Antigenic Mapping of an Avian HI Influenza Virus Haemagglutinin and Interrelationships of HI Viruses from Humans, Pigs and Birds

Transcription

1 J. gen. Virol. (1986), 67, Printed in Great Britain 983 Key words: influenza A viruses ( H l ) / haemagglutinin/ monoclonal antibodies Antigenic Mapping of an Avian HI Influenza Virus Haemagglutinin and Interrelationships of HI Viruses from Humans, Pigs and Birds By FRANCIS J. AUSTIN 1. AND ROBERT G. WEBSTER 2 1 Virus Research Unit, The Medical Research Council of New Zealand, University of Otago, P.O. Box 56, Dunedin, New Zealand and 2Department of Virology and Molecular Biology, St. Jude Children's Research Hospital, 332 North Lauderdale, P.O. Box 318, Memphis, Tennessee 38101, U.S.A. (Accepted 19 February 1986) SUMMARY Monoclonal antibodies to the haemagglutinin (HA) of the avian H1 influenza virus A/duck/Alberta/35/76 were used to construct an operational antigenic map of the HA molecule and to study the interrelationships of H1 viruses from different hosts. Haemagglutination inhibition tests between the monoclonal antibodies and variants selected by them provided evidence of four antigenic regions which overlap to varying degrees. Avian H1 influenza viruses displayed a spectrum of reactivities to the monoclonal antibody panel. Representatives of the epidemic strains of human H1 influenza viruses and early swine influenza viruses showed little or no reactivity with the monoclonal antibodies but swine influenza-like viruses isolated from pigs and humans in the last decade reacted with 11 of 17 antibodies. The antigenic similarity of these viruses to many avian isolates suggests that there has been a transfer of HA genetic information between mammalian and avian HI influenza viruses. INTRODUCTION Influenza A viruses of the H1N1 subtype are of interest for they represent one of the few subtypes that has been documented to spread from pigs to humans (Hinshaw et al., 1978). Influenza viruses of this subtype infect humans, swine and birds (Easterday, 1975; Webster et al., 1984) and one was probably responsible for the 1918 to 1919 Spanish influenza epidemic that resulted in at least 20 million deaths worldwide. The sensitivity of the scientific community to swine influenza was illustrated in 1976 when a national vaccine programme was initiated in the U.S.A. after a swine-like influenza virus was isolated from a military recruit (Neustadt & Fineberg, 1978). Our renewed interest in these viruses arises from genetic and antigenic analyses (Scholtissek et al., 1983; Hinshaw et al., 1984) indicating that some of the swine influenza in Europe may have originated from H1N 1 influenza viruses in avian species. This subtype has been shown to spread from pigs to domestic turkeys and to humans (Hinshaw et al., 1983), emphasizing the ability of H 1N 1 viruses to spread between species. Some of the genetic information in the vast pool of influenza virus genes present in aquatic birds may gain access to humans through pigs. In order to investigate further the antigenic relationships between the H1 influenza viruses from birds and mammals, monoclonal antibodies were prepared to the haemagglutinin (HA) of the prototype avian H1 influenza virus [A/duck/Alberta/35/76, H1N1 (dk/alberta)]. Earlier analysis with a small panel of monoclonal antibodies to A/N J/11/76 (X-53A) H1N 1 (Hinshaw et al., 1983, 1984) showed that although the antibodies reacted with two turkey (ty) isolates which were genetically, antigenically and biologically similar to swine (sw) isolates, they did not react with three duck isolates, including dk/alberta. This suggests that the number of epitopes shared between avian and mammalian H1N1 viruses is limited. Extensive antigenic analysis has been done on the human H1N1 viruses A/PR/8/34 and A/FM/1/47 (Gerhard et al., 1981 ; Caton et al., 1982; Raymond et al., 1983) showing similarity in SGM

