Deborah Lyn, 1 Mary B. Mazanec, 2,3 John G. Nedrud 3 and Allen Portner 1. Introduction

Transcription

1 Journal of General Virology (1991), 72, Printed in Great Britain 817 Location of amino acid residues important for the structure and biological function of the haemagglutinin-neuraminidase glycoprotein of Sendai virus by analysis of escape mutants Deborah Lyn, 1 Mary B. Mazanec, 2,3 John G. Nedrud 3 and Allen Portner 1. 1Department of Virology and Molecular Biology, St Jude Children's Research Hospital, P.O. Box 318, Memphis, Tennessee and Departments of 2 Medicine and 3pathology, Case Western Reserve University School of Medicine, Cleveland, Ohio 44106, U.S.A. To locate sites important for the structure and function of the haemagglutinin-neuraminidase glycoprotein (HN) of Sendai virus, the biological characteristics of antibody-selected escape mutants were correlated with mutations in the primary HN amino acid sequence. An escape mutant virus deficient only in neuraminidase function but with an HN content equal to that of the wild-type virus had an amino acid change at residue 184, implying that this position may be important for maintaining a functionally active enzymic site. In contrast, other escape mutant viruses with reductions in haemagglutination (eightfold) and neuraminidase activities (70 to 80%) had a sharply diminished HN content and substitutions either at residue 375, or double mutations at residues 279 and 461. The loss of biological activity with the concomitant loss of HN content suggests that these sites may be important for the processing and transport of HN, or in maintaining a structure resistant to proteolytic degradation; residue 451 was shown to have an undefined role in fusion activity. The monoclonal antibodies (MAbs) used to isolate the mutant viruses included those of the IgA and IgG classes and were divided into four operational groups based on their haemagglutination-inhibition pattern against the selected mutants. MAbs of the IgA class recognized epitopes overlapping with (group A) as well as epitopes distinct from (groups C and D) those recognized by the IgG class; group B included only IgG antibodies. The epitopes recognized by IgA antibodies may identify residues important for the secretory immune response to the HN molecule. Introduction Paramyxoviruses are a leading cause of respiratory infections in young children (Chanock & Parrott, 1965; Glezen et al., 1971). Sendai virus, a natural pathogen of mice, has been used extensively as a laboratory model to study paramyxoviruses. Infection of host cells by Sendai virus is initiated by the haemagglutinin-neuraminidase glycoprotein (HN) and fusion glycoprotein (F), which project from the enveloped surface of the virion. HN is a multifunctional molecule responsible for virus attachment to host and red blood cells (haemagglutinin), whereas the neuraminidase activity is thought to play a role in the release of progeny virus by cleavage of sialic acid (Scheid et al., 1972; Homma & Ohuchi, 1973; Scheid & Choppin, 1974; Nagai & Klenk, 1977). Biochemical and immunological studies have implied a role for HN in fusion, in addition to the attachment of the virus to host cell receptors (Ozawa et al., 1979; Miura et al., 1982; Orvell & Grandien, 1982; Merz & Wolinsky, 1983; Heath et al., 1983; Portner, 1984; Portner et al., 1987; Uchida et al., 1984). F is responsible for fusion of virus and host cell membranes during virus penetration into the host cell, cell-to-cell spread and haemolysis (Homma & Ohuchi, 1973; Scheid & Choppin, 1974, 1977; Nagai et al., 1976). Valuable information on the structure and function of the HN of Sendai virus has been obtained from escape mutants selected wth anti-hn neutralizing monoclonal antibodies (MAbs) and from sequence analysis of a temperature-sensitive (ts271) mutant (Thompson & Portner, 1987). Amino acids in the primary HN sequence of Sendai virus were identified as being responsible for haemagglutination (HA) activity, neuraminidase activity and haemolysis. Iorio et al. (1989) have related changes in the neuraminidase activity of Newcastle disease virus escape mutants to alterations in the HN content of the virus. Sequence analysis of the HN genes of these mutant viruses led to the identification of a domain in HN important for neuraminidase function. We have used a similar approach to isolate escape mutants of Sendai virus with anti-hn MAbs. The HA SGM

