Codominant mixtures of viruses in reference strains of influenza virus due to host cell variation

Influenza viruses grown in chicken eggs may comprise mixtures of variants, creating problems in establishing international reference strains and in preparing high growth reassortants. We therefore analyzed representative reference strains of H3N2 viruses from 1987 to 1989 by direct sequencing of HA1. Three of seven reference strains had different nucleotides at the same position in nucleotide gels, indicating the presence of codominant mixtures. These nucleotide duplications occurred at residues previously shown to code for amino acids associated with egg adaptation (156, 186, and 193 of HA1). Cloning of these viruses in chicken eggs permitted separation of the mixtures, and the majority of these cloned viruses could be distinguished with monoclonal antibodies. The remaining four reference strains were homogeneous and contained one of the two amino acids usually found at these residues in HA1 (e.g., 145 Glu or Lys, 186 Ser or Ile). Analysis of epidemic H3N2 isolates, for which mammalian cell and egg isolates are available for sequence analysis from the same patient, confirmed that multiple nucleotide changes can occur at residues associated with egg adaptation. When reference or epidemic strains containing codominant mixtures were passaged in eggs one to five times, one of the codominant strains usually became dominant. Similar results were obtained with Madin Darby Canine kidney (MDCK) cells, although the dominant virus usually differed from that in eggs. Attempts to reselect an influenza virus possessing sequence changes in HA1 characteristic of mammalian cells or the original human isolate (i.e., 156 Glu, 158 Glu, 186 Ser) by multiple passages in MDCK cells were not successful, but evidence was obtained that MDCK cells can provide a selective growth advantage. Thus, variants that are dominant in eggs are not necessarily dominant in MDCK cells. To preserve the original genotype of viruses used as reference strains, we recommend the following procedure: (i) cloning in chicken eggs of the candidate virus at a very early passage, (ii) selection and analysis of multiple clones with the same ferret polyclonal and/or monoclonal antibodies used in the initial screening, and (iii) selection of the isolate whose hemagglutinin molecule most closely resembles the clinical isolate.