Antigenic Variation between Human Respiratory Syncytial Virus Isolates

J. gen. Virol. (1986), 67, 863-870. Printed in Great Britain 863 Key words: RS virus/antigenic variation/phosphoprotein Antigenic Variation between Human Respiratory Syncytial Virus Isolates By H. B. GIMENEZ,* N. HARDMAN, H. M. KEIR AND P. CASH 1 Department of Biochemistry, Marischal College, University of Aberdeen, Aberdeen AB9 1AS and 1Department of Bacteriology, University of Aberdeen, Foresterhill, Aberdeen AB9 2ZD, U.K. (Accepted 7 February 1986) SUMMARY Three hybridoma antibodies, prepared against the RSN-2 strain of human respiratory syncytial (RS) virus, have been used to identify antigenic variation between 41 isolates of RS virus collected from widely separated geographical regions over a period of 29 years. One antibody was directed against an antigenic site on the virus fusion protein, VP70. This site was shared by 21 virus isolates tested and its recognition by the antibody was sensitive to the presence of 2-mercaptoethanol. The remaining two antibodies used react against the virus phosphoprotein, VPP32. Two independent sites were recognized on VPP32 by these antibodies. One antibody reacted with all of the virus isolates screened while the second reacted with only 21 out of the 41 virus isolates. On the basis of the variable epitope, two antigenic types of human RS virus were identified. The distribution of each antigenic group among 28 RS virus isolates from the Grampian Region, north-east Scotland, collected between 1982 and 1984 was determined. The reactivity of these antibodies was examined using immunofluorescence staining and by immunoblotting; the latter technique also revealed that the electrophoretic mobility of VPP32 varied in parallel with the variable antigenic site. INTRODUCTION Human respiratory syncytial (RS) virus is a major cause of acute respiratory infections in infants with a peak incidence at 2 months of age (Parrott et al., 1973). Infections in adults have also been documented (Hall et al., 1976). Extensive efforts have been made to develop effective vaccines using both killed and attenuated viruses (Chanock et al., 1982). As an essential preliminary step it is necessary to identify possible antigenic variation between different RS virus isolates which may influence the effectiveness of any prospective vaccine. Human convalescent sera show minimal antigenic differences between RS virus isolates in neutralization and complement fixation tests (Bennett & Hamre, 1962; Coates et al., 1963; Beem, 1967). However, hyperimmune animal antiserum, raised against a single virus type, can be used to detect antigenic variation (Coates et al., 1963, 1966). With the advent of monoclonal antibodies against specific RS virus proteins, antigenic variation has been documented for the phosphoprotein VPP32 (Gimenez et al., 1984) and the nucleoprotein VPN4 l:(ward et al., 1984). Recently, monoclonal antibodies against the virus glycoproteins, nucleoprotein and matrix protein have been used to study antigenic variation between human RS virus isolates. Anderson et al. (1985) identified three antigenic groups among 38 RS virus isolates using monoclonal antibodies against the virus glycoproteins and nucleoprotein. Mufson et al. (1985) demonstrated two subtypes among seven virus isolates on the basis of differences in the glycoproteins, nucleoprotein and matrix protein. The latter group of workers were unable to detect antigenic differences in the phosphoprotein using a single monoclonal antibody. In this report, we investigate the antigenic and structural variation of the virus phosphoprotein in 41 RS virus isolates. 0000-6951 1986 SGM

