How much of influenza fitness is set by its interferon antagonist?
NS1 is the product of the unspliced segment 8 transcript and is described in the prior literature as the dominant antagonist of the cellular response to influenza A virus infection. Viruses deleted for it are attenuated by more than four logs in animals while retaining wild-type virulence in hosts without antiviral signalling. Its documented activities include sequestration of double-stranded RNA away from OAS, RIG-I like receptors, PKR and MDA5, direct inhibition of RIG-I and PKR, and interference with host messenger RNA maturation and export through CPSF30, and several of these are strain specific. Two questions follow, and the two papers in this theme take one each. How much NS1 does a virus actually need, and how is NS1 function distributed across the sequence space the gene occupies in nature.
Chua 2013 is lab-led and answers the first question by accident of design. The study needed to silence NS1 without disturbing NEP, which it achieved by inserting tandem perfect target sites for host microRNAs into the intergenic region of a modified NS vector, using one ubiquitous microRNA and one restricted to haematopoietic cells, with a scrambled insert as the isogenic control and Dicer1-deficient cells to confirm dependence on the microRNA pathway. Losing more than ninety percent of NS1 protein and messenger RNA left titres comparable to control in lung epithelial cells, while an NS1-deleted virus was substantially impaired, and induction of the interferon-regulated gene Mx1 was unchanged in the silenced viruses and elevated only for the deletion virus. The same result held in primary lung fibroblasts and bone marrow-derived macrophages and in mice, in both interferon-competent and signalling-deficient animals.
The reading offered is that low levels of NS1 are sufficient to antagonise the host response, which the authors describe as surprising, and that the apparent overproduction of the NS1 transcript exists for a different reason, to pace the accumulation of NEP as its minor spliced product. NS1 on this account is regulated in a way that is incidental to its own function and load-bearing for the timing of another. The conclusion is bounded by its readouts, which are interferon-regulated gene induction and titre in the cells and mouse strain tested, and by the fact that silencing removes more than ninety percent of NS1 rather than all of it, so residual activity cannot be excluded.
Munoz-Moreno 2019 addresses the second question and is led by the García-Sastre laboratory, with tenOever listed under supervision and under writing review and editing, and the barcoded library design taken from the tenOever laboratory study of transmission bottlenecks by Varble 2014. It placed 56 natural NS1 sequences, spanning allele A and allele B and isolates from seven host groups between 1954 and 2013, into a common A/Puerto Rico/8/1934 backbone using a split NS segment that separates NS1 from NEP, each tagged with a neutral 22-nucleotide barcode and most carried by two independent barcodes. The pool was competed in MDCK cells, A549 cells, embryonated eggs and mice, and relative barcode abundance after replication defined a fitness landscape per host.
Fitness varied widely and varied by host. Allele B NS1 viruses were overrepresented in every substrate tested, which the split segment design attributes to NS1 rather than to segment 8 as a whole. Human H3N2 NS1 viruses were underrepresented everywhere except in human A549 cells. Network analysis found clusters sharing fitness profiles despite substantial amino acid divergence, so phylogenetic proximity predicts phenotype poorly for this gene. Loss of STAT1 in mice flattened much of the spread while loss of RAG1 did not, placing the selective pressure in innate signalling, although human H3N2 NS1 remained restricted without STAT1, which the authors read as an NS1-regulated restriction independent of interferon. No molecular mechanism for any individual difference is established, and the authors say the appropriate next step is protein interaction work.
The two papers do not form a lineage, and the honest account is that this theme is the least continuous in the area. Chua 2013 measures how much NS1 activity a single virus needs and finds the answer to be less than the virus makes. Munoz-Moreno 2019 measures how the activity of many NS1 proteins differs across hosts and finds it structured by allele and host rather than by phylogenetic distance. The methodological connection runs through the split NS segment and the barcode, both from this laboratory and both existing because segment 8 carries two products that have to be decoupled before either can be studied cleanly. That shared constraint is the substantive link, and it is a synthesis drawn from setting the two papers together rather than a claim in either.
Both papers also sit somewhat aside from the unifying argument of this area. NS1 is a dedicated antagonist protein, which is the kind of regulator the rest of the area shows influenza doing without for its own internal scheduling. What the pair contributes is negative and useful. The interferon antagonist is not where the virus keeps its timing information, and in Chua 2013 its abundance turns out to be in service of something else.
Chua 2013 supplies the finding that most NS1 is dispensable for antagonism in the settings tested, lab-led. Munoz-Moreno 2019 supplies the multi-host fitness landscape and is led by another laboratory, with this laboratory contributing the barcoded library method and supervision.
The theme shares segment 8 with the splicing theme and shares Chua 2013 with it outright, since the NS1 result and the NEP timer result come from the same set of viruses. Munoz-Moreno 2019 belongs also to the laboratory's work on viral populations and evolution, where the barcoded library method originates and where the question of what selection reads within a host is pursued further.
A library of 107 barcoded influenza A viruses differing only in their NS1 sequence, competed in dog cells, human cells, chicken eggs and mice, resolves NS1-driven fitness as a set of divergent and partly convergent evolutionary trajectories rather than a single ordered adaptation gradient.
The inefficient 5 prime splice site of influenza A virus segment 8 functions as a timing device, causing the nuclear export protein to accumulate slowly as a minor product of abundant NS1 transcription, with both raising and lowering that rate attenuating the virus through mistimed ribonucleoprotein export.