|Home | About | Journals | Submit | Contact Us | Français|
Hendra virus (HeV) is an emerging paramyxovirus capable of infecting and causing disease in a variety of mammalian species, including humans. The virus infects its host cells through the coordinated functions of its fusion (F) and attachment (G) glycoproteins, the latter of which is responsible for binding the virus receptors ephrinB2 and ephrinB3. In order to identify the receptor binding site, a panel of G glycoprotein constructs containing mutations was generated using an alanine-scanning mutagenesis strategy. Based on a predicted G structure, charged amino acids residing in regions that could be homologous to those in the measles virus H attachment glycoprotein known to be involved in its protein receptor interaction were targeted. Using a coprecipitation-based assay, seven single-amino-acid substitutions in HeV G were identified as having significantly impaired binding to both the ephrinB2 and ephrinB3 viral receptors: D257A, D260A, G439A, K443A, G449A, K465A, and D468A. The impairment of receptor interaction conferred a concomitant diminution in their abilities to promote membrane fusion when coexpressed with F. The G glycoprotein mutants were also recognized by three or more conformation-dependent monoclonal antibodies of a panel of five, were expressed on the cell surface, and retained their abilities to bind and coprecipitate F. Interestingly, some of these mutant G glycoproteins coprecipitated with F more efficiently than wild-type G. Taken together, these data provide strong biochemical and functional evidence that some of these residues could be part of a conformation-dependent, discontinuous, and overlapping ephrinB2 and -B3 binding domain within the HeV G glycoprotein.
The genus Henipavirus within the Paramyxoviridae is comprised of two newly emergent and highly virulent zoonoses, Hendra virus (HeV) and Nipah virus (NiV). Pteropid bats, commonly referred to as flying foxes, within the family Pteropodidae, are the natural hosts of HeV and NiV (reviewed in reference 15). HeV emerged in Australia in 1994 as a severe respiratory disease in horses and was also transmitted to humans (reviewed in reference 30). NiV was recognized later in peninsular Malaysia during a larger outbreak of encephalitis among pig farmers in 1998 and 1999 and was primarily transmitted to humans from infected pigs (reviewed in reference 12). Since their initial recognition, additional spillover events involving both HeV and NiV have continued to occur. HeV caused fatal infections of horses in Australia in 1999 (18), 2004 (1), and 2006 (29). In the 2004 incident, a veterinarian was also infected and recovered with seroconversion. NiV has caused at least four additional recognized outbreaks in Bangladesh and one in India since 2001 (reviewed in reference 15). Of particular note, the most recent episodes in 2004 and 2005 were associated with a higher incidence of acute respiratory distress syndrome in conjunction with encephalitis, person-to-person transmission, higher fatality rates (~75%), and apparently direct transmission from a natural reservoir to humans (2-4, 21). The broad host range, together with their zoonotic potential, high virulence, and unique genetic makeup, has set HeV and NiV apart from all other paramyxoviruses, and they are presently categorized as biological safety level 4 agents.
Paramyxoviruses, like HeV and NiV, infect cells through a pH-independent fusion event mediated by two membrane-expressed surface glycoprotein spikes; an attachment glycoprotein, which, depending on the particular virus, is designated either the hemagglutinin-neuraminidase protein (HN), the hemagglutinin protein (H), or glycoprotein (G), which has neither hemagglutinating nor neuraminidase activities (reviewed in reference 28), and a fusion glycoprotein (F), which facilitates the fusion between the virus and the host cell membrane (reviewed in reference 6). Early characterization of HeV and NiV found them to be genetically closely related, with identical host cellular tropisms, and not surprisingly, they have been recently shown to utilize the same cellular receptor, known as ephrinB2, for infection of host cells (5, 32). Additionally, ephrinB3, a related protein with a high degree of sequence homology, has also been shown to be a functional receptor for NiV (32). EphrinB2 and ephrinB3 are members of a family of surface-expressed glycoprotein ligands that bind to Eph receptors, a large family of tyrosine kinase receptors (14, 35). Eph receptors and ephrins play key roles during development, especially in the nervous and vascular systems. EphrinB2 is expressed in neurons, smooth muscle, arterial endothelial cells, and capillaries (reviewed in references 20, 37, and 41), and its identification as a functional henipavirus receptor has aided in understanding the tropism and subsequent pathogenic processes brought about by infection with these viruses in several host species (reviewed in reference 16).
