Egg-grown Virus Research Articles

Adaptation of Egg-Grown and Transfectant Influenza Viruses for Growth in Mammalian Cells: Selection of Hemagglutinin Mutants with Elevated pH of Membrane Fusion

A series of eight transfectant influenza viruses was generated by reverse genetics for studies of the palmitylated cysteine residues in the cytoplasmic tail of the hemagglutinin glycoprotein (HA). Following amplification of these viruses in MDCK cells we found that all had developed an elevated pH of membrane fusion—an unexpected result since previous mutant HA expression studies had shown that substitutions of the cysteine residues had no effect on fusion properties. Sequence analyses revealed that each of the viruses had at least one additional mutation in the ectodomain of HA which was responsible for the increase in fusion pH. Similarly, when we passaged egg-grown wild-type X-31 virus in three different lines of MDCK cells or in MDBK cells, high pH fusion mutants were selected within a few passages in every case. The locations of the substitutions in the HA structure are in or near the “fusion peptide” or at subunit interfaces throughout the length of the trimer—reminiscent of the changes selected in earlier studies on amantadine resistance. The observation that passage of certain viruses in mammalian cells can result in the selection of mutants with elevated fusion pH has potential implications both for reverse genetic experiments and, perhaps more importantly, for the choice of substrates for propagation of vaccine viruses.

African green monkey kidney (Vero) cells provide an alternative host cell system for influenza A and B viruses.

The preparation of live, attenuated human influenza virus vaccines and of large quantities of inactivated vaccines after the emergence or reemergence of a pandemic influenza virus will require an alternative host cell system, because embryonated chicken eggs will likely be insufficient and suboptimal. Preliminary studies indicated that an African green monkey kidney cell line (Vero) is a suitable system for the primary isolation and cultivation of influenza A viruses (E. A. Govorkova, N. V. Kaverin, L. V. Gubareva, B. Meignier, and R. G. Webster, J. Infect. Dis. 172:250-253, 1995). We now demonstrate for the first time that Vero cells are suitable for isolation and productive replication of influenza B viruses and determine the biological and genetic properties of both influenza A and B viruses in Vero cells; additionally, we characterize the receptors on Vero cells compared with those on Madin-Darby canine kidney (MDCK) cells. Sequence analysis indicated that the hemagglutinin of Vero cell-derived influenza B viruses was identical to that of MDCK-grown counterparts but differed from that of egg-grown viruses at amino acid positions 196 to 198. Fluorescence-activated cell sorting analysis showed that although Vero cells possess predominantly alpha2,3 galactose-linked sialic acid, they are fully susceptible to infection with either human influenza A or B viruses. Moreover, all virus-specific polypeptides were synthesized in the same proportions in Vero cells as in MDCK cells. Electron microscopic and immunofluorescence studies confirmed that infected Vero cells undergo the same morphological changes as do other polarized epithelia] cells. Taken together, these results indicate that Vero cell lines could serve as an alternative host system for the cultivation of influenza A and B viruses, providing adequate quantities of either virus to meet the vaccine requirements imposed by an emerging pandemic.

Comparison of 10 influenza A (H1N1 and H3N2) haemagglutinin sequences obtained directly from clinical specimens to those of MDCK cell- and egg-grown viruses

PCR was used to amplify and sequence the complete HA1 region of the haemagglutinin (HA)-encoding genes of 10 clinical isolates of influenza virus of the H1N1 or H3N2 subtypes. These sequences were compared to those obtained from viruses isolated from the same specimens after passage in eggs and MDCK cells. Amino acid substitutions in the egg-derived HA sequences were found in nine out of the 10 specimens analysed, whereas seven out of eight of the MDCK-derived HA sequences were identical to those in the corresponding original specimens. Changes in the H1 HA occurred at residues 77a, 196 (also found in the corresponding HA from the MDCK isolate), 225, 226 and 227; changes in the H3 HA occurred at residues 137, 156, 186, 248 and 276. In addition, we have shown that an amino acid change at residue 145 in the HA of the H3 subtype that was previously demonstrated to be egg-selected is now present in circulating strains.

