The Cellular DExD/H-Box RNA Helicase UAP56 Co-localizes With the Influenza A Virus NS1 Protein

UAP56, a member of the DExD/H-box RNA helicase family, is essential for pre-mRNA splicing and mRNA export in eukaryotic cells. In influenza A virus-infected cells, UAP56 mediates viral mRNA nuclear export, facilitates viral ribonucleoprotein complex formation through direct interaction with the viral nucleoprotein, and may indirectly affect antiviral host responses by binding to and/or facilitating the activation of the antiviral host factors MxA and PKR. Here, we demonstrate that UAP56 also co-localizes with the influenza A viral NS1 protein, which counteracts host cell innate immune responses stimulated by virus infection. The UAP56–NS1 association relies on the RNA-binding residues R38 and K41 in NS1 and may be mediated by single-stranded RNA. UAP56 association with NS1 does not affect the NS1-mediated downregulation of cellular innate immune pathways in reporter gene assays, leaving in question the exact biological role and relevance of the UAP56–NS1 association.


INTRODUCTION
UAP56, a member of the DExD/H-box RNA helicase family, is highly conserved from yeast to humans and plays a critical role in pre-mRNA splicing and mRNA nuclear export (Luo et al., 2001). The ATPase and unwinding activities of UAP56 are required for spliceosome assembly and maturation (Shen et al., 2007(Shen et al., , 2008. In addition, UAP56 functions as a component of the transcript export (TREX) complex to efficiently export spliced mRNAs to the cytoplasm (Jensen et al., 2001;MacMorris et al., 2003). URH49, a 90% homologous paralog of UAP56, has similar functions in mRNA processing (Pryor et al., 2004;Kapadia et al., 2006).
Influenza viruses replicate in the nucleus of infected cells and usurp the cellular nuclear export systems for the transport of viral mRNA from the nucleus to the cytoplasm. Accordingly, they rely on host factors such as UAP56/URH49 for efficient replication (Read and Digard, 2010;Wisskirchen et al., 2011b). UAP56/URH49 interacts with the influenza viral nucleoprotein (NP) to facilitate the formation of viral ribonucleoprotein complexes (Momose et al., 2001;Kawaguchi et al., 2011), and interacts with the cellular MxA protein (Wisskirchen et al., 2011a), an interferoninduced dynamin-like GTPase that restricts the replication of influenza A viruses (Pavlovic et al., 1992;Haller et al., 2015) by directly interacting with the viral NP protein (Manz et al., 2013). However, the significance of the UAP56/URH49-MxA interaction for the antiviral effect of MxA is unknown. UAP56/URH49-depletion has been shown to increase the accumulation of double-stranded RNA (dsRNA) in influenza virus-infected cells, resulting in the activation of protein kinase RNA (PKR; Wisskirchen et al., 2011b), a known antiviral factor (Gale and Katze, 1998). UAP56/URH49 may, therefore, be involved in regulating host antiviral responses in influenza virus-infected cells.
NS1 performs several roles in the influenza A viral life cycle (Wright et al., 2013). It is involved in the nuclear export of cellular and viral mRNAs (Fortes et al., 1994;Qiu and Krug, 1994) and is the major influenza A viral interferon (IFN) antagonist (Garcia-Sastre et al., 1998;Krug, 2015). It downregulates host antiviral responses by suppressing the activation of viral RNA receptor retinoic acid-inducible gene I (RIG-I; Pichlmair et al., 2006;Mibayashi et al., 2007;Opitz et al., 2007;Gack et al., 2009) and the downstream adaptor-protein IFN-beta promoter stimulator protein 1 (IPS-1; Mibayashi et al., 2007), resulting in reduced activation of IFN regulatory factor 3 (IRF3), a critical transcription factor for type-I IFN gene induction (Talon et al., 2000). Furthermore, NS1 inhibits PKR (Lu et al., 1995) and 2 -5 -oligoadenylate synthetase (OAS; Min and Krug, 2006), both of which are IFN-induced antiviral effector proteins. NS1 also affects viral replication through its PDZ domain binding motif (Jackson et al., 2008;Soubies et al., 2010) and by activating the PI3K/Akt pathway (Ehrhardt et al., 2006(Ehrhardt et al., , 2007Hale et al., 2006;Shin et al., 2007a,b,c). Moreover, NS1 binding to the 30-kDa subunit of the cellular cleavage and polyadenylation specificity factor CPSF30 leads to the nuclear accumulation of cellular mRNAs (including mRNAs encoding IFN) and hence to lower levels of these proteins (Nemeroff et al., 1998;Noah et al., 2003). Since NS1 and UAP56 affect similar cellular processes including pre-mRNA splicing, mRNA nuclear export, and the indirect suppression of antiviral responses, we asked whether UAP56 associates with NS1 and/or affects functions of NS1.