Measles virus (MeV) and canine distemper virus are the only other paramyxoviruses that have been shown to make use of host cell membrane-expressed proteins as receptors for cellular infection (CD46 and CD150/SLAM) (13, 17, 31). The domains of MeV H that engage CD46 and CD150 have been modeled to lie along the top of the globular head, forming conformation-dependent and overlapping binding sites composed primarily of residues in beta sheets 4 and 5 (22, 24, 25, 36). The structure of the henipavirus G has yet to be determined; however, the amino acid sequences of both HeV and NiV G proteins can accommodate a six-bladed beta-propeller structural model similar to other paramyxovirus H or HN glycoproteins (38, 40). Here, using this model and making inferences based on the information known about those receptor binding elements in MeV H, we conducted an alanine-scanning mutagenesis analysis of specific HeV G residues in locations we hypothesized might play a role in ephrinB2 receptor binding. Accordingly, we identified several amino acid residues that are critical for receptor binding, consisting of residues modeled to be in β1 and β4 in G. Specifically, two individual mutations of HeV G, K443A and K465A, led to defective binding of both human ephrinB2 and ephrinB3 as measured with a coprecipitation-based binding assay. These mutations also resulted in a significantly reduced ability to promote cell fusion when coexpressed with F. Five other mutations in HeV G also demonstrated partial defects in receptor binding and cell fusion promotion activities. Taken together, these data provide both biochemical and functional evidence of conformation-dependent, discontinuous receptor binding elements within the HeV G glycoprotein. These findings will aid our understanding of the binding and infection process, as well as help define important neutralization determinants of these important emerging pathogens.
(K. A. Bishop performed this work in partial fulfillment of the requirements of the Ph.D. program in Emerging Infectious Diseases of the Uniformed Services University [USU] of the Health Sciences.)
The following cell lines were obtained from the American Type Culture Collection (ATCC): HuTK−143B (ATCC CRL 8303), HeLa-CCL2 (ATCC CCL 2), and BSC-1 (ATCC CCL 26). HeLa-USU cells have been described previously (5). The 293T cells were obtained from G. Quinnan (USU). HeLa-USU, HeLa-CCL2, BSC-1, and 293T cells were maintained in Dulbecco's modified Eagle's medium (Quality Biologicals, Gaithersburg, MD) supplemented with 10% cosmic calf serum (HyClone, Logan, UT) and 2 mM l-glutamine. HuTK−143B cells were maintained in Eagle's medium (Quality Biologicals) supplemented with 10% cosmic calf serum (HyClone) and 2 mM l-glutamine. All cell cultures were maintained at 37°C in 5% CO2.
Conversion of specific residues of HeV G to alanine was performed via site-directed mutagenesis using the QuickChange II Site-directed Mutagenesis Kit (Stratagene, Cedar Creek, TX). The template for the reactions consisted of a C-terminal myc epitope-tagged version of HeV G in the vaccinia virus-based plasmid pMCO2 (10). All mutation-containing constructs were sequence verified.
A soluble, secreted, S-peptide-tagged version of human ephrinB2 was generated from a full-length cDNA (accession number NM_004093) (Origene, Rockville, MD) by PCR. The 5′ primer included an external SalI site, a Kozak sequence, and ephrinB2-specific sequence. The 3′ primer included ephrinB2 specific sequence, the codons for the S-peptide tag (KETAAAKFERQHMDS), a stop codon, and an external SalI site. The PCR generated a SalI-flanked product that encoded the first 233 amino acids of ephrinB2 fused in frame to the S-peptide tag. The PCR product was gel purified and cloned into TOPO (Invitrogen Corp., Carlsbad, CA) and subsequently subcloned into pMCO2. The final construct was sequenced, and protein expression was verified in cell lysates and supernatants by immunoprecipitation using the epitope tag and Western blotting with ephrinB2-specific polyclonal and monoclonal antibodies (MAbs). Recombinant vaccinia virus vKAB8 (human ephrinB2/s-tag) was generated and used to infect BSC-1 cells in roller bottles at a multiplicity of infection (MOI) of 5. Supernatant was collected from the roller bottles and clarified by centrifugation. EphrinB2 protein was purified via an S-bead column and eluted with 0.2 M sodium citrate, pH 2, with immediate neutralization by 1 M HEPES, pH 8. Binding of purified protein to henipavirus G and S agarose was verified, and the protein was aliquoted and stored at −80°C.
