Recombinant Wild-Type and Edmonston Strain Measles Viruses Bearing Heterologous H Proteins: Role of H Protein in Cell Fusion and Host Cell Specificity

Abstract

Wild-type measles virus (MV) isolated from B95a cells has a restricted host cell specificity and hardly replicates in Vero cells, whereas the laboratory strain Edmonston (Ed) replicates in a variety of cell types including Vero cells. To investigate the role of H protein in the differential MV host cell specificity and cell fusion activity, H proteins of wild-type MV (IC-B) and Ed were coexpressed with the F protein in Vero cells. Cell-cell fusion occurred in Vero cells when Ed H protein, but not IC-B H protein, was expressed. To analyze the role of H protein in the context of viral infection, a recombinant IC-B virus bearing Ed H protein (IC/Ed-H) and a recombinant Ed virus bearing IC-B H protein (Ed/IC-H) were generated from cloned cDNAs. IC/Ed-H replicated efficiently in Vero cells and induced small syncytia in Vero cells, indicating that Ed H protein conferred replication ability in Vero cells on IC/Ed-H. On the other hand, Ed/IC-H also replicated well in Vero cells and induced small syncytia, although parental Ed induced large syncytia in Vero cells. These results indicated that an MV protein(s) other than H protein was likely involved in determining cell fusion and host cell specificity of MV in the case of our recombinants. SLAM (CDw150), a recently identified cellular receptor for wild-type MV, was not expressed in Vero cells, and a monoclonal antibody against CD46, a cellular receptor for Ed, did not block replication or syncytium formation of Ed/IC-H in Vero cells. It is therefore suggested that Ed/IC-H entered Vero cells through another cellular receptor.

Measles virus (MV), a member of the family Paramyxoviridae, genus Morbillivirus, causes an acute exanthematous disease that kills approximately 1 million children each year in the world. MV contains two envelope glycoproteins, the hemagglutinin (H) protein, responsible for receptor binding, and the fusion (F) protein, mediating membrane fusion (reviewed in references 15 and 54).

Wild-type MV strains circulating in nature could be efficiently isolated from B95a cells, an adherent subline of the B95-8 marmoset B-lymphoblastoid cell line (19). In this report we use the term “wild-type” for MV strains that have been isolated and passaged exclusively in B95a cells or human B-lymphoid cells (2). Wild-type MV strains isolated from B95a cells have been shown to induce clinical signs resembling those of human measles, such as skin rashes, Koplik's spots, and leukopenia, in experimentally infected cynomolgus and squirrel monkeys (19, 21). On the other hand, the Edmonston (Ed) vaccine strain, which has been passaged in nonlymphoid cells, is no longer pathogenic in monkey models (1, 12, 52, 55). In addition, wild-type MV strains and the Ed strain show different host cell specificities in vitro. Wild-type MV strains replicate efficiently only in B95a cells and some lymphocyte cell lines, whereas the Ed strain can replicate in a variety of cell lines including B95a and Vero cells (19, 50).

It is well established that the Ed strain and its derivatives utilize CD46 as a cellular receptor (9, 29). However, there has been an accumulation of data suggesting that wild-type MV strains use a receptor other than CD46. Recently, Tatsuo et al. (51), Hsu et al. (17), and Erlenhoefer et al. (13) reported that signaling lymphocytic activation molecule (SLAM; also known as CDw150) is a cellular receptor for both wild-type MV and Ed strains and is highly expressed on the surfaces of B95a cells. It has also been shown by using vesicular stomatitis virus pseudotypes bearing MV envelope proteins that the entry step is a major determinant of cell tropism for the Ed strain and wild-type MV (50). Previous cell fusion experiments in which the H and F proteins were expressed from cDNA also indicated that the cell tropism of MV is determined by the H protein (16, 24, 48).

However, inconsistent observations have also been reported. It was reported that no amino acid change was predicted in the H protein between a B95a cell-grown wild-type strain (9403B) and its Vero cell-adapted strain (9403V) (45). Similarly, no amino acid change occurred in the H protein between B95-8 cell-grown wild-type strains (FV93 and BCL94) and their Vero cell-adapted strains (FV93 Vero and BCL94 Vero) (26). We determined the nucleotide sequences of the entire genomes of B95a cell-isolated (IC-B) and Vero cell-isolated (IC-V) strains from the same patient and found only two nucleotide differences with predicted amino acid differences in the P/V and M proteins and a 19-amino-acid deletion in the C protein, but no difference was found in the amino acid sequences of the H proteins (47). In addition, a recombinant Ed virus bearing wild-type (strain WTF) H protein was generated and shown to replicate in Vero cells (18). These observations are inconsistent with the notion that the host cell specificity of MV is simply determined by the entry process mediated by interaction between the H protein and its receptor. To investigate the role of H protein in MV cell specificity, we generated two recombinant viruses, a wild-type MV bearing the Ed H protein and an Ed strain bearing the wild-type H protein, by using the reverse genetics systems for MV (33, 46) and compared the growth and fusion activities of these viruses in Vero and B95a cells.