Influence of Host Cell-Mediated Variation on the International Surveillance of Influenza A (H3N2) Viruses

Growth of clinical specimens of influenza viruses in eggs can result in the selection of antigenic variants distinct from corresponding viruses grown in mammalian tissue culture. To evaluate the contribution of host cell selection on the antigenic diversity of human influenza isolates, as seen in annual surveillance studies, viruses grown in embryonated eggs were compared by antigenic and genetic analyses with their mammalian tissue culture-grown counterparts. Clinical specimens were gathered from around the world from late 1987 to 1990 and the antigenicity of isolated viruses was assessed by hemagglutination-inhibition assays using immune ferret sera as is currently performed for routine surveillance and the selection of vaccine strains. In addition, viruses were assessed using a panel of anti-H3 HA monoclonal antibodies. The extent of antigenic variation exhibited by the egg-grown strains was far greater than the relative antigenic homogeneity of the tissue culture-grown viruses. Nucleotide sequence analysis of HA1 gene PCR products of 28 MDCK cell and egg derived pairs allowed identification of amino acid substitutions responsible for the antigenic differences observed and the adaptation to growth in eggs. Among these substitutions was a change at amino acid position 186 of HA1 (Ser in tissue culture viruses and lie in egg-grown viruses) which was observed at relatively high frequency. Egg- and MDCK-grown pairs with this single amino acid difference were classified into distinct antigenic groups by ferret sera raised to WHO reference viruses. Given the additional antigenic diversity observed among egg-grown strains, considerable care should be taken in the selection of reference and vaccine strains grown in eggs. Rapid sequence comparisons of MDCK- and egg-grown viruses allow identification of variants arising through egg selection and will prove to be a useful adjunct to antigenic surveillance for the selection of reference and vaccine strains.

Selection of antigenically distinct variants of influenza C viruses by the host cell

Five strains of influenza C virus were isolated and passaged in the amniotic sacs of embryonated hens' eggs, or in the HMV-II line of human malignant melanoma cells, and were tested for reactivity with a panel of monoclonal antibodies to the hemagglutinin-esterase (HE) glycoprotein. It was observed with two strains (C/Yamagata/4/88, C/Yamagata/7/88) that the HE of virus passaged in HMV-II cells was antigenically distinguishable from that of virus cultivated in eggs. Virus clones obtained after repeated passages of these two strains in HMV-II cells all showed a significant increase in the ability to replicate in the cell culture compared to clones derived from viruses grown in eggs. No difference was seen, by contrast, in the ability to grow in eggs between HMV-II- and egg-derived virus clones. It was also found that HMV-II-grown viruses but not egg-grown viruses could agglutinate glutaraldehyde-fixed chicken erythrocytes at 23°. These observations, taken together, suggest that isolation and passage of influenza C virus in HMV-II cells sometimes result in selection of antigenically distinct variants which have an advantage in binding to the cell surface receptors. Sequence analyses of the HE genes revealed that compared to egg-grown viruses, HMV-II-adapted variant of the Yamagata/4/88 strain had a single amino acid substitution in the HE molecule at position 283 (Asp → Asn) and that of the Yamagata/7/88 strain had two substitutions at positions 212 (Glu → Lys) and 519 (Asn → Asp).

Immune response to inactivated influenza virus vaccine: antibody reactivity with epidemic influenza B viruses of two highly distinct evolutionary lineages

Vaccination of adults (healthy female employees potentially capable of transmitting influenza to high-risk persons; n = 104) in autumn 1990 with a trivalent influenza virus vaccine containing B/Yamagata/16/88 induced a low antibody response to B/Finland/150/90, a recent variant of B/Victoria/2/87-like viruses, as compared with the antibody response to B/Finland/172/91, a current variant in the lineage of B/Yamagata/16/88-like viruses. Up to the end of the epidemic season, the antibody status declined but was still significantly better than before the vaccination. The results suggest that the vaccine strain was appropriate for the outbreak of 1990 to 1991 in Finland, but may provide unsatisfactory protection against B/Victoria/2/87-like viruses. Evidence is given that use of Madin-Darby canine kidney (MDCK)-grown virus as an antigen in the haemagglutination inhibition test (HI) may provide more reliable information about the protective antibodies than use of untreated or ether-treated egg-grown viruses. Significantly higher postvaccination and postepidemic antibody titres were recorded among subjects who exhibited the antibody before vaccination than among seronegative subjects. A significantly higher response rate among initially seronegative people than among seropositive people was recorded for antibody to B/Finland/150/90, but no clear evidence was obtained that the pre-existing antibody could have had a negative effect on the antibody production.

