Histone Deacetylase 6 Inhibits Influenza A Virus Release by Downregulating the Trafficking of Viral Components to the Plasma Membrane via Its Substrate, Acetylated Microtubules

ABSTRACT

Mammalian cells produce many proteins, such as IFITM3, ISG15, MxA, and viperin, that inhibit influenza A virus (IAV) infection. Here, we show that a new class of host protein, histone deacetylase 6 (HDAC6), inhibits IAV infection. We found that HDAC6-overexpressing cells release about 3-fold less IAV progeny, whereas HDAC6-depleted cells release about 6-fold more IAV progeny. The deacetylase activity of HDAC6 played a role in its anti-IAV function as tubacin, a specific small-molecule inhibitor of HDAC6, increased the release of IAV progeny in a dose-dependent manner. Further, as visualized by electron microscopy, tubacin-treated cells showed an increase in IAV budding at the plasma membrane, the site of IAV assembly. Tubacin is a domain-specific inhibitor and binds to one of the two HDAC6 catalytic domains possessing tubulin deacetylase activity. This indicated the potential involvement of acetylated microtubules in the trafficking of viral components to the plasma membrane. Indeed, as quantified by flow cytometry, there was about a 2.0- to 2.5-fold increase and about a 2.0-fold decrease in the amount of viral envelope protein hemagglutinin present on the plasma membrane of tubacin-treated/HDAC6-depleted and HDAC6-overexpressing cells, respectively. In addition, the viral ribonucleoprotein complex was colocalized with acetylated microtubule filaments, and viral nucleoprotein coimmunoprecipitated with acetylated tubulin. Together, our findings indicate that HDAC6 is an anti-IAV host factor and exerts its anti-IAV function by negatively regulating the trafficking of viral components to the host cell plasma membrane via its substrate, acetylated microtubules.

IMPORTANCE Host cells produce many proteins that have the natural ability to restrict influenza virus infection. Here, we discovered that another host protein, histone deacetylase 6 (HDAC6), inhibits influenza virus infection. We demonstrate that HDAC6 exerts its anti-influenza virus function by negatively regulating the trafficking of viral components to the site of influenza virus assembly via its substrate, acetylated microtubules. HDAC6 is a multisubstrate enzyme and regulates multiple cellular pathways, including the ones leading to various cancers, neurodegenerative diseases, and inflammatory disorders. Therefore, several drugs targeting HDAC6 are under clinical development for the treatment of a wide range of diseases. Influenza virus continues to be a major global public health problem due to regular emergence of drug-resistant and novel influenza virus strains in humans. As an alternative antiviral strategy, HDAC6 modulators could be employed to stimulate the anti-influenza virus potential of endogenous HDAC6 to inhibit influenza virus infection.

INTRODUCTION

Influenza A virus (IAV) continues to cause significant morbidity and mortality worldwide through seasonal epidemics. Further, IAV has already caused a pandemic in this century, and the threat of another IAV pandemic remains real because of the regular emergence of novel avian IAV strains (e.g., H5N1, H6N1, H7N9, and H10N8) in humans (1–5). IAV targets the airway epithelium in the human respiratory tract to initiate the infection, subsequently causing an acute febrile respiratory illness commonly known as flu. To counter the infection, epithelial cells impose restrictions on various steps of the IAV life cycle, including entry, replication, assembly, and release. The antiviral state against invading viral pathogens is launched by the induction of interferon in infected cells, leading to the expression of several interferon-stimulated genes that restrict virus infection using various mechanisms (6). Interferon-induced proteins such as IFITM3, ISG15, MxA, and viperin have been described to pose restrictions on infection by IAV as well as other animal viruses (7–10). In turn, viruses have evolved their own mechanisms to antagonize the restriction imposed by antiviral host factors. These mechanisms include the degradation, mislocalization, sequestration, and downregulation of the activity of host restriction factors (6). Recently, we discovered that IAV downregulates the activity of the human enzyme histone deacetylase 6 (HDAC6) and induces caspase-mediated cleavage of HDAC6 polypeptide in epithelial cells, including primary human bronchial epithelial cells, the natural target of IAV infection (11, 12). These findings led to the hypothesis that human HDAC6 has an anti-IAV function.

