J. gen. Virol. (1981), 54, 243-251 243 Printed in Great Britain

Analysis of Antigenic Drift in the Haemagglutinin Molecule of Influenza B Virus with Monoclonal Antibodies

By R. G. WEBSTER* AND M. T. BERTON St. Jude Children's Research Hospital 332 North Lauderdale, P.O. Box 318, Memphis, Tennessee 38101, U.S~4.

(Accepted 7 Janua~ 1981)

SUMMARY Antigenic drift in the haemagglutinin (HA) molecule of influenza B viruses was studied with monoclonal antibodies. Antigenic drift occurred in each of the 12 different epitopes studied and there was evidence that at least two antigenically distinguishable influenza B virus strains can co-circulate during an epidemic. The frequency of antigenic variation in the HA of influenza B viruses was frequently less than 1 in 10 s and was approx. 1000-fold below that found in influenza A strains. Haemagglutination inhibition (HI) tests on antigenic variants selected with 12 different monoclonal antibodies suggested that the antigenic determinants could be subdivided into three partially overlapping groups. Many of the antigenic variants selected with monoclonal antibodies were distinguishable from the parental virus with post-infection ferret sera, suggesting that the majority of the variants that do occur could have epidemiological potential. INTRODUCTION Influenza B viruses cause epidemics of influenza in people at approx. 4- to 6-year cycles (Stuart-Harris & Schild, 1976) and have been associated with Reye's syndrome, an acute encephalopathy of children (Corey et al., 1976). Although antigenic drift has been shown to occur in these viruses (Chakraverty, 1972; Schild et al., 1973), no major antigenic shift has been reported and the minor antigenic changes occur less frequently than in influenza A viruses. The lack of antigenic shift in influenza B virus may be due to the absence of a gene pool of influenza B viruses in animals (Pereira, 1969), but the relatively slow antigenic drift in these viruses has not been explained. There is no a priori reason why antigenic drift in influenza B viruses should be different from influenza A viruses. The eight RNA segments code for at least eight polypeptides analogous to those of A viruses (Racaniello & Palese, 1979) and, as in influenza A, the RNA segments of influenza B viruses have common sequences at their ends (Skehel & Hay, 1978), although these are different from influenza A viruses. The only reported differences are that the gene segments of influenza B viruses are larger than for the influenza A viruses (Desselberger & Palese, 1978) and preliminary studies with three monoclonal antibodies to the haemagglutinin (HA) molecule of the prototype strain B/Lee/40 suggest a low frequency of isolation of antigenic variants (Lubeck et al., 1980). In comparison with the detailed information on antigenic drift in influenza A viruses (Laver et al., 1979; Webster & Laver, 1980), little or no information is available about influenza B viruses. We present here a study of antigenic drift in the HA molecule of influenza B viruses using monoclonal antibodies which were used to determine the frequency of antigenic variation, the number of non-overlapping antigenic determinants and to determine whether variants selected with monoclonal antibodies could be distinguished from wild-type viruses with post-infection ferret antisera.

0022-1317/81/0000-4417 $02.00 © 198] SGM 244 R. G. WEBSTER AND M. T. BERTON

METHODS Viruses. B/Hong Kong/8/73 (B/HK/8/73) and representative influenza B isolates (Table 1) were grown in the ailantoic cavity of 11-day-old embryonated hens' eggs. When necessary, the virus was concentrated by adsorption to and elution from chicken erythrocytes, and purified by banding on sucrose gradients as previously described (Laver & Webster, 1968). An inhibitor-resistant strain of B/HK/8/73 was prepared by growing the virus in embryonated eggs in the presence of horse serum and foetal calf serum from which immunoglobulins had been removed. The variants that were not inhibited by calf and horse serum were used to screen for the production of antibodies in hybridoma cells. Serological assays. Haemagglutination inhibition (HI) tests followed the method of Fazekas de St. Groth & Webster (1966), using sera or ascitic fluid treated with receptor-destroying enzyme (Robinson & Dowdle, 1959). Neuraminidase inhibition (NI) tests were performed according to Aymard-Henry et al. (1973), and indirect enzyme-linked immunosorbent assays (ELISA) according to Ruitenberg et al. (1976), as modified by van Wyke et al. (1980). Preparation of hybridoma cell lines. Hybridoma cell lines were prepared by fusion of the 8-azaguanine-resistant clone of MOPC-21 myeloma cells (P3/X63-AgS) with spleen cells of immunized Balb/c mice using polyethylene glycol (Kohler & Milstein, 1976) as previously described for influenza virus (Koprowski et al., 1977). The mice received two doses of 1000 HA units of purified B/HK/8/73 virus given 2 months apart. The culture fluids from the fused cells were screened by ELISA, HI and NI tests with inhibitor-resistant influenza viruses for the detection of antibodies. Cultures producing antibody to the HA or neuraminidase (NA) were cloned in soft agar and injected intraperitoneally into pristane-treated mice (Koprowski et al., 1977). Ascitic fluid was collected 7 to 10 days later and used in the assay. Additional cloning of the hybridoma lines provided preliminary evidence that the antibodies were monoclonal. The antibodies produced after each cloning reacted identically against a large panel of natural influenza variants in HI tests. The panel of monoclonal antibodies used in these studies originated from hybridomas prepared from different immunized mice. Post-infection ferret sera were prepared as previously described (Kendal et al., 1977). Selection of antigenic variants and determination of frequency. Antigenic variants were selected according to 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 l 1-day-old chick embryos. The viruses that grew in the presence of monoclonal antibodies were harvested and 'cloned' twice at limiting dilutions'in 11-day-old embryonated hens' eggs.

