The Road to Understanding NS1 Functions: The Non-Structural Protein that became Lindenmann's Reverse Interferon
When the influenza virus genome was originally mapped (Palese 1977; Shaw and Palese 2013), it became a challenge to assign a role for the non-structural proteins found in infected cells, but apparently not present in the virion. Among them was the non-structural protein 1 (NS1) (Lazarowitz et al. 1971), a highly expressed protein originated from the collinear mRNA transcribed from segment 8 in the influenza virus genome. The same segment also encoded another non-structural protein, NS2, generated by alternative splicing of segment 8 mRNAs and that was renamed NEP (nuclear export protein) after its functions became clearer and the protein was also found in virions (O'Neill et al. 1998). Some initial studies with temperature sensitive mutants (Koennecke et al. 1981; Shimizu et al. 1982) showed that segment 8 was required for the efficient replication of the virus, and hinted that NS1 might be playing a role in the regulation of viral gene expression, yet it was unclear how could be so given that NS1 was not necessary for in vitro replication of the viral segments, that only required the polymerase subunits and the nucleoprotein (Huang et al. 1990) (Fig. 1).
Among the first features that could be ascribed to NS1 was its ability to bind different species of RNA (vRNA, dsRNA, polyA tails and snRNA) (Hatada et al. 1992; Hatada and Fukuda 1992; Qiu and Krug 1994; Qiu et al. 1995) and, related to this, the NS1 was found to be a regulator of host gene expression by blocking splicing of pre-mRNA and export of polyadenylated mRNAs out of the nucleus (Alonso-Caplen and Krug 1991; Alonso-Caplen et al. 1992; Fortes et al. 1994;
Fig. 1 a–c Features, interactors, and structure of the influenza A virus NS1 protein. a Schematic representation of the primary structure of an NS1 monomer, highlighting its most representative features and the regions known to be required for its multiple interactions. b The NS1 dimer, adapted from the X-ray structure described for the H5N1 A/Vietnam/1203/2004 NS1 (Bornholdt and Prasad 2008; PDB ID: 3F5T). The RNA-binding domains (RBD) are shown interacting, while the effector domains (ED) are in a monomeric state. Both the linker and the disordered tail regions of the protein, not present on the original structure, are hinted as a dashed line. c Residues and regions involved in the best characterized interactions of NS1. The color code is the same as in (a) for comparison purposes
Qian et al. 1994; Qiu and Krug 1994). When a little later it was described that NS1 interacts with and inhibits the cellular polyadenylation machinery (Nemeroff et al. 1998), it became evident that this viral protein was efficiently shutting down host gene expression at a post-transcriptional level. The implications of this in viral replication remained unclear. At first it was described that the effects of NS1 on host mRNAs resulted in a steady pool of sequestered mRNAs in the nucleus available for the viral polymerase to cap-snatch and transcribe viral mRNA (Nemeroff et al. 1998). At this point the NS1 protein was mainly considered a posttranscriptional gene regulator, but the interferon pathway was not into the picture yet (Fig. 2).
History of interferon (IFN) research runs in parallel with that of influenza virus. It was treating cells with heat-inactivated influenza viral particles that Isaacs and Lindenmann (1957) first discovered the secretion of this cytokine with antiviral properties. Soon afterwards, Lindenmann reported that cells infected with a live influenza virus were not producing interferon upon subsequent treatment with the inactivated virus particles, an inhibitory phenomenon he denominated ''inverse interference'': some factor related to viral replication had to be inhibiting interferon production (Weber et al. 2004; Lindenmann 1960). The first interferonstimulated gene (ISG) found to prevent viral infections in vivo was identified for its ability to confer resistance to influenza virus, hence its naming as Mx, from ''mixovirus resistance'' (Lindenmann 1962; Horisberger et al. 1983). Another ISG, the antiviral translational repressor PKR, was found to be inhibited during influenza virus infection (Katze et al. 1986), and it was the first connection between the NS1 protein and the interferon response as it was shown that the PKR inhibitory effect was, at least in part, due to NS1 (Lu et al. 1995) (Fig. 3).