Histone deacetylases (HDACs) are a family of enzymes that catalyze the deacetylation of acetylated proteins, consequently influencing diverse cellular processes, including chromatin remodelling, signaling, RNA splicing, gene expression, cell cycle, and protein stability and transport (13, 14). Acetylation is a posttranslational modification of proteins and has been studied extensively in histones to gain understanding of chromatin structure and gene transcription. However, acetylation is now known to occur in a variety of mammalian proteins. A proteomic study has identified at least 3,600 acetylation sites in 1,750 nuclear and nonnuclear proteins, indicating a broader role of acetylation/deacetylation in nuclear and cytoplasmic functions of the cell (15). Acetylation/deacetylation is a dynamic and reversible process regulated by the competition of histone acetyltransferases (HATs) and HDACs (13–15). So far, 18 HDACs have been identified and classified into four main classes. HDAC6 is a member of class IIb and is one of the best-characterized deacetylases (14). Unlike other HDACs, HDAC6 mainly localizes to the cytoplasm and has unique substrate specificity for nonhistone proteins such as α-tubulin (a microtubule heterodimer), heat shock protein 90 (Hsp90; a molecular chaperone), and cortactin (an actin filament-binding protein) (16–18). Consequently, HDAC6 regulates multiple biological processes like cell migration, immune synapse formation, and degradation of misfolded proteins (14). Here, we demonstrate that HDAC6 is an anti-IAV host factor that inhibits IAV replication by negatively regulating the trafficking of viral components to the site of virus assembly via its substrate, acetylated microtubules.

MATERIALS AND METHODS

Cells, virus, and plasmid.

A549 and MDCK cells (kindly provided by Kevin Harrod, Lovelace Respiratory Research Institute) were grown and maintained in complete minimum essential medium (MEM) supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin, and l-glutamine (Invitrogen) at 37°C under a 5% CO₂ atmosphere. The influenza virus A/PR/8/34 (H1N1) strain (kindly provided by Kevin Harrod, Lovelace Respiratory Research Institute) was propagated in 10-day-old embryonated chicken eggs and titrated on MDCK cells. Human HDAC6 cloned in plasmid pcDNA3 (Invitrogen) (kindly provided by Tso-Pang Yao, Duke University) was prepared from Escherichia coli DH5α cells using a Qiagen plasmid purification kit.

Infection.

Cells were infected with IAV at a multiplicity of infection (MOI) of 0.5 to 1.0 PFU per cell. Virus inoculum was prepared in serum-free MEM and added to cell monolayers previously washed twice with serum-free MEM. For infection of MDCK cells, 1 μg/ml tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-trypsin (Sigma-Aldrich) was added to the virus inoculum. After 1 h of incubation at 35°C, the inoculum was removed, and cells were washed once with serum-free MEM. Fresh serum-free MEM was added, and cells were incubated again at 35°C. In some experiments, serum-free MEM was supplemented with tubacin (kindly provided by Ralph Mazitschek, Harvard University).

Western blotting (WB).

Cells were lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, and protease inhibitor cocktail). Equal amounts of proteins were resolved by 7% or 10% Tris-glycine SDS-PAGE and transferred onto nitrocellulose membrane (GE Healthcare). Membranes were probed with rabbit anti-HDAC6 (1:200; Santa Cruz), mouse anti-acetylated α-tubulin (1:1,000; Sigma-Aldrich), mouse anti-NP (1:1,000; Millipore), or goat anti-NP (1:1,000; kindly provided by Robert Webster and Richard Webby, St Jude Children's Research Hospital) or with rabbit anti-protein disulfide isomerase (PDI) (1:2,000; Sigma-Aldrich) antibody followed by horseradish peroxidase-conjugated or IRDye 680RD- or 800CW-conjugated anti-mouse, anti-goat, or anti-rabbit IgG antibody (1:5,000; Invitrogen or Li-Cor). Proteins were visualized by chemiluminescence or fluorescence, and images were acquired on an Odyssey Fc imaging system (Li-Cor). Images were exported as TIFF files and compiled in Adobe Photoshop Elements, version 9.