2 818 D. Lyn and others and neuraminidase activities of the derived mutant viruses were correlated with changes in their HN content and to mutations in the primary amino acid sequence. Analysis of a mutant virus deficient only in neuraminidase function identified residue 184 as essential for neuraminidase activity, which was not related to an alteration in HN content. On the other hand, virus escape mutants were isolated in which a loss of both the HA and neuraminidase activities were correlated with a reduction of HN content. The amino acid changes of these mutant viruses identified residues which may cause structural alterations in the HN molecule. This study also included MAbs of the IgA and IgG classes. Antibodies of the IgA class are locally secreted at the mucosal surfaces and are the predominant antibodies of the upper respiratory tract (Tomasi, 1983). These upper airway antibody secretions, rather than serum antibodies, are thought to play a major role in resistance to paramyxovirus reinfections (Smith et al., 1966; Mills et al., 1971). Therefore, we sought to identify the epitopes recognized by the IgA MAbs, which may provide insight into determinants of the secretory immune response to paramyxovirus infection. We constructed an operational map of the antibodies and compared the epitopes recognized by MAbs of the IgA class with those recognized by MAbs of the IgG class. One MAb of the IgA class recognized an epitope also recognized by the IgG class, whereas two other IgA antibodies identified epitopes distinct from those recognized by the IgG class. Overall, this report identified amino acids in the primary sequence of HN which are involved in virus neutralization, neuraminidase activity and the structural stability of HN. The IgA MAbs recognized two new locations on the primary amino acid sequence of HN responsible for virus neutralization. Methods Viruses. The Enders strain of Sendai virus was used in this study and served as the wild-type virus. Viruses were grown in 10-day-old embryonated chicken eggs for 3 to 4 days, the allantoic fluid was harvested and virus was purified as previously described (Portner et al., 1987). MAbs to the HN of Sendai virus. Hybridomas of the IgG class were prepared and screened as described by Portner (1981). The MAbs were determined to be specific for HN by Western blot analysis (data not shown). Immunoglobulin isotypes of the IgG MAbs were analysed by their reactivity with isotype-specific antisera in an ELISA (Boehringer Mannheim). MAbs of the IgA class, raised against a different strain of Sendai virus, had been characterized previously (Mazanec et al., 1987). Selection of escape mutant viruses. Escape mutant viruses were isolated by direct selection in embryonated eggs. Briefly, undiluted ascites fluid and serial 10-fold dilutions of virus were incubated for 1 h at room temperature and injected into embryonated chicken eggs. After 72 to 96 h, the allantoic fluid was collected from eggs which had HA activity. A mutant virus was considered to have escaped neutralization if there was no reactivity towards the selecting virus MAb in a haemagglutination inhibition (HAI) assay and if there was no reduction in p.f.u, in a neutralization test (Portner, 1984). If an MAb did not inhibit these functions, a stock of the allantoic fluid was used to prepare purified virus. Biological tests. Procedures to assay both the wild-type and escape mutant viruses for inhibition of HA, neuraminidase and haemolysin activities by the selecting MAbs were performed as described (Aymard-Henry et al., 1973; Portner et al., 1987). Neuraminidase assays were done using N-acetyl neuraminylactose (Sigma A3001) as the substrate. ELISA was carried out as recommended by van Wyke et al (1981). To determine the ELISA reactivity of the MAbs, ascites fluids were assayed against the purified wild-type virus or a selected mutant virus as antigen; reciprocal ELISA titres of the MAbs with wild-type virus were greater than SDS-PAGE was performed according to Laemmli (1970). RNA sequence analysis of the HN gene. Genomic RNA was obtained from purified virions as previously described (Bean et al., 1980). Synthetic oligonucleotides complementary to the Enders strain of Sendai virus (Gorman et al., 1990) were used to sequence the RNA by reverse transcription in dideoxynucleotide chain-terminating reactions (Sanger et al., 1977). The reactions were simultaneously carried out with the wild-type and mutant virus genomic RNAs. Results Reactivity of MAbs and selection of escape mutant viruses To select escape mutant viruses for functional and sequence analysis, we initially characterized a panel of MAbs specific for Sendal virus HN for their inhibitory effects against the HA, neuraminidase and haemolysis activities of wild-type virus (Table 1). The IgG antibodies had HAI titres of 800 or greater and inhibited neuraminidase activity by 83 to 98%. IgA MAbs 561 and 380 had HAI