Direct Isolation in Eggs of Influenza A (H1N1) and B Viruses with Haemagglutinins of Different Antigenic and Amino Acid Composition

Influenza A (H1N1) and influenza B viruses from clinical samples were isolated in the amniotic cavity of embryonated hens' eggs by classical techniques and propagated in the allantoic cavity. Virus progeny from different eggs which had been inoculated with virus material from the same clinical sample possessed antigenically distinguishable haemagglutinins (HAs). Virus progeny of some eggs possessed HAs which were serologically identical to those of virus isolated in parallel in mammalian (MDCK) cells. These egg-grown viruses possessing HAs with the antigenic phenotype of mammalian cell-grown viruses appeared to be antigenically related to epidemic influenza virus because post-infection human sera reacted to high titre with the virus HA. Specific nucleotide changes were detected in the HAs of the viruses isolated directly in eggs at positions 163 and 189 for influenza A (H1N1) viruses or positions 141 and 196 to 198 for influenza B viruses. Egg-isolated viruses which possessed the antigenic phenotype of mammalian cell-grown viruses retained glycosylation sites at positions 163 and 196. The viruses isolated directly in embryonated hens' eggs which possessed the HA antigenic phenotype and glycosylation sites of MDCK cell-grown virus can, unlike the latter viruses themselves, be used as candidate influenza vaccine viruses.

Now and Future Influenza Vaccines

Influenza is a modern day plague. In the young, the clinical picture is classical, but in the elderly, the disease may go unsuspected until complications such as pneumonia develop. Influenza A and B viruses are responsible, and these viruses mutate with great regularity. Antibodies to the HA and NA surface antigens of influenza viruses, both naturally and vaccine induced, are protective. The earliest influenza vaccines were crude, toxic, and ineffective. With modern purification techniques, the egg-grown viruses have been turned into safe, immunogenic, and effective killed-virus vaccines--whole virus and split virus. Surveillance permits the correct virus strains to be incorporated into each new vaccine. Those who have been experiencing the worst effects of influenza have been identified. These individuals need to be immunized each year. In the future, live influenza virus vaccines may offer the benefits of ease of administration and longer-lasting protection. Synthetic peptides, genetically engineered antigens, and even nonantigen (anti-idiotype) vaccines are possible, but such vaccines will require adjuvant enhancement. For the present, greater efforts must be made to use existing influenza vaccines.

A host-cell-selected variant of influenza B virus with a single nucleotide substitution in HA affecting a potential glycosylation site was attentuated in virulence for volunteers

An influenza B virus was passaged in man (virus A) and then in human embryo trachea (C) and into embryonated eggs (D) or directly into eggs (B). Virus A, B, and C had the same (cell-like) haemagglutinin phenotype on reaction with selected monoclonal antibodies while D had an "egg-like" phenotype. The viruses were administered at a dose of 1,000 TCD50 (for MDCK cells) by intranasal inoculation to groups of 27 or 28 volunteers. Viruses A, B, and C all produced disease in six to eight volunteers, whereas D produced no illness and only four volunteers were infected. The viruses shed by the volunteers were indistinguishable from those with which they were inoculated. The haemagglutinin genes of the viruses were sequenced and changes were detected indicating amino acid substitutions at position 196-198 in the attenuated egg-grown virus D whereby a potential glycosylation site present in the other viruses was lost.

Efficacy of inactivated influenza A virus (H3N2) vaccines grown in mammalian cells or embryonated eggs.

Influenza virus (H3N2) host cell variants isolated from a single infected individual were compared for their protective efficacies when used as formalin-inactivated purified whole virus vaccines in ferrets. A/Mem/12/85 virus grown in embryonated chicken eggs (egg-grown), which differs from A/Mem/12/85 grown in mammalian Madin-Darby canine kidney cells (MDCK-grown) by a single amino acid substitution in the hemagglutinin molecule, was shown to be distinguishable by immune ferret serum. Ferrets were immunized intramuscularly with intact inactivated MDCK- or egg-grown virus and were subsequently challenged with infectious virus grown in either host cell type. MDCK-grown-virus vaccine induced higher mean serum hemagglutination-inhibiting (HAI) and neutralizing antibody titers than did egg-grown-virus vaccine and induced superior protection of ferrets against subsequent challenge with infectious virus grown in either host cell type. These results suggest that human influenza viruses that are antigenically and structurally similar to viruses grown in mammalian cells may be more efficacious as vaccines than some variants selected in eggs.