A second human ephrinB2/FC construct was generated by PCR using direct (TAAAGCTTCCGCCATGGCTGTGAGAAGGGA) and reverse (TAGGATCCACTTCGGAACCGAGGATGTTGTTCCC) oligonucleotide primers, which generated a 695-bp fragment of the ephrinB2 extracellular domain (B2EC), and subcloned into pEF6-TOPO-myc vector (Invitrogen). The resulting pEF6-TOPO-myc-B2EC plasmid was digested with HindIII-BamHI, and a 687-bp fragment was subcloned into pCXFc vector, which is a pcDNA3.1/Zeo(+) (Invitrogen) vector containing the sequence of human Fc between BamHI and XbaI sites. The human ephrinB2/FC protein was produced into HEK-293 cells as a secreted protein by transfection of the pCXFc-B2EC with Lipofectamine 2000 (Invitrogen). Human ephrinB2/FC was purified from culture media by affinity chromatography using protein A-agarose. After sample loading, the protein A matrix was washed with phosphate-buffered saline (PBS), and absorbed ephrinB2/FC was eluted by 0.1 M glycine-HCl, pH 2.3, with immediate neutralization by 1 M Tris-HCl, pH 8.5. The eluted protein was dialyzed against 50 mM Tris-HCl, 150 mM NaCl, pH 8, buffer; the binding reactivity of the ephrinB2/FC to EphB4 was verified; and the protein was aliquoted and kept at −80°C. Murine ephrinB2/FC, human ephrinB1/FC, and human ephrinB3/FC were obtained from R&D Systems, Minneapolis, MN.
Subconfluent HeLa-USU cells were transfected with the various alanine mutation-containing G proteins or wild-type (WT) G using the Fugene transfection reagent (Roche, Indianapolis, IN). Cells were transfected with 3 μg total DNA per T-25-cm2 flask overnight, followed by infection with WT vaccinia virus (strain WR) at an MOI of 10. At 6 h postinfection, the cells were washed and incubated overnight with methionine- and cysteine-free minimum essential media (Invitrogen) containing 2.5% dialyzed fetal calf serum (Invitrogen), l-glutamine, and 100 μCi per ml of [35S]methionine-cysteine (Promix; Amersham Pharmacia Biotech, Piscataway, NJ). Approximately 16 h later, the cells were chased with complete medium for 1.5 h, and cell lysates were prepared using lysis buffer (100 mM Tris-HCl [pH 8.0], 100 mM NaCl, and 1% Triton X-100), clarified by centrifugation, and analyzed by immunoprecipitation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).
For coprecipitations of G with receptor, G-containing cell lysates were incubated with 3 to 5 μg murine EFNB2/FC, human EFNB2/FC, human ephrinB1/FC, human ephrinB3/FC, or human EFNB2/s-tag, followed by precipitation with either protein G-Sepharose (Amersham) or S agarose (EMD Biosciences Inc., Madison, WI). For immunoprecipitations with G-specific antibodies, 4 or 5 μl of a polyclonal antiserum, 3 μl purified MAb, or 5 μl concentrated hybridoma supernatant were incubated with G-containing lysate at 4°C for 1 h or overnight. Samples were washed twice with lysis buffer, followed by one wash with 100 mM Tris-HCl (pH 8.0), 100 mM NaCl, 0.1% sodium deoxycholate, and 0.1% SDS (DOC wash buffer). Samples were boiled in sample buffer with 2-mercaptoethanol, analyzed by SDS-PAGE, visualized by autoradiography, and quantified by densitometry.
For coprecipitation assays of G with F glycoproteins, the F- and G-encoding plasmids were cotransfected into HeLa-USU cells. The cells were then infected, and 16 to 18 h later, cell lysates were prepared as described above. Equivalent amounts of each lysate were precleared with protein G-Sepharose for 45 min at room temperature and then incubated at 4°C overnight with 5 μl rabbit polyclonal F1 antiserum or 5 μl rabbit polyclonal G antiserum and then precipitated with protein G-Sepharose, washed, and boiled with 2-mercaptoethanol as described above before being analyzed by SDS-PAGE and Western blotting under reducing conditions with mouse polyclonal G-specific antiserum at 1:20,000. Immunoprecipitated proteins were quantified by spot densitometry using AlphaEase Fc Software (Alpha Innotech, San Leandro, CA).