titres of 200 and 400, respectively, and showed variable inhibition of neuraminidase activity (86% and 63%, respectively). A single antibody (IgA), MAb 37, inhibited haemolysis (fusion) but did not block virus attachment (HAI titre of 10) or neuraminidase activity (5 % inhibition), suggesting a specific function of HN in the F-mediated fusion activity. The way in which the HN molecule is involved with fusion is still unknown. To relate specific regions on the primary amino acid sequence of HN to biological functions, we selected and sequenced escape mutant viruses using the MAbs listed in Table 1. To determine whether the isolation procedure had indeed selected virus escape mutants, the HN functions of the prospective mutant virions were tested for their susceptibility to inhibition by the selecting MAbs. HAI, neuraminidase inhibition (NI) and haemolysin inhibition (HLI) assays were performed and in all cases the mutant viruses were no longer susceptible to biological inhibition by the selecting antibody (Table 1). The ELISA was carried out to determine whether the lack of inhibition of mutant viruses was due to an

3 HN of Sendai virus 819 Table 1. Reactivity of anti-hn MAbs Selecting Antibody HAI NI HLI antibody class/isotype Virus fitre (~)* (~)* 561 IgA wtt /1C < IgA wt /7C < IgA wt /6C < IgG2a wt /7A IgG2a wt /6A IgG2a wt /6C IgG2b wt /7B < IgG2b wt /6A < IgG2b wt /5B IgG2b wt /2A < IgG1 wt B < IgG 1 wt /2A < * Assays were done using 1/100 dilutions of ascites and HLI tests. t wt. Wild-type Sendai virus fluid for the NI inability of the antibody to bind or whether the antibodies bound but did not inhibit biological activities. The ELISA reactivity of all the mutant viruses was reduced 1000-fold with the selecting MAb when compared to wild-type virus, indicating that the antibody had little or no reactivity against the selected mutant virus (data not shown). Thus, we selected escape mutant viruses that could be used to map MAb epitopes and which were suitable for functional and sequence analysis. To delineate the epitopes recognized by the selecting MAbs we constructed an operational map, in which HA1 assays were performed on each mutant virus against the panel of selecting MAbs (Table 2). The MAbs were categorized according to their pattern of inhibition with the escape mutant viruses and were thus divided into four different operational groups. The operational map also allowed us to define the epitopes recognized by the selecting antibodies and to compare the epitopes recognized by MAbs of the IgA class with those recognized by the IgG class. The MAbs that describe operational group A distinguished three overlapping epitopes and contained antibodies of both the IgG class and the IgA (MAb 380) class. Five MAbs (57, 66, 380, 42 and 39) in group A displayed identical reactivity patterns against the panel of escape mutants, suggesting that they recognized the same epitope. In contrast, MAbs 68 and 41 showed a different pattern and MAb 36 displayed a third pattern of reactivity with the virus panel; there was a nonreciprocal relationship between MAb 36 and the other antibodies in group A because the derived mutant virus 36/2A, reacted with the other MAbs in group A. However, because MAb 36 did not react with the other escape mutant viruses in group A, this, together with the functional and sequence analysis of mutant virus 36/2A (Tables 3 and 4), supports the placing of MAb 36 in group A rather than a separate group. Group B included only IgG MAbs but they appeared to recognize epitopes also recognized by MAbs from Group A. For example, except for MAbs 41 and 68, antibodies in group A did not inhibit HA of mutant virus 63/5B. Groups C and D consisted of single MAbs of the IgA class; the epitopes Table 2. HA1 reactions of escape mutant viruses with selecting antibodies Group MAb wt* 68/7B 66/6A 380/6C 42B 39/2C 41/6A 36/2A 63/5B 561/1C 37/7C A <t < < < < < 1600 < < < < < < < < < < < < < 1600 < , 400 < < < < < < 200 < < < < < < < 800 < < < < < < < 1200 < < < < < < < < < < < < < < < B < < C 561~ < 400 D 37~ < 320 < < < < < < * wt, Wild-type Sendai virus. t < Represents a titre of ~ 100. :~ MAbs of the!ga class, Virus