RESULTS

Specificity of the monoclonal antibodies The antibodies produced in the hybridoma lines were specific for the HA molecule of B/HK/8/73 viruses (Table 1) and did not react with influenza A viruses (results not shown). The 12 monoclonal antibodies used in this study were tested against a panel of influenza B isolates and each monoclone gave a different pattern of reactivity, establishing that the monoclonal antibodies were directed against different antigenic determinants (Table 1).

Antigenic drift in influenza B viruses and the co-circulation of antigenically distingu&hable viruses Monoclonal antibodies to the HA molecule of B/HK/8/73 showed that antigenic drift had occurred in influenza B isolates (Tables 1, 2) and that the B/HK/8/73 strain is more closely Table 1. Antigenic drift in influenza B viruses detected with different monoelonal antibodies to the haemagglutinin (HA) of B/Hong Kong/8/73*

Monoclonal antibody Haemagglutinafion inhibition (HI) antibody fitre to influenza B virus: to B/HK/8/73 r ~ haemagglutinin B/Lee/40 B/Bonn/43 B/Sweden/2/49 B/Melb/4/53 B/Eng/847/73 B/Mere/I/77 B/Mere/3/77 B/Hann/58/79 B/HK/8/73 16/1 <~ 10000 3800 160 510 < < 40 21000 313/2 < 1300 1300 1000 320 < < < 6600 430/1 < 7100 5000 280 260 < 110 320 20000 331/5 1280 14000 7100 320 90 < < < 14000 82/1 < 590 2500 170 450 320 550 2700 7100 173/1 < 85 1300 < < < < < 2700 282/1 < 70 160 < 11000 < < < 10000 365/1 < 2500 110 < 260 < < < 7600 74/1 < 65 < < < 2300 11000 11000 17000 t~ 21/6 < 1900 1600 < 260 510 5000 2900 6200 174/1 4500 6300 22000 600 280 630 320 350 35000 290/3 < 5000 5000 4100 540 450 5000 5000 8700

r,,,a * The figures give the reciprocal of the dilution inhibiting 3 out of 4 HA units of virus in the HI assay. <, HI titre less than 20.

Table 2. Antigenic drift in recent influenza B viruses and the co-circulation of antigenieally distinguishable influenza B viruses*

