How does a two-protein segment produce its products in the right order and ratio?
Influenza A virus has eight segments driven by comparable promoters, and yet it must order nuclear entry, genome replication, nuclear export and assembly in time. It has no transcriptional clock, so it cannot delay a protein by transcribing its segment later. Segment 8 is bicistronic. The unspliced transcript encodes the interferon antagonist NS1, and a 5 prime splice site used only ten to fifteen percent of the time yields the nuclear export protein NEP, which is required for export of nucleoprotein-associated genomes and has also been implicated in complementary RNA levels, host tropism and small viral RNA production. Chua 2013 asks whether that inefficiency is functional or incidental. A second problem sits underneath it. Influenza splicing has never been reconstituted in vitro and has been argued to be noncanonical, so the mechanism that executes it, and therefore the thing that sets its rate, was unidentified. Oishi 2023 takes that as its entry point.
Chua 2013 separated the two segment 8 products and moved each in both directions. Inserting perfect target sites for host microRNAs into a modified NS vector silenced NS1 by more than ninety percent without touching NEP, with silencing following the microRNA content of the cell and reversed in cells lacking Dicer1. That manipulation had little effect on titres or on interferon-regulated gene induction in cell lines, in primary lung fibroblasts and macrophages, or in mice, which the paper describes as surprising given that NS1-deleted virus is heavily attenuated. NEP was then lowered by a 2A recoding arrangement and by small interfering RNA, each of which cost titre, and raised by an extra copy from a replication-incompetent vector and by optimising the 5 prime splice site so that NEP became the dominant product. Both increases cost about two logs in culture, and the splice-optimised virus barely replicated in mice. Nucleoprotein appeared in the cytoplasm as early as five hours with that virus while remaining nuclear in the matched parental control, and the NEP-reduced virus showed the converse delay.
The reading the authors propose is that the poor splice site is a molecular timer. NS1 is overproduced not as waste but because NEP is generated as its minor spliced byproduct, so NEP concentration rises slowly and crosses the threshold for ribonucleoprotein export only after replication has proceeded. Coupling the timer to NS1 is advantageous on this account because excess NS1 is tolerated while excess polymerase, nucleoprotein or surface protein would not be. What the paper established is that the amount of NEP and the rate at which it accumulates determine the outcome, and that the splice site sets that rate. The step at which mistimed export becomes lethal is not identified, and no experiment restores correct timing to rescue the titre defect.
Oishi 2023 came at the same processing step with a heterologous tool. A screen of codon-optimised archaeal RNA binding proteins identified the kink-turn binding protein L7Ae, which eliminated the spliced products M2 and NS2 while the unspliced M1 and NS1 accumulated. The effect reproduced without infection and without viral polymerase, required nuclear L7Ae, and disappeared against the splicing-independent 2A virus built in the Chua line of work, which is the decisive genetic test that the target is splicing. Only archaeal orthologues had the activity, and the archaea-defining residues contributed most, while substitutions that abolish canonical box C and D binding did not, which argues that the viral substrate deviates from a canonical kink turn. Host transcriptome, proteome and an intron-containing reporter were largely unchanged. Sensitivity mapped by chimeric minigenes to roughly twenty nucleotides of intron upstream of the 3 prime splice site plus about nine nucleotides of coding sequence, in both segment 7 and segment 8. The same protein cost influenza A virus two to three logs and influenza B virus one to two logs, and blocked splice product formation in infectious salmon anemia virus while leaving vesicular stomatitis virus untouched.
The decade between the two papers changes the question from whether splice site efficiency matters to what executes the splicing and how tightly the arrangement is constrained. Chua 2013 showed the timer can be broken by raising or lowering the rate. Oishi 2023 showed the virus cannot easily route around an inhibitor of the step, since twenty passages under selection produced no true escape, only a segment 8 G60A variant that raises splicing efficiency at a severe fitness cost and is outcompeted by wild type within two passages without the inhibitor. Taken together the two papers argue that splice site suboptimality is the mechanism by which the virus schedules its cycle, and that this is a family-level property rather than a feature of influenza A virus alone, which is a synthesis visible from the pair rather than a claim either makes alone.
The mechanism remains unresolved. Oishi 2023 states its account as a postulate, that a kink-turn-like structure forms transiently during lariat formation and that L7Ae occludes it. No canonical kink turn was found in the viral genomes, crosslinking immunoprecipitation produced no footprint, and the mapping data are equally consistent with an indirect effect. The laboratory has continued this line with work on a conserved hairpin in segment 8. Readers should note the declared conflict, since tenOever is a co-founder of Archean Biologics and an author on a patent covering commercialisation of L7Ae.
Chua 2013 established the timer and built the reagent set that titrates segment 8 output in either direction. Oishi 2023 separated orthomyxovirus splicing from host splicing experimentally and extended the property across three genera of the family. Both are lab-led.
The theme meets the small viral RNA work through NEP, which Perez 2010 and Perez 2012 identify as required for svRNA production and which Chua 2013 places as the likely trigger of the transcription to replication switch. It meets the NS1 theme through the same segment, since the finding in Chua 2013 that ninety percent of NS1 is dispensable for antagonism sets up the question that Munoz-Moreno 2019 addresses from the sequence-diversity side. Oishi 2023 also carries the laboratory's wider interest in borrowing molecules from other domains of life as antiviral tools, which sits outside this area.
Archaeal members of the L30 protein family, but not their orthologues from any other domain of life, block production of the spliced influenza A virus, influenza B virus and isavirus transcripts without measurably disturbing host splicing, identifying a shared and apparently noncanonical feature of orthomyxovirus splicing.
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.