MATERIALS AND METHODS

Viruses and cells.

IC323 virus, corresponding to the IC-B strain of wild-type MV (47), was recovered from plasmid p(+)MV323 encoding the antigenomic IC-B sequence as reported previously (46). B95a cells, an adherent marmoset B-cell line (19), were grown in Dulbecco's modified essential medium (DMEM) supplemented with 10% fetal calf serum (FCS). Vero (African green monkey kidney) cells (56) were grown in Eagle's minimal essential medium supplemented with 5% FCS and 10% tryptose phosphate broth (Becton Dickinson, Sparks, Md.). Vero cells expressing human SLAM (Vero/hSLAM cells) (32) were grown in DMEM supplemented with 10% FCS and 400 μg of G418/ml for stable expression of the transfected hSLAM gene product. 293-3-46 cells expressing the MV nucleoprotein (N), phosphoprotein (P), and T7 RNA polymerase (33) were grown in DMEM supplemented with 10% FCS and 1.2 mg of G418/ml.

Construction of expression plasmids and DNA transfection.

To generate plasmids expressing the F and H proteins of the IC-B strain, total RNA was isolated from B95a cells infected with the IC-B strain and cDNAs were synthesized with reverse transcriptase and a random 9-mer DNA primer. Regions spanning the F and H genes were amplified by PCR using primers IBf5 (TCTCACTCGAGATCATGGGTCTCAAGGTGAAC) and IBf3 (TGTACTGAGCTCTCAGAGCGACCTTACATAGG) for the F gene and primers IBh5 (TGTGTGGTACCACAATGTCACCACAACGAGAC) and IBh3 (GACGAGAGCTCCTATCTGCGATTGGTTCCATC) for the H gene. The amplified F and H genes were subcloned into the pCAGGSP7 vector (5), which is a derivative of pCAGGS (30), resulting in pCA-IC-F after XhoI and SacI digestions for the F gene and pCA-IC-H after KpnI and SacI digestions for the H gene. The nucleotide sequences of the coding regions of pCA-IC-F and pCA-IC-H were confirmed by the dideoxy method using an ABI 377 sequencer (Applied Biosystems, Foster City, Calif.). To generate a plasmid expressing the H protein of the Ed strain, the H protein coding region was amplified by PCR with the proofreading KOD polymerase (Toyobo, Osaka, Japan), primers XhoHA1 (CGCTCGAGGTGCAAGATCATCCACAATG) and XhoHA2 (CGCTCGAGTGGTTCACTAGCAGCCCTAT), and p(+)MV017 as a template. The DNA was subcloned into the pCAGGSP7 vector after XhoI digestion, resulting in pCA-Ed-H. For cell-cell fusion experiments, subconfluent B95a, Vero, or Vero/hSLAM cells in 6-well plates were cotransfected with equal amounts (1 μg each) of the F and H expression plasmids by using DOSPER transfection reagent (Roche Diagnostics, Mannheim, Germany) according to the protocol of the supplier. Photographs were taken 24 h after transfection by using a microscope (Zeiss, Jena, Germany) equipped with an HR250 digital camera (Fuji, Tokyo, Japan).

Construction of full-length plasmids with exchanged H genes.

Plasmid p(+)MV2A, carrying the full-genome cDNA of the Ed strain, has been described previously (33). To exchange the H gene of p(+)MV2A with that of wild-type MV (the IC-B strain), a SpeI-SpeI fragment containing both the F and H genes was excised from p(+)MV2A and subcloned into plasmid pLITMUS28 (New England Biolabs, Beverly, Mass.). The PacI-HindIII fragment was excised from this plasmid and replaced with the corresponding fragment of the IC-B strain synthesized by reverse transcription-PCR (RT-PCR) from the virion RNA. The fragment synthesized by RT-PCR was completely sequenced to rule out any amino acid change. Then the SpeI-SpeI fragment of the resulting plasmid was excised and introduced into p(+)MV2A between the SpeI sites to yield p(+)MV001, from which a chimeric Ed virus with the H protein of IC-B, designated Ed/IC-H, would be produced. Plasmid p(+)MV323, carrying the full-genome cDNA of the IC-B strain, has been described previously (46). To exchange the H gene of p(+)MV323 with that of the Ed strain, a PacI-SpeI fragment containing the H gene was excised from p(+)MV323 and replaced with the corresponding fragment synthesized by RT-PCR from the virion RNA of the Ed-B strain (36). The resultant plasmid, p(+)MV017, would produce a chimeric IC-B virus bearing the H protein of the Ed strain (IC/Ed-H).