4 820 D. Lyn and others Table 3. Functional analysis of escape mutant viruses HA* NA HN/NP P.f.u. Group Virus (%) (%)t ratio:~ ( 10-7) wt A 57/7A /7B /6A /6C B /2C l/6A /2A B 50/6C 63/5B C 561/1C D 37/7A * HA activity of virus as a percentage of that of wild-type virus using 20 Ixg/ml protein. t Neuraminidase activity of virus as a percentage of that of wild-type virus using 20 ~tg/ml protein. :~ Determined from Fig. 1. wt, Wild-type Sendai virus. they recognized did not overlap operationally with those recognized by MAbs of the IgG class. Biological characterization of the selected escape mutant viruses To characterize the selected escape mutant viruses we performed a functional analysis of their HA and neuraminidase activities. The biological activities of the HN of escape mutant viruses were compared to those of wild-type virus using equivalent amounts of viral protein (Table 3). Mutant viruses in group A had 88 to 100 K wild-type HA activity, but there was a variable reduction in neuraminidase function (20 to 100K of wild-type activity). The most dramatic change in the biological activity of mutant viruses to group A was that of mutant virus 41/6A, which had an 80% reduction in neuraminidase but no reduction in HA activity. In contrast, mutant viruses 50/6C, 63/5B (group B) and 561/1C (group C) had an eightfold reduction in HA as well as a 70 to 80% reduction in neuraminidase activity compared to wildtype virus. In group D, mutant virus 37/7A had a neuraminidase activity increased by 59 % while retaining the wild-type HA function. Analysis of HN content of mutant viruses with reduced HN biological activity The reduced HA and neuraminidase activities of some escape mutants may reflect viruses having a diminished HN content. Alternatively, the mutant viruses may have amino acid substitutions that alter the structure of the HN molecule which subsequently affects the HN biological activity. To differentiate between these possibilities, equivalent amounts of purified virus were electrophoresed by SDS-PAGE (Fig. 1). The HN content of group B (lanes 3 and 4) and group C mutant viruses (lane 5) was significantly reduced compared to that of wild-type virus (lane 1). To quantify the reduction of HN content in the mutant viruses relative to wild-type virus, the Coomassie blue-stained bands in the gel were scanned using whole band analysis (Biolmage analyser; Millipore). The HN content of the mutant viruses was estimated as a ratio of the absorbance of the HN band against that of the nucleocapsid protein (NP). The HN/NP ratio of mutant viruses 63/5B, 50/6C (group B) and 561/1C (group C) was reduced 2.5- to sixfold compared to wild-type virus (Table 3). Other group A and group D mutant viruses had an HN content comparable to wild-type virus (data not shown). The diminished HA and neuraminidase functions of group B and C mutant viruses apparently reflect a reduction in HN. However, the infectivity of mutant viruses in groups B and C was similar to wild-type virus (Table 3), suggesting that the residual biological activity of the mutant viruses was contributed by the remaining HN molecules and that this was sufficient for infectivity. The reduction in the HN content is a unique escape mutant virus phenotype, implying that the amino acid substitutions in these mutant HNs has altered their structure sufficiently to restrict HN incorporation into virions. Although the neuraminidase activity of the group A mutant virus 41/6A was reduced by 80%, it had an HN/NP ratio (0.31) comparable to that of wild-type virus (Fig. 1, lane 2 and Table 3) and the same HA activity. Thus, the diminished neuraminidase activity observed in mutant virus 41/6A was similar to that found in mutant viruses of groups B and C, but this could not be attributed to a reduced HN content of the virus (Fig. 1 and Table 3). Mutant virus 41/6A may therefore have an amino acid substitution that alters the HN molecule sufficiently to change the enzymic activity. Nucleotide sequence analysis of escape mutant viruses To relate the specific changes in biological function to the primary amino acid sequence, the entire HN gene of each derived mutant virus was sequenced (Table 4). In group A, mutant viruses selected with five different MAbs (57, 68, 66, 380 and 42) had amino acid changes at residue 279, of arginine to either glycine, serine or isoleucine. These mutant viruses had a neuraminidase activity of 50 to 70 % of that of the wild-type virus (Table 3) and amino acid substitutions which resulted in a change of a positive charge to neutrality. This group of MAbs included both IgA (380) and IgG antibodies