Monoclonal HI antibody titre to influenza B virus: e~ antibody to B/HK/8/73 B/HK/ B/Eng/ B/Hannover/B/Hannover/B/Hannover/B/Hannover/B/Singapore/B/Singapore/ B/Ohio/ B/Canada/ B/Oregon/ haemagglutinin 8/73 65/76 9/78 13/78 58/79 63/79 222/79 263/79 10/80 1/80 5/80~ 16/1 21000 <$ < < 100 < < < < < < 313/2 6600 < < < < < < < < < < 430/1 20000 < < < 200 < < < < < < ~° 331/5 14000 < < < < < < < < < < 82/1 7100 200 < 100 800 < 100 < 100 < < 173/1 2700 < < < < < < < < < < 282/1 10000 < < < < < < < < < < 365/1 7600 < < < < < < < < < < 74/1 17000 6400 < 800 12800 < 6400 < 6400 800 400 21/6 6200 3200 < 400 3200 < 1600 < 3200 200 200 174/1 35000 200 < 200 400 < 200 < 100 < < 290/3 8700 400 < 400 3200 < 400 < 800 100 < to * The figures give the reciprocal of the dilution inhibiting 3 out of 4 HA units of virus in the HI assay performed in microtitre trays. 4~ "l" Virus isolated from a Reye's syndrome patient. :~ <, HI titre less than 100. 246 R. G. WEBSTER AND M. T. BERTON Table 3. Frequency of antigenic variants in different antigenic areas of the HA of B/Hong Kong/8/73 Monoclonal antibody Infectivitytitre in to B/HK/8/73 r ~" haemagglutinin Saline Monoclonalantibody Frequency* 16/1 8.69 0.64~" --8.05 21/6 8.13 0.75 --7.38 74/1 8.00 0-0 <--8.00 82/1 8.19 0.0 <--8.19 150/3 8.25 0.0 <--8.25 174/1 7.81 0.0 <--7.81 282/1 8.00 0.16 --7.84 313/2 7.35 0.08 --7.27 331/5 8.00 0.00 <--8.00 365/1 8.00 0.16 --7.84 419/2 7-94 0.0 <--7-94 430/1 7.20 0.0 <--7.20 * Frequency of antigenic variants is expressed as log~0 of the ratio of the virus infectivity titre in the presence and absence of monoclonal antibodies. ~" Viruses which grew in the presence of monoclonal antibodies were antigenic variants. related to B/Bon/43 and B/Sweden/49 than to recent influenza B isolates. It must be emphasized that only 12 monoclonal antibodies were used in this study and probably do not represent all of the different epitopes on the HA of B/HK/8/73. Among recent influenza B isolates, only three of the 12 different monoclonal antibodies showed significant cross- reactions (Table 2). Influenza B viruses isolated during the same outbreak were examined with a panel of monoclonal antibodies to determine if they were antigenically homogeneous (Table 2). Although significant antigenic drift had occurred in the influenza B isolates, it is apparent that at least two different antigenic variants were co-circulating in Hannover during the outbreaks of 1978 and 1979 and in Singapore in 1980. In each case, one of the variants showed some antigenic reactivity in at least five antigenic sites with the B/HK/8/73 strain, whereas the co-circulating strain showed no reactivity with B/HK/8/73 at the 12 different antigenic determinants tested. Among the more recent isolates, B/Singapore/222/79 (the current vaccine strain) is similar to B/England/65/76 and to the isolates from 1980 (Table 2). An influenza B strain isolated from a Reye's syndrome patient (B/Oregon/5/80) did not show any antigenically distinguishable properties with this panel of monoclonal antibodies.

Frequency of antigenic variation in influenza B viruses Previous studies using a clonal preparation of influenza A virus have shown that the frequency of variant HA molecules was between 1 in 104.5 and 1 in 107 (with a mean of approx. 1 in 105) when estimated with a single monoclonal antibody (Yewdell et al., 1979; Webster & Laver, 1980) and was less than 1 in 101° when monoclonal antibodies from two non-overlapping antigenic determinants were used in combination. All influenza A variants so far examined show a single amino acid substitution in the HA1 polypeptide of the HA molecule. The frequency of antigenic variants in cloned preparations of influenza B/HK/8/73 was determined with the above panel of monoclonal antibodies to the HA (Table 3). Each variant was not inhibited in HI tests by the monoclonal antibody used in its selection. The results show that the frequency of variation ranged from approx. 1 in 107.3 to levels that were undetectable. One explanation for the low levels of antigenic variants could be that the Antigenic areas on influenza B haemagglutinin 247 antibody preparations were not monoclonal but contained mixtures of antibodies to different non-overlapping antigenic areas. This possibility was ruled out by re-cloning the antibody- secreting hybridoma cells and re-testing the frequency of variation; the results did not change. Another possible explanation could be that different cloned preparations of influenza B/HK/8/73 may show significant differences in the frequency of antigenic variants. Four different cloned preparations of B/HK/8/73 were therefore examined and were found to give similar results to those given above.

