trans-Complementation Allows Recovery of Human Respiratory Syncytial Viruses That Are Infectious but Deficient in Cell-to-Cell Transmission

Abstract

Human respiratory syncytial virus (HRSV) expresses three transmembrane glycoproteins: small hydrophobic protein SH, attachment protein G, and fusion protein F. The genes encoding SH and G can be deleted from the HRSV genome and infectious virus recovered. In contrast, HRSVs lacking the F gene or a functional replacement thereof have not been reported. To generate a system with which to study the roles of the viral transmembrane glycoproteins, including F, in the HRSV life cycle, we generated a cell line expressing a heterologous viral glycoprotein for complementation of glycoprotein function in trans. We previously demonstrated that the baculovirus GP64 protein or a chimeric form of GP64 carrying the 12 C-terminal amino acids of the HRSV F protein (GP^64/F) can efficiently mediate HRSV infectivity and improve its stability, when expressed from an engineered HRSV genome. Here, we report the development of a stably transfected Vero cell line (Vbac) constitutively expressing the GP^64/F protein. From the Vbac cell line, viruses that lacked the SH and F open reading frames (ORFs) or the SH, G, and F ORFs could be recovered from cDNAs. These viruses, designated RSΔsh,f and RSΔsh,g,f, respectively, had place-keeper ORFs inserted in place of the deleted ORFs to maintain authentic transcription levels. In the Vbac cell line, RSΔsh,f and RSΔsh,g,f could be amplified to near wild-type-level titers, and the resulting viruses were infectious to Vero and HEp-2 cells. After entry into Vero or HEp-2 cells, however, neither virus RSΔsh,f nor virus RSΔsh,g,f was able to spread in the infected cultures. Growth analyses showed that infectious virions were not produced in Vero or HEp-2 cells infected with RSΔsh,f and RSΔsh,g,f. Combined, these data provide direct evidence that HRSV F is an essential viral protein required for cell-to-cell transmission and demonstrate that complementation with the GP64 protein in trans constitutes a powerful tool for the study of the role of individual HRSV transmembrane glycoproteins in virus assembly, morphogenesis, and pathogenesis. In addition, the ability to generate infectious but nonspreading viruses may provide an alternative approach for the development of safe and stable HRSV vaccine candidates.

Human respiratory syncytial virus (HRSV) is a pneumovirus classified within the Paramyxoviridae family. HRSV is a major cause of severe lower respiratory tract disease in infants, children, immunosuppressed individuals, and the elderly, and a vaccine is not yet available (4, 7, 20). The 15,222-nucleotide, negative-sense, single-stranded RNA genome of HRSV contains 10 genes encoding 11 known proteins, 3 of which have been characterized as transmembrane glycoproteins: SH, a small hydrophobic protein of unknown function; G, a mucinous protein thought to provide an attachment function; and F, a membrane fusion protein that is responsible for the syncytia observed in infected cell cultures (11, 14, 28). The requirements of the SH and G proteins for HRSV infectivity have been examined in some detail. A cold-adapted HRSV was isolated in Vero cells and found to have a large deletion that included most of the SH and G genes (12). Furthermore, recovery of virus from an engineered cDNA lacking the SH gene showed that the absence of SH did not affect the ability of HRSV to replicate in cell culture and only minimally impacted the yield of infectious virus in vivo (3, 25, 29). In contrast, HRSVs lacking the G gene (which could be recovered in Vero cells) were significantly impaired in their ability to replicate in most cell types and in mice, cotton rats, and humans, indicating that the G protein plays an important role both in vitro and in vivo (25, 27). In a helper virus-dependent mini-replicon system, the F protein was found to be the minimum glycoprotein requirement for production of particles (26). This finding appears consistent with the fact that the F protein induces membrane fusion, a function believed to be essential for HRSV entry. F has also been shown to bind host cell surface molecules, such as glycosaminoglycans, and rhoA, and recently, the F2 subunits of HRSV and bovine respiratory syncytial virus (BRSV) were shown to be determinants of species-specific virus entry into cells (5, 19, 22). Given the above findings, it is not unexpected that viruses lacking the F gene or a functional equivalent have not been described. Glycoprotein substitution studies indicate that the G and F proteins can, in cell culture, be functionally replaced by a variety of heterologous transmembrane glycoproteins, including the BRSV G and F proteins, the bovine parainfluenza virus HN and F proteins, the vesicular stomatitis Indiana virus (VSIV) G protein, and the GP64 protein of baculovirus Autographa californica multiple nucleopolyhedrovirus (2, 17, 18, 24). However, due to the expression of a heterologous replacement protein gene from within the engineered HRSV genome, viruses with glycoprotein gene substitutions cannot be used to directly test the requirement for the HRSV proteins in question.

