Human embryonic kidney 293T, A549, and MDCK cells were obtained from ATCC. Cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS and 1% penicillin/streptomycin in an incubator at 37°C with 5% CO₂. The influenza viruses used in this study, A/environment/Qinghai/1/2008(H5N1), A/Beijing/05/2009(H1N1), and A/Chicken/Hangzhou/174/2013(H7N9), were isolated and stored in Institute of Zoology, Chinese Academy of Sciences. The Viruses were propagated in 10-day-old specific pathogen-free embryonated chicken eggs. All experiments with IAVs were carried out in biosafety level 3 containment laboratories approved by the Chinese Academy of Science.

The full-length NP, PB1, PB2, PA, HA, NA, NS, M [A/environment/Qinghai/1/2008(H5N1)] was cloned into the pHW2000 vector, respectively. The full-length coding sequence of NP was cloned into the pcDNA3.1-Myc-His and pET-28a vectors. Lysine (K) to alanine (A) or arginine (R) mutation in the NP gene was introduced by using the QuikChange site-directed mutagenesis kit (Stratagene) following the manufacturer’s protocol. The full-length coding region sequences of human HDAC1, HDAC2, HDAC3, and HDAC8 were amplified from the HEK293T cDNA and then inserted into the pcDNA3.0-HA and pGEX-4T-1 vectors. All primers used in this study are shown in Table S1 in Supplementary Material. All plasmids were confirmed by Sanger sequencing at BGI (Shenzhen, China).

Glutathione S-transferase (GST) fusion protein GST-HDAC1 and 6His fusion protein 6His-NP were expressed in E. coli strain BL21 host cells induced by 0.5 mM IPTG for 4 h at 28°C. After cell lysis and centrifugation, the GST-HDAC1 and 6His-NP proteins were purified with glutathione-sepharose 4B and nickel-nitrilotriacetic acid (Ni-NTA) column (QIAGEN, Gaithersburg, MD, USA) as reported previously (31, 32), respectively. The purified proteins were concentrated and dialyzed against PBS. The recombinant proteins were kept at −80°C until use.

To purify the acetylated NP protein, 293T cells were transfected with NP-6His and then treated with 1 µM trichostatin A (TSA) for 24 h, cell extracts were purified by Ni-NTA column as described previously (33). In brief, cell extracts were incubated with Ni-NTA beads for 4 h at 4°C. Beads were then washed with buffer A (20 mM Tris-HCl (pH8.0), 0.5 M Nacl, 10 mM Imidazole), and the bound proteins were eluted with buffer B (20 mM Tris-HCl (pH8.0), 0.5 M Nacl, 300 mM Imidazole). The eluent NP proteins were then concentrated and dialyzed against PBS.

The deacetylation assay was performed as described previously (34). Briefly, the acetylated NP was incubated with GST-HDAC1 in HDAC assay buffer (25 µM Tris-HCl (PH8.0), 137 mM NaCl, 2.7 mM KCl, 1 mM MgCl₂, and 60 µM NAD⁺) in a final volume of 50 µl for 60 min at 30°C. The proteins were resolved by 12% SDS-PAGE and analyzed by Western blot.

The antibodies used for immunoblotting include: mouse monoclonal antibodies (mAbs) c-Myc antibody (sc-56634) and β-actin (C4) (sc-47778) (both purchased from Santa Cruz Biotechnology, Dallas, TX, USA); rabbit polyclonal antibodies against NP (GTX125989) was purchased from GeneTex (Taiwan); anti-TBK1 (ab40676), and anti-TBK1 (phosphor S172) (ab109272) antibodies were obtained from Abcam (Cambridge, MA, USA). Rabbit polyclonal antibody against acetylated-lysine (9441), mouse mAb against HDAC1 (10E2) (5356), p-IRF3 (4947), IRF3 (4302), p-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (4370), p44/42 MAPK (Erk1/2) (9102), and rabbit mAb HA-Tag (c29F4) (28) were from Cell Signaling Technology (Beverly, MA, USA).