Fusion between F and G glycoprotein-expressing effector cells and permissive target cells was measured by a reporter gene assay in which the cytoplasm of one cell population contained vaccinia virus-encoded T7 RNA polymerase and the cytoplasm of the other contained the Escherichia coli lacZ gene linked to the T7 promoter. β-Galactosidase (β-Gal) is synthesized only in fused cells (7, 8, 33). Plasmids encoding HeV F and each alanine mutant of G or no DNA (control/mock transfection) were transfected into HeLa-USU cells and allowed to express overnight as described above. 293T cells served as receptor-positive target cells. Vaccinia virus-encoded proteins were produced by infecting cells at an MOI of 10 and incubating the infected cells at 31°C overnight. Cell fusion reactions were conducted with the various cell mixtures in 96-well plates at 37°C. Typically, the ratio of envelope glycoprotein-expressing cells to target cells was 1:1 (2 × 105 total cells per well; 0.2-ml total volume). Cytosine arabinoside (40 μg/ml) was added to the fusion reaction mixture to reduce nonspecific β-Gal production (7). For quantitative analyses, Nonidet P-40 alternative was added (0.5% final concentration) at 2.5 or 3.0 h, and aliquots of the lysates were assayed for β-Gal at ambient temperature with the substrate chlorophenol red-d-galactopyranoside (Roche Diagnostics Corp.). Assays were performed in duplicate, and fusion results were calculated and expressed as rates of β-Gal activity (change in optical density at 570 nm per minute × 1,000) (9) in an MRX microplate reader (Dynatech Laboratories, Chantilly, VA). The individual cell fusion reactions mediated by each mutant were converted to percentages of WT fusion activity by dividing the substrate cleavage rate corresponding to wells containing each mutant by the rate for wells with WT G and multiplying by 100.
Various effector cell populations coexpressing HeV F, along with either mutant or WT HeV G, were prepared as described above for the reporter gene cell fusion assay. Prior to mixing the various effector populations with target cells, aliquots were made of each effector cell population (1 × 106 cells each). These aliquots were washed once with PBS to remove the Dulbecco's modified Eagle's medium and then incubated with fluorescein isothiocyanate (FITC)-conjugated polyclonal antisera (Cell Signaling Technology, Inc., Danvers, MA), specific for the myc epitope tag on each G glycoprotein, in PBS with 3% goat serum on ice for 1 h. Samples were washed three times before being fixed in paraformaldehyde (1.6% at 4°C) and then analyzed on a Beckman Coulter Epics XL flow cytometer.
The reporter gene assay was conducted as described earlier for each mutant or WT effector cell population, and then the raw β-Gal readings were normalized to compensate for the difference between each mutant and the WT G glycoprotein's surface expression as measured by the anti-myc staining described above. These numbers were then divided by that of the WT and multiplied by 100 to obtain the predicted percentage of WT fusion activity that each mutant would be expected to demonstrate when a mutant G glycoprotein was expressed on the surfaces of effector cells at equivalent levels to the WT. Most of the mutants (11 of 17) were found to be surface expressed at between 50% and 120% of WT G. Four mutants were found to be surface expressed at between 20 and 50% of WT G, and two mutants, K443A and E254A, were surface expressed at ~10% of WT. The cell surface expression levels of each G glycoprotein mutant construct were also analyzed by flow cytometry on other occasions, using several MAbs and a FITC-conjugated anti-myc MAb (Invitrogen), as well as by cell surface biotinylation and probing assays, and these additional analyses yielded results similar to those of the anti-myc staining and flow cytometry analyses. In addition, a subset of mutants and WT G were tested for surface expression in the presence and absence of HeV F coexpression, using the polyclonal and monoclonal anti-myc antibodies, in order to determine whether any variations in surface expression levels were evident. We found no differences in the cell surface expression levels of WT or any mutant G glycoprotein, whether or not they were coexpressed with the F glycoprotein.
In order to identify residues within the G glycoprotein that are important for engaging the ephrinB2 and ephrinB3 receptors, an alanine-scanning mutagenesis strategy targeting specific charged amino acids in the predicted globular-head domain of the molecule was conducted. We first chose to target a series of charged amino acid residues within G predicted to correspond to homologous regions within the MeV H attachment glycoprotein known to be involved in its protein-receptor interaction. The initial panel of G glycoprotein mutants consisted of 13 single-amino-acid point mutations and one double mutation covering a range of charged residues in what is predicted to be beta sheets 1, 4, and 6 and two loops, the β3/β4 loop and the β5/β6 loop (locations are based on a model proposed by Yu et al. (40).