RNase Susceptibility Assay
HEK293T cells were transfected with pCAGGS vectors encoding WSN-NS1 and FLAG-tagged UAP (pCAGGS-FLAG-UAP56) or with the pCAGGS control vector. At 48 h post-transfection, cells were lyzed in lysis buffer containing 2 mM MgCl 2 and cell lysate was treated without or with 40 U/mL RNase III (BioLabs), 200 µg/mL RNase A (Thermo Scientific), or 40 U/mL RNase H (Invitrogen) at 37 • C for 20 min. The lysates were then incubated with anti-FLAG antibody-conjugated magnetic beads (4 • C, overnight), and co-precipitated proteins were analyzed by immunoblotting.

Indirect Immunofluorescent Analysis
A549 cells were infected with WT WSN or WSN-NS1-R38A-K41A mutant virus at a multiplicity of infection (MOI) of three. At the indicated time points post-infection, cells were fixed with 4% paraformaldehyde in PBS and permeabilized with 0.1% Triton X-100 in PBS. Mouse anti-UAP56 antibody (LS-C172345; LSBio) and rabbit anti-NS1 antibody (PA5-32243; Thermo Fisher) were used as primary antibodies. Alexa 488-conjugated anti-rabbit and Alexa 594-conjugated anti-mouse antibodies (Life Technologies) were used as secondary antibodies. Slides were mounted in mounting media with DAPI, and analyzed by using LSM510 META (Carl Zeiss). The immunofluorescence co-localization of NS1 and UAP56 was assessed by Pearson's correlation coefficient of red-and green-pixels, calculated by using the FIJI "coloc2" function 1 (Schindelin et al., 2012).

Reporter Assays to Assess the Inhibitory Effect of NS1 on Innate Immune Responses
To assess NS1's ability to suppress IFN-β promoter activity, HEK293T cells were transfected with p125Luc (a plasmid encoding firefly luciferase under the control of the IFN-β promoter; kindly provided by Dr. T. Fujita, Kyoto University, Japan), and with pRL-TK (a plasmid encoding Renilla luciferase under the control of the herpes simplex virus thymidine kinase promoter; Promega) as an internal transfection control. Cells were co-transfected with a constitutively active form of RIG-I (pCAGGS-RIG-I N; Yamayoshi et al., 2015), without or with increasing amounts of the protein expression plasmids for WSN-NS1 (pCAGGS-NS1) and UAP56 (pCAGGS-UAP56). The cells were cultured at 37 • C, and 24 h later lyzed to measure Firefly-luc and Renilla-luc (internal control) activity by using the Dual-Glo Luciferase assay system (Promega). Firefly luciferase activity was normalized to Renilla luciferase values. The pCAGGS control vector was added as needed to ensure that all wells of cells were transfected with the same amount of DNA.
To measure NS1's ability to interfere with the activation of IFN-stimulated response elements (ISREs), HEK293T cells 1 https://fiji.sc/ were transfected with a plasmid expressing firefly luciferase under the control of an ISRE (pISRE-Luc) and with pRL-TK (internal control) with increasing amounts of protein expression plasmids for WSN-NS1 (pCAGGS-NS1) and UAP56 (pCAGGS-UAP56). At 24 h post-transfection, the cells were treated without or with IFN-β (10,000 U/mL) and cultured for 24 h. Then, the cells were lyzed and the luciferase activity of the Fireflyluc and internal control Renilla-luc were analyzed as described above.