Recombinant viruses were generated by plasmid-based reverse genetics as previously described (35, 36). In brief, The pHW2000-NP plasmid (as wild-type) or NP gene with mutation K103A and K103R was co-transfected with the other seven plasmids pHW2000-PB1, -PB2, -PA, -HA, -NA, -NS, and -M genes from A/environment/Qinghai/1/2008(H5N1) virus into MDCK cells and 293T cells. 24 h posttransfection, the medium was replaced with DMEM plus 1 µg/ml tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin. The cells were further cultured for 72 h, and the supernatant containing the recombinant viruses was harvested. The K103A and K103R mutations of NP in the generated viruses were confirmed by sequencing.

MDCK cell monolayers (5 × 10⁶ cells at 100% confluence in a 6-well plate) were infected with different dilutions of virus for 1 h at 37°C, with shaking every 15 min. The virus inoculums were then removed, whereas the infected cells were washed by PBS, and covered with agar overlay medium (DMEM supplemented with 1 µg/ml TPCK-treated trypsin and 0.6% low-melting point agarose) and then incubated at 37°C for 2–3 dpi. Visible plaques were counted and virus titers were determined. All data were expressed as means of three independent experiments.

To characterize, the single growth curves of rWT, rK103A, and rK103R viruses on A549 cells, we first titrated the viruses on MDCK cells and then used them to infect A549 cells at an MOI of 1. After 1 h incubation at 37°C in 5% CO₂, the medium were replaced with DMEM with 2% FBS, and the cell-free supernatants were harvested at 3, 6, 9, and 12 h postinfection. Virus amounts in the supernatant were determined using MDCK cells. For the multiple growth curves, the A549 cells were infected with rWT, rK103A, or K103R virus at an MOI of 0.1, and the procedures were the same as those for the single growth curve experiments.

To determine the polymerase activities of the ribonucleoprotein complexes, 293T cells were co-transfected with 1 µg of each pHW2000 plasmid encoding the polymerase subunits PB2, PB1, PA, and NP (WT or mutants), together with 1 µg pPOL1-Luc and 50 ng pRL-TK. Renilla luciferase was used as an internal control. The transfected cells were incubated at 37°C and harvested at 36 h posttransfection, and luciferase levels were measured by using the Dual Luciferase Assay System (Promega, Madison, WI, USA) according to the manufacturer’s protocol.

Total intracellular RNA was extracted by using TRIzol (Invitrogen, Carlsbad, CA, USA). The RT reaction was carried out by using the GoScript Reverse Transcription System (Promega, Madison, WI, USA). For synthesizing the cDNA, we used the influenza A-specific primer uni-12 (5′-AGCAAAAGCAGG-3′) and random primer.

Each cDNA sample was analyzed in triplicate on an ABI 7500 Real-Time Detection System (Applied Biosystems) using SYBR Green (CWBio, Beijng, China) according to the manufacturer’s protocol. Endogenous housekeeping genes (β-actin or GAPDH) were used as internal standards. Cycle conditions were as follows: 95°C for 2 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min.

Cells were lysed in RIPA buffer (50 mM Tris-HCl pH 7.6, 150 mM NaCl, 0.1% SDS, 1% Triton-100) supplemented with protease inhibitor (Roche, Basel, Swiss) and 300 nM TSA. Protein concentrations of the extracts were determined by bicinchoninic acid assay. NP was purified with protein A Mag Sepharose (GE Healthcare, Little Chalfont, UK) prebound to a mouse Myc-tagged mAb for 10 h at 4°C. The immunocomplexes and the whole cell lysate were resolved by SDS-PAGE. Proteins were transferred to nitrocellulose membrane and blocked with 5% nonfat milk in TBS with 0.1% Tween 20 (TBST). The membrane was probed with primary antibody followed by secondary antibody at room temperature for 2–4 h. The membrane was stripped with stripping buffer before being probed again with another antibody.

The production of lentivirus expressing HDAC1–shRNA\vector was purchased from Sigma-Aldrich (TRCN0000195103 and TRCN0000004818). The clone of short hairpin HDAC1 (shHDAC1) used for knockdown of HDAC1 was validated. The lentivirus-negative control was generated from pLKO.1, in which the shRNA sequence was replaced by the following scrambled sequence: 5′-AGACGTGGTTTTTTTTGCTAGCTTGC-3′. A549 cells were infected with shHDAC1 lentivirus for 48 h, and then the medium was replaced with selective medium containing puromycin (2 µg/ml). After the cells were incubated for 14 days, individual clones were obtained, expanded, cryopreserved, and cultured in 6-well plates. The total cell lysate was collected to check the knockdown efficiency of shHDAC1 by Western blot.