Previously, we had demonstrated a detectable interaction between soluble ephrinB2 with either HeV or NiV G using a coprecipitation assay (5). Here, we sought to use this assay to examine our panel of HeV G mutants as a preliminary screening method for G and ephrinB2 or ephrinB3 binding. Since the extracellular domains of the murine and human ephrinB2 share 97% sequence identity (11), a commercially available recombinant murine ephrinB2/FC protein was used for the initial screening of the mutants. After a 1.5-h chase, metabolically labeled cell lysates of the various alanine-substituted mutants or WT HeV G were coprecipitated with either the murine ephrinB2/FC or human ephrinB3/FC protein, and the results are shown in Fig. Fig.1.1. Since neither HeV or NiV G can bind and coprecipitate ephrinB1, a commercially available human ephrinB1/FC protein was used as a negative control for the specificity of the G glycoprotein-receptor interaction. In addition, since the receptor interaction is specific for the G glycoprotein of HeV, the HeV F protein was also used to demonstrate the specificity of binding as a further negative control. In parallel, identical amounts of lysates were also precipitated directly with rabbit polyclonal antisera against HeV G to control for the expression levels of the various mutant glycoproteins. The precipitated proteins were washed and analyzed by SDS-PAGE and autoradiography. As demonstrated by the control immunoprecipitations, most of the 14 HeV G mutants were expressed to reasonably similar levels in comparison to WT G, with the exception of the E254A mutant, which was notably poorly expressed (Fig. (Fig.1,1, bottom row). Notably, two G glycoprotein mutants, K443A and K465A, demonstrated the most defective binding to both ephrinB2 and ephrinB3, with both mutants having approximately 75% reduction in binding compared to the WT. In addition, several other mutants, D257A, D260A, and D468A, appeared to possess moderately impaired abilities to bind ephrinB2 and ephrinB3 as well, although these reductions in binding were not as dramatic as those seen for K443A and K465A. As expected, none of the mutants or WT G bound and coprecipitated with ephrinB1, and the HeV F protein precipitated only with F-specific antisera.
Using a predicted structure of HeV G (40) as a guide, residues K443, K465, and D468 all likely reside in the fourth beta sheet of the globular-head region as illustrated in Fig. Fig.2.2. Because K443A and K465A demonstrated such substantially impaired receptor binding capacity, we constructed several additional mutations in G within this general region. The analysis of the additional G glycoprotein mutants revealed two further mutations (G439A and G449A) that possessed moderately impaired ephrinB2 and ephrinB3 binding phenotypes (Fig. (Fig.1)1) and hence further delineated an important domain within G for receptor interaction, although neither mutation completely abrogated receptor binding.
We also wanted to confirm that these HeV G mutations that impaired murine ephrinB2 binding also abrogated human ephrinB2 binding. In these studies, two human ephrinB2-derived constructs, an S-peptide-tagged version and an FC-tagged version, were employed. The mutants D257A, D260A, G439A, K443A, G449A, K465A, and D468A, as well as WT G, were expressed in HeLa-USU monolayers, and the resulting lysates were subjected to coprecipitation with the two tagged forms of human ephrinB2, followed by Western blotting with G-specific antisera. Using two alternative methods of coprecipitation, the patterns of binding of HeV and NiV G to either of these two human ephrinB2 constructs were remarkably similar to those obtained with the murine ephrinB2 protein (Fig. (Fig.33 and and1,1, respectively). For the sake of simplicity, only the data obtained with the S-peptide-tagged version are shown. Specifically, the K465A and K443A mutants of HeV G were the most significantly impaired in binding human ephrinB2, with reductions in binding of over 75%. In addition, the D257A, D260A, G439A, G449A, and D468A mutants also showed reduced binding activity, which correlated very well with the results obtained using murine ephrinB2.
Although it seemed unlikely that a single-amino-acid alteration could result in major conformational alterations in the G glycoprotein, the profound deficiencies in receptor binding demonstrated by several of the mutants, such as K443A and K465A, merited some additional examination of their overall structural and functional integrity. For these purposes, we made use of available conformation-dependent MAbs specific for the G glycoprotein and assessed their abilities to bind the G glycoprotein mutants, and the results are summarized in Table Table1.1. The mutants D260A, K443A, G449A, and K465A were all examined using six conformation-dependent MAbs to at least five distinct epitopes, including two Mabs known to block receptor binding and thought to bind overlapping epitopes on or near the receptor binding site of G (42) and another MAb that is thought to bind G at a region analogous to the SLAM-binding site of MeV H (38). D260A and G449A were recognized by all six MAbs, although in some cases to a lesser extent than the WT. K443A and K465A were recognized by three out of six MAbs. The three MAbs that are implicated in binding to G on or near the receptor binding site, 101, 102.4, and H2.1, did not bind K443A and K465A and exhibited reduced binding to D260A and G449A. The mutants D257A, G439A, and D468A were each recognized by five of five MAbs employed, although the G439A mutant exhibited less reactivity with 101 and 102.4 than WT G. These binding data provide several important types of information. First, all of the mutants identified with impaired receptor binding are recognized by at least three or more conformation-dependent MAbs, providing strong evidence that they are not globally disrupted in structure. Second, they further confirm the importance of several amino acid residues in receptor binding, because their alteration not only disrupts ephrinB2 and ephrinB3 interaction, but also results in loss of recognition by neutralizing MAbs known to bind to the G glycoprotein and to compete for receptor binding (42).
