A rabbit polyclonal antibody against M1 and a rat polyclonal antibody against NEP were prepared as previously described (21). A rabbit monoclonal antibody against β-actin (CST; 8547) was purchased. HeLa cells present in this study were obtained from ATCC. HeLa cells and HeLa-GFP-emerin cells (22) were grown in Dulbecco’s minimal essential medium (DMEM) containing 10% bovine fetal calf serum and incubated at 37°C in 5% CO2. pPol-I-M and pPol-I-NS were kindly gifted by Dr. Fumitaka Momose (Kitasato University). pCAGGS-PB1, pCAGGS-PB2, pCAGGS-PA, and pCAGGS-NP were prepared as previously reported (23). For the construction of plasmid expressing N-terminal or C-terminal Renilla luciferase-fused NEP, FLAG-tagged NEP cDNA, which was amplified with primers, 5’-CTGCTAGCGCCGCCACCATGGACTACAAGGACGACGATGACAAGATGGACCCAAACACTGTGTC-3’ and 5’-CCGGAATTCTTAAATAAGCTGAAACGAGAAAG-3’, was cloned into pcDNA3.1-NR, pcDNA3.1-CR, or pcDNA3.1-NR-CR plasmids (24). The NES motif of NEP was mutated by the PCR-based site-directed mutagenesis using primers, 5’-GGACGTCAAAAACGCAGTCGGAGTCC-3’ and 5’- TCTGCTGTGTGTCCTGGAAGAGAAGG-3’. For the construction of plasmid expressing delNES mutant harboring additional NES motif at the C-terminus (exoNES), NEP cDNA was amplified with primers, 5’-CCAAGGATCCCATATGGACCCAAACACTGTGTCAAGC-3’ and 5’-CCAAGGATCCAATAAGCTGAAACGAGAAAGTTC-3’. Influenza A/Puerto Rico/8/34 (A/PR/8/34) virus was grown at 35.5°C for 48 h in allantoic sacs of 11-days-old embryonated eggs, and then the infected allantoic fluids were collected and stored at -80°C until use.

Indirect immunofluorescence assays and fluorescence in situ hybridization (FISH) assays were carried out as previously described (25). Briefly, cells were fixed with 4% PFA for 10 min. After being washed with PBS, cells were permeabilized with 0.5% Triton X-100 in PBS for 5 min. After incubation in PBS containing 1% bovine serum albumin for 30 min, the coverslips were incubated with each antibody for 1 h and then further incubated with Alexa Fluor 488- and 568-conjugated secondary antibodies, respectively (Life Technologies). FISH assays were performed after indirect immunofluorescence assays using an RNA probe complementary to the segment 8 virus genome. PLA was carried out using Duolink In Situ PLA kit (Olink Bioscience) according to the manufacturer’s protocol. The number of the PLA signals was measured with reconstructed 3D images from z-stack series using IMARIS software (Carl Zeiss). Images were acquired by a confocal laser scanning microscopy (LSM700; Carl Zeiss) using ×63 Apochromat objective (NA = 1.4).

HeLa cells were transfected with the pcDNA3.1 plasmid encoding N-terminal (NR; 1-229 a.a.) and/or C-terminal (CR; 230-311 a.a.) fragment of Renilla luciferase fused to NEP. At 24 h post-transfection, cells were collected and lysed in a cell lysis buffer for Renilla luciferase (Promega) by three freezing-thawing cycles. Renilla luciferase activity was measured according to the manufacturer’s instruction (Promega). Relative luminescence intensity was measured for 10 sec with a Mini Lumat (Berthold Technologies).