Analysis of antigenic variants selected with monoclonal antibodies Since the frequency of variants in the above studies was very low, it was difficult to obtain antigenic variants for detailed analysis. We circumvented this problem by inoculating large numbers of embryonated eggs with undiluted virus preparations containing dilutions of monoclonal antibodies; in this way a small number of variants were isolated (Table 4). To obtain information on the number of non-overlapping antigenic areas on the HA molecule of the influenza B virus, variants produced were analysed with the 12 different monoclonal antibodies (Table 4). Although the reactivity patterns suggested three groups, there was overlap between each, suggesting that this panel of monoclonal antibodies was not able to detect more than one non-overlapping antigenic area. Two monoclonal antibodies (174/1 and 290/3) reacted with all of the variants and since antigenic variants have so far not been obtained with these monoclonal antibodies, we do not know if they recognize a different antigenic area.

Analysis of antigenic variants with antisera Previous studies with antigenic variants of influenza A viruses selected with monoclonal antibodies have shown that even though the variants were clearly distinguishable with monoclonal antibodies, only a minority could be distinguished from the parental strain with antisera (Gerhard & Webster, 1978; Webster & Laver, 1980). To determine if the antigenic variants of B/HK/8/73 could be distinguished from parental virus with antisera, they were assayed in HI tests with post-infection ferret sera and with a hyperimmune goat antiserum (Table 5). The results show that six of the ten variants were significantly different from the parental virus; with ferret antisera, five of the variants had HI antibody titres which were more than 10-fold different from the parental virus. Some of the antigenic variants (16/1 and 313/2) gave higher titres with the antisera, suggesting that some mutations actually enhanced the combination of some antibodies with the HA molecule.

DISCUSSION Influenza B viruses are different from influenza A viruses as they show antigenic drift (Schild et al., 1973) but not antigenic shift, and to date a non-human reservoir of B viruses has not been conclusively demonstrated. Monoclonal antibodies to the HA molecule of influenza B viruses have been used in this study to investigate more precisely antigenic drift in these viruses. Antigenic drift occurred in each of the epitopes defined by the 12 different monoclonal antibodies in influenza B viruses isolated from people since 1973. The sensitivity of the monoclonal antibodies permitted detection of antigenically distinguishable viruses from the same influenza B outbreak in Hannover in 1978 and 1979 and in Singapore in 1979. The frequency of antigenic variation in the HA molecule of B/HK/8/73 was less than 1 in l0 s with most monoclonal antibodies. This figure is 100- to 1000-fold below that found with influenza A viruses (Yewdell et al., 1979), suggesting that the HA of influenza B viruses is less variable than that of influenza A virus. There are several possible explanations for the lower frequency of antigenic variation in influenza B viruses that may be related to fundamental differences between influenza A and B viruses or to the failure of the monoclonal antibodies to to o0

Table 4. Reactivity patterns of antigenic variants of B/Hong Kong/8/73 selected with monoclonal antibodies to the HA molecule*

Variant of B/Hong Kong/8/73 selected with monoclonal antibody: A r' ., 16/1~" 313/2 430/1 282/1 365/1 365/1 74/l 74/1 21/6 21/6 HK/8/73 Im Group Monoclonal antibody VI$ V1 V3 V1 V2 V3 V1 V2 V2 V3 Parent O

+ + + + t 16/1 ------+ + + + Ia 313/2 - - - + + + + + + + + m 430/1 - - - + + + + + + + + l. 331/5 - - - + + + + + + + + -t

[ 82/1 - -- -- + + - + + + + + ) 173/1 - - - + - + - _ + + + > Ib ] 282/1 .... + ------+ + + Z L 365/1 ------+ .... + + +

74/1 + + + + + -- -- _ + + + ,..] Ic 21/6 + + + + + + .... +

( 174/1 + + + + + + + + + + + ~° ? 290/3 + + + + + + + + + + + O * Reactivity patterns were determined in HI tests: +, titre with variant identical to that with parental virus; -, titre with variant at least 10-fold below that with parental virus. Z "j" Monoclonal antibody preparation number. $ Antigenic variant number. Antigenic areas on influenza B haemagglutinin 249 Table 5. Reactions of antigenic variants ofB/Hong Kong/8/73 with antisera*

HI antibody titre with antisera to B/HK/8/73 Antigenic variant selected with Hyperimmune Ferret-31 Ferret-2 monoclonal antibody goat antiserum post-infection serum post-infection serum 16/1 V1 3200 640 2560 313/2 V 1 6400 640 1280 430/1 V3 3200 640 640 282/1 V1 1600 640 1280 74/1 V1 3200 20 160 74/1 V2 3200 20 160 365/1 V2 800 80 80 365/1 V3 400 <20 40 21/6 V2 400 20 80 21/6 V3 400 20 80 B/HK/8/73 parent 1600 320 640