In order to generate a system with which to study the roles of the viral transmembrane glycoproteins in HRSV infectivity and pathogenesis and to directly test whether the F protein is essential for the propagation of HRSV, we generated a cell line expressing a heterologous viral glycoprotein for complementation in trans. GP64, the major envelope protein of baculovirus Autographa californica multiple nucleopolyhedrovirus, is an excellent candidate for the generation of such a cell line, because (i) it contains both attachment and membrane fusion functions within one protein; (ii) it significantly enhances the stability of HRSV infectivity when it is used as a replacement for the SH, G, and F proteins by expression from within the genome; (iii) it can be expressed efficiently from promoters recognized in the cell nucleus; and (iv) in contrast to the VSIV G protein-expressing cells, GP64-expressing cells show no signs of cytopathic effect (1, 9, 13, 18). In the present report, we describe the generation of a cell line constitutively expressing a chimeric form of the GP64 protein and the use of this cell line to recover from cDNA HRSVs lacking all or two of the three HRSV transmembrane glycoproteins. The resulting engineered viruses allowed us to examine the ability of the virus to spread in Vero and HEp-2 cell cultures in the absence of the F protein or a functional replacement. The results of this study indicate that the complementation approach constitutes a powerful system with which to study the role of the HRSV transmembrane glycoproteins in the biology of the virus and to evaluate the use of infectious, nonspreading HRSVs for the generation of safe vaccine candidates.

MATERIALS AND METHODS

Cells and Abs.

Vero 76 (Vero) and HEp-2 cells were acquired from the American Type Culture Collection and grown in standard growth medium containing 3% (for Vero cells) or 5% (for HEp-2 cells) fetal bovine serum. Vero cells were used for the generation of a cell line (Vbac) expressing chimeric glycoprotein GP^64/F (see below). Monoclonal antibodies (MAbs) 6 and 19 were provided by Geraldine Taylor (Institute for Animal Health, Compton, United Kingdom), MAb L9 was provided by Ed Walsh (University of Rochester School of Medicine, Rochester, N.Y.), and MAbs AcV1 and -5 were provided by Gary Blissard (Boyce Thompson Institute, Cornell University, Ithaca, N.Y.). The anti-VSIV G Ab was a mouse ascitic fluid acquired from the American Type Culture Collection. The anti-vaccinia virus Ab was a polyclonal serum harvested from rabbits infected with vaccinia virus.

Production and characterization of a cell line expressing a chimeric GP64 protein (Vbac).

An open reading frame (ORF) expressing chimeric protein GP^64/F was constructed previously (18) and cloned into the expression vector pRc/CMV (Invitrogen). The resulting plasmid (pRc-GP^64/F), expressing both the GP^64/F and a neomycin resistance gene, was transfected into Vero cells at passage 5 by using Lipofectamine 2000 (Invitrogen). Cells were incubated at 33°C. The following day, transfected cells were detached and plated in six-well plates at various densities. At 48 h posttransfection, 1 mg of G418 sulfate (Geneticin; Invitrogen)/ml was added to the medium, and cells were further incubated at 33°C. G418-resistant colonies were selected between 18 and 34 days later, expanded, and tested for GP^64/F expression by using anti-GP64 MAb AcV5 and indirect immunofluorescence (IF). A number of GP^64/F-positive colonies were subjected to two rounds of subcloning by plating detached cell suspensions at 1 cell/well in 96-well plates, with continued incubation at 33°C in medium containing G418 sulfate. Following the two rounds of subcloning, a final GP^64/F-positive cell line (116-25-1) was selected and designated Vbac, passage 1. Throughout this work, Vbac cells were maintained in standard growth medium containing 5% fetal bovine serum, 1 mg of G418 sulfate/ml, and 25 mM HEPES at 33°C. For characterization of the Vbac cell line, cells were attached to glass coverslips, chilled, and incubated with MAb AcV1 diluted 1:20 in phosphate-buffered saline (PBS)-0.1% bovine serum albumin for 1 h on ice. Unbound Ab was removed by four washes with cold PBS, and cells were fixed with cold 4% paraformaldehyde (PF) for 10 min and then transferred to room temperature; and incubated another 15 min in 4% PF. Cells were washed; blocked with a solution containing PBS, 2% bovine serum albumin, and 0.1% cold-water fish skin gelatin for 1 h at room temperature; and incubated with a goat anti-mouse Ab carrying an Alexa-594 fluorescent conjugate (Molecular Probes) for 50 min at room temperature. Cells were washed with PBS, stained for 10 min with Hoechst reagent (Molecular Probes) to visualize nuclei, mounted, and photographed on an Olympus IX70 inverted microscope. For flow cytometry, Vbac cells were detached in 5 mM EDTA, chilled, and incubated with Abs as described above (without treatment with Hoechst reagent). After the last wash, cells were resuspended in 500 μl of PBS and analyzed in a FACSCalibur flow cytometer (Becton Dickinson), using 50,000 events per sample.