864 H. B. GIMENEZ AND OTHERS METHODS Cells and virus. The BS-C-1 and HEp-2 cell lines were maintained in Eagle's medium (Glasgow modification, Flow Laboratories) supplemented with 10~ newborn calf serum. For the MRC-5 cell line the medium was also supplemented with l ~o non-essential amino acids. The human RS virus isolates RS A2, RS Long and virus isolates from Glasgow, U.K. (RSG-1503, -4988, -2528, -2520, -2818 and -2707) and Newcastle, U.K. (RSN-2, -1599, -7463, -7335 and RSS-2) were kindly supplied by Professor C. R. Pringle (University of Warwick, Coventry, U.K.). The isolates RS A2 and RS Long were originally isolated in Australia and U.S.A. respectively. The RS virus isolates from Aberdeen, U.K. (designated RSF-4 to RSF-54) were isolated by one of us (H.B.G.) from infants (< 6 months of age) suffering from bronchiolitis and admitted to the Royal Aberdeen Children's Hospital. Nasopharyngeal secretions from infected infants were inoculated directly onto monolayers of HEp-2 cells. The cells were overlaid with Eagle's medium containing 5~o foetal calf serum and additional antibiotics (penicillin and streptomycin). Syncytia were normally observed after 5 to 13 days incubation at 34 C. If necessary the growth medium was replaced with fresh medium at 7 days post-infection. Demonstration of the presence of RS virus was by immunofluorescence staining using bovine RS virus antiserum (Wellcome Laboratories). The RS virus isolates RSN-2, RSN-1599, RSN-7463, RSN-7335, RSG-1503, RSG-4988, RS A2 and RS Long were grown in BS-C-1 cells. RSS-2 was grown in MRC-5 cells and the remaining viruses grown in HEp-2 cells. Immunofluorescence staining. Cell monolayers were grown on coverslips and infected with RS virus isolates (m.o.i. approximately 0.01 p.f.u./cell). Uninfected cells were processed in parallel. Indirect immunofluorescence staining was performed as described by Faulkner et al. (1976). For the detection of intracellular antigen the cell monolayer was fixed by a brief immersion in acetone at - 20 C, air-dried and then stained. For the detection of surface antigen the cell monolayers were stained directly without fixation. Cells were fixed immediately before mounting on to the microscope slides. Source ofhybridoma antibodies and human sera. The source of the monoclonal antibodies 3-5 and 4-14 has been documented previously (Gimenez et al., 1984). The hybridoma antibody designated 4-15 was prepared from the spleen of BALB/c mice inoculated intraperitoneally with 100000 p.f.u, of density gradient-purified RSN-2 virus inactivated with u.v. light, with a boost inoculation of live RSN-2 virus 3 days before extraction of the spleen. The hybridomas were prepared from these mice and screened as described previously (Gimenez et al., 1984). Human serum containing antibodies to RS virus as shown by the complement fixation test (titre 1:512) was kindly supplied by Dr M. A. J. Moffat, Department of Bacteriology, University of Aberdeen. Neutralization assay. To test for neutralization with the hybridoma 4-15, 0-2 ml of undiluted tissue culture fluid was mixed with 0-2 ml of virus suspension containing 105 p.f.u, of virus. After incubation for 1 h at 18 C the virus was assayed using the plaque assay described by Gimenez & Pringle (1978). The titres were compared with virus treated in parallel With HAT medium. Preparation of virus proteins. For intracellular proteins, cell monolayers in 60 mm Petri dishes were infected with RS virus at an approximate m.o.i, of 0-01 p.f.u./cell. After 1 h adsorption at room temperature the cell monolayers were overlaid with Eagle's medium containing 5 ~ foetal calf serum and incubated at 34 C. When formation of syncytia was extensive (normally 5 to 7 days post-infection) the cells were harvested by washing three times with 0.01 i-tris-hcl ph 7.4, and lysed in 0-2 ml of protein dissociation buffer (Gimenez et al., 1984). Pelleted virus was prepared as described previously (Gimenez et al., 1984). SDS-PAGE and immunoblotting. The procedures used for SDS-PAGE and immunoblotting were as described previously (Gimenez et al., 1984) except that the antibodies were used at the following dilutions: human serum at 1 : 2000, ascites fluids for hybridomas 3-5 and 4-14 at 1 : 85 and 1 : 60 respectively and the tissue culture fluid from the hybridoma 4-15 was