Rescue of infectious viruses from cloned cDNA.

Infectious recombinant Ed virus (rEd) was recovered from plasmid p(+)MV2A by the standard method using 293-3-46 helper cells (33) and was propagated in B95a cells. The infectious chimeric MVs (IC/Ed-H and Ed/IC-H) were recovered from plasmids p(+)MV017 and p(+)MV001, respectively, by the modified method for recovering wild-type MV (46). Briefly, 5 μg of the plasmid carrying the full-length cDNA of the MV genome and 10 ng of pEMC-La (33) were introduced into subconfluent 293-3-46 cells in 3.5-cm-diameter dishes by using CellPhect calcium phosphate transfection reagent (Amersham Pharmacia Biotech, Piscataway, N.J.). The next day, the transfection medium was replaced with 2 ml of DMEM and incubated for 24 to 30 h. Then B95a cells were overlaid onto the transfected 293-3-46 cells to rescue and propagate MV. Syncytia which appeared 1 to 3 days after B95a overlay were picked up under a microscope and transferred into fresh cultures of B95a cells.

Immunoprecipitation.

B95a cells in 3.5-cm-diameter dishes were infected with 10⁴ 50% tissue culture infective doses (TCID₅₀) of MV and incubated for 24 h. After starvation by incubation with methionine-free DMEM for 30 min, infected cells were labeled with 20 μCi of [³⁵S]methionine (Perkin-Elmer Life Sciences, Boston, Mass.) for 2 h and lysed in radioimmunoprecipitation assay buffer. MV-specific proteins were immunoprecipitated with rabbit serum against MV strain Toyoshima (14) and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by fluorography.

Virus growth and syncytium formation in cell culture.

Monolayer cultures of B95a, Vero, and Vero/hSLAM cells in 3.5-cm-diameter dishes were infected with recovered viruses at a multiplicity of infection (MOI) of 0.01 TCID₅₀/cell. At various times, cells were scraped, and harvested cells with media were subjected to three cycles of freezing and thawing. The infectivity titer was determined by the TCID₅₀ in B95a cells.

ABC staining.

Infected monolayer cells on 3.5-cm-diameter dishes were fixed with 1% glutaraldehyde for 10 min at room temperature, washed with phosphate-buffered saline lacking magnesium and calcium [PBS(−)], and incubated with rabbit anti-MV antiserum (14). After a wash with PBS(−), cells were incubated with biotin-conjugated goat serum against rabbit immunoglobulin G (IgG) (Vector Laboratories, Burlingame, Calif.), washed with PBS(−), and stained with the Elite ABC staining kit (Vector Laboratories) in the presence of Co²⁺ and Mn²⁺.

Inhibition of MV infection with monoclonal antibodies.

Monolayers of Vero cells in 24-well plates were incubated for 60 min at 37°C with 10 μg of the anti-CD46 monoclonal antibody M75 or M160/ml (42), 5 μg of the anti-SLAM monoclonal antibody IPO-3 (Kamiya Biomedical, Seattle, Wash.)/ml, or 10 μg of the affinity-purified anti-moesin monoclonal antibody 38/87 (NeoMarkers, Fremont, Calif.)/ml in Eagle's minimal essential medium supplemented with 5% FCS; then they were infected with MV at 10⁴ TCID₅₀ per well and incubated for 60 min at 37°C. The cells were washed three times and incubated for 2 or 4 days in the presence of 10 μg of M75 or M160/ml, 5 μg of IPO-3/ml, or 10 μg of 38/87/ml.

Fluorescence-activated cell sorter (FACS) analysis.

Vero cell suspensions were recovered from monolayer cultures by incubation with 0.05% trypsin and 0.02% EDTA, centrifuged after addition of medium containing 10% FCS, and infected with MV at an MOI of 1 TCID₅₀/cell. Infected cells were stained with the anti-H monoclonal antibody B5 (39) and fluorescein isothiocyanate-conjugated anti-mouse IgG, fixed with 3.7% paraformaldehyde, and analyzed by FACScalibur (Becton Dickinson, San Jose, Calif.).

RESULTS

Cell-cell fusion assay in a transient-expression system.