5 HN of Sendai virus 821 t -- ~,~ Table 4. Nucleotide sequence analysis of escape mutant viruses Amino acid change Group Virus Codon change (position/change) A 57/7A AGA~AGC 279/R~S 68/7B AGA~AGC 279/R~S 66/6A AGA~GGA 279/R--,G 380/6C AGA~GGA 279/R~G 42B AGA~ATA 279/R~I 39/2C GAC-~AAC 277/D~N 41/6A TCT~CCT 184/S~P 36/2A ACC~ATC 280/T~I B 50/6C AGA~GGA 279/R~G AAA~GAA 461/K-,E 63/5B AGA-,GGA 279/R~G AAA-GAA 461/K~E C 561/1C GGA~GAA 375/G~E D 37/7A CCT-,TCT 451 / P~S Fig ~ SDS-PAGE showing protein composition of wild-type and escape mutant viruses. Each lane represents 15 ~tg of virion protein visualized by Coomassie blue staining. The viruses shown in lanes 1 to 5 are wild-type, 41/6A (group A), 50/6C, 63/5B (group B) and 561/1C (group C), respectively. which recognized the same amino acid residue. MAb 39 selected an escape mutant virus which had a mutation which changed aspartic acid to asparagine at position 277. Although the amino acid change resulted in the substitution of a positive with a negative charge, this mutant virus had only a 20 ~o reduction in neuraminidase activity compared to that of the wild-type virus. Mutant virus 36/2A had the same HN biological activities as wild-type virus and a change at residue 280 of threonine for isoleucine. Sequence analysis of mutant virus 41/6A identified a serine to proline substitution at residue 184. Since mutant virus 41/6A was deficient only in neuraminidase activity, which was independent of viral HN content, this substitution may be essential for neuraminidase function (Table 3). It also shows that despite a substantial reduction in neuraminidase activity (80~), HA was not affected, suggesting that the sites for cell binding and neuraminidase are not identical. Position 184 is about a hundred amino acids away from the changes found in mutant viruses selected with the other MAbs in group A (Table 4), indicating that antibodies in this group recognize a discontinuous epitope. It is therefore likely that the amino acid residues at positions 184 and 279 are situated close to each other on the folded HN. The group B mutant viruses 50/6C and 63/5B have two amino acid changes at positions 279 and 461. Mutant viruses with a single amino acid change at position 279 (group A) had the same HA activity and HN content as wild-type virus. In contrast, other mutant viruses with a single amino acid substitution at residue 461 had an increased neuraminidase activity that was not correlated with a change in HN content (Gorman et al., 1991). The presence of two mutations may, therefore, concomitantly contribute to the lowered HN content (Fig. 1 and Table 3). In addition, the amino acid substitutions resulted in the loss of two positive charges and the gain of a negative charge, which could cause conformational changes that affect the incorporation of HN into the virion. In group C, mutant virus 561/1C had a glycine to glutamic acid change at position 375. Since mutants in groups B and C had a loss of biological activity that correlated with a loss of HN content, positions 279 and 461 together, and residue 375, may be important for folding the molecule into a form which is resistant to proteolytic degradation. Mutant virus 37/7A (group D) had an amino acid substitution of proline to serine at position 451. This amino acid mutation was implicated in fusion because the selecting MAb, 37, did not prevent virus binding or neuraminidase activities but inhibited haemolysis activity. Also, groups C and D only included IgA MAbs which selected mutations at sites distinct (residues 375 and 451) from those detected by IgG MAbs (Table 4). Therefore, the IgA MAbs identified neutralization epitopes that have not been described with IgG MAbs. Whether these residues are unique to IgA MAbs remains to be determined by analysis of additional antibodies. A comparison of the operational map in this report and a previous topological map which delineated four (I to IV) non-overlapping antigenic sites (Portner, 1984; Portner et al., 1987) by competitive binding assays has

6 822 D. Lyn and others provided further understanding of the structure and function of the HN molecule. Antibodies that define groups A and B in this study apparently recognize the same overlapping epitopes previously defined as site I by MAbs used to generate the topological map (Thompson & Portner, 1987) because the selected escape mutant virus HNs from both studies had mutations at positions 184, 277 and 279. In addition, sequence analysis of an escape mutant virus HN from group B identified a new mutation at amino acid residue 461, suggesting that antigenic site I encompasses a larger area on the primary sequence than was previously thought. The group C MAb identified a mutation at amino acid residue 375, a position not previously identified on the topological map by competitive binding. However, based on the MAb inhibition data (both HA and neuraminidase activities are inhibited by MAbs that define group C and antibodies that define antigenic site I), we tentatively place group C, and thus residue 375, topologically in antigenic site I. Antigenic site II, previously shown to be involved in fusion (Portner et al., 1987) and subsequently mapped to amino acid residue 420 (Thompson & Portner, 1987), appears