used undiluted. RESULTS Characterization of hybridoma antibodies The two monoclonal antibodies designated 3-5 and 4-14 react specifically against the virus phosphoprotein VPP32 (Gimenez et al., 1984). A third hybridoma antibody (4-15), prepared as described in Methods, was characterized as follows. The antibody reacted against an antigen on the surface of virus-infected cells, but not uninfected cells, as shown by immunofluorescence staining of unfixed, live cells (Fig. 1), suggesting that it may be specific for a virus surface protein, one of the virus glycoproteins for example. However, the antibody failed to react by immunoblotting against any virus proteins treated with SDS and 2-mercaptoethanol (2-ME). It has been reported that the glycoprotein VP70, the virus fusion protein, consists of two components (VGP48 and VGP26) held together by disulphide bonds (Gruber & Levine, 1983; Walsh et al., 1985). In view of these data, we tested the specificity of 4-15 against purified RS virus particles treated by each of three methods: (a) the proteins were heated at 100 C for 5 min

RS virus antigenic variation 865 Fig. 1. Immunoflaorescence staining of RS virus-infected BS-C-1 cells using the 4-15 antibody. Indirect immunofluorescence staining of live, unfixed cells was performed as described in Methods. in the presence of 2~ (w/v) SDS; (b) proteins were not treated with either SDS or 2-ME but separated electrophoretically in the presence of 0-1 ~ (w/v) SDS; (c) the proteins were heated at 100 C for 5 min in the presence of 2 ~ (w/v) SDS and 5 ~ (v/v) 2-ME. Prote.ins treated by each of these methods were resolved by SDS-PAGE and transferred to nitrocellulose membranes. Duplicate strips of nitrocellulose membrane were reacted with either the hybridoma antibody or human serum. The latter was used to locate the RS virus proteins (Fig. 2). Treatment of virus particles with 2~ SDS and 5~ 2-ME showed VGP48 to be present, but VP70 absent (Fig. 2c, lane 5). The VGP26 component of VP70 was not resolved under the electrophoretic conditions used. With 2~ SDS alone, VP70 but not VGP48 was detected (Fig. 2a, lane 1). Both VP70 and the 145000 mol. wt. dimeric form of VP70 (Walsh et al., 1985) were detected when SDS and 2- ME were omitted from the sample preparation buffer (Fig. 2b, lane 3). When the 4-15 antibody was used with these three virion preparations, the antibody reacted with the native VP70 and the 145000 mol. wt. protein (Fig. 2, lanes 2 and 4). Using the neutralization assay described in Methods, the tissue culture fluid from two independent clones of the 4-15 hybridoma reduced the titre of the RSN-2 virus by 25-fold. Antigenic variation in RS virus isolates The three antibodies described above were used to detect antigenic variation between 41 RS virus isolates (Table 1). The data previously obtained by Gimenez et al. (1984) for four virus isolates using the antibodies 3-5 and 4-14 have been included for comparison. The place and date of isolation of the viruses are presented in Methods and Table 1. Cells were grown on coverslips and infected with the respective virus. Infected cells were processed for immunofluorescence staining either as acetone-fixed preparations for antibodies 3-5 and 4-14 or as unfixed live cells for antibody 4-15. All preparations of the virus isolates produced comparable levels of staining with bovine RS virus antiserum and with the cross-reacting 3-5 antibody (see below). Twenty-one virus isolates were screened for reaction against the antibody 4-15; all gave surface immunofluorescence staining. Comparison of 39 isolates using the monoclonal antibodies 3-5 and 4-14 demonstrated antigenic variation for VPP32. All virus isolates reacted with 3-5, but the reactivity against 4-14 was variable (Table 1). Nine RS virus isolates gave clear