2 984 F. I. AUSTIN AND R. G. WEBSTER the number and location of antigenic sites on H1 and H3 influenza viruses. Similar studies have not been done on avian H 1 viruses. The available evidence suggests that antigenic variation may be different in avian strains and that antigenic drift as we know it in human influenza viruses may not occur. Instead, the mutant and parent viruses may co-exist. Analysis of H1N1 viruses with monoclonal antibodies to the HA of dk/alberta provides evidence for the co-circulation of variants of avian influenza viruses, and that at least one of the four antigenic sites on avian H1 viruses is shared with swine influenza-like strains from humans and pigs. METHODS Viruses and disruption. The majority of the H 1 influenza viruses were from the repository at St. Jude Children's Research Hospital [including isolates from New Jersey (N J), Wisconsin (WIS), Iowa (IA), New York (NY), Tennessee (TN), Oklahoma (OK), Michigan (MICH), Missouri (MO), Nevada (NEV), Kansas (KS), Memphis (MEM), Cambridge, Mass., U.S.A. (CAM) and Bavaria (Bav)]; some were isolations from wild ducks in New Zealand (NZ) and Australia (Vict) (Tables 2 to 4). The viruses were grown in embryonated chicken eggs and were purified by adsorption to and elution from chicken erythrocytes followed by differential centrifugation and sedimentation through a sucrose gradient (10 to 40~ sucrose, 0.15 M-NaC1) (Laver, 1969). Beta-propiolactone inactivation of antigens, which was necessary before their international shipment, prevented haemagglutinationinhibition (HI) reactions between some antigenic determinants and their antibodies. The reactivity could be restored by disrupting the antigens with Tween-ether (Kida et al., 1982). Serological tests. HA titrations and HI tests were performed in microtitre plates with receptor-destroying enzyme-treated sera (Palmer et al., 1975). Neuraminidase titrations and neuraminidase inhibition (NI) tests were done by the procedure of Aymard-Henry et al. (1973). The method for indirect ELISA was that used by Kida et al. (1982). Plates were coated with 200 haemagglutinating units (HAU)/well of disrupted purified virus as antigen. Monoclonal antibodies and polyclonal antisera. BALB/c mice were injected intraperitoneauy with 5000 HAU of ether-disrupted dk/alberta virus as described (Kida et al., 1982). After 1 to 2 months, a booster dose of the same antigen was injected intravenously and fusion was carried out 4 days later. Monoclonal antibodies to the HA of dk/alberta were prepared by the method of K6hler & Milstein (1976) using SP2/O-Ag-14 cells (Shulman et al., 1978). The hybridoma cells were screened for antibody production by HI, NI and ELISA tests. Those producing monoclonal antibodies were cloned in soft agar and grown as ascites in BALB/c mice. Polyclonal antisera were prepared by immunizing goats with purified HA subunits. Antigenic variants. Single-step antigenic variants were selected as described previously (Gerhard & Webster, 1978). Briefly, monoclonal hybridoma antibody plus cloned parent virus were incubated together for 30 min at 20 C. This mixture was inoculated into ll-day-old chick embryos. The viruses that grew in the presence of monoclonal antibodies were cloned twice in Madin Darby canine kidney (MDCK) cell cultures (Tobita et at., 1975) either by limiting dilutions in microtitre trays or by plaquing. RESULTS Operational mapping of the dk/alberta HA To determine the number of distinct antigenic sites on the dk/alberta HA molecule we selected antigenic variants with a panel of 17 monoclonal antibodies. These variants were tested for reactivity with the monoclonal antibodies in HI assays. Antigenic sites can be operationally defined in such an analysis by the ability of the antigenic changes in a certain group of variants to prevent the binding of a corresponding group of monoclonal antibodies, but at the same time to have no effect on the binding of other antibodies which recognize different antigenic sites on the same molecule. At least four different reactivity patterns were discernible from this analysis (Table 1) but some overlap existed between the epitopes defined by each reactivity group. The clearest resolution was between groups I and III. The epitope defined by monoclonal antibody no. 53, which gave a different reactivity pattern from the rest, formed group IV and showed overlap with group III. All or most of the monoclonal antibodies in group I recognized antigenic determinants on recent H1 viruses from swine and on human strains that had originated from swine. The epitope recognized by monoclonal antibody no. 53 appeared to be absent from the recent human and swine isolates, but was present on the classical swine influenza virus from 1930 (see Table 3).

