within this gate.

Statistical Analysis.

Means of replicate experiments, each performed in triplicate, were compared by using an unpaired, 2-tailed Student's t test assuming unequal variance. P values <0.05 were considered to be statistically significant.