Construction of cDNAs and recovery of infectious HRSVs by using the Vbac cell line.

Engineered cDNAs were generated based on a cDNA of the A2 strain of HRSV, which was constructed as described previously (17, 18). Briefly, by using standard cloning techniques, two shuttle vectors containing three ORFs were constructed. For the generation of virus RSΔsh,g,f, the shuttle vector contained the ORFs encoding marker proteins enhanced with green fluorescent protein (GFP) (Clontech), chloramphenicol acetyltransferase (CAT), and β-glucuronidase (GUS). For the generation of virus RSΔsh,f, the shuttle vector contained the ORFs encoding GFP, HRSV G, and GUS. The translation context of the HRSV G gene in this shuttle vector was altered to enhance the ratio of the amount of G protein anchored to the membrane to the amount of G protein secreted (10, 21); by PCR and conventional cloning techniques, the three nucleotides preceding as well as the first nucleotide following the G start codon were modified to read ACC-ATG-G. This modification resulted in the change of the second amino acid of the G protein from serine to alanine. Surface expression of the altered G protein and its ability to incorporate into virus-induced filaments were confirmed by confocal microscopy (data not shown).

The ORFs contained in each of the shuttle vectors were separated by authentic HRSV intergenic junctions and flanked by the unique restriction sites FseI and AscI. These shuttle vectors were cloned into a cDNA backbone that lacked the SH, G, and F ORFs and that also contained FseI and AscI restriction sites, yielding cDNAs that varied from wild-type (wt) HRSV only in the content of the ORFs at gene positions 6, 7, and 8 (pRSΔsh,g,f and pRSΔsh,f, respectively). Modified areas of the engineered cDNAs were verified by nucleotide sequencing. Infectious viruses were recovered from cDNA as described previously (17), with the following exceptions: (i) in the initial transfection, we included 0.1 μg of a plasmid encoding G^vsv, a chimeric VSIV G protein containing the HRSV F protein cytoplasmic tail domain (17), and (ii) instead of Vero cells, generated viruses were amplified in Vbac cells. At passage 2, recovered virus suspensions were incubated with anti-vaccinia virus Abs (1:50 dilution) and anti-VSIV G Abs (ascitic fluid at a 1:100 dilution) for 1 h at room temperature to neutralize any modified vaccinia Ankara virus that might have carried over independently or via the VSIV G protein transiently expressed in the HEp-2 cells initially transfected. These preincubated virus suspensions were used for the production of passage 3 stocks, which were subsequently titrated by immunoplaque assay as previously described (17), with the following exceptions: (i) Vbac cells were used instead of Vero cells and (ii) to visualize plaques, anti-HRSV nucleocapsid Ab (MAb 6) was used for virus RSΔsh,g,f and anti-HRSV G Ab (L9) was used for virus RSΔsh,f. In addition, for both viruses, plaques were scored based on GFP expression prior to fixation; GFP-expressing plaques overlapped precisely with those identified by immunostaining, as was observed previously (17). Viral RNAs were harvested from cells infected with the engineered virus stocks at passage 3, amplified by reverse transcription-PCR, and verified by nucleotide sequence analysis across cloning junctions and in modified areas. Passage 3 stocks were used for the experiments described below.

Glycoprotein expression.

Protein expression by the engineered viruses was examined by cell enzyme-linked immunosorbent assay (CELISA) as previously described (15, 18). Briefly, Vero or Vbac cells infected with the engineered viruses at a multiplicity of 2 were fixed with 4% PF for 15 min at 28 h postinfection and permeabilized by incubation with 0.1% sodium dodecyl sulfate for 5 min at room temperature. Cells were then incubated with Abs against GP64 (AcV5), HRSV G (L9), HRSV F (MAb 19), or HRSV N protein (MAb 6) as a control; washed; and incubated with a horseradish peroxidase-conjugated secondary goat anti-mouse Ab (Pierce). Next, cells were washed and incubated in 1 ml of 3,3′,5,5′-tetramethylbenzidine substrate (Pierce). At various times after the addition of the substrate, 100-μl aliquots were collected and added to 2 M sulfuric acid in a 96-well plate to stop the reaction, and the optical density at 450 nm was determined in an ELISA plate reader.

Virus transmission in cell culture.

Vero or Vbac cells were plated in six-well plates at approximately 60% confluence, infected at a multiplicity of 0.05 for 1.5 h, and then washed and incubated at 33°C. On days 1, 2, 4, 6, and 8 postinfection (data for day 8 are not shown), one well for each condition was fixed in 4% PF for 15 min and examined for GFP expression on an Olympus IX70 inverted fluorescence microscope. As determined previously, GFP expression provided an accurate indicator of virus infectivity that correlated with PFU (17). Fixed cell cultures were photographed at ×200 magnification.