3 Table 1. Reactivity patterns of antigenic variants of dk/alberta selected with monoclonal antibodies t-,i OO Monoclonal Group antibody I II III IV 53 * Monoclonal antibody HI titre against variant compared with titre against parental eightfold reduction; - represents over eightfold reduction. Antigenic variants selected with monoclonal antibodies no _, ± _ ± ± _ _ ± _ _ - _ _ - _ ± _ t ± _.... _..... ± ± ±_.... virus: represents zero- to twofold reduction; represents four- to

4 Pooled monoclonal antibodies against dk/alberta > < 100 < > "4 Z t3,-4 Human Swine Avian Virus strain* Table 2. PR/8/34 (HI N 1) FM/I/47 (HINI) N J/8/76 (H 1N 1) WIS/263/76 (H 1N 1) USSR/90/77 (H 1N 1) England/333/80 (H t N 1) Chile/l/83 (H 1N 1) sw/la/15/30 (H 1N 1 ) sw/wis/l/67 (H 1N 1) sw/tn/3/77 (H 1N 1) dk/alberta (H 1N 1 ) mall/ok/7/77 (H 1 N 1 ) pint dk/ok/8/77 (H1NI) bl dk/mich/235/77 (HIN3) dk/mich/77 (H 1N l) mall/n Y/6861/78 (H 1NS) dk/nz/160/76 (HI N3) dk/n Z/176/76 (H 1 N 3) Analysis of HI influenza A viruses with polyclonal sera in HI tests Polyclonal antisera against ( PR/8/34 FM/I/47 sw/ia/15/ ~ <20 < < < < < < < < 20 < < 20 < < 20 < 20 <20 < 20 < 20 < 20 < 20 < 20 dk/alberta O * Intact virus was used as antigen in HI tests. 1" HI titre; homologous reactions are in bold type.

5 Avian H1 influenza viruses 987 The HI influenza viruses The H1 influenza viruses occurred in humans from 1918 to 1957 and from 1977 to the present. Additionally, these viruses occur in pigs and birds. The extent of cross-reaction between these viruses was examined by HI assays with polyclonal antisera (Table 2). Antisera to the human and swine H1 viruses showed antigenic relationships between A/PR/8/34, A/FM/I/47 and A/sw/IA/15/30, but there was only minimal cross-reaction with the avian (mallard, pintail, black duck and duck) H1 viruses. Polyclonal antiserum to an avian H1 virus (or pooled monoclonal antibodies to dk/alberta), on the other hand, showed cross-reactions between H1 viruses from human, swine and avian sources. The results confirm that H1 viruses from humans, pigs and birds are members of the same subtype, and this is most clearly illustrated with an antiserum to an avian strain (dk/alberta). These results indicate asymmetrical cross-reactions between the HAs of mammalian and avian H1 influenza viruses. The similarity in crossreaction obtained with the dk/alberta polyclonal antiserum and the monoclonal antibody pool suggests that the majority of epitopes on this avian H1 virus were recognized by the individual monoclonal antibodies. However monoclonal antibodies that recognized epitopes on A/USSR/90/77 and A/England/333/80 were missing from the panel. Analysis of ill influenza viruses with monoclonal antibodies Analyses of the human, swine and avian H1 influenza viruses with monoclonal antibody to the HA ofdk/alberta were done to define the interrelationships of these viruses. The panel of 17 monoclonal antibodies to the HA of dk/alberta were all different as judged by their interaction with naturally occurring variants (Tables 3 and 4) or with laboratory-selected