We previously determined the complete nucleotide sequence of the genome of the IC-B strain (47). Predicted amino acid differences between the H proteins of the IC-B and Ed strains are shown in Table 1. The Ed H protein (Ed-H) has a tyrosine residue at position 481 which has been shown to be important for cell fusion in HeLa cells, hemadsorption, and CD46 down-regulation (16, 24, 48), while the H protein of the IC-B strain (IC-H) has an asparagine residue at this position. As expected, the IC-B strain lacks hemadsorption activity (20), and cotransfection of the IC-H and IC-F plasmids did not induce cell fusion in HeLa cells expressing CD46 (48). On the other hand, a serine-to-glycine change at position 546 in the wild-type H proteins which occurred during adaptation of wild-type MVs to Vero cells has also been reported to be important for the acquisition of binding to the CD46 receptor (35, 43). However, IC-H still retains a serine residue at this position. Ed-H also possesses the serine residue at this position, although this virus had adapted well to Vero cells (36).

TABLE 1.

Predicted amino acid differences in the H protein between the MV Edmonston strain and the wild-type IC-B strain

Amino acid position	Amino acid in strain:
	Edmonstona	IC-Bb
174	T	A
176	T	A
211	G	S
235	E	G
243	R	G
252	Y	H
276	L	F
284	L	F
296	L	F
302	G	R
334	Q	R
390	I	N
416	D	N
446	S	T
481	Y	N
484	N	T
575	Q	K

^aDDBJ/EMBL/GenBank accession no. Z66517.

^bDDBJ/EMBL/GenBank accession no. AB016162.

Wild-type MVs isolated from B95a cells can replicate only in B95a cells, several lymphocyte cell lines, and peripheral blood mononuclear cells of monkeys and humans. They neither replicate nor induce syncytia in Vero cells. In contrast, the Ed strain can replicate in a variety of cell lines including B95a and Vero cells. In addition, the Ed strain induces large syncytia in infected Vero cells. To examine the role of the H protein in the efficiency of cell-cell fusion, the F and H genes of the wild-type (IC-B) strain and the H gene of the Ed strain were cloned into the eukaryotic expression vector pCAGGS (30). Because the amino acid sequences of the F proteins of IC-B and Ed were identical (47), we used plasmid pCA-IC-F, synthesized from the IC-B strain, in all experiments. When B95a cells were cotransfected, each combination of H and F genes, pCA-IC-H/pCA-IC-F and pCA-Ed-H/pCA-IC-F, induced cell-cell fusion (Fig. 1A and D). On the other hand, in Vero cells, transfection with the combination of pCA-Ed-H and pCA-IC-F induced cell fusion (Fig. 1E) but transfection with pCA-IC-H/pCA-IC-F did not (Fig. 1B). When Vero/hSLAM cells (32) were transfected with either the pCA-IC-H/pCA-IC-F or the pCA-Ed-H/pCA-IC-F combination, cell-cell fusion was induced efficiently (Fig. 1C and F). In these experiments both the H and F proteins were found in each syncytium, and single cells expressing both proteins were not observed by immunofluorescent staining (data not shown). The apparent lack of cell-cell fusion in Vero cells cotransfected with the IC-H- and IC-F-expressing plasmids would be attributable to the absence of a high-affinity receptor (SLAM) for IC-H on the surfaces of Vero cells.

FIG. 1.

Cell-cell fusion induced after transfection of B95a, Vero, and Vero/hSLAM cells with plasmids expressing IC-B or Ed glycoproteins. B95a cells were transfected as detailed in Materials and Methods with either pCA-IC-H plus pCA-IC-F (A) or pCA-Ed-H plus pCA-IC-F (D). Vero cells were transfected with either pCA-IC-H plus pCA-IC-F (B) or pCA-Ed-H plus pCA-IC-F (E). Vero/hSLAM cells were transfected with either pCA-IC-H plus pCA-IC-F (C) or pCA-Ed-H plus pCA-IC-F (F). Cells were photographed 24 h after transfection by phase-contrast microscopy.

Construction and recovery of recombinant MVs bearing heterologous H proteins.

To investigate the role of the H protein in the context of virus infection, we generated a recombinant chimeric wild-type MV possessing Ed-H (IC/Ed-H) and a chimeric Ed virus bearing the wild-type H protein (Ed/IC-H) from the respective cloned cDNAs. MV strains recovered were propagated in B95a cells for further studies.

The identity of the H protein of each MV was confirmed by immunoprecipitation. As shown in Fig. 2, IC-H and Ed-H can be distinguished from each other by their molecular sizes due to an additional glycosylation site in IC-H (38). Interestingly, the amount of M protein detected in rEd-infected B95a cells was much smaller than that in IC323-infected cells. It is unlikely that this difference is due to poor reactivity of the antibody (anti-Toyoshima strain), because both the Ed and Toyoshima strains belong to genotype A and there are only two amino acid differences between the M proteins of the two strains. To confirm the identity of each recombinant MV, the protein-coding regions of the H, P/C/V, and M genes of IC/Ed-H and Ed/IC-H, as well as the region of the N gene corresponding to the C-terminal half of its product, were amplified by RT-PCR and sequenced. All sequence data so far determined were identical to those of the original plasmids, confirming the recovery of the desired recombinant viruses.