866 H. B. GIMENEZ AND OTHERS (a) (b) (c) 1 2 3 4 5 145K-- VP70~ -- VGP90 --VGP48 VPN41 ~VPP32 --VPM27 Fig. 2. Reaction of the 4-15 antibody with the proteins of RS virus. Proteins in pelleted virus were treated by one of three methods (a to c) described in the text. For each method, duplicate strips of nitrocellulose membrane processed for immunoblotting, either with human serum (lanes 1, 3 and 5) or with 4-15 antibody (lanes 2, 4 and 6). immunofluorescence staining, as previously described for the 4-14 antibody against the RSN-2 isolate (Gimenez et al., 1984). Of the remaining isolates, 18 showed no detectable staining with 4-14 and 12 gave a low level of staining. The reactions of the antibodies 3-5 and 4-14 against 36 RS virus isolates were investigated using immunoblotting. Proteins in lysates from RS virus-infected cells were prepared, resolved by SDS-PAGE and transferred to nitrocellulose membranes. Replicate strips were reacted either with antibody 3-5 or with 4-14. Representative data for eight RS virus isolates are presented in Fig. 3. All of the virus polypeptides were located when nitrocellulose strips were reacted with human serum (data not shown). All of the virus isolates reacted with the 3-5 antibody (Fig. 3 a). A non-specific reaction with a host protein was observed for both antibodies when virus-infected cell lysates were used as the antigen, but not when pelleted extracellular virus was used (compare Fig. 3 a, lanes 3 and 4). Immunoblotting with virus isolates which gave positive immunofluorescence staining with 4-14 demonstrated that the same antibody reacted with VPP32 (Fig. 3b, lanes 6 and 8). Conversely, RS virus isolates which were negative for immunofluorescence staining with 4-14 also failed to react under the immunoblotting conditions (Fig. 3 b, lanes 2 to 5, 7 and 9). Virus isolates which showed the low level of staining with the 4-14 antibody were found to be positive by immunoblotting. From these data the latter group were considered as showing a positive reaction with the 4-14 antibody.' Consequently, two antigenic groups (I and II) of RS virus were recognized, based on the pattern of reaction with the monoclonal antibody 4-14 (Table 1).

- - ND - - ND RS virus antigenic variation 867 Group I II Table 1. Reaction of RS virus isolates with hybridoma antibodies lmmunofluorescence Immunoblotting VPP32 Virus Year of ~. x mol. wt. isolate isolation 4-15 3-5 4-14 3-5 4-1 ( 10-3) * RSN-7335 1978 + + RSS-2 1976 + + RSG-1503 1972 + + RSG-4988 1976 + + RSG-2520 1982 + + RSG-2818 1982 ND~ + RSG-2707 1982 ND + RS A2 1961 ND + RS Long 1956 ND + RSF-4 1982 + + RSF-5 1982 + + RSF-7 1982 + + RSF-8 1982 + + RSF-9 1983 + + RSF-13 1983 + + RSF-17 1984 ND ND RSF-32 1984 + + RSF-44 1984 + + RSF-54 1984 + + RSN-2 1972 + + RSN-1599 1975 + + RSN-7463 1978 + + RSG-2528 1982 + + RSF-15 1984 ND ND RSF-20 1984 + + RSF-22 1984 ND + RSF-24 1984 ND + RSF-25 1984 ND + RSF-26 1984 ND + RSF-29 1984 ND + RSF-33 1984 ND + RSF-34 1984 ND + RSF-35 1984 + + RSF-36 1984 ND + RSF-37 1984 ND + RSF-38 1984 + + RSF-39 1984 ND + RSF-41 1984 ND + RSF-45 1984 ND + RSF-46 1984 ND + RSF-52 1984 ND % *Molecular weight values ~om Cash et al. (1977). t ND, No data. Low imrnunofluorescence staining. ND ND -- ND ND ND - -- + -- 35 - - ND ND ND + -- 35 -- ND ND ND + + + 31 ND ND ND ND + + 32 +$ + + 32 +$ + + 32 +5 + + 32 +$ + + 32 We have examined 28 RS virus isolates collected from infants in the Grampian Region (north-east Scotland) between 1982 and 1984 (Table 1). The numbers of virus isolates belonging to antigenic groups I and II was determined by immunofluorescence staining and immunoblotting with the 4-14 antibody (Table 1). During 1982 and 1983 all of the RS virus isolates were group I, whereas in 1984 there was co-circulation of both antigenic variants with four out of 22 isolates from group I and 18 from group II. Comparison of RS virus isolates collected from patients in Glasgow also presented evidence for co-circulation of the two antigenic groups during 1982. RS virus isolates from Newcastle collected between 1972 and 1978 showed representatives of each of the antigenic groups, although these were found in separate years (Table 1).