The siRNA duplexes against HDAC1 and the negative control siRNA (siCONT) were purchased from Invitrogen (Shanghai, China). The sequences of the two HDAC1 siRNAs are as follows: CUGUACAUUGACAUUGAUAdTdT (siHD1-1) and CAGCGAUGACUACAUUAAA dTdT (siHD1-2). The siRNAs transfection was performed using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instruction.

Student t-test (where two groups of data were compared) or two-way ANOVA followed by Bonferroni post-test analysis (where more than two groups of data were compared) was used for statistical comparisons. Data were expressed as means ± SDs. A P value of 0.05 or less was considered statistically significant.

Class I HDACs (i.e., HDAC1, HDAC2, HDAC3, and HDAC8) play a critical role in the life cycle of influenza virus (28). To determine which of these HDACs binds to NP, we co-transfected each Hemagglutinin (HA)-tagged HDAC with Myc-tagged NP. At 36 h after transfection, the cell lysates were subjected to immunoprecipitation with Myc-tagged antibody. Immunoblot analysis, using HA-tagged antibody, showed that NP interacted with HDAC1 but not with HDAC2, HDAC3, or HDAC8 (Figure 1A). The viral NP came from H5N1, H1N1, and H7N9 viruses in infected HEK293T cells was also demonstrated to interact with endogenous HDAC1 by the co-immunoprecipitation by anti-NP antibody (Figure 1B). And the binding was validated by the nucleus co-localization of NP and HDAC1 in H5N1, H1N1, and H7N9 viruses infected A549 cells (Figure 1C). The co-localization is quantified using Pearson’s correlation (r) (0.8, 0.6, and 0.8 for H5N1, H1N1, and H7N9, respectively). To better define, the direct interaction between NP and HDAC1, a Ni-NTA pulldown assay was carried out using 6His-NP and GST-HDAC1 expressed in E. coli strain BL21. As shown in Figure 1D, NP interacted with the recombinant HDAC1 in vitro. Furthermore, HDAC assay in vitro showed that the acetylation level of NP (Ace-NP) was decreased about 40% when the acetylated NP was incubated with GST-HDAC1 (Figure 1E). Taken together, our results demonstrated the NP protein interacted with the HDAC1 in vivo and in vitro, and the cellular NP was deacetylated by recombinant HDAC1 in vitro.

Above results showed that NP interacts with HDAC1, but whether NP is an acetylated protein or the actual modification sites of NP were need to explore. To identify the acetylated amino acid residues within the NP, 293T cells were transfected with NP plasmids with a Myc-tagged. Transfected cell proteins, that were either treated or non-treated with TSA (300 ng/ml) for 24 h, were purified using Myc-tagged antibody. TSA was used to suppress the activity of HDACs and expose the potential acetylation sites. The immunoprecipitation complexes were then resolved by SDS-PAGE and stained with Coomassie blue. A visible band with a molecular mass of 55 kD (Figure 2A), was extracted and subjected to liquid chromatography-tandem mass spectrometry, and results confirmed this band as the NP of influenza A/environment/Qinghai/1/2008(H5N1) virus (Figure 2B). Four acetyl-lysine residues, K103, K227, K229, and K470, were detected without TSA treatment, and two additional acetyl lysine residues, K91 and K198, were detected after TSA exposure (Figure S1A in Supplementary Material). Further analysis of the spectra indicated that the peptide containing K103 was acetylated at this position (Figure 2C).

To further determine the functional acetylation sites on NP, we performed site-directed mutagenesis to replace the K residue at one of six sites with arginine (R), which will remove the acetylation if there is any. We transfected HEK293T cells with the NP mutants (K91R, K103R, K198R, K227R, K229R, and K470R) and purified them by using Myc-tagged antibody and detected the Ace-NP using the Ace-lysine antibody. The results showed that K103R reduced the 30% level of NP acetylation compared with that of the WT protein (Figures 2D,E). With reference to the three-dimensional crystal structure of the NP (Protein Data Bank accession No. 2Q06), it showed that K103 was located on the surface of the NP (Figure S1B in Supplementary Material), indicating the K103 residue had a critical role. Together, our data suggest that the K103 is an acetylation site of NP.