FIG. 2.

Immunoprecipitation analysis of MV-specific proteins of the parental and recombinant viruses. B95a cells infected with IC323 (lane 1), IC/Ed-H (lane 2), Ed/IC-H (lane 3), or rEd (lane 4) were labeled with [³⁵S]methionine, and MV-specific proteins were immunoprecipitated with rabbit antiserum against MV and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Syncytium formation and growth of recombinant MVs in B95a, Vero, and Vero/SLAM cells.

First, we examined syncytium formation by recombinant viruses in B95a, Vero, and Vero/SLAM cells. In B95a cells, all viruses induced large syncytia (Fig. 3A to D). In Vero cells, IC323 did not induce any syncytia (Fig. 3E) over a period of 5 days, as reported previously (19, 46), while rEd induced large syncytia (Fig. 3H) 1 to 2 days postinfection (p.i.), as reported previously (33). IC/Ed-H and Ed/IC-H induced small syncytia 4 to 5 days p.i. (Fig. 3F and G). In Vero/hSLAM cells, all viruses induced large syncytia 1 to 2 days p.i. (Fig. 3I to L). Interestingly, syncytia induced by IC323 (Fig. 3I) and IC/Ed-H (Fig. 3J) in Vero/hSLAM cells were relatively smaller than those induced by Ed/IC-H (Fig. 3K) and rEd (Fig. 3L). All syncytia were positive for staining with mouse monoclonal antibodies against MV and fluorescein isothiocyanate-conjugated anti-mouse IgG, whereas single cells expressing MV proteins but not involved in syncytium formation were scarcely observed (data not shown).

FIG. 3.

Syncytium formation induced by recombinant viruses. B95a cells (A to D), Vero cells (E to H), and Vero/hSLAM cells (I to L) were infected with IC323 (A, E, and I), IC/Ed-H (B, F, and J), Ed/IC-H (C, G, and K), or rEd (D, H, and L). Cells were photographed under a microscope 2 days after infection (A to D, H, and I to L) or 4 days after infection (E to G).

Next, we compared the replication kinetics of the viruses in B95a, Vero, and Vero/hSLAM cells. In B95a cells, all viruses replicated efficiently (Fig. 4A). In Vero cells, rEd replicated efficiently (Fig. 4B). The infectious rEd titer seemingly decreased from day 3 p.i. (Fig. 4B) due to extensive cell fusion in the entire culture caused by rEd, since fused cells detached from dishes and would no longer support virus replication. It should be noted that IC/Ed-H and Ed/IC-H replicated efficiently, though slowly, and reached maximum titers of approximately 10⁵ TCID₅₀/ml on day 4 p.i. IC323 hardly replicated in Vero cells (Fig. 4B), as reported previously (19, 46). In Vero/hSLAM cells, all viruses replicated more efficiently (Fig. 4C) than in B95a cells, and the replication kinetics of IC323 were comparable to those of rEd. Vero/hSLAM cells were extremely sensitive to MV, and huge syncytia consisting of several hundred cells developed within 1 to 2 days after infection and expanded to almost the entire monolayer (Fig. 3K and L), subsequently detaching from dishes. Therefore, titers for some viruses in Vero/hSLAM cells reached their peaks at day 1 and then declined, even at an MOI of 0.01.

FIG. 4.

Replication kinetics of the parental and recombinant viruses. B95a cells (A), Vero cells (B), and Vero/hSLAM cells (C) were infected with IC323 (open circles), IC/Ed-H (open triangles), Ed/IC-H (solid triangles), and rEd (solid circles) at an MOI of 0.01 TCID₅₀/cell. Cells and media were harvested at days 1, 2, 3, and 4 after infection, and infectivity titers were determined by the TCID₅₀ by using B95a cells.

The observations that Ed/IC-H replicated efficiently (Fig. 4B) and induced syncytium formation (Fig. 3G) in infected Vero cells suggested that the H protein of the wild-type MV is capable of mediating receptor binding and virus entry into Vero cells, although IC-H coexpressed with IC-F failed to induce cell-cell fusion in Vero cells (Fig. 1B).

Binding of Ed/IC-H to Vero cells.