To visualize the interaction of NEP with M1 in the formation of vRNP export complexes, we performed in situ proximity ligation assays (PLA) with rat polyclonal anti-NEP and rabbit polyclonal anti-M1 antibodies in IAV-infected HeLa cells expressing GFP-emerin, a marker of the nuclear envelope. In the antibody-based PLA system, when a pair of DNA primer-conjugated secondary antibodies is in close proximity, the primers hybridize with a connector oligo and serve as a template for rolling circle amplification, which can generate DNA targets for a fluorescently labeled DNA probe (Figure 1A and Supplementary Figure 1). Strong punctate PLA signals (red) were observed in the nucleus and the cytoplasm of infected cells (Figure 1B). The number of PLA signals in the nucleus and cytoplasm were counted from the reconstructed z-stack images, respectively. We found that the number of PLA signals increased in both the nucleus and cytoplasm along with the progression of infection, but at 11 h post-infection, most PLA signals were observed in the cytoplasm (Figure 1C). We next examined the close proximity between NEP and M1 in the presence of leptomycin B (LMB), a potent inhibitor of the interaction between CRM1 and its cargo protein through NES motif (Figure 2). At 5 h post-infection, before the nuclear export of vRNP, infected HeLa cells were treated with 10 nM LMB for 2 h and subjected to indirect immunofluorescence assays (Figure 2A) and in situ PLA with anti-NEP and anti-M1 antibodies (Figures 2B, C). The intracellular localization of NEP was unchanged by the addition of LMB, possibly due to the passive diffusion of NEP in the nucleus and cytoplasm (Figure 2A, green). In contrast, M1 was mainly localized to the cytoplasm and cell periphery in the absence of LMB, but the level of M1 protein increased in the nucleus by LMB treatment (Figure 2A, red). The nuclear accumulation of M1 may be due to the inhibition of NEP-mediated nuclear export of M1, which has an NLS motif and is actively imported to the nucleus. In contrast, the localization pattern of NEP was not changed by LMB treatment, possibly due to passive diffusion of NEP by the low molecular weight (Figure 2A).

LMB treatment significantly reduced the number of PLA signals not only in the cytoplasm but also in the nucleus (Figures 2B and C), although M1 accumulated in the nucleus (Figure 2A). This suggests that NEP requires the interaction with CRM1 via the NES motif of NEP to assemble the vRNP export complexes (NEP-M1-vRNP). The expression levels of NEP and M1 were comparable with LMB treatment for 2 h (Figure 2D).

NEP has a NES motif consisting of multiple hydrophobic residues at amino acid positions 12 to 21 in the N-terminal mobile domain (10, 11). We next examined the interaction of M1 with NEP harboring point mutations in the NES motif (delNES mutant). The hydrophobic residues critical for NES function were mutated to hydrophilic residues to prevent the interaction of CRM1 with NEP (Figure 3A). A mutant virus defective in NES function cannot be recovered using the reverse genetics system because the NEP-mediated nuclear export of vRNP is an essential process for virus production. Thus, cells were transfected with pPol I plasmids, containing M and NS segments (pPol-I-M and pPol-I-NS), and expression plasmids for PB1, PB2, PA, and NP (Figure 3B). In this system, M and NS segments are replicated by viral polymerases and NP, then M1/M2 and NS1/NEP genes are expressed from the reconstructed viral replicons. Therefore, as in infected cells at 11 h post-infection, the synthesized virus genome was successfully exported to the cytoplasm and localized as punctate structures in the recycling endosomes at 24 h post-transfection (26) (Figure 3B; WT). PLA signals were observed at the periphery of endosomal puncta, although clear colocalization was not found (Figure 3B; enlarged panel). This may be due to the fact that PLA signal is produced by the rolling-circle amplification of short DNA primers conjugated with secondary antibodies, indicating that the pattern of PLA signals is possibly different from the exact localization of the target proteins. The delNES mutation was introduced in NEP ORF without any amino acid changes in NS1 ORF. The virus genome was found in the nucleus in cells transfected with delNES mutant (Figures 3B, C), indicating that NEP is responsible for the nuclear export of virus genome in an NES motif-dependent manner. Unexpectedly, PLA signals between delNES mutant and M1 were mainly observed in the cytoplasm and increased to about 1.6-fold those of WT (Figures 3B, D). Further, this was not complemented by the insertion of an exogenous NES (N-ILLRMSKMQL-C) at the C-terminus of the delNES mutant (Figures 3B–D). These findings indicate that the endogenous NES motif of NEP intrinsically inhibits the interaction with M1 possibly in the cytoplasm.

