The NS1 protein of influenza A virus has been implicated in a plethora of functions related to the regulation of gene expression in infected cells. Some of these described functions, such as the inhibition of mRNA polyadenylation, splicing, and nucleocytoplasmic transport, would result in inhibition of gene expression and possibly in shutoff of host protein synthesis (4
). Some others, such as the inhibition of PKR (2
) and of IFN-α/β-related pathways (28
), would result in sustained protein synthesis during viral infection, which otherwise should induce the host IFN-α/β antiviral system. Thus, in the absence of NS1 protein expression, activation of PKR in virus-infected cells would lead to the phosphorylation and inactivation of the translational factor eIF-2α, resulting in protein synthesis inhibition (11
). We have investigated here the overall effects of the NS1 protein on cellular and viral gene expression. Transfection experiments resulted in the enhancement of reporter gene expression when reporter plasmids were cotransfected with an NS1 expression plasmid (Fig. to ). This effect did not appear to be promoter specific, since it was seen with both simian virus 40- and cytomegalovirus-based reporter plasmids.
The NS1-mediated enhancement of reporter plasmid expression was found to occur at a posttranscriptional level (Fig. ). Interestingly, a translational enhancement activity of the NS1 protein has already been proposed (1
). However, it was believed that this effect is specific for a subset of influenza virus mRNAs (8
). While we cannot exclude the possibility that the NS1 protein preferentially enhances the translation of some or of all viral mRNAs, our results show that the NS1 protein, in the absence of any other influenza virus protein, functions as a general enhancer of protein translation. In fact, the effects of NS1 expression in reporter gene activity are comparable to or even more dramatic than the effects of a known translational enhancer molecule, the adenoviral VAI RNA (Fig. and ). Further experiments will be required to completely understand the effects of NS1 in viral and cellular translation, especially as relates to its ability to interact with the eukaryotic initiation factor 4GI (1
Three different lines of indirect evidence suggest that the NS1-mediated translational enhancement is at least partly a result of the inhibition of PKR activity in transfected cells. First, coexpression of VAI RNA and NS1 protein does not have a cooperative effect in reporter gene expression (Fig. ). Second, expression of NS1 partly overcomes the translational inhibition induced by overexpression of PKR (Fig. ). Third, a mutant NS1 impaired in its ability to inhibit PKR is also compromised in translational enhancement activity (Fig. ). These results are also in agreement with the known anti-PKR activity of the NS1 protein of influenza A virus (2
). However, at this point we cannot rule out the involvement of other cellular proteins besides PKR in the translational enhancement mediated by the NS1 protein.
Only recently has it become possible to study the effects of the NS1 protein on viral expression by using recombinant mutant viruses containing selected mutations in their NS1 genes (7
). We have taken advantage of different influenza A virus recombinants containing deletions in their NS genes to study the effects of the NS1 protein on protein expression during influenza A virus infection. A number of different viruses, including influenza A viruses, induce cellular protein synthesis shutoff while maintaining viral protein synthesis in infected cells. It has been postulated that the NS1 protein affects cellular mRNA processing and that this protein may be responsible for the influenza virus-induced protein synthesis shutoff (4
). Therefore, we investigated the overall levels of cellular protein synthesis in cells infected with wild-type virus, with NS1-99 virus lacking the NS1 carboxy-terminal effector domain postulated to be involved in inhibition of mRNA processing, and with delNS1 virus. In order to distinguish effects due to NS1 protein expression from those due to low levels of viral replication, we used IFN-deficient Vero cells, in which the levels of replication of wild-type and NS1 mutant viruses are comparable (7
). Since no major differences in overall protein expression were found among the three viruses, we conclude that the NS1 protein does not play a major role in host cell protein synthesis shutoff. Clearly, some other viral product(s) plays a more prominent role than the NS1 protein in the shutoff of cell protein expression in influenza virus-infected cells. Our results are in agreement with recent findings from Zürcher et al. (34
) showing identical levels of protein synthesis shutoff in cells infected with mutant influenza A viruses expressing truncated NS1 proteins. These mutants express NS1 proteins of 81 and 156 amino acids which lack the effector domain. However, Enami and Enami (9
) recently reported impaired protein synthesis shutoff in MDBK cells infected with an influenza A virus mutant expressing an NS1 protein of 110 amino acids. The reason for this discrepancy could be that this NS1 truncated protein has unusual characteristics in MDBK cells. Furthermore, this group did not have the benefit of working with a virus lacking the entire NS1 gene, and no experiments were done in IFN-deficient cell systems.