868 H. B. G I M E N E Z (a) (a) 1 AND OTHERS 2 3 4 5 6 7 8 9 1 2 3 4 5 6 VPP32~ 7 8 9 ~VPP32 Fig. 3. Reaction of monoclonal antibodies 3-5 and 4-14 with RS virus isolates. The virus antigen used for lanes 2, 3 and 6 was prepared from pelleted virus. Cell lysates from RS virus-infected cells were used for the remaining lanes which were prepared, analysed by SDS-PAGE and transferred to nitrocellulose membrane as described in Methods. Duplicate membranes were immunoblotted either with antibody 3-5 (a) or with 4-14 (b). Lanes 1, uninfected cells; lanes 2, RS A2; lanes 3, RS Long; lanes 4, RSS-2; lanes 5, RSN-7335; lanes 6, RSN-2; lanes 7, RSG-4988; lanes 8, RSN-1599; lanes 9, RSG-1503. Polypeptide analysis o f R S virus isolates Cash et al. (1977) compared the protein profiles of 11 isolates of human RS virus by S D S P A G E and demonstrated variation in the electrophoretic mobility of VPP32. By this criterion, three groups of RS virus were identified according to the relative mobility of VPP32. W e have examined an additional 29 isolates by SDS P A G E and observed the same size classes for VPP32. Representative d a t a showing the different electrophoretic mobilities of VPP32 are presented in Fig. 3 for eight virus isolates. There was a correlation between the reactivity of VPP32 with antibody 4-14 and its electrophoretic mobility. Viruses possessing a slow migrating VPP32 (apparent mol. wt. 35000) failed to react with antibody 4-14 (for example, Fig, 3, lanes 7 and 9) whereas viruses with VPP32 of greater electrophoretic mobility (apparent mol. wt. 32 000 or less) reacted with 4-t4 (for example, Fig. 3, lanes 6 and 8). One RS virus isolate (RSF-44) has been cloned and the clones showed the same pattern as the uncloned virus (data not shown). Cash et al. (1977) examined two RS virus isolates with fast migrating VPP32 (apparent tool. wt. 31000). One of these viruses (RSN-1599; Fig. 3, lane 8) reacted with 4-14 in the immunoblot assay. The d a t a for 35 virus isolates, which were c o m p a r e d with respect to both the antigenicity and electrophoretic mobility of VPP32 are summarized in Table 1. Differences were also noted in the electrophoretic mobilities of VGP90, VGP48 and VPM27 between RS virus isolates (H. B. Gimenez, unpublished data).

RS virus antigenic variation 869 DISCUSSION Three hybridoma antibodies were used to determine the extent of antigenic variation between 41 RS virus isolates. Two monoclonal antibodies (3-5 and 4-14) were specific for VPP32, as shown by immunoprecipitation and immunoblotting with virus proteins (Gimenez et al., 1984). This initial work indicated that one of these antibodies (4-14) recognized an epitope which varied between four human RS virus isolates. The present study has extended this to a total of 41 RS virus isolates, including a collection of 28 viruses isolated from a single locality. The two antibodies defined two epitopes on the VPP32 protein: one was conserved between isolates (defined by antibody 3-5), and the second was a variable site defined by antibody 4-14. The reaction of the antibodies with the virus isolates was established using two methods: immunofluorescence staining of virus-infected cells, or immunoblotting of virus proteins resolved by SDS-PAGE. Although the former method is the most widely used it is difficult to interpret the results in cases with a low level of immunofluorescence. Anderson et al. (1985) also pointed out that some of the epitopes recognized by their monoclonal antibodies showed varying sensitivities to fixation among different RS virus isolates. The immunoblotting technique was most useful with virus isolates which gave low levels of immunofluorescence staining since these were readily typed by this method. On the basis of the variable antigenic site we distinguished two groups (I and II) of RS virus according to whether or not they reacted with 4-14. Although the viruses which reacted with 4-14 (group II) may have represented a fairly homogeneous group, at least with respect to this epitope, those not reacting with 4-14 might be a heterogeneous group; a number of different mutations at the epitope may be expected