The M1 binding domain is located at the C-terminal domain of NEP and consists of the antiparallel hairpin structure with hydrophobic and hydrophilic interfaces (13). The endogenous NES motif may have additional functions other than nuclear export, because the exogenously added NES motif did not result in the functional nuclear export of the virus genome (Figure 3). Thus, we hypothesized that the NES motif in the N-terminal domain associates with the M1 binding domain intramolecularly to inhibit the interaction of NEP with M1. To address this possibility, we next performed split Renilla luciferase complementation assays to examine the close proximity of N-terminal and C-terminal domains of NEP (27). The N-terminal (NR; 1-229 a.a.) and C-terminal (CR; 230-311 a.a.) fragments of Renilla luciferase were fused to the N-terminus and C-terminus of NEP, respectively, with a linker sequence (GGGSGGGS) (Figure 4A). At 24 h post-transfection, the cell lysates were prepared and subjected to the luciferase assays. Although the luciferase activity was not detected in the cell lysates transfected with NR-NEP or NEP-CR, the complementation of luciferase activity was observed by expressing NEP harboring NR and CR (NR-NEP-CR) (Figure 4A). Co-transfection of NR-NEP and NEP-CR did not show the complemented luciferase activity, suggesting that the close association of N- and C-termini of NEP is mediated by the intramolecular interaction rather than the intermolecular interaction. We also found that the level of luciferase complementation in the lysates expressing the NES-mutated NEP with NR and CR (NR-delNES-CR) was reduced to about 10% of that of WT NEP (Figure 4A).

The hydrophobic surface of the amphiphilic hairpin structure in the M1-binding domain is thought to be exposed to solvent (13). Thus, it is possible that the hydrophobic NES motif in the N-terminal domain associates intramolecularly with the M1-binding domain through the hydrophobic interaction (Figure 4B). To address this possibility, we examined the split Renilla luciferase complementation assays with NEP proteins mutated at the hydrophobic residues (I76, I80, V83, L106, V109, and I113) aligned to the exposed surface of the antiparallel amphiphilic hairpin structure (Figure 4C). I76T/I80T and I80T/L106T partially impaired the complemented luciferase activity, while I80T/V83T did not. Furthermore, I76T/L106T and V109T/I113T, reduced the intramolecular interaction to less than 20% of WT NEP. Notably, the expression levels of mutated NEP proteins were unchanged compared to WT NEP (Figure 4D). To elucidate the biological significance of the intramolecular interaction of NEP, we next examined the intracellular localization of NEP harboring I76T/L106T mutations. HeLa cells were transfected with pPol-I-M and either WT or I76T/L106T-mutated pPol-I-NS, and expression plasmids for PB1, PB2, PA, and NP (Figure 5). WT NEP was diffusively localized in the nucleus and the cytoplasm (Figure 5A, upper panels and Figure 5B). Since M1 was also localized in the nucleus, but it was predominantly observed in the cytoplasm (Figure 5A). In contrast, the I76T/L106T mutant was localized to the cell periphery with M1 (Figure 5A, middle panels and Figure 5B), and M1 was not observed in the nucleus. Although a milder phenotype than I76T/L106T, similar results were obtained for V109T/I113T mutant (Figure 5A, lower panels and Figure 5B). We also examined the close proximity of M1 with either I76T/L106T or V109T/I113T mutant by in situ PLA (Figures 5C, D). Although the number of PLA signals between M1 and NEP mutants defective in the intramolecular interaction tended to be reduced compared to NEP WT, a significant amount of NEP mutants was still associated with M1 (Figure 5D).