Growth curves.

Vero, Vbac, and HEp-2 cells were infected in six-well plates at a multiplicity of 0.25 for 1.5 h at 33°C, washed twice, and then incubated in 1.5 ml of normal growth medium at 33°C. At days 0, 2, and 4 postinfection, the supernatants of infected cells were drawn off, and an equal volume of fresh growth medium was added to the cells. Cells were then scraped into the medium, followed by gentle pipetting. The titers of both the supernatant and cell-derived virus suspensions in Vbac or Vero cells were determined by a 50% tissue culture infective dose (TCID₅₀) assay based on GFP expression from the viral genome, as previously described (17, 18). To determine the total virus titer, supernatant and cell-derived titers were combined. To compare the infectivities of the above-described harvested virus samples to Vero and Vbac cells, frozen virus suspensions from days 0, 2, and 4 postinfection were titrated by the TCID₅₀ assay as described above or were examined by flow cytometry for their ability to enter cells. For the latter procedure, samples were diluted 10 times in normal growth medium and adsorbed to Vero or Vbac cells by using conditions identical to those used for the growth curves (1.5 h at 33°C). At 20 h postinfection, cells were trypsinized, fixed in 4% PF, and analyzed in a FACSCalibur flow cytometer (Becton Dickinson), using 50,000 events per sample.

RESULTS

Generation of a cell line expressing the baculovirus GP64 protein.

In order to recover an HRSV in which the SH, G, and F glycoprotein genes were deleted and not functionally replaced, a cell line which could provide a heterologous replacement protein for complementation in trans was generated. As a replacement protein, we used a chimeric form of the baculovirus envelope protein GP64, consisting of the ecto- and transmembrane domains of GP64 coupled to the 12 C-terminal amino acids of the HRSV F protein (GP^64/F) (Fig. 1A). GP^64/F was previously shown to be expressed and functional in Vero and HEp-2 cells, and cells expressing GP^64/F showed no signs of cytopathic effect (18). Moreover, the infectivity of an SH/G/F-depleted HRSV containing GP^64/F expressed from the viral genome was substantially more stable than a wt HRSV.

FIG. 1.

Characterization of Vbac, a cell line constitutively expressing the GP^64/F protein. (A) The GP^64/F protein consists of the ecto- and transmembrane domains of the baculovirus GP64 protein coupled to the 12 C-terminal amino acids of the HRSV F protein. TM, transmembrane domain. (B) Analysis of GP^64/F surface expression by indirect IF. Vbac cells (passage 8) or Vero cells were incubated on ice with anti-GP64 MAb AcV1, washed, and fixed with 4% PF. Cells were then incubated with an Alexa-594-conjugated secondary Ab (red), stained with Hoechst reagent to visualize nuclei (blue), and examined with fluorescence microscopy (magnification, ×400). (C) Flow cytometry analysis. Vbac cells (passage 8) and Vero cells were detached from plates with EDTA, fixed, and incubated with anti-GP64 Abs as described above. Next, cells were incubated with an Alexa-488-conjugated secondary Ab and examined by flow cytometry.

A plasmid constructed to express both GP^64/F and a neomycin resistance gene was used to transfect early-passage Vero cells. After incubation at 33°C in G418-containing medium, resistant colonies were selected, expanded, and tested for expression of GP^64/F by IF with anti-GP64 MAb AcV5. Cells from colonies expressing the highest levels of GP^64/F were then subjected to two rounds of subcloning via limiting dilution and screening by indirect IF. Finally, one cell line (clone 116-25-1) constitutively expressing GP^64/F was selected and designated Vbac, passage 1; these cells were slightly larger and had a slightly lower cell division rate than the Vero cells from which they were derived. Figure 1B and C show the expression of GP^64/F at the surfaces of Vbac cells at passage 8 compared to that of the parental Vero cell line. GP^64/F expression was detected by incubation of cells at 4°C with anti-GP64 MAb AcV1, followed by fixation with 4% PF. After fixation, cells were incubated with an Alexa-594-conjugated secondary Ab, stained with Hoechst reagent (to visualize nuclei), and photographed on a fluorescence microscope (Fig. 1B). Vbac or Vero cells were also detached from plates with 5 mM EDTA, fixed and labeled as described above, and examined for surface GP^64/F expression by flow cytometry (Fig. 1C). The data of Fig. 1B and C show that most cells within the Vbac population expressed the GP^64/F protein at the cell surface. The expressed protein appeared evenly distributed across the plasma membrane, but expression levels among cells varied (Fig. 1C). The long-term stability of GP^64/F expression in the absence of G418 was not investigated. However, when Vbac cells were maintained in medium containing G418, the level and distribution of GP^64/F expression remained constant for at least 10 passages.