to affect the binding of the antibody. This would be particularly true for viruses isolated from widely separated locations on different dates. We also examined viruses collected between 1982 and 1984, from north-east Scotland. Group I virus isolates were demonstrated during all 3 years whereas viruses in group II were only identified during 1984. Although only a few RS virus isolates were available from 1982 and 1983, our data may be explained by antigenic drift in RS virus or by the introduction of new antigenic virus strains into the Grampian area. Support for the latter hypothesis derives from the identification of RS viruses that reacted with 4-14 which had been isolated in localities outside the Grampian area earlier than 1984. The detailed study of these viruses will prove a fruitful source of material with which to test for the origin of this antigenic difference. The results we describe above taken together with those of Anderson et al. (1985) clearly show that multiple antigenic variants of RS virus co-circulate in a limited geographical area over a limited period of only a few months. The influence that this may have on the epidemiology of RS virus and development of future vaccines remains to be determined. Only limited comparisons between our data and those of Anderson et al. (1985) and Mufson et al. (1985) can be made since antigenic sites on different proteins were examined. However, there is some overlap in the RS virus isolates chosen for analysis in the three investigations (RS Long, RS A2, RSN-2 and RSN-1599). The results place RSN-2 and RSN-1599 into a single group which is separate from that of RS Long and RS A2. The latter two viruses are closely related antigenically. Although the data are limited, the possibility exists that RS virus isolates may fall into two groups characterized by a number of antigenic differences as proposed by Mufson et al. 0985). During the course of determining the reactions of antibodies 3-5 and 4-14, extensive use was made of the immunoblotting technique with virus proteins resolved by SDS-PAGE. We demonstrated an empirical relationship between the electrophoretic mobility of VPP32 and the reaction of the protein with the 4-14 antibody. The structural basis for this correlation is unknown. It is possible that the mutation(s) affecting the mobility of VPP32 also affected the epitope recognized by the antibody. The shift in the apparent molecular weight of VPP32 could be accounted for by the addition (or deletion) of a number of amino acids which also affected the 4-14 epitope. However, dejong et al. (1978) demonstrated that the substitution of a single hydrophilic amino acid by a hydrophobic amino acid can alter the mobility of a protein analysed by SDS-PAGE. If such a substitution were to occur at the 4-14 epitope then it might also alter the recognition of the site by antibody. Alternatively, since VPP32 is a phosphoprotein (Cash et al., 1979), it may be that the level of phosphorylation might affect both the recognition of the 4-

870 H. B. GIMENEZ AND OTHERS 14 epitope and the electrophoretic mobility of VPP32. Further structural studies of VPP32 will be necessary to distinguish these alternatives; these studies will be greatly aided by the nucleotide sequence information on the VPP32 gene which is now available (Satake et al., 1984; Lambden, 1985). We have also characterized an antigenic site on VPV0 that is recognized by the hybridoma antibody 4-15. A comparison of 21 isolates, belonging to either antigenic group I or II, showed that the antigenic site was conserved between these viruses. The location of the antigenic site on VPV0 was unknown. It is possible that the antigenic site involves both of the components of VP70. Thus, treatment with 2-ME, which dissociates these proteins (Gruber & Levine, 1983), would eliminate this antigenic determinant. A second possibility is that the site resides on one of the two proteins but that its conformation is maintained by disulphide bonds. A more detailed analysis of VPV0 will be required to determine which of these two possibilities is correct. We would like to thank Professor T. H. Pennington for