variants (Table 1). Analysis of the human prototype H1 viruses with this panel of monoclonal antibodies revealed that only two antibodies, both in group II, reacted with A/PR/8/34 and none reacted with A/FM/1/47 or A/USSR/90/77 (Table 3). Similarly, the classical Shope swine influenza virus A/sw/IA/15/30 reacted with only two (no. 88 and no. 53; groups III and IV) of the 17 monoclonal antibodies, and another early swine strain, A/sw/Cambridge/39, did not react with any. H 1N 1 influenza viruses isolated from pigs in 1967, 1975 and 1980 reacted with more than half of the monoclonal antibodies in the panel representing groups I, II and III. The reactions of the 1967 isolate differed from those of the three viruses isolated later. The reactivities of the swine influenza-like viruses isolated from humans in 1976 and 1982 were essentially identical with those which were circulating in pigs at that time (e.g. A/sw/WIS/8/80) (Table 3). An exception was the A/N J/11/76 reassortant X-53A which was selected with antiserum to A/sw/Cambridge/39 (Kilbourne, 1978). It was significantly different antigenically from the A/N J/8/76 parental virus and reacted with only one monoclonal antibody (group III). Analysis of avian and recent swine influenza-like H 1 influenza viruses from humans, pigs and birds leaves no doubt that these viruses are related antigenically (Table 4). The viruses could be divided into groups based on their reactivity patterns. Viruses isolated from wild ducks in North America in 1976 to 1978 which reacted with all or most of the antibodies in groups I, II and IV made up group A. Group 13 contained the H1 viruses from mammals (humans and pigs) and turkeys. They reacted with most of the group I antibodies, some in group II and monoclonal antibody no. 88 in group III. The turkey isolates in this group (A/ty/KS/4880/80 and A/ty/MO/1/81) have been shown to originate probably from pigs (Hinshaw et al., 1983). The remaining group contains H 1 viruses isolated from wild ducks in Australia (Victoria), Germany (Bavaria), New Zealand and North America. They reacted with varying numbers of groups I, II and III monoclonal antibodies. The similarity of the monoclonal antibody reactivity patterns of the viruses isolated from humans, pigs and turkeys (three different patterns among l0 viruses) contrasts with the diversity of the reactions of the wild duck viruses (10 patterns among 13 viruses).

6 88 53 oo oo o,..] Z,--t r/l Virus dk/alb/35/76 PR/8/34 FM/I/47 USSR/90/77 NJ/8/76 NJ/I 1/76 X-53 NJ/I 1/76 X-53A WIS/263/76 MEM/4/82 NEV/101/82 sw/ia/15/30 sw/cam/39 sw/wls/l/67 sw/tn/l/75 sw/w1s/8/80 sw/wls/629/80 Table 3. Reactions of swine and human HIN1 influenza A viruses with monoclonal antibodies to dk/alberta I Monoclonal antibody group and number ~k * > <t < < < < < < < 2 < < < < < < < < < < < < < < < < < < > > > > > < < < < < < < < < > > < < < < < < < < < < < < < < < < < < < < 32 4 < > > > 128 > # II III IV 4 > < < < < < < < < < 16 < 2 8 < < 4 < < 2 < 2 24 < 2 4 < < 2 4 < < < > < < 2 < < 32 < < 24 < *HItitre x " <, HI titre less than 200.