To further determine, whether K103 site was an acetylation site that regulated by HDAC1. HEK293T cells were transfected with NP-WT and NP-K103R expression plasmids with or without HDAC1. The results showed that HDAC1 co-expressed with the NP-WT, and the acetylation was downregulated 0.7-fold compared with that of the control group (Figure 2F). However, the acetylation of NP-K103R was not downregulated (Figure 2G). Collectively, these data suggest that the K103 of NP is an acetylation site regulated by HDAC1.

Sequence alignment showed that K103 is highly conserved among different IAV subtypes (Figure 3A). To assess the roles of K103 in virus life cycle, we introduced K103A/R substitution to mimic the K103 deacetylation state of NP protein. First, we tested the replication function of K103A and K103R by using a mini-replicon system (containing luciferase as a reporter) that assays vRNA replication activity. The results (Figure 3B) showed that the substitution of K103 by A and R increased the RNP activity by 1.4-fold and 1.2-fold, respectively, compared with that of WT. Therefore, these results suggested that K103 mutant increased the RNP activity.

We next examined whether the K103 mutation could affect the growth of the viruses. Recombinant viruses, incorporating the NP-rK103A/R substitution, were produced by reverse genetics techniques (35), and the biological properties of the mutant viruses were examined. The plaque sizes of the NP-rK103A/R mutant viruses were similar to that of the WT virus (Figure 3C). QRT–PCR was performed to detect vRNA expression. 24 h postinfection, vRNA levels of the NP-K103A and NP-K103R mutant viruses were 8-fold and 2.5-fold higher, respectively, than that of the WT virus (Figure 3D). Growth analysis suggested that the virus titer of NP-rK103A/R mutant virus was twofold at 12 h and NP-rK103A fourfold at 48 h postinfection, compared to the rWT virus (Figures 3E,F). Collectively, these data suggest that K103 is a key acetylation site and that the K103A and K103R mutants show increased virus replication levels in vitro than that in rWT.

Mutation of K103 may lead to other “gain-of-function” that it may act as a nuclear import/export signal. Therefore, the intracellular distribution of mutant NP expressed from the recombinant viruses was examined by IFAs. Though the NP of rWT, rK103A, and rK103R displayed cytoplasm (C) and nuclear (N) localization, the nuclear fraction of the two mutants was significantly increased (Figures 4A,B). The increased nuclear fraction of rK103A and rK103R may be correlated with the enhancement of viral growth. In addition, when HDAC1 was knockdown in A549 cells, the cytoplasm fraction of NP was significant increased compared with control cells (Figures 4A,B), indicating that HDAC1 facilitated the nuclear retention of NP. Furthermore, depletion of HDAC1 expression resulted in the decrease of the RNP activity (Figure 4C). However, there was no difference between the WT and K103 mutants, suggesting that HDAC1 activity was important for virus replication but this was independent of the deacetylation of the NP K103 residue.

To determine the role of HDAC1 in influenza virus replication, we used shRNA to silence the expression of endogenous HDAC1. Immunoblotting results showed that, among the five shHDAC1 clones, clones 1 and 2, had the best depletion efficiency (Figure 5A). Therefore, these two clones were used for further studies. To verify that HDAC1 is involved in IAV infection, we infected HDAC1 depletion cells with rWT, rK103A, or rK103R influenza virus at an MOI of 1. At 24 h postinfection, the culture medium was collected and divided into two parts; one part was analyzed by Western blot and the other was analyzed by plaque assay to determine virus titers. Depletion of HDAC1 resulted in decreased NP and a lower virus titer (Figures 5B,C). The result was also verified in transient HDAC1 knockdown cells using siRNA (siHD1#1 and siHD1#2) (Figures 5D,E). Intriguingly, overexpression of HDAC1 in A549 cells only slightly increased virus production and virus titer (Figures 5F–H). It is possible that the endogenous HDAC1 was enough for the replication of virus and the effect of extraneous HDAC1 was limited. Taken together, these results suggest that HDAC1 can regulate IAV production.