his comments on the manuscript and Professor C. R. Pringle for providing virus isolates. We are also grateful to Drs P. Saul, E. Hennassy and the nursing staffof the Royal Aberdeen Children's Hospital for their collaboration in collecting the nasopharyngeal secretions. This work was supported by a grant from the Medical Research Council. REFERENCES ANDERSON, L. J., HIERHOLZER, J. C., TSOU, C., HENDRY, R. M., FERNIE, B. F., STONE, Y. & MCINTOSH, K. (1985). Antigenic characterisation of respiratory syncytial virus strains with monoclonal antibodies. Journal of Infectious Diseases 151, 626-633. BEEM, M. (1967). Repeated infections with respiratory syncytial virus. Journal of Immunology 98, 1115-1122. BENNETT, C. R. & HAMRE, D. (1962). Growth and serological characteristics of respiratory syncytial virus. Journal of InJectious Diseases 110, 8 16. CASH, P., WUNNER W. H. & PRINGLE, C. R. (1977). A comparison of the polypeptides of human and bovine respiratory syncytial viruses and murine pneumonia virus. Virology 82, 369-379. CASH, P., PRINGLE, C. R. & PRESTON, C. M. (1979). The polypeptides of human respiratory syncytial virus: products of cell-free protein synthesis and post-translational modifications. Virology 92, 299-313. CHANOCK, R. M., KIM, H. W., BRANDT, C. D. & PARROTT, R. H. (1982). Respiratory syncytial virus. In Virallnfectionsof Humans, pp. 471 489. Edited by A. S. Evans. New York: Plenum Press, COATES, H. V., KENDRICK, L. & CHANOCK, R. M. (1963). Antigenic differences between two strains of respiratory syncytial virus. Proceedings of the Society Jbr Experimental Biology and Medicine 112, 958-964. COATES, H. V., ALLING, D. W. & CHANOCK, R. M. (1966). An antigenic analysis of respiratory syncytial virus isolates by a plaque reduction neutralisation assay. American Journal of" Epidemiology 83, 299-313. DEJONG, W., ZWEERS, A. & COHEN, L. H. (1978). Influence of single amino acid substitutions on electrophoretic mobility of sodium dodecyl sulfate-protein complexes. Biochemical and Biophysical Research Communications 82, 532-539. FAULKNER, G. P., SHIRODARIA, P. V., FOLLETT, E. A. C. & PRINGLE, C. R. (1976). Respiratory syncytial virus ts mutants and nuclear immunofluorescence. Journal of Virology 20, 487 500. GIMENEZ, H. B. & PRINGLE, C. R. (1978). Seven complementation groups of respiratory syncytial virus temperature sensitive mutants. Journal of Virology 27, 459-464. GIMENEZ, H. B., CASH, P. & MELVIN, W. T. (1984). Monoclonal antibodies to human respiratory syncytial virus and their use in comparison of different virus isolates. Journal of General Virology 65, 963-971. GRUBER, C. & LEVINE, S. (1983). Respiratory syncytial virus polypeptides. III. The envelope-associated proteins. Journal of General Virology 64, 825-832. HALL, C. B., GEIMAN, J. M., BIGGAR, R., KOTOK, D. I., HOGAN, P. M. & DOUGLAS, R. G. (1976). Respiratory syncytial virus infections within families. New England Journal of Medicine 294, 414 419. LAMBDEN, V. R. (1985). Nucleotide sequence of the respiratory syncytial virus phosphoprotein gene. Journal of General Virology 66, 1607-1612. MUFSON, M. A., ORVELL, C., RAFNAR, B. & NORRBY, E. (1985). Two distinct subtypes of human respiratory syncytial virus. Journal of General Virology 66, 2111 2124. PARROTT, R. H., KIM, H. W., ARROBIO, J. O., HODES, D. S., MURPHY, B. R., BRANDT, C. D., CAMARGO, E. & CHANOCK, R. M. (1973). Epidemiology of respiratory syncytial virus infection in Washington, D.C. II. Infection and disease with respect to age, immunologic status and sex. American Journal of Epidemiology 98, 289-300. SATAKE, M., ELGANO, N. & VENKATESAN, S. (1984). Sequence analysis of the respiratory syncytial virus phosphoprotein gene. Journal of Virology 52, 991-994. WALSH, E. E., BRANDRISS, M. W. & SCHLESlNGER, J. J. (1985). Purification and characterization of the respiratory syncytial virus fusion protein. Journal of General Virology" 66, 409-415. WARD, K. A., EVERSON, J. S., LAMBDEN, P. R. & WATT, P. J. (1984). Antigenic and structural variation in the major nucleocapsid protein of respiratory syncytial virus. Journal of General Virology 65, 1749-1757. (Received 15 November 1985)