We next examined whether the decreased receptor binding capacities of the various HeV G mutants correlated with a decrease in the glycoproteins' fusion promotion activities when coexpressed with the F glycoprotein. Each of the HeV G mutant constructs was cotransfected with HeV F into an effector cell population and then tested for its ability to promote membrane fusion with a receptor-bearing partner cell population, and the results are shown in Fig. Fig.4A.4A. The HeV G mutants K443A and K465A demonstrated the most profound defect in cell fusion, with levels of 10.8 and 15.3% of the level of fusion measured with WT HeV G, respectively, when 293T cells served as receptor-positive partner cells. The overall deficiencies measured in cell fusion promotion activity by the various mutants were qualitatively consistent with the receptor binding data in that the mutants with defects in receptor binding (Fig. (Fig.1)1) were the same as those exhibiting defects in fusion promotion activity (Fig. (Fig.4A).4A). In addition, the mutants D257A, D260A, and D468A, which were all identified as having moderate reductions in receptor binding ability (Fig. (Fig.1),1), also showed moderate reductions in fusion promotion ability. Two G mutants that exhibited only moderate receptor binding defects, G439A and G449A, did not possess significant fusion-promoting defects. The remaining HeV G mutants showed variable effects in fusion-promoting activities and are summarized in Table Table2,2, along with their locations within the predicted G structure. It was also noted that some G mutants possessed a hyperfusion-promoting phenotype, such as D564A, which exhibited an average fusion promotion ability of 34% more than the WT (Fig. (Fig.4A4A).
As an additional means of assessing the conformational integrity and function of the mutant G glycoprotein panel, each was also assessed for its relative cell surface expression level in comparison to WT G. Each G glycoprotein mutant construct or WT G was expressed in at least two independent experiments, and cell surface expression was measured by flow cytometry using antibodies directed against a common myc epitope tag within each construct, as well as by cell surface biotinylation assays as detailed in Materials and Methods. All the G glycoprotein mutants were found to be expressed on the surfaces of cells at levels similar to or slightly less than that of WT G, with the exception of E254A and K443A, each of which was expressed at ~10% of WT G. Further, we also conducted these cell surface expression assays in both the presence and absence of the partner glycoprotein, F, and it was further determined that F coexpression did not affect or influence the relative cell surface expression levels of the various mutants in comparison to WT G.
Since the cell fusion assay is dependent not only on the intrinsic fusion promotion activities of the individual glycoproteins, but also on the relative expression levels of each glycoprotein on the surfaces of cells, and since in this assay each G glycoprotein construct or WT G was expressed by transient transfection (the efficiency of which can vary from mutant to mutant), we next sought to normalize the cell fusion data to account for any significant variations in cell surface expression. Therefore, each mutant G or WT G glycoprotein was coexpressed with F in an individual effector cell population, as described above, for use in the reporter gene cell fusion assay and also simultaneously assayed for G glycoprotein cell surface expression levels in parallel in the same population of cells.