7 III 88 > 128 IV tx t,,i Group A B C Virus isolate* dk/alb/35/76 (H 1N 1) mall/ok/7/77 (H 1 N 1) pint dk/ok/8/77 (H 1N 1) bl dk/mich/235/77 (HIN3) dk/mich/77 (H 1 N 1) mall/ny/6861/78 (HIN5) ty/ks/4880/80 (H 1 N 1) ty/mo/l/81 (H 1 N 1) sw/wis/629/80 (H 1N 1) NEV/101/82 (HINI) NJ/8/76 (H 1 N 1) WIS/263/76 (H 1N 1) MEM/4/82 (H 1N 1) sw/tn/3/77 (H 1N 1) sw/wls/8/80 (H 1N 1) sw/wls/1/67 (H 1N 1) dk/vict/2/80 (H 1N l) dk/bav/2/77 (H l N 1) dk/nz/160/76 (H 1 N3) dk/nz/176/76 (HIN3) dk/vict/23/81 (H 1 N?) mali/alb/10/76 (HINI) mall/alb/285/83 (HINI) Table 4. Analysis of swine-like HI influenza A viruses with monoclonal antibodies to the HA Reactivity in HI tests with monoclonal antibody (group and number) A t ~" >128 >128 >128 > ~ * Virus antigens were ether-disrupted. thititre x 10 -'. represents HI titre greater than or equal to 200; no entry means HI titre less than 200. II A A > 128

8 990 F. J. AUSTIN AND R. G. WEBSTER DISCUSSION The purpose of this study was to establish an antigenic map of the HA of an avian H1 influenza virus and to determine the antigenic interrelationships of the H 1 viruses from humans, pigs and birds. It also provides the opportunity to examine antigenic drift in the same influenza virus subtype from different hosts. Antigenic mapping of the human H1 viruses A/PR/8/34 (Gerhard et al., 1981 ; Caton et al., 1983) and A/FM/1/47 and the related A/USSR/90/77 strain (Kendal et al., 1978; Nakajima et al., 1983; Raymond et al., 1983) revealed that the globular head of the HA molecule has four antigenic regions which correspond with those found on the H3 subtype (Webster & Laver, 1980; Wiley et al., 1981). Operational antigenic mapping of the HA of dk/alberta showed that it also has four antigenic regions in which the antigenic determinants were recognized by eight, five, three and one monoclonal antibodies respectively (Table 1). Studies are in progress to locate the four antigenic sites on the primary amino acid sequence of the molecule. Antigenic analysis of the avian H 1 viruses with the panel of monoclonal antibodies showed a spectrum of activity ranging from dk/alberta, which by definition reacted with all 17 monoclonal antibodies, through some which reacted with only five antibodies, to some which were not inhibited by any (data not shown). Most of the variation between isolates occurs in the antigenic determinants which are recognized by groups II, III and IV antibodies. Some of the heterogeneity is associated with the geographical location from which the virus was obtained, but even in a single area, such as Alberta in North America and Victoria in Australia, antigenic variants were co-circulating. This could arise by antigenic drift and cocirculation of stable antigenic variants. By contrast, all of the swine H 1 isolates from pigs and humans fall clearly into two categories, those which react minimally or not at all with the dk/alberta monoclonal antibodies, and those which react with all or most of the group I antibodies and some in groups II and III. The first category comprises two early isolates (A/sw/IA/15/30 and A/sw/Cambridge/39) and the X-53A variant of A/N J/11/76. The second category contains swine influenza-like H1 viruses which were isolated from humans and pigs between 1967 and 1982 (Table 3). The recent isolates differ from the earlier ones in at least three of the four antigenic sites. The antigenic similarities among the HAs of swine influenza-like viruses isolated from different species and from different geographical areas suggest that they have a common origin. Two possible ways in which this could have occurred are independent parallel antigenic drift of the H1 HA in different species or transfer of virus genes between species. While independent antigenic drift cannot be excluded, the probability of it occurring and producing such similar antigenic spectra is not high. On the other hand interspecies transmission of swine influenza virus to turkeys and humans has occurred naturally (Hinshaw et al., 1983) and