Construction of cDNAs and recovery of engineered viruses.

To generate HRSVs with a modified glycoprotein gene content, two cDNAs were constructed from the A2 strain of HRSV by using methods previously described (17). In one cDNA, the SH, G, and F ORFs were deleted and replaced with the ORFs of three marker proteins, GFP, CAT, and GUS, respectively (pRSΔsh,g,f) (Fig. 2A). This was done to keep the number of genes as in wt HRSV in order to maintain authentic transcription levels. This HRSV cDNA thus encoded no viral envelope proteins. In the second cDNA, the SH and F ORFs were replaced with GFP and GUS, respectively, but the HRSV G protein ORF remained in its native position (pRSΔsh,f). Though altered in the coding potential of genome positions 6, 7, and 8, both cDNAs generated retained fully wt intergenic junctions throughout. Infectious virus was recovered from cDNAs as described previously (Fig. 2B) (17). Briefly, the engineered cDNAs were transfected into HEp-2 cells already infected with modified vaccinia virus Ankara-T7 (30) along with plasmids encoding the nucleocapsid protein (N), phosphoprotein (P), polymerase (L), and transcription factor M2-1 (8). To ensure that the resulting viruses could be passaged from the infected/transfected HEp-2 cells, we included an additional plasmid encoding G^vsv, a chimeric VSIV G protein from which the cytoplasmic tail domain was exchanged with that of the HRSV F protein (17). Supernatants from infected/ transfected HEp-2 cells were transferred onto Vbac cells, and GFP-expressing spreading foci were observed upon incubation at 33°C. The viruses harvested from the first round of Vbac cells, designated RSΔsh,g,f and RSΔsh,f (Fig. 2), were further amplified by one additional passage in Vbac cells, as described in Materials and Methods. Final virus stocks (passage 3) were harvested and examined by reverse transcription-PCR across the modified areas, followed by sequence analysis (data not shown). In Vbac cells, both viruses yielded stock titers of approximately 2 × 10⁶ PFU/ml, approaching titers generally reported for wt HRSV.

FIG. 2.

Composition of engineered HRSVs and recovery protocol. (A) From a cDNA of the wt HRSV A2 strain, two cDNAs were constructed with an altered glycoprotein gene content. In construct RSΔsh,g,f, all three HRSV transmembrane glycoprotein ORFs (SH, G, and F) were deleted and replaced with those of GFP, CAT, and GUS, respectively. Construct RSΔsh,f is similar to RSΔsh,g,f; however, the HRSV G protein ORF remained in its native position. (B) Infectious HRSVs were recovered from these cDNAs as previously described (17) with the following modifications: a plasmid encoding a chimeric VSIV G protein, G^vsv (17), was included in the transfection of HEp-2 cells, and amplification of recovered viruses was done in Vbac instead of Vero cells. Passage 2 viruses were incubated with anti-vaccinia (α-Vacc) and anti-VSIV G (α-VSIV G) virus Abs as described in Materials and Methods and then used for the generation of passage 3 virus stocks. Passage 3 (pass 3) stocks were titrated in Vbac cells, yielding titers of 2 × 10⁶ PFU/ml.

Protein expression by the engineered HRSVs.

To examine the glycoprotein expression by each of the engineered viruses, infected Vero cells were incubated at 33°C and then fixed and permeabilized at 28 h postinfection. Fixed cells were incubated with Abs against GP64 (AcV5), HRSV G (L9), HRSV F (MAb 19), or HRSV N protein (MAb 6) as a control, and protein expression was measured by CELISA (Fig. 3) as previously described (17). A previously described virus, RSΔsh,g,f/GP^64/F, was included as a control (18). RSΔsh,g,f/GP^64/F lacks the SH, G, and F ORFs, but in contrast to virus RSΔsh,g,f generated in this study, it contains the GP^64/F ORF within its genome at position 7 (the ORF order at positions 6, 7, and 8 is GFP-GP^64/F-GUS). As a positive control for F expression, a previously engineered virus that expresses the HRSV F protein was included (data not shown). Both of the newly constructed viruses, RSΔsh,g,f and RSΔsh,f, were able to infect Vero cells, as evidenced by the expression of GFP from the viral genomes (data not shown) and by the expression of the HRSV N protein (Fig. 3). In addition to the N protein, virus RSΔsh,f expressed the HRSV G protein but not F or GP64; virus RSΔsh,g,f did not express any transmembrane glycoproteins, as expected. In Vbac cells, both uninfected and infected, the GP^64/F protein was detected at a level similar to that seen in Vero cells infected with control virus RSΔsh,g,f/GP^64/F. These results show that the glycoprotein expression by the engineered viruses was in agreement with the genome content shown in Fig. 2A and that high levels of GP^64/F were present in the Vbac cell line even after infection with the engineered HRSVs.