Overexpression of HDAC6.

Cells were transfected with plasmids using Lipofectamine 2000 (Invitrogen) by following the manufacturer's guidelines. Briefly, 4 × 10⁵ cells were grown to 80 to 90% confluence in 12-well culture plates (Corning). Routinely, 1 μg of plasmid DNA and 2 μl of Lipofectamine 2000 were diluted separately in OptiMEM I medium (Invitrogen), mixed together, and incubated for 20 to 30 min at room temperature. DNA-Lipofectamine 2000 complex was then added to the cells. Cells were incubated at 37°C for 48 h before infection.

Knockdown of HDAC6 expression.

Predesigned small interfering RNA (siRNA) oligonucleotides targeting the HDAC6 gene (catalogue number sc-35544) and a nontargeting control (catalogue number sc-37007) were purchased from Santa Cruz Biotechnology. siRNA oligonucleotides (at 50 nM) and 2 μl of Lipofectamine RNAiMax (Invitrogen) were diluted separately in OptiMEM I medium (Invitrogen), mixed together, and incubated for 20 to 30 min at room temperature. RNAiMax-siRNA complex was then mixed with 2 × 10⁵ cells and transferred to 12-well plates. Cells were incubated at 37°C for 72 h before infection.

Virus release assays.

The culture medium from infected cells was harvested, cleared of cell debris by low-speed centrifugation, and divided into two parts. One part was subjected to protein precipitation by trichloroacetic acid (TCA) (Calbiochem), whereas the other part was mixed with 0.3% bovine serum albumin (BSA) and titrated on MDCK cells, followed by a microplaque assay. For protein precipitation, ice-cold TCA was mixed with culture medium at a final concentration of 20%, and the mixture was incubated on ice for 30 min. The mixture was then centrifuged at 20,000 × g and 4°C for 30 min. The supernatant was removed carefully, and the pellet was washed twice with ice-cold acetone. The pellet was air dried and directly suspended in SDS-PAGE sample buffer (50 mM Tris-HCl, pH 6.8, 2% SDS, 30% glycerol, 10% dithiothreitol, and 0.04% bromophenol blue). Proteins were resolved by SDS-PAGE, and viral NP was detected by WB. For the microplaque assay, confluent monolayers of MDCK cells were infected with 10-fold serial dilutions of the culture medium. The viral inoculum was removed, and cells were overlaid with serum-free MEM containing 1 μg/ml TPCK-trypsin and 0.8% Avicel (RC-581; FMC Biopolymer). After 18 to 20 h of incubation, the overlay was removed, and the cells were fixed with 4% formalin and subsequently permeabilized with 0.5% Triton X-100. Cells were then stained with mouse anti-NP antibody (1:1,000), followed by horseradish peroxidase-conjugated anti-mouse IgG antibody (1:1,000). Plaques were developed by the addition of a substrate solution containing 0.4 mg/ml 3-amino-9-ethylcarbazole (Sigma-Aldrich) in 0.05 M sodium acetate buffer (pH 5.5) and 0.03% hydrogen peroxide (Calbiochem).

EM.

Cells were fixed in 3% glutaraldehyde and stained with 1% osmium tetroxide and 1% uranyl acetate. Samples were then dehydrated and embedded in Embed epoxy resin. Ultrathin sections were cut and visualized under a Philips CM100 BioTWIN transmission electron microscope (EM). Images were acquired, exported as TIFF files, and compiled in Adobe Photoshop Elements, version 9. The number of virions per micrometer of the plasma membrane was calculated from the 4.1- by 5.5-μm-diameter images of multiple cells.