Following preparation of the various effector cell populations but prior to mixing them with target cells, an aliquot of 1 × 106 cells was taken from each effector cell population, washed in PBS to remove residual cell culture medium, and then stained with FITC-conjugated anti-myc antibody for analysis by flow cytometry. The cell fusion assay was then carried out on the same populations of effector cells. The reporter gene cell fusion data were later normalized as a means to account for any significant differences in cell surface expression levels using the cell surface expression data derived from the flow cytometry analysis as detailed in Materials and Methods. The normalized cell fusion results are shown in Fig. Fig.4B.4B. The mutants D257A, D260A, K443A, G449A, K465A, and D468A, all of which were found to exhibit decreased binding to receptors (Fig. (Fig.11 and and3),3), all possessed a decreased ability to promote fusion with a permissive target cell population, following fusion data normalization to account for cell surface expression level variations (Fig. (Fig.4B).4B). The mutant D564A, which was noted to possess a hyperfusion-promoting activity in previous fusion assays (Fig. (Fig.4A),4A), retained this phenotype following normalization of the data; thus, its phenotype is likely not a result of G glycoprotein overexpression on the cell surface. The mechanism(s) which underlies these apparent hyperfusion promotion phenotypes is unclear. Interestingly, two additional mutants (D470A and G439A) also appear to have increased fusion-promoting phenotypes in comparison to the WT (Fig. (Fig.4B),4B), even though the G439A mutation confers a measurable decrease in the ability of G to bind receptor (Fig. (Fig.11 and and33).
There were also three mutants, E254A, K261A, and K560A, that exhibited slightly decreased fusion promotion ability after normalization (Fig. (Fig.4B)4B) without significant receptor binding defects (Fig. (Fig.11).
In addition to measuring the reactivities of the G glycoprotein mutants to conformation-dependent MAbs in immunoprecipitations and on the surfaces of cells, we also assessed whether they retained their abilities to associate with their partner F glycoprotein as a measure of their conformational integrity. The strength of the interaction between an F glycoprotein and its partner attachment glycoprotein can vary depending on the viral system being examined, but a biochemical interaction has been demonstrated with several paramyxoviruses (23, 26, 34, 39). The G glycoprotein mutants D257A, D260A, G439A, K443A, G449A, K465A, and D468A and WT HeV G were cotransfected with HeV F and analyzed by coprecipitation with a polyclonal antiserum raised against F, followed by Western blotting with a polyclonal antiserum raised against G (Fig. (Fig.5A,5A, top row). The lysates were also immunoprecipitated with G-specific antiserum, followed by blotting with a second G-specific antiserum in order to control for differing expression levels of the various mutants (Fig. (Fig.5A,5A, bottom row). Both HeV F (F control) and HeV G (G control) were expressed singly in the absence of the partner glycoprotein to illustrate the specificity of the coprecipitation reaction. Under these conditions, each of the mutants retained the ability to bind and coprecipitate with F. In fact, when quantified by densitometry (Fig. (Fig.5B),5B), mutants D257A, D260A, K443A, G449A, and D468A all demonstrated an approximately twofold-increased amount of G bound to F in comparison to the WT G.
Using a site-directed mutagenesis strategy, we have identified conformation-dependent receptor binding elements in the predicted globular-head region of the HeV G envelope glycoprotein. The regions identified here as important for receptor engagement consist of residues located in several strands of putative beta sheet 4, as well as two residues residing in beta sheet 1. Our data are complementary to the findings of Guillaume et al. (19), who identified putative receptor binding residues in NiV G residing in the globular-head region in areas similar to those found to be critical for MeV H receptor binding, although no biochemical or immunological characterizations were made with those mutant glycoproteins. In addition, our data are also consistent with earlier findings of Negrete et al. (32), where a deletion mutant of NiV G composed of residues 437 to 464 (Δ437-474 NiV G) had lost the ability to bind to the surfaces of receptor-positive 293T cells. According to a proposed structural model of HeV G (38, 40), residues 437 to 464 lie within β4, and we have identified five residues important for receptor binding located within this region, namely, G439, K443, G449, K465, and D468, which are crucial to HeV G's ability to bind ephrinB2 or ephrinB3. Thus, we speculate that the binding defect of the Δ437-474 NiV G construct is likely due to the absence of amino acid residues G439, G449, K443, K465, and D468, all of which were found to be important for receptor binding in the present study.
In addition, we found that differences observed in the MAb reactivities of the various G glycoprotein mutants also correlated with their receptor binding phenotypes. Of the six MAbs, three are predicted to bind G at or near the domain(s) of the glycoprotein responsible for receptor binding. Accordingly, it was these three MAbs that showed the most striking differences in the reactivities of the receptor-binding-defective mutants compared to WT G. Specifically, H2.1, 101, and 102.4 showed considerable decreases in their abilities to recognize the mutants G439A, K443A, and K465A, as well as smaller but still noticeable effects on their abilities to recognize D260A and G449A (Table (Table1).1). The correlation between loss of ephrinB2 and ephrinB3 binding ability and the loss of reactivity to MAbs 101 and 102.4, which were previously shown to block receptor binding (42), supports our conclusion that the domains of G identified in this study are indeed likely to be involved in receptor binding. In addition, the correlation between losses in receptor binding ability and the decrease in reactivity to H2.1 by these same mutants further supports our hypothesis that HeV G and MeV H may have similarities in the locations of their respective receptor binding sites, as this antibody is thought to bind G in regions analogous to the SLAM-binding domain of MeV H (38).