FIG. 3.

Transmembrane glycoprotein expression by HRSVs lacking the SH, G, and F or SH and F ORFs. Protein expression in Vero and Vbac cells infected with the engineered viruses was examined at 28 h postinfection with CELISA as previously described (18). As a comparison, Vero cells infected with virus RSΔsh,g,f/GP^64/F were included. RSΔsh,g,f/GP^64/F contains the GP^64/F gene within its genome as described previously (18). Bars represent the average values from duplicate samples. α-N, anti-HRSV N protein Ab; α-G, anti-HRSV G protein Ab; α-F, anti-HRSV F protein Ab; α-GP64, anti-GP64 protein Ab; OD450, optical density at 450 nm.

Virus transmission in cell culture.

Having a system in which infectivities of glycoprotein depleted HRSVs are solely dependent on the GP^64/F protein provided by the complementing cell line allowed us for the first time to monitor HRSV transmission in cell culture in the absence of the SH, G, and F proteins (and without the presence of heterologous viral entry and exit proteins). Vero or Vbac cells were plated at ∼60% confluence and infected with viruses RSΔsh,g,f and RSΔsh,f at a multiplicity of 0.05. Cells were incubated at 33°C and examined for GFP expression at daily intervals (Fig. 4). GFP expression from the same position within the genome of an infectious engineered HRSV was previously shown to be an accurate indicator of infectivity that correlated with PFU (17, 18). To examine the spread to neighboring cells, the infected cultures were photographed using fluorescence microscopy on days 1, 2, 4, and 6 postinfection (Fig. 4). As a control, Vero cells infected with a virus containing the GP^64/F ORF within its genome (RSΔsh,g,f/GP^64/F) were also examined. Viruses RSΔsh,g,f and RSΔsh,f spread rapidly in Vbac cells; by day 6, all cells in the cultures displayed green fluorescence. The rate of spread was similar to that of control virus RSΔsh,g,f/GP^64/F in Vero cells (Fig. 4, bottom panels). In contrast, no significant spread to neighboring cells was detected in Vero cell cultures infected with virus RSΔsh,g,f. Although a moderate increase was observed in the number of green cells between days 1 and 2, no increase in the number of cells displaying green fluorescence was observed after day 2. The intensity of fluorescence in infected Vero cells increased over time, resulting by day 4 in single cells that fluoresced very brightly compared to infected Vbac cells. Similar to the results for virus RSΔsh,g,f, a virus expressing the HRSV G protein as its only transmembrane glycoprotein (RSΔsh,f) was also unable to spread to neighboring cells in Vero cultures. Thus, the G protein alone was not sufficient to allow cell-to-cell transmission in Vero cultures, nor did it appear to alter the rate of virus transmission in Vbac cells compared to that of virus RSΔsh,g,f.

FIG. 4.

Virus transmission in cell culture. Vero and Vbac cells were plated at ∼60% confluence, infected with viruses RSΔsh,g,f and RSΔsh,f at a multiplicity of 0.05, and incubated at 33°C. At days 1, 2, 4, and 6 postinfection, cells were fixed with 4% PF and examined for GFP expression with fluorescence microscopy (magnification, ×200). As a comparison, Vero cells infected with the previously described virus RSΔsh,g,f/GP^64/F (18) were included.

These results demonstrate that, in contrast to the SH and G proteins, the F protein is critical for HRSV transmission in Vero cell cultures and that expression of the G protein alone does not rescue the transmission deficiency. The reason for the small increase in the number of fluorescing Vero cells between days 1 and 2 has not been investigated, but it may be caused by the fact that GFP expression at day 1 was not sufficiently strong to be detected in all infected cells. Alternatively, a limited amount of virus spread may have occurred through cell division, as cells were plated at 60% confluence and HRSV does not shut down host cell metabolism.

Production of infectious progeny virus.