to be the domain that the group D MAb recognizes on the HN molecule. The group D MAb also inhibits fusion and identifies a mutation at residue 451, suggesting that this position and position 420 are located close to each other in the folded HN structure. Thus, previous findings, along with those of this study, provide evidence that the cell binding and neuraminidase sites are complex and are contributed by different areas of the HN molecule brought together by protein folding. Discussion The purpose of this study was to identify amino acids important for the structure and function of the HN molecule. By comparing the biological characteristics of escape mutant viruses to changes in the HN content and the primary amino acid sequence, we can propose areas of the molecule which may be structurally important for the HN functions. Of the group A mutant viruses, 41/6A had the most significant reduction (80%) in neuraminidase activity, which was not attributed to a diminished HN content. The mutation mapped to amino acid residue 184, which suggests that this position may be located close to the neuraminidase site. Iorio et al. (1989) have suggested a domain in Newcastle disease virus between positions 171 and 178 that is essential for neuraminidase activity. This region corresponds to amino acids 191 to 198 of Sendai virus, close to the change at position 184. A sialidase-deficient mutant of mumps virus had an amino acid change at residue 181 (Waxham & Arnowski, 1988), which aligns with position 195 of Sendai virus. By comparing five different paramyxoviruses, a conserved hexapeptide region (NRKSCS) has been proposed by Jorgenson et al. (1987) as the neuraminidase site. In Sendai virus this region aligns with positions 254 to 259 and we propose that site 184 may be located close to this area in the threedimensional structure. No escape mutant viruses have been selected in this region (positions 254 to 259), but group A MAbs recognized areas in the vicinity of the proposed site. This is consistent with data describing the localization of antigenic sites to areas surrounding, but not including, the sialidase site of the influenza A virus neuraminidase glycoprotein (Colman et al., 1983). Mutant viruses selected by the other group A antibodies had closely spaced amino acid changes at positions 277, 279 and 280, in which the position and nature of the amino acid mutation appeared to influence the neuraminidase activity. Thompson & Portner (1987) selected escape mutants of Sendai virus which also had amino acid mutations at residues 277 and 279. Escape mutants have also been isolated for human parainfluenza virus type 3 (van Wyke Coelingh et al., 1987) with mutations which align well with this region in Sendai virus. This region may therefore represent a highly immunodominant area of the HN molecule which is recognized by both IgA and IgG antibodies (Tables 2 and 4). Mutant viruses identified by MAbs in group B (63/5B and 50/6C) and group C (561/1C) have amino acid mutations which appear to be responsible for structural alterations of the HN molecule. The significant reduction in the HN content of these mutants is apparently responsible for the decrease in HA and neuraminidase activities. The amino acid changes in these mutants may disrupt the normal transport to the cell surface, and processing of, the HN protein or the mutations may result in aberrant folding of the HN molecule, which subsequently leads to proteolytic degradation. Sequence analysis of a temperature-sensitive (ts271) mutant of Sendai virus identified three mutations at positions 262, 264 and 461. Interestingly, the HN of this ts271 mutant was rapidly degraded at restrictive temperatures (Tuftereau et al., 1985). Mutant viruses in group B have a substitution at residue 279, close to the changes at positions 262 and 264, as well as a mutation at position 461. Although these mutant viruses are not temperaturesensitive (data not shown), the mutations at these positions result in a reduced HN content in the virus, as did the change at residue 375 (group C). Thus, mutations around position 262, at position 375, and positions 279 and 461 together may be critical for the incorporation of HN into virions. The lysine change at position 461 is located between a conserved area of cysteine, proline and