Extensive heterogeneity in the hemagglutinin of egg-grown influenza viruses from different patients

We establish that the cultivation of influenza (H3N2) virus from any infected individual in chicken embryos (eggs) can result in the isolation of viruses with antigenic and/or structural heterogeneity in the hemagglutinin (HA) molecule. This variability contrasted sharply with the apparent lack of antigenic alterations in the HA of influenza viruses isolated from patients in Madin Darby canine kidney (MDCK) cells. The most common subpopulation of egg-grown influenza viruses had the same phenotype as MDCK cell-grown virus and may best represent the virus circulating in humans. It should be considered the optimal strain for use in vaccine and epidemiologic studies.

Adsorption of LLCMK2 cell-grown Sendai virus onto human red blood cells and its release from the virus adsorbed cells.

An early stage of virus adsorption was studied in a system of Sendai virus metabolically labeled with [3H]leucine in LLCMK2 cells and of human red blood cells (RBCs). The efficiency of viral release from the virus-bound RBCs by incubation at 37 C depended on the number of virus particles which had been used for adsorption onto the RBC at 4 C. When 7.8 x 10(2) virus particles were previously adsorbed onto the RBC at 4 C, most of the viruses were dissociated from the RBC at 37 C. In the case of adsorption of 3 to 12 virus particles per RBC, however, most of the viruses were not dissociated from the RBC by incubation at 37 C. Such RBC-bound viruses were released by incubation with various bacterial neuraminidases (Clostridium perfringens, etc.) or with a large number of LLCMK2 cell-grown Sendai virus (LLCMK2-Sendai) particles, but not released by treatment with hemagglutinin-neuraminidase protein (Sendai-gp) isolated from egg-grown Sendai virus.

Antigenic and structural characterization of multiple subpopulations of H3N2 influenza virus from an individual

Influenza viruses grown in embryonated chicken eggs frequently possess antigenically distinguishable hemagglutinin (HA) compared to virus from the same source grown in mammalian cell culture. To further investigate the extent of variation among viruses from an individual, viruses were isolated from throat washes collected over a 48-hr period during infection with influenza virus designated A/Mem/6/86 (H3N2). Viruses were isolated from limit dilutions in eggs and mammalian Madin-Darby canine kidney (MDCK) cells and the antigenic, structural, and receptor-binding properties of these viruses were determined. Viruses which could be isolated in MDCK cells were present at 10- to 100-fold higher frequency in the original sample than viruses which could be isolated in eggs. The HA of virus clones isolated in MDCK cells were antigenically and structurally identical. In contrast, viruses from the same source, selected at limit dilution in eggs, could be divided into three distinct subpopulations based on the distinguishable antigenic and structural characteristics of their HA molecules. The three groups of egg-grown viruses could be distinguished from each other, and from MDCK cell-grown viruses, not only by a panel of anti-HA monoclonal antibodies, but also by immune ferret sera raised to H3N2 virus strains of recent years and sera raised to the different egg-grown clones themselves. Of these groups, group 1 and group 2 egg-grown viruses each represented a minor subpopulation of viruses which could be isolated in eggs, while viruses of the third antigenic phenotype were the most frequently isolated in eggs. Amino acid substitutions in the HA of egg-grown viruses occurred in antigenic and receptor-binding sites of the molecule. Group 1 viruses each possessed two amino acid substitutions in their HA molecules at residues 193 and 229 in HA1. Group 3 viruses, which displayed altered receptor specificities compared to MDCK cell-grown viruses and other egg-grown viruses, possessed a single amino acid substitution at residue 145 in HA1. The HA of the group 2 egg-grown viruses appeared structurally identical, yet displayed marked differences in antigenic and receptor-binding properties, compared to viruses isolated in MDCK cells. These results demonstrate that multiple, distinct subpopulations of virus can be isolated from a single patient during an infection with influenza and highlights the potential problems in selecting the most appropriate virus for epidemiological and vaccine purposes since selection could result in the use of viruses that are not representative of those which predominate in a human population.