The above data, showing both the expression of GFP and an inability to spread in Vero cells, suggested that viruses RSΔsh,g,f and RSΔsh,f entered Vero cells and replicated their genomes but did not produce infectious virions. To examine infectious virion production, the yields of infectious virus produced in Vbac, Vero, and HEp-2 cells were compared (Fig. 5). Cells were infected with viruses RSΔsh,g,f, RSΔsh,f, and RSΔsh,g,f/GP^64/F at a multiplicity of 0.25, washed two times, and then incubated at 33°C. Immediately after infection (day 0), or at 2 and 4 days postinfection, supernatants and cells were harvested separately and titrated on Vbac cells by the TCID₅₀ assay based on the expression of GFP from the viral genome as described previously (17). Supernatant and cell-associated viral titers were combined to determine the total virus yield. In Vbac cells, viruses RSΔsh,g,f and RSΔsh,f replicated to titers similar to that of the control virus RSΔsh,g,f/GP^64/F (Fig. 5A), which contains the GP^64/F gene within its genome and has previously been shown to yield higher titers than a virus containing the wt G and F proteins in Vero cell cultures (18). A comparison of supernatant and cell-associated titers revealed that, for each virus, the majority of infectivity was associated with the cell fraction, similar to what occurs with wt HRSV (data not shown). In contrast to Vbac cells, when viruses RSΔsh,f and RSΔsh,g,f were used to infect Vero and HEp-2 cells, the yield of infectious virus was approximately 2 orders of magnitude lower than in the Vbac cells. This reduction in infectious virus yield was consistent with the results of Fig. 4, confirming the importance of the F protein for virus infectivity. However, even in the absence of viral transmembrane glycoproteins, there was an approximate 10-fold increase in titers of viruses RSΔsh,g,f and RSΔsh,f grown in Vero and HEp-2 cells when titrated in Vbac cells. This was unexpected given the transmission deficiency shown in Fig. 4. It was unlikely that GP^64/F from input virus was responsible for the observed increase, as no GP^64/F could be detected by CELISA in Vero cells infected with viruses RSΔsh,g,f and RSΔsh,f at 28 h postinfection (Fig. 3); moreover, analysis by fluorescence microscopy of infected Vero cells fixed at 26 h postinfection with anti-GP64 Abs confirmed the absence of GP64, as shown by CELISA (data not shown).

FIG. 5.

Growth curve analysis. Vbac, Vero, and HEp-2 cells were infected with virus RSΔsh,g,f, RSΔsh,f, or RSΔsh,g,f/GP^64/F at a multiplicity of 0.25 and incubated at 33°C. Immediately after viral adsorption (day 0) and at days 2 and 4 postinfection (pi), supernatants and cells were harvested separately and titrated on Vbac cells by a TCID₅₀ assay based on the expression of GFP, as described previously (17). Supernatant and cell-associated viral titers were combined to determine the total virus yield.

Because different cell types were used for the virus titrations shown in Fig. 4 and 5 (Vero and Vbac, respectively), we investigated whether the low level of infectivity derived from Vero and HEp-2 cells infected with viruses RSΔsh,g,f and RSΔsh,f, seen in Fig. 5, might be an effect specific to assay on Vbac cells. Selected samples used to determine the titrations shown in Fig. 5 were reexamined. If the engineered viruses were able to produce infectious virions in Vero and HEp-2 cells, their infectivity should be measurable in either GP^64/F- or non-GP^64/F-expressing cells. Therefore, the infectivities of the stored growth curve samples of virus RSΔsh,g,f and RSΔsh,g,f/GP^64/F from HEp-2 and Vero cells (Fig. 5) were compared by titration on Vbac cells or Vero cells, using a TCID₅₀ assay (Fig. 6A). In addition to the TCID₅₀ assay, parallel samples of infected cells were examined with flow cytometry for the expression of GFP as a measure of viral entry (Fig. 6B), as described previously (17). The results shown in Fig. 6A confirmed that, when virus RSΔsh,g,f grown in Vero and in HEp-2 cells was titrated on Vbac cells, modest increases in infectivity were observed. In contrast, when it was titrated on Vero cells, RSΔsh,g,f infectivities of both Vero and HEp-2 cells declined immediately after infection, while those of control virus RSΔsh,g,f/GP^64/F did not. Flow cytometry analysis yielded similar results (Fig. 6B). In this assay, virus samples parallel to those used to obtain the results shown in Fig. 6A were diluted 10-fold and used to infect Vero or Vbac cells, and the number of GFP-expressing cells was counted. The percentage of RSΔsh,g,f virions that entered Vbac cells was found to increase from 0.01 to 1.54% (grown in Vero cells) or from 0.01 to 0.43% (grown in HEp-2 cells) from day 0 to 4 postinfection, while no entry into Vero cells was observed (Fig. 6B). Combined, these data show that virus RSΔsh,g,f grown in Vero or HEp-2 cells did not produce infectious virions. The data also suggest that Vbac cells were able to bind and internalize virions produced in the absence of viral transmembrane glycoproteins, presumably mediated by the GP^64/F protein expressed at the Vbac cell surface.

FIG. 6.

Production of infectious virions in Vero and HEp-2 cells and comparison of titers in Vero and Vbac cells. (A) Samples of viruses RSΔsh,g,f and RSΔsh,g,f/GP^64/F from Vero and HEp-2 cells from the experiments whose results are shown in Fig. 5A were titrated on Vbac or Vero cells by a TCID₅₀ assay as described previously (17). Values represent the averages from duplicate samples. pi, postinfection. (B) Tenfold dilutions of the samples used for panel A were adsorbed to Vbac or Vero cells, and GFP-expressing cells were counted with flow cytometry at 20 h postinfection as a measure of viral entry, as previously described (17).