* The figures give the reciprocal of the dilution inhibiting 3 out of 4 HA units of virus. detect antigenic variants. However, we consider it unlikely that the mutation frequency in the RNA coding for the influenza B HA is less than that in influenza A viruses. Other suggestions are: (i) the majority of the mutations are lethal, i.e. there may be more functional restrictions on the conformation of the influenza B haemagglutinin molecule; (ii) influenza B transcription is less error prone than transcription in influenza A viruses; (iii) the RNA coding for the HA of influenza B viruses is partially duplicated; (iv) influenza B virus populations contain a high frequency of phenotypically mixed particles; (v) the monoclonal antibodies are not detecting single amino acid changes, so that most of the variants isolated seem to contain more than one amino acid substitution. Sequence analysis of the gene coding for the HA of influenza B viruses will resolve some of these possibilities. Alternative explanations such as differences between cloned virus preparations and/or that the antibodies were not monoclonal have been excluded experimentally. The low frequencies of antigenic variation with B/HK/8/73 are in agreement with the low frequencies obtained by Lubeck et al. (1980) with B/Lee/40. Reactivity patterns of the antigenic variants established in HI tests suggested that there was only a single overlapping antigenic area on the HA molecule of influenza B viruses, although this area could be further subdivided (Table 4). Antigenic variants selected with monoclones in group Ia were not inhibited by eight of the different monoclonal antibodies and were indistinguishable from parental virus with ferret antisera. On the other hand, most of the antigenic variants selected with antibodies in groups Ib and Ic were significantly different from the parental virus with ferret antisera. This suggests that these antigenic determinants are perhaps biologically important in antigenic drift, while the determinants affected in the variants in group Ia may be epidemiologicaUy irrelevant. Analysis of antigenic variants of influenza A viruses selected with monoclonal antibodies shows that the majority of the variants could not be distinguished from the parental strain with rabbit or ferret antisera (Gerhard & Webster, 1978; Webster & Laver, 1980). Only one of 30 influenza A variants examined could be distinguished with ferret sera and this variant showed a single amino acid substitution of glycine to aspartic acid at residue 144 in the HA1 polypeptide. In contrast, six out of ten of the antigenic variants of influenza B viruses examined were approx. 10-fold different from the parental virus with ferret sera. These results suggest that, although the frequency of variation in influenza viruses is lower than in influenza A viruses, more of the variants would have epidemiological potential. The relatively small number of different monoclonal antibodies and the small number of variants used in these studies, as well as the difficulty in isolating antigenic variants, could 250 g.G. WEBSTER AND M. T. BERTON mean that we have too few monoclonal antibodies to fully map the antigenic determinants on the HA of influenza B viruses. Nevertheless, the results do suggest that the number of non-overlapping antigenic areas on influenza B viruses is likely to be small. We have no knowledge as to whether the changes in these variants are actually in the antibody binding site or are removed from the actual site and result in large conformational changes in the molecule. Sequence studies on the RNA coding for the HA molecule of influenza B virus, in conjunction with detailed analysis of the antigenic areas and their relationship with the tertiary structure, will provide the necessary information for understanding the molecular basis of antigenic drift in influenza B viruses and the difference from influenza A viruses.

This work was supported by Contract AI 42510 and Research Grant AI 08831 from the National Institute of Allergy and Infectious Diseases, Childhood Cancer Center support grant CA 21765 and by ALSAC. The excellent technical assistance of Alford Pointer and Sue Anne Jenkerson is acknowledged.