7 HN of Sendai virus 823 glycine residues, and the glycine substitution at position 375 (group C) is partially conserved among the paramyxoviruses, suggesting that these regions are structurally important. The MAb in group D inhibited haemolysin activity but did not block HA or neuraminidase activity. The mutant virus selected with this MAb had a mutation at position 451. The involvement of HN with fusion is poorly understood, but Portner et al. (1987) have suggested that cell binding may induce a conformational change in HN necessary for F-mediated fusion activity. Thompson & Portner (1987) have identified position 420 in Sendai virus HN as being involved with fusion. This mutation is close to the change at position 451, suggesting that this region of the molecule is critical for the fusion-related function. The MAbs (IgA and IgG) used to select escape mutant viruses were divided into four operational groups, in which antibodies from groups A, B and C inhibited HA and neuraminidase activities whereas the MAb in group D inhibited haemolysin activity. The four groups defined by our operational map correlated with the relative positions of the amino acid substitutions in the selected mutant viruses. For example, in group A MAbs 57, 66, 380 and 42 exhibited the same reactivity pattern (Table 2) and the selected mutant viruses had amino acid changes at position 279. However, we cannot exclude the remote possibility that a mutation in some other gene besides HN was responsible for the selection and/or the phenotype of the neutralization escape mutants. Antibodies of the IRA class appear to be important in resistance to paramyxovirus reinfections and therefore the epitopes recognized by IRA MAbs may identify important regions for the development of host immunity. One of the IgA antibodies (380) recognized an epitope which overlapped that recognized by the IgG MAbs of group A (Tables 2 and 4). Antibodies from groups C and D identified amino acid residues (375 and 451) which appear to be restricted to MAbs of the IgA class because the panel of IgG MAbs used in this study (Table 4) and previous studies (Thompson & Portner, 1987) did not select mutant viruses with amino acid changes at these positions. The epitopes recognized by IRA MAbs may identify sites on the HN molecule important for the secretory immune response and the difference in specificity could reflect differences in antigen processing after mucosal and systemic immunization. Ms Ruth Ann Scroggs provided skilful technical assistance in the preparation and screening of the hybridomas of the IgG class. Synthetic oligonucleotides were provided by the Molecular Resource Center of St Jude Children's Hospital (CORE Grant CA-21765). This work was supported by the National Institute of Health Research Grants AI-11949, HL and HL-02002, the American Lebanese Syrian Charities of St Jude Children's Research Hospital and a Parker B. Francis Fellowship Award to M.B.M. References AYMARD-HENRY, M. T., COLEMAN, M. T., DOWDLE, W. R., LAVER, W. G., SCHILD, G. C. & WEBSTER, R. G. (1973). Influenza virus neuraminidase and neuraminidase-inhibition test procedures. Bulletin of the World Health Organization 48, BEAN, W. J., SRIRAM, G. & WEBSTER, R. G. (1980). Electrophoretic analysis of iodine-labeled influenza virus RNA segments. Analytical Biochemistry 102, CHANOCK, R. M. & PARROTT, R. H. (1965). Acute respiratory disease in infancy and childhood: present understanding and prospects for prevention. Pediatrics 36, COLMAN, P. M., VARGHESE, J. N. & LAVER, W. G. (1983). Structure of the catalytic and antigenic sites in influenza virus neuraminidase. Nature, London 303, GLEZEN, W. P., LODA, F. A., CLYDE, W. A., JR, SENIOR, R. J., SHEAFER, C. I., CONLEY, W. G. & DENNY, F. W. (1971). Epidemiologic patterns of acute lower respiratory disease of children in a pediatric group practice. Journal of Pediatrics 78, GOR~N, W. L., GILL, D. S., SCROGGS, R. A. & PORTNER, A. (1990). The hemagglutinin-neuraminidase glycoproteins of human parainfluenza virus type 1 and Sendal virus have high structure-function similarity with limited antigenic cross-reactivity. Virology 175, GORMAN, W. L., TAKAHASH, T., SCROGGS, R. A. & PORTNER, A. (1991). Identification of amino acid positions associated with neuraminidase activity of the hemagglutinin-neuraminidase glycoproteins of Sendal virus. Virology 180, HEATH, T. D., MARTIN, F. J. & MACHER, B. A. (1983). Association of ganglioside-protein conjugates into cell and Sendal virus: requirement for the HN subunit in viral fusion. Experimental Celt Research 149, HOMMA, M. & OHUCHI, M. (1973). Trypsin action on the growth of Sendal virus in tissue-culture cells. Structural difference of Sendal virus grown in eggs and tissue culture cells. Journal of Virology 12, IORIO, R. M., SYDDALL, R. J., GLICKMAN, R. I., RIEL, A. M., SHEEHAN, J. P. & BRATT, M. A. (1989). Identification of amino acid residues important to the neuraminidase activity of the HN glycoprotein of Newcastle disease virus. Virology 173, 