Research on NS1 up to this point was limited to biochemical assays, transfection studies, and the use of temperatures sensitive mutants. The development of reverse genetics techniques allowed directed and selective manipulation of the influenza virus genome (reviewed by Palese et al. 1996), and revealed the identity of Lindenmann's ''inverse interferon'' encoded by influenza virus. Recombinant viruses carrying truncated forms of NS1 (Egorov et al. 1998) or devoid altogether of the NS1 gene (delNS1) (García-Sastre et al. 1998) were severely attenuated in IFN competent cells while still growing efficiently in Vero cells, which cannot produce a/ß interferons. Furthermore, the delNS1 virus was lethal in STAT1-/mice lacking a transcription factor required for IFN action, and induced higher activation of ISG promoter than its wild type counterpart. For the first time it was postulated that the NS1 protein was a virally encoded antagonist of the interferon response, and thus became the first of such inhibitors to be described for a negative-stranded RNA virus. While interferon-inhibitor accessory proteins had by then been described for other viruses (Weber et al. 2004), the limited genome size and small number of proteins encoded by influenza viruses had made it difficult to believe that one of them could be nonessential and mostly devoted to oppose the host innate immune response, as the delNS1 virus demonstrated. The exploration of the mechanisms by which NS1 exerts its inhibitory effect advanced in parallel with the increasing knowledge on the innate immune response pathways: the discovery
Fig. 2 Overview of the functions of the influenza A virus NS1 protein in the cytoplasm and nucleus of infected cells. NS1 is the main IFN antagonist of influenza virus, and also plays a critical role in the viral takeover of the cellular gene expression machinery. NS1 inhibits the IFN pathway at a pre-transcriptional level by blocking the activation of the RIG-I sensing and signaling platform, and at a post-translational level by repressing the activation of the anti-viral genes PKR and OAS. In the nucleus, NS1 inhibits the correct processing and export of mRNA, thus hindering all cellular gene expression including that of IFN-related genes. At the same time, NS1 enhances the production of viral mRNA, and in turn the specific translation of the latter by interacting with several translation-related factors. As a result of all these functions, NS1 renders the cell unable to properly respond to the viral infection or to alert neighboring cells of the danger. The reported activation of PI3K by NS1 may change the expression profile of the cell, inhibiting premature apoptosis or inducing other unknown changes on the cellular environment
Fig. 3 a–c Structural versatility of the NS1 protein. a Variability of the orientation of the ED respect to the RBD on the crystal structures of different NS1 proteins: A/Vietnam/1203/2004 (H5N1)(Bornholdt and Prasad 2008), A/blue-winged teal/MN/993/1980 (H6N6) and A/bluewinged teal/MN/993/1980 carrying a deletion on the linker region similar to that of the H5N1 protein (Carrillo et al. 2014). b Homodimerization of the NS1 ED. Depiction of one of the variable helix-helix interfaces between the two ED, showing the central, pivotal position of tryptophan 187 (Adapted from Kerry et al. 2011; PDB ID: 3O9S). c Proposed oligomerization of NS1 upon binding to dsRNA (Aramini et al. 2011). While the RBD on each NS1 dimer binds to the central dsRNA molecule, neighboring effector domains interact through the W187-mediated helix-helix interface promoting a spiral oligomerization around the RNA
of the cytoplasmic helicase RIG-I as the main sensor of influenza virus infection (Yoneyama et al. 2004) pointed toward a possible target for pre-transcriptional repression of innate immunity, and ever since numerous publications have addressed how NS1 limits RIG-I signaling, as discussed below.
With its role as an interferon antagonist clearly established, the actual strategy used by NS1 remained unclear. Some experimental setups pointed toward a prevalence of the originally described general gene expression shut off (Nemeroff et al. 1998) as the main mechanism to prevent antiviral inducible gene expression, while others highlighted the pre-transcriptional inhibition of the RIG-I/IRF3 axis (Pichlmair et al. 2006; Mibayashi et al. 2007). This apparent dichotomy was resolved thanks to growing evidence of a strain-dependent behavior of the NS1 protein (Kochs et al. 2007). Indeed, the subtleties of NS1 performance were soon revealed as complex as the ever-growing array of functions it has been related to; including a very well structurally-characterized—but functionally obscureactivation of the PI3K signaling pathway (Hale et al. 2006), a critical regulator of cell fate. In the following paragraphs, we will summarize the broad scope of the NS1 functions, their regulation and the—still not completely understood—effects they have in influenza virus pathogenesis.