To further elucidate the molecular mechanism how HDAC1 regulates IAV virus production, we analyzed the gene expression profile of cells infected with rWT, rK103A, or rK103R influenza virus at an MOI of 1. The NP mRNA level was decreased in HDAC1-depleted cells compared to shCONT cells (Figure 6A). The HDAC1 mRNA level was downregulated when A549 control cells were infected with IAV (Figure 6B). The relative Ace-NP was increased in HDAC1-knockdown cells compared with the control cells (Figure 6G). It has been reported that type I IFN pathway that played a critical role in antivirus activity (37). Interestingly, gene expression analysis using qRT-PCR showed that IFNα mRNA levels were higher in HDAC1-depleted cells in response to the IAV infection (Figure 6C). The secretion of the IFNα in the culture supernatants was increased in HDAC1-depleted cells induced by IAV infection (Figure S2A in Supplementary Material). Results from siRNA-mediated knockdown of HDAC1 further confirmed these results (Figures S2B–E in Supplementary Material). To determine whether IRF3 has a role in producing IFN type I in the HDAC1-depleted cells after IAV infection, we measured the phosphorylation level of IRF3 in these cells. The results showed that phosphorylation of IRF3 were increased in HDAC1-depleted cells infected with IAVs compared with the control cells (Figure 6F). As TBK1 has been implicated in IRF3 activation in response to viral infections, we then test whether TBK1 is involved in IRF3 activation, thus leading to type I IFN production following infection with IAV. As shown in Figure 6F, the phosphorylation of TBK1 was increased in HDAC1-depleted cells infected with IAV compared with the control cells. Collectively, these results suggest that knockdown of HDAC1 suppresses IAV replication through activation of TBK1-IRF3 signaling pathway.

In addition, the TLR3 and IL6 mRNA levels were increased in HDAC1-depleted cells infection with IAV (Figures 6D,E). However, the mRNA levels of P53, TNF-α, and ILF16 mRNA levels did not differ significantly between the HDAC1-depleted and the control cells (Figure S3 in Supplementary Material). Moreover, the phosphorylation of ERK was increased in HDAC1-depleted cells infected with IAVs compared with the control cells (Figure 6F). On the contrary, the phosphorylation of AKT was decreased in HDAC1-depleted cells (Figure S4 in Supplementary Material). Collectively, knockdown of HDAC1 suppressed IAV replication through activation of the TBK1-IRF3 and ERK signaling pathways.

To further determine the dynamic change of HDAC1 level after IAV infection, we analyzed the time course of HDAC1 expression profile in H5N1 infected cells at an MOI of 0.5. The cell lysates were harvested on 2, 4, 8, 12, 24, 48, and 72 h postinfection and detected HDAC1 and NP expression by Western blot. We found that the level of HDAC1 polypeptide was reduced at the early stage of IAV infection and reached a valley at 8 h (Figures 7A,B). However, the HDAC1 expression level gradually came back up to the basal level from 12 to 48 h and reached a plateau at 48 h (Figures 7A,B). Meanwhile, a time-dependent increase in the NP level was also observed. The host cells produced IFN-a at 8 h postinfection, and the level was increased in a time-depend manner (Figures 7A,B). We speculated that downregulation of HDAC1 at the early stage in IAV-infected cells might be a self-host protective response to the virus infection, which may be critical to activate the IFN-α expression.

To further determine the interaction between HDAC1 and NP, we transfected HEK293T cells with different concentration of NP. NP exhibited a dose-dependent effect on the HDAC1 level, which was downregulated with the increasing NP expression (Figures 7C,D). Based on our data and others, we propose a possible schematic model: in the NP WT, K103 of NP was an acetylation site regulated by HDAC1. Two K103 NP mutants, K103A and K103R, lacking of regulation of K103 acetylation by HDAC1, enhanced the RNPs activity and replication efficiency of the recombinant viruses in vitro. In addition, depletion of HDAC1 inhibited influenza A virus replication via two paths: one was decreased the RNP activity, and the other was activated TBK1-IRF3 signaling, which was independent the NP K103 site (Figure 7E).