DISCUSSION

Recovery of HRSV lacking the SH, G, and F protein ORFs and containing no functional replacements.

While many aspects of HRSV biology have been well studied, neither the assembly and morphology of the infectious unit nor the mechanism by which HRSV spreads from cell to cell is well understood. Early work with the Long strain of HRSV showed significant virus transmission in cell cultures in the presence of neutralizing quantities of pooled sera from convalescent-phase patients, suggestive of a mode of cell-to-cell transmission independent of the transmembrane glycoproteins (23). Subsequent work with a helper virus-dependent mini-replicon system, as well as recovery of viruses lacking the SH and G genes, showed that F protein alone was sufficient to allow virus propagation in Vero cells, although virus lacking the G gene was severely restricted in other cell types and animal models (12, 25-27). Such studies pointed to an important role for the F protein in HRSV propagation. We developed a cell line that expresses a heterologous viral entry and exit protein in order to complement recovery of an HRSV lacking the SH, G, and F protein ORFs and containing no functional replacement ORFs, so that propagation of HRSV in the absence of both homologous and heterologous viral transmembrane glycoproteins could be studied. In previous work, Vero cells proved exceptional compared to other cultured cell lines in that they allowed relatively efficient amplification of an HRSV lacking the G gene, while such a virus was severely restricted in other cell types and in vivo (25, 27). In contrast to the efficient propagation of virus lacking the G gene in Vero cells, the results of this study show that the spread of HRSVs lacking all transmembrane glycoprotein ORFs is impaired in Vero as well as in HEp-2 cells. This demonstrates that the F protein is a critical factor in the ability of HRSV to spread within cell cultures. A virus containing the HRSV G gene as its only transmembrane glycoprotein behaved identically to one lacking all three transmembrane glycoproteins, indicating that the G protein alone was not sufficient to allow virus transmission.

GP^64/F-complementing cell line as a tool for studies of the HRSV transmembrane glycoproteins.

The importance of the F protein is most likely related to its membrane fusion function (28), which allows the virus to cross the plasma membrane to achieve virus entry. In addition, the F protein may provide other functions that are critical for virus propagation, such as functions in assembly, budding, and morphology. A supportive role for the F protein in cell attachment has already been demonstrated (5), and recently a “chaperoning” of F function by the G protein has been suggested. Having demonstrated efficient recovery of HRSVs lacking a functional membrane fusion protein, the Vbac line constitutes a powerful system to study the roles and relative contributions of the transmembrane glycoproteins in the above-mentioned aspects of HRSV biology. Independence from functional SH, G, and F proteins allows for the recovery of HRSVs with any desired glycoprotein content, so that both the roles of individual proteins and the importance of interactions between them can be studied. Importantly, this can be done at wt protein levels due to the replacement of deleted genes with place keeper genes, maintaining authentic transcription and replication levels (17). Moreover, viruses functionally complemented by the GP^64/F protein have the potential to be significantly more stable, based on previous work showing enhanced stability at 4°C (18), further facilitating HRSV studies.

Trans-complementation as an alternate approach for vaccine development.

To date, it has proven challenging to find an acceptable balance between immunogenicity and the safety of live HRSV vaccine candidates (6, 16). The ability to generate infectious virus by complementation with a heterologous viral protein from a cell line may provide new opportunities to address vaccine safety. With the Vbac cell system, viruses that are impaired in cell-to-cell transmission yet should not be attenuated for replication and thus expression of viral antigens can be generated. Also, trans-complementation removes the dependence on functional G and F protein for virus propagation and stock production, allowing one to include selective G and F protein domains or epitopes (or nonfunctional versions of the F protein) into an engineered genome. Alternatively, when included in the engineered HRSV genome, the GP^64/F protein conferred onto the virus a temperature-sensitive phenotype, with reduced propagation at 37°C (18). These possibilities, together with enhanced HRSV stability via the use of the GP^64/F protein as shown previously (18), may provide new advantages in our ability to find a balance between safety and efficacy and to improve the stability of HRSV, two key problematic areas in the design of live attenuated vaccines.

Conclusions.

A cell line complementation system based on a heterologous viral protein, baculovirus GP64, allowed the recovery of HRSVs that lack any transmembrane glycoprotein gene. Cell culture experiments with the resulting engineered viruses showed that these viruses are competent to infect cultured Vero and HEp-2 cells, express their genes, and replicate but that they cannot spread. Further, these studies provided direct evidence that the F protein is critical for cell-to-cell transmission of HRSV. The efficiency with which glycoprotein gene-deficient HRSVs were recovered demonstrates that the approach of providing viral proteins in trans constitutes a powerful system to study the role of the individual HRSV SH, G, and F proteins, as well as interactions between them, in HRSV biology.