UAP56 Associates With NS1
First, we tested whether UAP56 associates with NS1. Human embryonic kidney HEK293T cells were transfected with protein expression vectors encoding FLAG-tagged UAP56 and the NS1 protein of A/WSN/33 (H1N1; WSN) virus. Overexpressed WSN-NS1 co-precipitated with UAP56 ( Figure 1A), indicating an association between these two proteins. We next tested whether the UAP56-NS1 co-precipitation was affected by the overexpression of MxA. NS1 did not directly interact with MxA ( Figure 1B), and increasing amounts of MxA did not abrogate the UAP56-NS1 co-precipitation ( Figure 1C).
We further examined the UAP56-NS1 association in virusinfected cells. FLAG-tagged UAP56 or a control vector was transfected into HEK293T cells, followed by infection with WSN virus at a MOI of one. At different time points after infection, proteins co-precipitated with UAP56 were analyzed for NS1 ( Figure 1D). As expected, NS1 co-precipitated with UAP56, demonstrating that NS1 associates with UAP56 in WSN virusinfected cells.

UAP56 Associates With NS1 Proteins of Different Influenza Viruses
To determine whether UAP56 associates with NS1 proteins from different influenza A viruses, we tested UAP56 coprecipitation with the NS1 proteins of pandemic 1918 virus (A/Brevig Mission/1/1918, H1N1; BM/1/1918), a representative of highly pathogenic avian H5N1 influenza viruses (A/Vietnam/1203/2004, VN1203), and a representative of the recently emerged avian H7N9 viruses (A/Anhui/1/2013, AH1). The NS1 proteins of all three viruses co-precipitated with UAP56 (Figures 2A-C), demonstrating that the UAP56 association with NS1 is not specific to WSN virus.
The RNA-Binding Residues in NS1 Are Critical for Its Association With UAP56 To examine the amino acid residues that mediate NS1-UAP56 association, we next tested WSN-NS1 mutants encoding R38A-K41A (defective binding to RNA), F103L-M106I (defective binding to CPSF30), or lacking the PDZ domain-binding motif ( PDM) for their binding affinity for UAP56 in a coimmunoprecipitation assay (Figure 3A). Co-precipitation of UAP56 with the NS1-R38A-K41A mutant was severely reduced (Figure 3B), demonstrating that the RNA-binding residues  in NS1 are critical for NS1-UAP56 association. By contrast, mutation of the CPSF30-binding residues or deletion of the PDZ domain binding motif did not abrogate the WSN-NS1 association with UAP56, although deletion of the PDZ domain binding motif reduced NS1's association affinity for UAP56 relative to that of WT NS1 (Figure 3B). To test the possibility that UAP56-NS1 association is mediated by nucleic acids, we examined the UAP56-NS1 association in cell lysates treated with RNase III (specific for dsRNA), RNase A (specific for ssRNA), and RNase H (specific for DNA-RNA duplexes). Treatment with RNase III or RNase H did not appreciably affect the UAP56-NS1 interaction ( Figure 3C). By contrast, RNase A treatment significantly reduced the extent of the UAP56-NS1 association (Figure 3C), establishing that this association is mediated by ssRNA.

UAP56 and NS1 Co-localize Near the Nuclear Membrane
To analyze the intracellular localization of UAP56 and NS1, we next infected human A549 cells with WSN virus and performed an indirect immunofluorescence analysis with antibodies to NS1 and UAP56. In non-infected cells, UAP56 was detected predominantly in the nucleus but was also present in the cytoplasm (Figure 4A, left panel). In WSN virus-infected cells, NS1 was expressed in the nucleus and cytoplasm ( Figure 4A). Further analysis showed that WT NS1 co-localized with UAP56 near the nuclear membrane (Figures 4B,C). On the other hand, in cells infected with WSN-NS1-R38A-K41A mutant virus, which encodes RNA-binding defective NS1, the mutant NS1 was dominantly expressed in the cytoplasm with limited co-localization with UAP56 (Figures 4D,E). Quantitative co-localization analysis confirmed NS1-UAP56 association in WT WSN-infected cells, whereas this co-localization was abrogated in cells infected with WSN-NS1-R38A-K41A mutant virus ( Figure 4F).