The NS1 protein has also been implicated in the inhibition of mRNA splicing during influenza A virus infection (10
). Specifically, it was shown that expression of the NS1 protein results in the inhibition of splicing of the M-specific mRNA in transfection assays, resulting in high levels of unspliced M1 mRNA and low levels of spliced M2 mRNA (17
). However, in our experiments no changes in the levels of M1/M2 protein expression were found in cells infected with delNS1 virus, demonstrating that the NS1 protein does not have a major impact on the levels of M1 and M2 protein expression in virus-infected cells.
Despite all of the proposed functions of the carboxy-terminal domain of the NS1 protein in overall cellular mRNA processing, we were unable to detect major changes on overall protein expression in IFN-deficient Vero cells infected with delNS1 virus or with a recombinant virus expressing an NS1 protein lacking the effector domain. In fact, only when COS-7 cells were used did we find that the presence of the NS1 protein is required for preventing a general inhibition of viral protein expression in infected cells. These results are consistent with a role of NS1 in inhibiting antiviral responses in cells having an intact IFN system. Thus, no major effects on viral protein expression are expected when the NS1 protein is expressed in infected Vero cells unable to produce IFN and therefore unable to induce high levels of PKR and other antiviral proteins encoded by IFN-stimulated genes. However, infection of COS-7 cells with delNS1 virus most likely results in the stimulation of IFN synthesis, leading to the transcriptional induction of IFN-stimulated genes, such as PKR, and to the inhibition of protein synthesis. In fact, activation of PKR readily occurs when IFN-competent substrates are infected with delNS1 virus or with NS1 temperature-sensitive (ts
) mutant viruses at a nonpermissive temperature (2
). The NS1-99 virus displayed an intermediate phenotype between delNS1 and wild-type viruses in COS-7-infected cells. Thus, while HA and M1 protein expression was reduced in cells infected with this virus, NP expression was comparable to that observed in cells infected with wild-type virus. Similar results have been reported with other NS1 mutant viruses (7
). Interestingly, we found that the NS1-99 protein was poorly expressed in virus-infected cells, suggesting that this truncated protein is unstable (data not shown). Therefore, we also used in these experiments NS1-126 virus, expressing a stable NS1 protein of 126 amino acids lacking an effector domain (33
). The results clearly show that this mutant virus induces a wild-type pattern of viral and cellular protein expression in infected COS-7 cells. Thus, the effect seen in NS1-99 virus-infected cells most likely reflects low levels of NS1 protein expression. Nevertheless, we cannot exclude the possibility that a domain between amino acid residues 99 and 126 is required for efficient expression of the HA and M1 proteins. In any case, it is intriguing that the expression of the late HA and M1 proteins is more sensitive to NS1 regulation than the expression of the early NP protein. A slight defect in M1 protein expression was also observed in Vero cells infected with NS1-99 and delNS1 viruses (Fig. ). These effects might be attributed to activation of basal levels of PKR in Vero cells and/or to decreased or lack of interaction between NS1 and eIF-4GI, resulting in a reduced translational rate for the M1 mRNA.
The necessity for a virus to take over the cellular synthetic machinery in order to replicate its own genome is balanced by the necessity that the cell maintain a basal level of synthesis sufficient to permit its survival until the end of the replication cycle. The NS1 protein of influenza A virus appears to inhibit PKR and other dsRNA-activated pathways to maintain high levels of viral protein synthesis. This effect is mainly due to the amino-terminal, RNA-binding domain of this protein. Although it is still possible that the expression of selected cellular genes might be inhibited by the NS1 carboxy-terminal (effector) domain, we have disproved a major role of the NS1 protein and its effector domain in the overall inhibition of protein synthesis in cells transfected with an NS1 expression plasmid or in influenza A virus-infected cells. Therefore, we conclude that non-NS1-mediated mechanisms are responsible for the observed generalized host shutoff of cell protein synthesis during influenza virus infection.