Egg-grown and tissue-culture-grown variants of influenza A (H3N2) virus with special attention to their use as antigens in seroepidemiology

A field strain of influenza A (H3N2) virus isolated in embryonated eggs during the 1984-5 influenza outbreak (A/Finland/13/85E) was compared in an antigenic analysis with virus from the same clinical specimen isolated in MDCK cell cultures (A/Finland/13/85M). The M-virus appeared to be more sensitive to haemagglutination-inhibiting antibodies against heterologous viruses than did the E-virus. The results of propagation and plaque purification experiments support the hypothesis that a single clinical specimen may consist of distinct antigenic variant subpopulations promoted selectively by the host during isolation procedures. Receptor-binding properties are discussed as a possible explanation for this selectivity. A set of 471 paired sera consisting of pre-epidemic and post-epidemic specimens taken from the same subjects in 1984-5 was studied for haemagglutination-inhibiting antibodies to six influenza A (H3N2) virus strains, including the E-virus and the M-virus from A/Finland/13/85. Of the antigens used, the M-virus detected significant antibody increases more frequently than did the E-virus (10.0 v. 5.9%). The superiority of the M-virus may rest primarily in its ability to pick out anamnestic antibody responses. Irrespective of this cross-reactivity, pre-epidemic antibody to the M-virus was fairly well associated with protection. In the set of sera (230 specimens) collected in summer 1985 to represent different age groups, the antibody status against the M-virus was significantly better than the status against the E-virus. The results suggest that, at least in some instances, antibody to MDCK-grown virus is a more accurate indicator of the immune status of a community than antibodies to egg-grown virus variants.

Host cell-mediated variation in H3N2 influenza viruses

The influence of the host cell on the selection of antigenic variants of influenza A H3N2 viruses and the relevance of host cell selection to the induction of immunity by these viruses have been investigated. Influenza viruses were isolated from human clinical samples during a single epidemic, were passaged in mammalian Madin-Derby Canine Kidney (MDCK) cells or in embryonated hens eggs, and were tested for antigenic variability in the hemagglutinin (HA) molecule with a panel of monoclonal antibodies. In many cases, the HA of virus cultivated in eggs was antigenically distinct from the HA of virus from the same individual grown in mammalian cells. Viruses recovered from different individuals were antigenically similar to each other when grown in mammalian cell lines yet were antigenically heterogeneous when cultivated in eggs. The HA genes of viruses isolated from different individuals during the epidemic were shown, by sequence analysis, to differ from each other by five or six amino acid residues. Sequence analyses of the HA genes of MDCK cell-grown and egg-grown virus obtained from the same individual demonstrated that the molecular changes between antigenically distinct HAs of MDCK cell- and egg-grown A/Mem/12/85 virus involved a single amino acid substitution at residue 156 in HA1, which lies at the tip of the HA molecule and immediately adjacent to the receptor-binding site. However, the amino acid sequences of HAs from MDCK-grown and egg-grown viruses (A/Mem/2/85) isolated from a second individual were identical although these viruses exhibited antigenic differences when examined with anti-HA monoclonal antibodies. Therefore, single amino acid changes in the HA molecule may not be the sole cause of antigenic changes in the HA observed between pairs of MDCK cell-grown and egg-grown viruses and genes other than that encoding the HA may contribute to the host cell-mediated antigenic variation of these viruses. Nevertheless, antigenic differences between viruses grown in eggs and MDCK cells did not influence their ability to protect, since ferrets infected with either live egg-grown or MDCK-grown virus were protected equally well from challenge with virus grown in either host cell type.

Immunological studies on the antigenic determinants of Sendai virus HN glycoprotein involved in haemagglutinating activity

Mouse monoclonal anti-HN antibodies were produced against the HN glycoprotein from egg-grown Sendai virus to investigate the role of each biologically active site of the HN molecule. The monoclonal anti-HN antibodies were divided into 7 groups on the basis of their inhibitory reactivities to hemolytic, neuraminidase and hemagglutinating activities. Antibodies in the first and third groups inhibited to the same extent both hemagglutinating and neuraminidase activities. Antibodies in the second and fourth groups blocked the hemagglutinating activity, but not the neuraminidase activity. These findings suggest that there are 2 different types of antigenic determinant involved in both hemagglutinating and neuraminidase activities. One determinant is associated with both activities, whereas the other determinant is involved in each individual biological activity. Furthermore, antibodies in the third group inhibited the hemagglutinating activity of virions only, while antibodies of the fourth and fifth groups inhibited the isolated HN glycoprotein only. These results suggest that the antigenic determinants associated with the hemagglutinating activity may be located both in the protein embedded in the lipid bilayer and on the outside of the envelope in Sendai virus. In addition, there appear to be several antigenic determinants embedded in the lipid bilayer, since the antibodies of the fourth and fifth groups reacting with such determinants blocked the hemagglutinating activity differently than the neuraminidase activity. These results suggest that the antigenic determinants associated with the neuraminidase activity may also be more than one. Taken together, the antigenic determinants involved in the biological activities of HN glycoprotein are suggested to be two or more.