To study how Ed/IC-H binds to Vero cells, we first carried out FACS analysis. However, the binding of Ed/IC-H to Vero cells could not be measured by FACS analysis (data not shown). On the other hand, binding of rEd to Vero cells was easily detected by FACS analysis, as reported previously (9, 29). Therefore, the plaque efficiencies of Ed/IC-H on B95a and Vero cells were determined comparatively as an indication of the binding of Ed/IC-H to Vero cells. Since plaques induced by Ed/IC-H on Vero cells were too tiny and fuzzy to be counted accurately by the standard neutral-red staining method, plaques were stained with anti-MV antiserum and the ABC staining kit as described in Materials and Methods. Neutral-red staining and ABC staining gave identical plaque formation efficiencies in most MV strains (data not shown). When the same amounts of Ed/IC-H virus stock were inoculated into B95a and Vero cells, almost the same numbers of plaques developed in the two types of cells (Fig. 5). The average numbers of plaques of Ed/IC-H forming in four different dishes of B95a cells and in four different dishes of Vero cells were 183.8 and 90.5, respectively. These results strongly suggested that Ed/IC-H must bind to the Vero cell surface, leading to the establishment of infection, and that the efficiency of binding to Vero cells was comparable to that to B95a cells. Since the IC323 virus induced neither single virus-infected cells (detectable by immunofluorescent staining) nor plaques (detectable by ABC staining) in Vero cells, the efficiency of binding of IC323 to Vero cells could not be measured by this method.

FIG. 5.

Plaque formation efficiencies of Ed/IC-H in B95a and Vero cells. B95a and Vero cells in 6-well dishes were infected with the same amount of Ed/IC-H and overlaid with agarose-containing medium. B95a cells were stained with neutral red. Vero cells were immunologically stained as described in Materials and Methods with rabbit anti-MV antiserum and an ABC staining kit.

Inhibition of MV infection by monoclonal antibodies.

Next, we examined the receptor usage of each virus in Vero cells. It has been reported that monoclonal antibodies against CD46, a cellular receptor for the Ed strain, efficiently inhibited replication and syncytium formation by the Ed strain (16, 25, 28, 41, 42). We therefore tested the effect of the monoclonal antibodies on Ed/IC-H infection. As shown in Fig. 6A, monoclonal antibody M75 (10 μg/ml) against the short-consensus-repeats 2 (SCR2) region of CD46 efficiently inhibited syncytium formation induced by rEd. Similarly, M75 inhibited syncytium formation induced by IC/Ed-H (Fig. 6B). These results strongly suggested that rEd and IC/Ed-H, bearing the Ed H protein, utilized CD46 on Vero cells as a cellular receptor. In contrast, M75 did not inhibit syncytium formation by Ed/IC-H (Fig. 6C). Another monoclonal antibody against CD46, M160, which recognizes the SCR3 region of CD46 but does not inhibit syncytium formation by the Ed strain (42), also did not inhibit syncytium formation by Ed/IC-H (Fig. 6D).

FIG. 6.

Inhibition of syncytium formation by monoclonal antibodies against MV receptors. Vero cells were pretreated with monoclonal antibodies against the SCR1 region of CD46 (M75) (A to C), against the SCR3 region of CD46 (M160) (D), against SLAM (IPO-3) (E), or against moesin (38/87) (F). Pretreated cells were then infected with rEd (A), IC/Ed-H (B), or Ed/IC-H (C to F). Cells were incubated in the presence of monoclonal antibodies and photographed at day 2 (A) or day 4 (B to F) p.i.

It has been reported that SLAM, a receptor common to wild-type and laboratory strains, is not expressed in Vero cells (51). We also confirmed this both by immunofluorescent staining of Vero cells with a monoclonal antibody to SLAM (IPO-3) and by RT-PCR using RNA extracted from Vero cells and several pairs of primers with sequences conserved between human and monkey (tamarin) SLAM genes (data not shown). In fact, monoclonal antibody IPO-3 against SLAM did not inhibit Ed/IC-H strain-induced syncytium formation in Vero cells (Fig. 6E). The same amount (5 μg/ml) of IPO-3 completely inhibited syncytium formation by Ed/ICH on B95a cells (data not shown). In addition, a monoclonal antibody against moesin (38/87), which has been reported to inhibit Ed replication (10) but for which inconsistent results have also been reported (7, 8), did not inhibit syncytium formation by Ed/IC-H (Fig. 6F), at least in this experimental condition.

Yields of progeny virus in Vero cells were then determined for the chimeric MVs by TCID₅₀ using B95a cells. As shown in Fig. 7, whereas M75 inhibited the replication of rEd and IC/Ed-H by more than 2 log units, the replication of Ed/IC-H was not influenced by M75, M160, IPO-3, or 38/87.

FIG. 7.

Inhibition of MV replication by monoclonal antibodies against MV receptors. Vero cells were pretreated with monoclonal antibodies, infected with MV, and incubated as described in the legend for Fig. 6. Cells and media were harvested at day 2 p.i. (rEd and rEd plus M75) or day 4 p.i. (all others), and infectivity titers were determined by TCID₅₀ using B95a cells.