UAP56 Does Not Interfere With NS1-Mediated Suppression of RIG-I-Dependent IFN-β Induction and ISG Stimulation
To test whether overexpression of UAP56 affects the immune suppressive functions of NS1, we transfected HEK293T cells with a reporter plasmid expressing firefly luciferase under the control of an IFN-β promoter. These cells were also transfected with a vector expressing a constitutively active form of RIG-I, and with vectors expressing increasing amounts of NS1 and UAP56 (Figure 5A). WSN NS1 suppressed RIG-Idependent IFN-β promoter activation, but this activity was not significantly affected by increasing amounts of co-expressed UAP56.
Next, we tested the effect of increasing amounts of UAP56 on NS1's ability to suppress the induction of ISGs. Cells were transfected with a reporter plasmid possessing the firefly luciferase gene under the control of a promoter with an ISRE, and with increasing amounts of plasmids expressing NS1 or UAP56. Twenty-four hours later, the cells were stimulated with IFN-β for 24 h, and then assayed for luciferase expression. NS1 expression suppressed the activation of IFN-β-activated genes, but co-expression of UAP56 did not appreciably alter this effect ( Figure 5B).  Twenty-four hours later, the cells were lyzed and firefly and Renilla luciferase activities were measured. The firefly luciferase activity was normalized by the internal control value. The total amount of transfected vector was adjusted in all wells using the pCAGGS control vector. Data are shown as the mean ± SD (n = 3; biological replicates). UAP56 and NS1 expression levels in cells were examined by immunoblotting. (B) UAP56 overexpression does not affect NS1's ability to suppress ISRE-driven gene expression. HEK293T cells were transfected with plasmids expressing firefly luciferase under the control of an ISRE promoter element, pRL-TK luciferase (as an internal control), and increasing amounts of pCAGGS-NS1 (0, 1, or 10 ng) and pCAGGS-UAP56 (0, 3, 10, or 30 ng). Twenty-four hours later, the cells were mock-treated or treated with IFN-β (10,000 U/mL), and cultured for 24 h. Luciferase activities were measured and data were analyzed as described in A. Data are shown as the mean ± SD (n = 3; biological replicates). UAP56 and NS1 expression levels in cells were examined by immunoblotting. Finally, we assessed influenza virus titers in Vero cells, which are defective in IFN-α/β secretion. As shown in Figures 5C,D, UAP56 downregulation also restricted WSN virus growth in Vero cells. WSN-NS1-R38A-K41A mutant virus, which encodes RNA-binding defective NS1, was also restricted by UAP56 downregulation. We also examined the growth of these viruses in the hepatocarcinoma cell lines Huh7.0 and Huh7.5 (a derivative of Huh7.0 cells which carry a defective form of RIG-I). As shown in Figures 5E-G