Interestingly, we also found that all of the mutants that were defective in receptor binding were still able to interact with and coprecipitate with their partner glycoprotein, HeV F (Fig. (Fig.5).5). Surprisingly, the majority of these mutants that demonstrated decreased receptor binding were found to exhibit an apparent increase in their abilities to coprecipitate with HeV F. This suggests that there is a subtle conformational difference between a mutant G that cannot bind receptor and the WT G. Historically, there have been two competing mechanistic models of paramyxovirus glycoprotein-mediated membrane fusion (reviewed in reference 27). One model suggests that F and G interact only after receptor binding takes place, and presumably, receptor binding triggers a conformational change in G that facilitates this F interaction. This interaction would be the fusion-promoting activity of G, and subsequently, the F glycoprotein becomes fusion activated, inserts its fusion peptide into target membranes, and facilitates the membrane fusion process. The second model suggests that interaction of the F and G envelope glycoproteins preexists and is independent of any receptor binding event and that it is receptor binding that triggers conformational change in G, which may or may not release F but nevertheless triggers the fusion activity of F. Our present data support this second model, in which F and G interact prior to receptor binding, not only in that WT G coprecipitates with F in the absence of receptor (HeLa-USU cells are receptor negative) (5), but also in that G glycoprotein mutants that possess significant defects in receptor binding can still coprecipitate with F to levels equivalent to or greater than that of WT G. Although we favor the interpretation that we have removed important residues in G for engaging receptor, an alternative explanation could be that some of these mutant G glycoproteins are adopting a prereceptor-bound conformation that is more favorable for F binding and less favorable for receptor binding.
The effects of several mutations in HeV G on ephrinB2 and ephrinB3 binding did translate into measurable effects on their abilities to promote cell fusion, as predicted. The overall trend observed from the present experiments was that reduced receptor binding capacities of individual G mutants correlated qualitatively with reductions in cell fusion measurements, although we cannot exclude the possibility that the reductions in cell fusion promotion activity observed for some G mutants (D257A, D260A, K443A, G449A, and D468A) could be caused, at least in part, by those mutants' increased abilities to bind HeV F, in addition to their decreased receptor binding phenotypes. Further experiments will be needed to dissect out the contribution of each of these effects to the fusion process. Taken together, the various mutations in HeV G identified here, which impaired its functional activities, appear to be specific for receptor binding and were not due to gross conformational defects, loss of an ability to interact with F, or a lower cell surface expression phenotype.
Notably, in addition to showing that, like NiV G, HeV G could engage ephrinB3 (Fig. (Fig.1)1) and use it as a functional receptor in cell fusion (data not shown), we also noted that the same residues within HeV G that appeared critical for binding ephrinB2 were also important for binding ephrinB3. Mutation of residues D257, D260, G439A, K443, G449, K465, and D468 of HeV G to alanine resulted in defects in ephrinB3 binding, similar to the binding pattern observed with ephrinB2. Together with data from Negrete et al. indicating that addition of soluble ephrinB2 can inhibit ephrinB3-dependent entry by NiV and vice versa (32), the present observations strongly suggest that each G glycoprotein binds to two different members of the ephrin ligand family via a common receptor binding domain. Further study of additional HeV, as well as NiV, G mutations will help to better delineate the domain(s) of G involved in receptor engagement, and it will also be important to confirm these observations in the context of virus infection in future experiments. In summary, our data provide strong biochemical and functional evidence of a conformation-dependent discontinuous ephrinB2 and ephrinB3 binding domain within the henipavirus G glycoprotein and will aid our understanding of the binding and infection processes of these important emerging pathogens.
We thank Michael Flora and all his staff at the Biomedical Instrumentation Center (BIC) of the USU for primer synthesis and sequencing, Karen Wolcott and Kateryna Lund at the BIC for their expertise in flow cytometry and analysis, and John White (CSIRO) and Dimiter Dimitrov (NCI), Frederick, MD, for MAbs.
This work was supported by NIH grant AI054715 to C.C.B.
The views expressed in this article are solely those of the authors, and they do not represent official views or opinions of the Department of Defense or The Uniformed Services University of the Health Sciences.
Published ahead of print on 21 March 2007.