Immunological studies on the antigenic determinants of Sendai virus HN glycoprotein involved in haemagglutinating activity

Mouse monoclonal anti-HN antibodies were produced against the HN glycoprotein from egg-grown Sendai virus to investigate the role of each biologically active site of the HN molecule. The monoclonal anti-HN antibodies were divided into 7 groups on the basis of their inhibitory reactivities to hemolytic, neuraminidase and hemagglutinating activities. Antibodies in the first and third groups inhibited to the same extent both hemagglutinating and neuraminidase activities. Antibodies in the second and fourth groups blocked the hemagglutinating activity, but not the neuraminidase activity. These findings suggest that there are 2 different types of antigenic determinant involved in both hemagglutinating and neuraminidase activities. One determinant is associated with both activities, whereas the other determinant is involved in each individual biological activity. Furthermore, antibodies in the third group inhibited the hemagglutinating activity of virions only, while antibodies of the fourth and fifth groups inhibited the isolated HN glycoprotein only. These results suggest that the antigenic determinants associated with the hemagglutinating activity may be located both in the protein embedded in the lipid bilayer and on the outside of the envelope in Sendai virus. In addition, there appear to be several antigenic determinants embedded in the lipid bilayer, since the antibodies of the fourth and fifth groups reacting with such determinants blocked the hemagglutinating activity differently than the neuraminidase activity. These results suggest that the antigenic determinants associated with the neuraminidase activity may also be more than one. Taken together, the antigenic determinants involved in the biological activities of HN glycoprotein are suggested to be two or more.

Biological activity of monomeric Sendai virus HN glycoprotein

Two HN glycoproteins of different molecular size were isolated from egg-grown Sendai virus by ion-exchange chromatography on DEAE-Sepharose CL-6B, and gel filtration on Sephadex G-200 at pH 7.2. The larger HN glycoprotein showed both hemagglutinating and neuraminidase activities, whereas the smaller one possessed only the neuraminidase activity. Analysis of their molecular weights showed that the larger HN glycoprotein is a homodimeric molecule of the smaller one. These observations suggest that multi- or divalent binding-sites are required for the expression of hemagglutinating activity and that the monomeric HN molecule itself possesses the enzyme activity.

Biological activity of monomeric Sendai virus HN glycoprotein

Two HN glycoproteins of different molecular size were isolated from egg-grown Sendai virus by ion-exchange chromatography on DEAE-Sepharose CL-6B, and gel filtration on Sephadex G-200 at pH 7.2. The larger HN glycoprotein showed both hemagglutinating and neuraminidase activities, whereas the smaller one possessed only the neuraminidase activity. Analysis of their molecular weights showed that the larger HN glycoprotein is a homodimeric molecule of the smaller one. These observations suggest that multi- or divalent binding-sites are required for the expression of hemagglutinating activity and that the monomeric HN molecule itself possesses the enzyme activity.

Analysis of antigenic determinants on internal and external proteins of influenza virus and identification of antigenic subpopulations of virions in recent field isolates using monoclonal antibodies and immunogold labelling

An electron microscopic immunogold labelling technique employing monoclonal antibodies has been applied to the antigenic analysis of influenza A and B viruses. Reassortant influenza A H3N2 viruses containing haemagglutinin molecules from viruses isolated between 1968 and 1982 were analysed with a panel of monoclonal antibodies raised against viruses which appeared over the same period. The immunogold labelling technique clearly demonstrated the antigenic drift in the haemagglutinin molecule that occurred between 1968 and 1982. When the technique was applied to the examination of viruses from a more geographically restricted influenza epidemic in a semi-closed community, antigenic variants were found. Furthermore the technique enabled the identification of distinct antigenic variant subpopulations within a single clinical isolate. Analysis of the HA of MDCK cell or egg grown virus by this procedure provided data to support the hypothesis that the host cell exerts selective pressure on subpopulations of virus resulting in the emergence of antigenic variants.