DISCUSSION
UAP56/URH49 facilitates the efficient replication of influenza A viruses through its roles in pre-mRNA splicing, mRNA nuclear export, and the indirect suppression of cellular antiviral responses (Luo et al., 2001); these functions are also affected by the viral NS1 protein (Fortes et al., 1994;Qiu and Krug, 1994). Here, we demonstrated that the cellular protein UAP56 associates with the influenza viral protein NS1. NS1 also co-precipitated with FLAGtagged URH49 (data not shown), suggesting that UAP56 and URH49 exhibit redundancy. Because we were unable to obtain anti-URH49 antibodies of high quality, we focused on UAP56 in this study. Since UAP56 interacts with MxA (Wisskirchen et al., 2011a), which suppresses influenza A virus replication (Pavlovic et al., 1992;Haller et al., 2015), we tested the possibility that UAP56-NS1 association is disturbed by MxA. NS1-MxA co-precipitation was not detected, and MxA overexpression did not affect the UAP56-NS1 co-precipitation. These results demonstrate that UAP56-NS1 association is not affected by MxA.
UAP56 was co-precipitated with the NS1 proteins of different influenza viruses including the pandemic 1918 virus, a representative of highly pathogenic H5N1 viruses which are now enzootic in poultry populations in Southeast Asia and have to date caused at least 860 human infections 2 , and a representative of the H7N9 viruses that are circulating in poultry in China and have caused five waves of human infections 3 to date. These results demonstrated that UAP56 association with NS1 is not specific to WSN virus.
Several amino acid residues in NS1 have been shown to interact with host proteins and/or affect the biological functions of NS1 (Marc, 2014). Among them are the RNA-binding residues R38 and K41 (Wang et al., 1999), the CPSF30-binding residues 103 and 106, and the PDZ domain binding motif (located at the C-terminus of most influenza A virus NS1 proteins) (Jackson et al., 2008;Liu et al., 2010;Zielecki et al., 2010;Fan et al., 2013). The PDZ domain binding motif is not present in the VN1203 and AH1 NS1 proteins, whereas the WSN and BM/1/1918 NS1 proteins encode PDZ domain binding motifs with the sequences RSEV and KSEV, respectively; the latter contributes to the virulence of the BM/1/1918 virus (Jackson et al., 2008). We tested UAP56 co-precipitation with WSN-NS1 mutants encoding R38A-K41A (defective binding to RNA; Wang et al., 1999), F103L-M106I (defective binding to CPSF30; Kochs et al., 2007;Ping et al., 2011), or lacking the PDZ domain binding motif ( PDM). Co-precipitation of the NS1-R38A-K41A mutant with UAP56 was severely reduced, demonstrating that the RNAbinding residues in NS1 are critical for NS1-UAP56 association. Arginine at position 38 and a basic amino acid at position 41 are highly conserved among influenza A virus NS1 proteins, likely explaining why the UAP56-NS1 association is conserved among the NS1 proteins of different virus origins.
In an indirect immunofluorescence analysis, UAP56 was predominantly localized in the nucleus but was also present in the cytoplasm in non-infected cells, consistent with previous studies (Kota et al., 2008;Wisskirchen et al., 2011a). WT NS1 co-localized with UAP56 near the nuclear membrane in WT-WSN virus-infected cells, whereas mutant NS1 was predominantly expressed in the cytoplasm with limited colocalization with UAP56 in WSN-NS1-R38A-K41A mutant virus. Quantitative co-localization analysis showed that NS1-UAP56 association in WT WSN-infected cells was abrogated in cells infected with WSN-NS1-R38A-K41A mutant virus, consistent with our in vitro co-immunoprecipitation analysis.
Among its many functions, NS1 suppresses RIG-I-dependent IFN-β induction and the activation of ISGs Garcia-Sastre et al., 1998;Rajsbaum et al., 2012), therefore we tested the effect of UAP56 overexpression on the immune suppressive functions of NS1 in a reporter gene assay. WSN NS1 suppressed RIG-I-dependent IFN-β promoter activation, but this activity was not significantly affected by increasing amounts of co-expressed UAP56. Also, in a reporter gene assay under the control of a promoter with an ISRE, NS1 expression suppressed the activation of IFN-β-activated genes, which was not significantly altered by co-expression of UAP56. Together, these data indicate that UAP56 does not significantly affect the immune suppressive functions of NS1. Wisskirchen et al. (2011b) demonstrated that siRNA-mediated downregulation of UAP56 reduced influenza virus titers in IFNcompetent cells. UAP56 downregulation also restricted WSN virus growth in Vero cells, which are defective in IFN-α/β secretion, suggesting that the supportive function of UAP56 in influenza A virus replication does not require a functional IFN system. WSN-NS1-R38A-K41A mutant virus was also restricted by UAP56 downregulation, further suggesting that the UAP56-NS1 association was not related to UAP56's essential role in influenza virus replication. We also examined the growth of these viruses in a cell line that carries a defective form of RIG-I. WSN virus growth was also restricted by UAP56 downregulation in this cell line, indicating that UAP56's supportive function in influenza A virus replication is not mediated by NS1's suppressive effect on RIG-I. The limited replication of WSN-NS1-R38A-K41A upon UAP56 downregulation further suggests that the UAP56-NS1 association is not essential for UAP56 function.
The UAP56-NS1 association may be related to the regulation of mRNA splicing and/or nuclear export, consistent with the known roles of NS1 (Fortes et al., 1994;Qiu and Krug, 1994;Satterly et al., 2007) and UAP56 (Jensen et al., 2001;Herold et al., 2003;MacMorris et al., 2003). This assumption is further supported by the co-localization of these two proteins near the nuclear membrane (Figures 4B,C), while NS1 counteracts the host immune responses mainly in cytoplasm. It is also supported by our finding that the RNA-binding residues of NS1 (i.e., NS1-R38/K41) are required for the UAP56-NS1 association. However, a functional relationship between UAP56 and NS1 has yet to be shown.
It was previously reported that influenza B virus NS1 colocalizes with nuclear mRNA splicing speckles in the nucleus, and that this localization pattern is different from that of influenza A NS1 (Schneider et al., 2009). We found that the ATPbinding/-hydrolyzing residues in UAP56, which are essential for UAP56 co-localization with splicing complexes, are not required for the co-precipitation with influenza A virus NS1 (Supplementary Figure S1). These results suggest that the association of UAP56 with influenza A NS1 may be functionally different from that of UAP56 with influenza B virus NS1. The biological importance and exact role of the UAP56-NS1 association in the life cycle of influenza A viruses remain to be assessed.