REFERENCES AYMARD-HENRY, M., COLEMAN, M. T., DOWDLE, W. R., LAVER, W. G., SCHILD, G. C. & WEBSTER* R. G. (1973). Influenza neuraminidase and neuraminidase inhibition test procedures. Bulletin of the World Health Organization 48, 199-202. ClJAKRAVERTV, P. (1972). Antigenic relationships between influenza B viruses. Bulletin of the World Health Organization 45, 755-766. COREY, L., RUBIN, A. J., HA'I'I'WlCK, M. A., NOBLE, G. R. & CASSIDY, E. (1976). A nationwide outbreak of Reye's syndrome. Its epidemiologic relationship to influenza B. American Journal of Medicine 61, 615-625. DESSELBEROER* U. & PALESE, P. (1978). Molecular weights of RNA segments of influenza A and B viruses. Virology 88, 398-399. FAZEKAS DE ST. GROTH, S. & WEBSTER. R. G. (1966). Disquisitions on original antigenic sin: evidence in man. Journal of Experimental Medicine 124, 331-345. GERHARD, W. & WEBSTER. R. G. (1978). Antigenic drift in influenza A viruses. I. Selection and characterization of antigenic variants of A/PR/8/34 (HONI) influenza viruses with monoclonal antibodies. Journal of Experimental Medicine 148, 383-392. KENDAL, A. P., NOBLE, G. R. & DOWDLE, W. R. (1977). Swine influenza viruses isolated in 1976 from man and pig contain two co-existing subpopulations of antigenically distinguishable hemagglutinins. Virology 82, 111-121. KOHLER* G. & MILSTEIN, C. (1976). Derivation of specific antibody-producing tissue culture and tumor lines by cell fusion. European Journal of Immunology 6, 51 I-519. KOPROWSKI, H., GERHARD, W. & CROCE, C. M. (1977). Production of antibodies against influenza virus by somatic cell hybrids between mouse myeloma and primed spleen cells. Proceedings of the National Academy of Sciences of the United States of America 74, 2985-2988. LAVER, W. G. & WEBSTER. R. G. (1968). Selection of antigenic variants of influenza viruses. Isolation and peptide mapping of their hemagglutinating proteins. Virology 34, 193-202. LAVER, W. G., AIR, G. M., WEBSTER. R. G., GERHARD, W., WARD, C. W. & DOPHEIDE, T. A. A. (1979). Antigenic drift in type A influenza virus: sequence differences in the bemagglutinin of Hong Kong (H3N2) variants selected with monoclonal hybridoma antibodies. Virology 98, 226-237. LUBECK, M. 0., SCHULMAN, J. L. & PALESE, P. (1980). Antigenic variants of influenza viruses: marked differences in the frequencies of variants selected with different monoclonal antibodies. Virology 102, 458-462. PEREIRA, H. ~. (1969). Influenza: antigenic spectrums. Progress in Medical Virology 11, 46-69. RACANIELLO, V. R. & PALESE, P. (1979). Influenza B virus genome: assignment of viral polypeptides to RNA segments. Journal of Virology 29, 361-373. ROBrNSON, R. Q. & DOWDLE, W. g. (1959). Influenza viruses. In Diagnostic Proceduresfor Viral and Rickettsial Infection, pp. 414-433. Edited by E. D. Lennette and J. J. Schmidt. Washington, D.C.: American Public Health Association. RUITENBERG, E. J., STEERENBERG, P. A., BROSI, B. J. M. & BUGS, J. (1976). Reliability of the enzyme-linked immunosorbent assay (ELISA) for the serodiagnosis of Trichinella spirales infection in conventionally raised pigs. Journal of Immunological Methods 10, 67-83. SCmLD, G. C., PEREIRA, M. S., CHAKRAVERTY, P., COLEMAN, M. T., DOWDLE, W. R. & CHANGE, W. K. (1973). Antigenic variants of influenza B virus. British Medical Journal IV, 127-131. SKEHEL, J. J. & HAY, A. J. (1978). Nucleotide sequences at the 5' termini of influenza virus RNAs and their transcripts. Nucleic Acids Research 5, 1207-1219. STUART-HARmS, C. H. & SCHILD, O. C. (1976). Influenza: The Viruses and the Disease, pp. 1-242. Littleton, Massachusetts: Publishing Sciences Group. Antigenic areas on influenza B haemagglutinin 251

VAN WYKE, K. L., HINSHAW, V. S., BEAN, W. J., JR & WEBSTER, R. G. (1980). Antigenic variation of influenza A virus nucleoprotein detected with monoclonal antibodies. Journal of Virology 35, 24-30. WEBSTER, R. G. & LAVER, W. G. (1980). Determination of the number of nonoverlapping antigenic areas on Hong Kong (H3N2) influenza virus hemagglutinin with monoclonal antibodies and the selection of variants with potential epidemiological significance. Virology 104, 139-148. YEWDELL, J. W., WEBSTER, R. G. & GERHARD, W. (1979). Antigenic variation in three distinct determinants of an influenza type A hemagglutinin molecule. Nature, London 279, 246-248.

(Received 11 September 1980)