Passaging of an influenza A(H1N1)pdm09 virus in a difluoro sialic acid inhibitor selects for a novel, but unfit I106M neuraminidase mutant.

An influenza A(H1N1)pdm09 and an influenza B virus were passaged in 3-fluoro(eq)-4-guanidino difluoro sialic acid (3Feq4Gu DFSA), an inhibitor of the influenza neuraminidase (NA) to determine whether resistant variants could be selected. 3Feq4Gu DFSA is a mechanism-based inhibitor, forming a covalent link to Y406 in the NA active site. Given its similarity to the natural substrate, sialic acid, we predicted resistant variants would be difficult to select. Yields of both viruses decreased with passaging, so that after 12 passages both viruses were only growing to low titers. Drug concentrations were decreased for another three passages. There was no difference in NA sensitivity in the MUNANA fluorescence-based assay, nor in plaque assays for the passaged virus stocks. All influenza B plaques were still wild type in all assays. There were isolated small diffuse plaques in the P15 pdm09 stock, which after purification had barely detectable NA or hemagglutinin (HA) activity. These had a novel non-active site I106M substitution in the NA gene, but unexpectedly no HA changes. The I106M may impact NA function through steric effects on the movement of the 150 and 430-loops. The I106M viruses had similar replication kinetics in MDCK cells as wild type viruses, but their ability to bind to and infect CHO-K1 cells expressing high levels of cell-bound mucin was compromised. The I106M substitution was unstable, with progeny rapidly reverting to wild type by three different mechanisms. Some had reverted to I106, some had V106, both with wild type NA and HA properties. A third group retained the I106M, but had a compensating R363K substitution, which regained almost wild type NA properties. These viruses now agglutinated chicken red blood cells (CRBCs) but unlike the I/V106, they rebound after elution at 37 °C. There were no mutations in the HA, but each phenotype correlated with the NA sequence. We propose that the activity in the I106M mutant is insufficient to remove carbohydrates from the virion HA and NA, sterically limiting HA access to CRBC receptors, thus resulting in poor HA binding. Crown