DISCUSSION

To examine the role of the H protein in MV-induced cell fusion, we first carried out transient expression experiments. The results presented in Fig. 1 were consistent with our current knowledge of receptor usage by different MVs. Since B95a cells express SLAM, and since both wild-type and Ed strains can use SLAM as a receptor (51), it is reasonable that cell-cell fusion occurred efficiently in B95a cells expressing IC-H and IC-F (Fig. 1A) or Ed-H and IC-F (Fig. 1D). Efficient induction of cell-cell fusion in Vero cells expressing Ed-H and IC-F (Fig. 1E) could be consistent with the facts that Vero cells express CD46 and the Ed strain can use CD46 as a receptor (9, 29). Interestingly, it has been reported that extensive cell-cell fusion occurred in Vero cells when they were cotransfected with a wild-type (strain WTF) H plasmid (pCG-WTFBH) and an Ed-F plasmid (pCG-EdF) (18). Since pCG-Ed-F encodes one amino acid difference (M94V) from the published sequence of Ed, we constructed an analogous Ed-F plasmid by amplifying the F coding region directly from the antigenomic Ed plasmid p(+)MV2A. However, cell-cell fusion was not observed in Vero cells when they were cotransfected with pCA-IC-H and the Ed-F plasmid containing the M94V substitution (data not shown). There were 22 amino acid differences between the IC-B and WTF H proteins (18, 47). Any of the amino acid differences in the H protein might affect the efficiency of cell-cell fusion in Vero cells.

To investigate the role of the H protein in the context of the viral infection process, we generated the recombinant chimeric MVs IC/Ed-H and Ed/IC-H. IC/Ed-H replicated well in Vero cells (Fig. 4B) and induced syncytia (Fig. 3F). These results indicated that replacement with Ed-H could confer efficient replication in Vero cells on the wild-type IC323 virus. Since Ed-H can bind to CD46, IC/Ed-H would enter Vero cells by recognizing CD46 expressed on Vero cells as a receptor. IC/Ed-H induced relatively small syncytia in Vero cells (Fig. 3F), although rEd induced large syncytia in Vero cells (Fig. 3H). This result indicated that any viral gene(s) other than the H gene affected fusion efficiency in the case of our chimeric viruses. This was in contrast to the report that the fusogenicity of the H chimeras of canine distemper virus was predominantly dependent on the H protein regardless of the viral backbone used (53). Although IC323 hardly replicated in Vero cells, it was able to replicate in Vero cells when SLAM was expressed on the cell surfaces (Fig. 4C). Extensive syncytia were induced in the infected cell culture (Fig. 3I). These results, taken together, indicated that both the H protein and its cellular receptor are important determinants for the host cell specificity and cell-cell fusion activity of MV, supporting a previous view (50).

It was interesting that Ed/IC-H also replicated in Vero cells (Fig. 4B) and induced syncytia (Fig. 3G), although IC-H along with IC-F did not induce cell-cell fusion in Vero cells in transient expression experiments (Fig. 1B). However, this result is not surprising, since it is well known that the fusion of the virus envelope with the cell membrane is related to but distinguishable from the fusion between cell membranes leading to syncytium formation (22). One may expect that Ed/IC-H could enter Vero cells in a very inefficient manner, but once it entered the cells, it could replicate more efficiently, overcoming its inefficient entry, because the replication machinery of the Ed internal proteins had been well adapted to Vero cells. Therefore, we compared the plaque-forming efficiencies of Ed/IC-H on B95a and Vero cells. As shown in Fig. 5, Ed/IC-H gave almost the same PFU in both types of cells, showing that entry of Ed/IC-H into Vero cells was as efficient as that into B95a cells. Since IC/Ed-H did replicate in Vero cells (Fig. 4B) and IC323 replicated well in Vero/hSLAM cells (Fig. 4C), the machinery of the IC323 backbone was considered to work efficiently for viral RNA synthesis in Vero cells. These results, therefore, suggested that the above hypothesis is unlikely. If the H protein of IC323 binds to Vero cells and the RNP complex consisting of the IC323 N, P, and L proteins is active in Vero cells, why did IC323 fail to grow in Vero cells?

