|dc.description.abstract||Influenza A virus (IAV) remains a seemingly undefeatable public health threat owing to its frequent antigenic drift and shift. Seasonal epidemics, occasional pandemics, and avian/zoonotic IAV outbreaks cause significant morbidity and mortality among human and other hosts, leading to substantial global economic burdens. Along the path of host-virus coevolution, mammalian hosts have evolved sophisticated cellular networks to combat viral infection, among which the innate immune system constitutes the first line of defense. Its rapid detection of the invading virus initiates an effective antiviral response which in turn primes a specific adaptive immunity. Therefore, an in-depth understanding of the interplay between IAV and host innate immune system not only provides knowledge on IAV pathogenesis but also sheds light on the development of novel approaches to combatting IAV infection.
The innate immune recognition of IAV relies on various families of pattern recognition receptors, among which the retinoic acid-inducible gene I (RIG-I), a founding member of the RIG-I-like receptor (RLR) family, is indispensable for the IAV-induced type I interferon (IFN) response. It has been well characterized as a cytoplasmic sensor that recognizes short double-stranded (ds) RNA harboring a 5’-triphosphate moiety. Interestingly, IAV bears a single-stranded RNA genome and does not generate a detectable amount of dsRNA during infection. Over the last decade, the panhandle structure, which is formed from the partial complementarity of the genomic RNA extremities of negative-strand RNA viruses, has been proposed to meet the dsRNA requirement for RIG-I activation. However, it remains unsubstantiated the contribution of panhandle structure in the genomic context of IAV to RIG-I activation. Therefore, in the first part of this thesis, I investigated whether the IAV panhandle structure is directly involved in RIG-I activation and type I IFN induction. Using reconstituted IAV genomic RNA with coding region truncations, it was demonstrated that the viral genomic coding region is dispensable for RIG-I-dependent IFN induction in primary alveolar macrophages. The in vitro-synthesized IAV panhandle RNA directly binds to RIG-I with 1:1 stoichiometry and stimulates RIG-I ATPase activity in vitro and RIG-I-dependent IFNβ promoter activation in DF-1 cells. These lines of evidence demonstrate a direct involvement of the IAV panhandle structure in RIG-I activation. Since the wild-type panhandle structure exhibits an imperfect double-strandedness containing wobble base pairs, mismatch, and bulged nucleotide, it was of particular interest to determine how these elements contribute to RIG-I activation. Elimination of these destabilizing elements from either the 5’ or 3’ arm of the panhandle structure enhances RIG-I binding and ATPase activity in vitro, and promotes RIG-I-dependent IFN induction in DF-1 cells and primary macrophages. Given that the IAV panhandle structure also serves as the viral promoter region, the promoter activity of panhandle-stabilized variants was monitored in an attempt to correlate with their RIG-I stimulatory activity. The panhandle-stabilizing mutations dramatically impair viral transcription, but only moderately affect viral replication. Collectively, these results indicate from an evolutionary perspective that the IAV panhandle promoter region adopts a nucleotide composition that is suboptimal for RIG-I activation and optimal for a balanced viral RNA synthesis, both of which are beneficial for efficient viral infection.
Compared to most of the other RNA viruses replicating in the cytoplasm, IAV has a unique site of replication – the nucleus. This nuclear-replicating strategy underscores a long-standing puzzle as to how the cytoplasmic RNA sensor RIG-I senses a nuclear-replicating virus such as IAV. In the second part of this thesis, I investigated the spatiotemporal detection of IAV by RIG-I during infection. Notably, a trace of nuclear-resident RIG-I was identified which contributes to nuclear IAV sensing. Using NS1 deletion viruses that are defective of IFN antagonism, it was shown that RIG-I activation during a single-cycle infection intimately associates with nuclear vRNA accumulation and that vRNA nuclear export is dispensable for RIG-I activation. Cellular fractionation followed by co-immunoprecipitation further demonstrates that endogenous nuclear RIG-I constantly associates with nuclear vRNP during the course of IAV infection. Ectopically expressed nuclear RIG-I efficiently mediates IFN and interferon-stimulated gene (ISG) induction in response to nuclear viral replication recapitulated by IAV RNP reconstitution. Complementation of RIG-I KO A549 cells with the nuclear RIG-I restores the canonical mitochondrial antiviral-signaling protein (MAVS)-dependent signaling axis inducing IFN, which in turn establishes an effective antiviral state restricting IAV infection. Strikingly, the exclusive signaling specificity conferred by nuclear RIG-I is reinforced by its inability to sense cytoplasmic-replicating Sendai virus and appreciable sensing of hepatitis B virus pregenomic RNA in the nucleus. Collectively, these results identify the existence of nuclear RIG-I which is actively involved in IAV sensing along with its cytoplasmic counterpart, and refine the RNA sensing paradigm for nuclear-replicating viruses by revealing a previously unrecognized subcellular milieu for RIG-I-like receptor sensing.
The identification of nuclear-resident RIG-I provides advanced insight into the spatiotemporal detection of IAV from the host perspective; however, the physiological viral ligands activating RIG-I during IAV infection remain underexplored. It is also of interest to speculate that the two different cellular pools of RIG-I may associate with distinct but overlapping groups of viral agonists. Other than full-length viral genomes, the viral defective interfering (DI) genomes and unknown aberrant RNAs are potential RIG-I agonists, which are proposed to be erroneously generated by the viral polymerase under cellular constraints impeding ongoing viral replication. In the third part of this thesis, I interrogated the origins of RIG-I activating viral RNA under two such constraints that may potentiate a malfunctioning viral polymerase. Using chemical inhibitors that inhibit viral protein synthesis, it was shown that the incoming but not de novo synthesized DI genomes specifically contributing to cytoplasmic RIG-I activation. In comparison, deprivation of viral nucleoprotein (NP) leads to the production of aberrant viral RNA species activating nuclear RIG-I, but their nature is likely to be distinct from DI RNA. Moreover, nuclear RIG-I activation in response to NP deprivation is not adversely affected by expression of the nuclear export protein (NEP), which diminishes the generation of a major subset of aberrant viral RNA but facilitates the accumulation of small viral RNA (svRNA). Accordingly, a specialized ligand characteristic activating nuclear RIG-I is proposed in which aberrant viral RNA or svRNA forms intermolecular RNA duplexes with their full-length vRNA or cRNA template. Collectively, these results indicate the existence of fundamentally different mechanisms of RIG-I activation under cellular constraints that impede ongoing IAV replication.
Overall, the three parts of this thesis establish the central role of IAV panhandle signatures in spatiotemporal RIG-I activation during IAV infection. Understanding the full spectrum of physiological RIG-I ligands and their relative contributions to IFN induction will facilitate the development of antiviral interventions to better control influenza infections.||