transmission of duck influenza viruses to pigs has been achieved experimentally (Hinshaw et al., 1981). Also the H 1 virus which is currently causing an epidemic of swine influenza in Europe contains genes of avian H1 virus origin (Scholtissek, et al., 19~3). The two co-circulating antigenically and biologically distinguishable variants of the swine influenza-like isolate A/NJ/11/76 (Kilbourne, 1978) reacted in different ways with the monoclonal antibodies. The reassortant X-53 reacted with 11 antibodies representing three antigenic regions, but X-53A reacted only with antibody no. 88, which also reacts with A/sw/IA/15/30. This suggests that although a variant with avian influenza virus HA antigenic determinants is predominant in the wild-type A/N J/11/76 virus, another HA variant which is antigenically similar to the prototype swine influenza virus was co-circulating with it. One possible explanation of these observations would be that between 1930 and 1967 either an avian HI influenza virus was transmitted to pigs or the HA gene of an avian H1 virus was introduced into swine H 1 influenza viruses. The antigenic differences between A/sw/WIS/1/67 and the post-1975 isolates as exemplified by A/N J/8/76 could be due either to subsequent antigenic drift in pigs or to the transfer of avian H 1 influenza virus HA genetic information occurring more than once. The finding that the antigenic reactivities of the HA of swine influenza-like viruses isolated from pigs and humans during the last decade were virtually identical compared with the variable

9 Avian H1 influenza Viruses 991 reactivities of the avian isolates may be relevant to understanding the mechanisms by which genetic information is transmitted between influenza viruses that infect different host species. One might speculate that mammalian viruses have a specific constellation of antigenic determinants on the HA and when an opportunity occurs for an avian influenza virus to infect a mammal, physiological pressures in the new host select a variant with the correct constellation of determinants. Evidence for such a host selection mechanism at the cellular level has been presented by Schild et al. (1983). We thank Dr A. W. Hampson, Melbourne, for providing viruses, and Lisa Newberry, Mary Ann Bigelow and Sue Green for providing excellent technical assistance. This work was supported by the Medical Research Council of New Zealand and in part by U.S. Public Health Research Grant AI 08831, AI 52586, and AI from the National Institute of Allergy and Infectious Diseases, Cancer Center Support (CORE) Grant CA 21765, a grant from the American Cancer Society (no. CD-253), and American Lebanese Syrian Associated Charities. REFERENCES AYMARD-HENRY, M., COLEMAN, M. T., DOWDLE, w. R., LAVER, W. G., SCHILD, G. C. & WEBSTER, R. G. (1973). Influenzavirus neuraminidase and neuraminidase-inhibition test procedures. Bulletin of the World Health Organization 48, CATON, A. J., BROWNLEE, G. G., YEWDELL, J. W. & GERHARD, W. (1982). The antigenic structure of the influenza virus A/PR/8/34 hemagglutinin (HI subtype). Cell 31, CATON, A. J., RAYMOND, F. L., BROWNLEE, G. G., YEWDELL, J. W. & GERHARD, W. (1983). Antigenic variation in influenza virus. Transactions. Biochemical Society 11, EASTERDAY, B. C. (1975). Animal influenza. In The Influenza Viruses and Influenza, pp Edited by E. D. Kilbourne. New York: Academic Press. GERHARD, W. & WEBSTER, R. G. (1978). Antigenic drift in influenza A viruses. I. Selection and characterization of antigenic variants of A/PR/8/34 (HON 1) influenza virus with monoclonal antibodies. Journal of Experimental Medicine 148, GERHARD, W., YEWDELL, J., FRANKEL, M. E. & WEBSTER, R. (1981). Antigenic structure of influenza virus haemagglutinin defined by hybridoma antibodies. Nature, London 290, HINSHAW, V. S., BEAN, W. J., WEBSTER, R. G. & EASTERDAY, B. C. (I978). The prevalence of influenza viruses in swine ~ and the antigenic and genetic relatedness of influenza viruses from man and swine. Virology 