The MV M protein might be involved in this question. Paramyxovirus M proteins play an important role in virus assembly (reviewed in reference 23). For MV, it has been shown that the M protein regulates cell-cell fusion (3, 4) and virus budding (27) by interacting with the F and H proteins and that it inhibits viral gene transcription (44). There are four amino acid substitutions (P64S, E89K, R175G, and A209T) in the M protein between IC323 and rEd. One possible explanation is that the M protein of rEd, and not that of IC323, might cause a structural alteration in the F and/or H protein of Ed/IC-H, leading to binding of Ed/IC-H to Vero cells. At present, we have not proven this possibility, because the 24 different monoclonal antibodies against the F and H proteins tested so far failed to detect any structural differences (data not shown). On the other hand, whereas the binding of rEd to Vero cells was measured by FACS analysis, the binding of Ed/IC-H, IC/Ed-H, and IC323 to Vero cells could not be precisely measured by this method (data not shown). Since MV is highly cell associated, and repeated cycles of freezing and thawing are required for preparing MV virus stock, there would be a large amount of the H protein bound to the disrupted cell membranes in virus preparations. If the H protein has strong affinity to cellular receptors, as, for example, Ed-H and CD46, the FACS analysis, a highly sensitive biochemical method, would give a positive result. Another possible explanation is that the M protein of rEd might have a stronger effect on the virus-cell fusion-promoting activity of the viral glycoproteins than the M protein of IC323. IC323 might bind to Vero cells but could not fuse the envelope with the plasma membrane. For many paramyxoviruses, coexpression of the F and H/HN proteins is required for syncytium formation, although in some paramyxoviruses the F protein alone causes syncytium formation. It has been hypothesized that after binding of paramyxoviruses to cellular receptors, H/HN proteins would undergo a specific conformational change, which in turn could trigger a conformational change in the F protein (reviewed in reference 23). When MV binds to cell surfaces through high-affinity receptors, successive conformational changes in the H and F proteins would occur efficiently. On the other hand, when MV binds to cells through an unidentified low-affinity receptor, the conformational change in the H protein might occur insufficiently to trigger the subsequent conformational change in the F protein. The rEd M protein might efficiently transmit the conformational change occurring in the H protein to the F protein. As shown in Fig. 2, IC323 synthesized more M proteins than did Ed/IC-H in B95a cells. Although we did not measure the amount of the M protein in each MV virus particle, the difference in the amount of M protein might also be involved in altering the structures of the H and F proteins and, as a result, in the virus envelope-cell membrane fusion-promoting activity. In this context it should be noted that there are two nucleotide differences in the P/C/V and M genes and no difference in the H gene between the IC-V and IC-B (IC323) strains, which were isolated from the same patient in Vero cells and B95a cells, respectively (47). The predicted amino acid differences are in the P, V, and M proteins, and there is a 19-amino-acid deletion in the C protein of the IC-V strain. The IC-V M protein, possessing a histidine residue at position 64, might have properties similar to those of the rEd M protein, which could confer efficient replication in Vero cells on the IC-V strain. Recombinant wild-type MV bearing chimeric M proteins might provide further insight into the role of the M protein in determining the host cell specificity of MV.

What is the receptor for Ed/IC-H on Vero cells? To identify such a cellular receptor, we carried out inhibition experiments with monoclonal antibodies against the MV receptors reported so far. However, neither syncytium formation (Fig. 6C) nor viral growth (Fig. 7) of Ed/IC-H was inhibited by the monoclonal antibody (M75) against CD46 (42). These results suggested that Ed/IC-H could enter Vero cells by a CD46-independent pathway. In addition, SLAM is not expressed on Vero cells, and in fact the monoclonal antibody (IPO-3) against SLAM inhibited neither syncytium formation (Fig. 6E) nor viral replication by Ed/IC-H (Fig. 7), indicating that entry of Ed/IC-H into Vero cells occurs independently of the SLAM-mediated pathway (51). To date, we have not obtained clear-cut results by treating Vero cells with proteases (pronase and trypsin) and sialidase (from Vibrio cholerae and Clostridium perfringens), due to cytotoxic effects of these reagents at high concentrations. Generation of monoclonal antibodies against membrane antigens of Vero cells is now under way.

In human measles, it is well known that MV spreads to numerous organs other than lymphoid tissues, including the skin, conjunctivae, kidney, lungs, gastrointestinal tract, respiratory mucosa, genital mucosa, and liver (15), where the expression of SLAM is unlikely. A SLAM-independent entry pathway(s) might be important for MV to infect such organs.

The roles of the paramyxovirus glycoproteins in pathogenesis have been studied by generating recombinant viruses bearing heterologous H proteins (6, 11, 31, 40, 49). However, the reverse genetics system based on the Ed vaccine strain is not suitable for this purpose, because Ed has been too attenuated to cause clinical signs in infected monkeys (1, 12, 52, 55). IC/Ed-H, derived from a pathogenic wild-type IC323 strain, and its reciprocal version, Ed/IC-H, would be useful for examining the role of the H protein in MV pathogenesis.

It has been reported that the H protein of contemporary MV strains circulating in the world has diverged from that of the Ed vaccine strains (34, 37, 38, 43). However, it is not known whether these changes in the H protein accompany changes in antigenicity and pathogenicity. By exchanging the H protein of IC323 with those of old or contemporary MV isolates, it would be possible to compare the roles of H proteins in antigenicity and pathogenicity.