19(~204. JORGENSEN, E. D., COLLINS, P. L. & LOMEDICO, P. T. (1987). Cloning and nucleotide sequence of Newcastle disease virus hemagglutininneuraminidase mrna: identification of a putative sialic acid binding site. Virology 156, 12~4. LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, London 227, MAZANEC, M. B., NEDRUD, J. G. & LAMM, M. E. (1987). Immunoglobulin A monoclonal antibodies protect against Sendai virus. Journal of Virology 61, MERZ, D. C. & WOLINSgY, J. S. (1983). Conversion of non-fusing mumps virus infections to fusing infections by selective proteolysis of the HN glycoprotein. Virology 131, MILLS, J., VANKIRK, J. E., WRXGrrr, P. F. & CnANOCK, R. M. (1971). Experimental respiratory syncytial virus infection of adults: possible mechanisms of resistance to infection and illness. Journal of Immunology 107, MIURA, N., UCHIDA, T. & OKADA, Y. (1982). HVJ (Sendai virus)- induced envelope fusion and cell fusion are blocked by monoclonal anti-hn protein antibody that does not inhibit hemagglutination activity of HVJ. Experimental Cell Research 141, NAGAI, Y. & KLENK, H.-D. (1977). Activation of precursors to both glycoproteins of NDV by proteolytic cleavage. Virology 77, NAGAI, Y., KLENK, H.-D. & ROTT, R. (1976). Proteolytic cleavage of the viral glycoproteins and its significance for the virulence of Newcastle disease virus. Virology 72, ORVELL, C. & GRANDIEN, M. (1982). The effects of monoclonal antibodies on biological activities of structural proteins of Sendal virus. Journal of Immunology 129, OZAWA, M., ASANO, A. & OKADA, Y. (1979). Biological activities of glycoproteins of HVJ (Sendal virus) studied by reconstitution of

8 824 D. Lyn and others hybrid envelope and by concanavalin A-mediated binding: a new function of HANA protein and structural requirement for F protein in hemolysis. Virology 99, PORTNER, A. (1981). The HN glycoprotein of Sendai virus: analysis of site(s) involved in hemagglutinating and neuraminidase activities. Virology 115, PORTNER, A. (1984). Monoclonal antibodies as probes of the antigenic structure and functions of Sendai virus glycoproteins. In Nonsegmented Negative Strand Viruses, pp Edited by D. H. L. Bishop & R. W. Compans. Orlando: Academic Press. PORTNER, A., SCROGGS, R. A. & METZGER, D. W. (1987). Distinct functions of antigenic sites of the HN glycoprotein of Sendal virus. Virology 158, SANGER, F., NICKLEN, S. & COULSON, A. R. (1977)DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences, U.S.A. 74, SCHEID, A. & CItOPPIN, P. W. (1974). Identification of biological activities of paramyxovirus glycoproteins: activation of cell fusion, hemolysis, and infectivity by proteolytic cleavage of an inactive precursor protein of Sendai virus. Virology 57, SCHEID, A., CALIGUIRI, L. A., COMPANS, R. W. & CHOPPIN, P. W. (1972). Isolation of paramyxovirus glycoproteins: association of both hemagglutination and neuraminidase activities with the larger SV5 glycoprotein. Virology 50, SMITH, C. B., PURCELL, R. H., BELLANTI, M. D. & CHANOCK, R. M. (1966). Protective effect of antibody to parainfluenza type 1 virus. New England Journal of Medicine 275, THOMPSON, S. D. & PORTNER, A. (1987). Localization of functional sites on the hemagglutinin-neuraminidase glycoprotein of Sendai virus by sequence analysis of antigenic and temperature-sensitive mutants. Virology 160, 1-8. TOMASI, T. B., JR (1983): Mechanism of immune regulation at mucosal surfaces. Reviews of Infectious Diseases 5, $784-$792. TUFFEREAU, C., PORTNER, A. & ROUX, L. (1985). The role of haemagglutinin-neuraminidase glycoprotein cell surface expression in the survival of Sendal virus-infected BHK-21 cells. Journal of General Virology 66, UCmDA, T., UEDA, S., NAKANISm, M., MIURA, N. & OKADA, Y. (1984). HN glycoprotein of HVJ (Sendai virus) enhances the selective cytotoxicity of diphtheria toxin fragment A-containing liposomes on subacute sclerosing panencephalitis virus-infected cells. Experimental Cell Research 152, VAN WYKE, K. L., BEAN, W. J., JR & WEBSTER, R. G. (1981). Monoclonal antibodies to the influenza A virus nucleoprotein affecting RNA transcription. Journal of Virology 39, VAN WYKE COELINGH, K. L., WINTER, C. C., JORGENSEN, E. n. & MURPHY, B. R. (1987). Antigenic and structural properties of the hemagglutinin-neuraminidase glycoprotein of human parainfluenza type 3: sequence analysis of variants selected with monoclonal antibodies which inhibit infectivity, hemagglutination, and neuraminidase activities. Journal of Virology 61, WAXHAM, M. N. & ARONOWSKI, J. (1988). Identification of amino acids involved in the siatidase activity of the mumps virus hemagglutinin-neuraminidase protein. Virology 167, (Received 21 September 1990; Accepted 2 January 1991)