84, HINSHAW, V. S., WEBSTER, R. G., EASTERDAY, B. C. & BEAN, W. J. (1981). Replication of avian influenza A viruses in mammals. Infection and Immunity 34, HINSHAW, V. S., WEBSTER, R. G., BEAN, W. J., DOWNIE, J. & SENNE, D. A. (1983). Swine influenza-like viruses in turkeys: potential source of virus for humans? Science 220, HINSHAW, V. S., ALEXANDER, D. J., AYMARD, M., BACHMANN, P. A., EASTERDAY, B. C., HANNOUN, C., KIDA, H., LIPKIND, M., MACKENZIE, J. S., NEROME, K., SCHILD, G. C., SCHOLTISSEK, C., SENNE, D. A,, SHORTRIDGE, K. F., SKEHEL, J. J. & WEBSTER, R. G. (1984). Antigenic comparisons of swine-influenza-like H1NI isolates from pigs, birds, and humans: an international collaborative study. Bulletin of the World Health Organization 62, KENDAL, A. P., NOBLE, G. R., SKEHEL, J. J. & DOWDLE, W. R. (1978). Antigenic similarity of influenza A (H1NI) viruses from epidemics in to "Scandinavian" strains isolated in epidemics of Virology 89, KIDA, H., BROWN, L. E. & WEBSTER, R. G. (1982). Biological activity of monoclonal antibodies to operationally defined antigenic regions on the hemagglutinin molecule of A/Seal/Massachusetts/I/80 (H7N7) influenza virus. Virology 122, KmBOURNE, E. D. (1978). Genetic dimorphism in influenza viruses: characterization of stably associated hemagglutinin mutants differing in antigenicity and biological properties. Proceedings of the National Academy of Sciences, U.S.A. 75, KOHLER, G. & MILSTEIN, C. (1976). Derivation of specific antibody-producing tissue culture and tumour lines by cell fusion. European Journal of Immunology 6, LAVER, W. G. (1969). Purification of influenza virus. In Fundamental Techniques in Virology, pp Edited by K. Habel & N. P. Salzman. New York: Academic Press. NAKAJIMA, S., NAKAJIMA, K. & KENDAL, A. P. (1983). Identification of the binding sites to monoclonal antibodies on A/USSR/90/77 (H1N 1) hemagglutinin and their involvement in antigenic drift in H IN 1 influenza viruses. Virology 131, NEUSTADT, R. E. & HNEBERG, H. V. (1978). The Swine Flu Affair: Decision-Making on a Slippery Disease. Washington: U.S. Department of Health, Education and Welfare. PALMER, D. F., COLEMAN, M. T., DOWDLE, W. R. & SCHILD, G. C. (1975). Advanced Laboratory Techniques for lnfluenza Diagnosis. Immunology Series No. 6. Washington: U.S. Department of Health, Education and Welfare. RAYMOND, F. L., CATON, A. J., COX, N. J., KENDAL, A. P. & BROWNLEE, G. G. (1983). Antigenicity and evolution amongst recent influenza viruses of H1N1 subtype. Nucleic Acids Research 11,

10 992 F.J. AUSTIN AND R. G. WEBSTER SCHILD, G. C., OXFORD, J. S., DE JONG, J. C. & WEBSTER, R. G. (1983). Evidence for host-cell selection of influenza virus antigenic variants. Nature, London 303, SCHOLTISSEK, C., BURGER, H., BACHMANN, P. A. & HANNOUN, C. (1983). Genetic relatedness of hemagglutinins of the H1 subtype of influenza A viruses isolated from swine and birds. Virology 129, SHULMAN, M., WILDE, C. D. & KSHLER, G. (1978). A better cell line for making hybridomas secreting specific antibodies. Nature, London 276, TOBITA, K., SUGIURA, A., ENOMOTO, C. & FURUYAMA, M. (1975). Plaque assay and primary isolation of influenza A viruses in an established line of canine kidney cells (MDCK) in the presence of trypsin. Medical Microbiology and Immunology 162, WEBSTER, R. G. & LAVER, W. G. (1980). Determination of the number of nonoverlapping antigen areas on Hong Kong (H3N2) influenza virus hemagglutinin with monoclonal antibodies and the selection of variants with potential epidemiological significance. Virology 104, WEBSTER, R. G., HINSHAW, V. S., NEAVE, C. W. & BEAN, W. J. (1984). Pandemics and animal influenza viruses. In The Molecular Virology and Epidemiology of Influenza, pp Edited by C. H. Stuart-Harris & C. Potter. London: Academic Press. WILEY, D. C., WILSON, I. A. & SKEHEL, J. J. (1981). Structural identification of the antibody-binding sites of Hong Kong influenza haemagglutinin and their involvement in antigenic variation. Nature, London 289, (Received 4 December 1985)