Flu has Mx'd-A-Lot to evade MxA

A large body of research has focused on how the adaptive immune system drives viral evolution. In the adaptive immune response, T-cells and antibodies target specific viral components and this evolutionary pressure selects for viral mutations that allow the virus to avoid detection. In addition to this adaptive system, the innate immune response generally protects organisms from foreign substances. Upon infection, proteins called interferons are produced and instigate the expression of restriction factors, proteins that target viruses and prevent them from carrying out their function. MxA is a major human restriction factor that has been shown to inhibit the influenza virus by binding the viral nucleoprotein (NP). The details of how MxA binding to NP inhibits influenza virus replication remain unclear.

A complete understanding of human influenza evolution would include understanding which viral mutations are required to evade both the adaptive and innate immune response. Recently, the development of a method called deep mutational scanning has allowed researchers to create “libraries” of thousands of protein variants, where each variant typically has a single mutation. Researchers in the Bloom Laboratory (Basic Sciences Division) extended these methods to generate libraries of mutant viruses and then measure the virus’ ability to replicate. In a recent publication in PLoS Pathogens, they report the results of their research testing which variants of human influenza can and cannot escape inhibition by MxA. “To understand how hard or easy it is for a virus to evade a restriction factor, we need to know what mutations that virus can make to avoid being recognized by the restriction factor,” said lead author Orr Ashenberg. “In our work, we developed a comprehensive and quantitative approach to identify all such mutations.”

Schematic of the workflow of the screen. A library of influenza variants was used to infect different lines of canine cells (MDCK-SIAT1 cells) which either did or did not express human MxA. Following 48 hours of growth in these cells, viral RNA was harvested and deep sequenced. After growth in cells expressing MxA, many variants were not represented in the final population of viruses after 48 hours. Therefore, these lost variants caused sensitivity to MxA. Image courtesy of Orr Ashenberg.

The researchers had previously created a library of human influenza viruses including nearly every amino acid variation in the nucleoprotein (NP) that supports viral growth. In order to identify “escape” mutations that allow influenza to evade inhibition by MxA, they wanted to compare the growth of these different influenza viruses in cells expressing versus not expressing human MxA. To do this, they chose to use a canine cell line because its particular MxA variant had been previously shown not to inhibit the growth of human influenza. Therefore, they could create canine cell lines which either did not express or which constitutively expressed human MxA.

The scientists measured viral titers at 24, 48, and 72 hours post-infection and found that the cells expressing human MxA had lower viral titers than cells without MxA, indicating reduced viral replication. This confirms that human influenza virus is indeed inhibited by active MxA in this experimental context. They isolated the viral RNA at 48-hours post-infection and used overlapping paired-end sequencing to determine the identity and frequency of the thriving variants. They defined mutations that were more frequently found in viruses infecting cells that were not expressing MxA as sites that affect sensitivity to MxA.

Figure 4 from the publication. (A) The 29 sites that were differentially selected in experimental versus control selection. They are ordered from highest to lowest differential selection from left to right, with the position of the amino acid written on the horizontal axis. Red indicates low similarity to avian influenza. (B) The 29 sites highlighted on the crystal structure of nucleoprotein (PDB-3ZDP). In this case, blue indicates sites that were previously identified to confer resistance to MxA and the purple site 283 was found previously and in this study. Red sites were newly identified as important for MxA resistance by this work. Image from the publication.

Ashenberg and his colleagues identified 29 amino acid positions that were differentially selected against by MxA. That is, the lack of growth of viruses carrying those mutations in cells expressing MxA exceeded the background growth frequencies under the control selection (cells not expressing MxA). The amino acids that were the most differentially selected against were mutations deviating from the wildtype (baseline) sequence of nucleoprotein. Therefore, this shows that these 29 sites in the influenza nucleoprotein confer resistance to human MxA. Interestingly, 26 of 29 of these sites are identical in human and avian influenza sequences. “This shows that even though human influenza has more resistance mutations than avian influenza, both of them are pretty resistant compared to what is probably possible,” said Dr. Bloom.

“Previous studies of virus escape from restriction factors has focused on the mutations that happened during host jumps,” continued Dr. Bloom. “The cool thing about Orr's work is that he examined how all mutations to influenza affect its resistance to host restriction.” Such systematic and quantitative work is critical to increase our understanding of how influenza evolves resistance to the host immune system, enabling more effective drug design.

Ashenberg O, Padmakumar J, Doud MB, Bloom JD. 2017. “Deep mutational scanning identifies sites in influenza nucleoprotein that affect viral inhibition by MxA.” PLoS Pathogens. 13(3): e1006288.

This research was supported by the National Institutes of Health, the Burroughs Wellcome Fund, and the PhRMA Foundation.