Confocal microscopy.

Cells were fixed in 4% methanol-free paraformaldehyde and permeabilized with 0.2% Triton X-100. Cells were stained with mouse anti-acetylated α-tubulin antibody (1:100) followed by Alexa 594-conjugated rabbit anti-mouse IgG antibody (1:100; Invitrogen). Cells were then stained with fluorescein isothiocyanate (FITC)-conjugated mouse anti-NP (1:50; Millipore) or goat anti-hemagglutinin (HA) (A/PR/8/34) (1:100; kindly provided by Robert Webster and Richard Webby, St Jude Children's Research Hospital), followed by Alexa 488-conjugated rabbit anti-goat IgG antibody (1:100; Invitrogen). Alternatively, cells were stained with mouse anti-PB2 antibody clone F5-59 (NR-31694; obtained through BEI Resources, NIAID, NIH), followed by Alexa 594-conjugated rabbit anti-mouse IgG antibody. Cells were then stained with Alexa 488-conjugated mouse anti-acetylated α-tubulin antibody. Cells expressing severe acute respiratory syndrome (SARS) virus membrane (M) protein from plasmid pCAGGS (kindly provided by Carolyn Machamer, The Johns Hopkins University School of Medicine) were stained with mouse anti-M antibody (kindly provided by Kevin Harrod, Lovelace Respiratory Research Institute), followed by Alexa 488-conjugated goat anti-mouse IgG antibody. Cells were then stained with rabbit anti-βCOP antibody (Affinity Bioreagents), followed by Alexa 594-conjugated goat anti-rabbit IgG antibody. Finally, cells were stained with the DNA-binding dye Hoechst (5 μg/ml; Invitrogen), and coverslips were mounted in mounting medium (Sigma-Aldrich). Fluorescent images were acquired by sequential scanning on a Zeiss LSM 710 confocal laser scanning microscope. Acquired images were analyzed in Zen software (Zeiss) or ImageJ, version 1.48v (NIH), or exported as TIFF files and compiled in Adobe Photoshop Elements, version 9. Colocalization of red and green pixels was quantitated using the colocalization indices plug-in for ImageJ from Kouichi Nakamura (Kyoto University). A threshold of 150 was set for both red and green channels to reduce the nonspecific noise in the images. The percent pixel overlap was calculated using the Manders' overlap coefficient obtained from individual analyses (19).

Flow cytometry.

Cells were dissociated into single-cell populations using cell dissociation solution (Sigma-Aldrich). Cells were stained with goat anti-HA antibody, followed by Alexa 488-conjugated rabbit anti-goat IgG antibody. Finally, cells were fixed with 4% paraformaldehyde. Ten thousand cells were read on a FACSCalibur flow cytometer (BD Biosciences), and the data were analyzed using FlowJo software (Treestar).

Coimmunoprecipitation.

Cells were lysed in immunoprecipitation buffer (10 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10% glycerol, 0.1% SDS, 0.5% sodium deoxycholate, and protease inhibitor cocktail) on ice for 1 h. Insoluble material was removed by centrifugation at 14,000 × g and 4°C for 1 h. Cleared lysates were then incubated with Dynabeads protein G (Invitrogen) conjugated to normal mouse IgG, rabbit anti-PDI, mouse anti-acetylated α-tubulin, or mouse anti-NP antibody for 30 min at room temperature. The beads were then washed three times with phosphate-buffered saline, and proteins were released from beads by incubation in SDS-PAGE sample buffer for 5 min at 95°C. The proteins were resolved by SDS-PAGE, and acetylated α-tubulin or viral NP was detected in respective samples by WB using mouse anti-acetylated α-tubulin antibody or goat anti-NP antibody followed by Quick Western kit IRDye 680RD (Li-Cor)- or IRDye 800CW-conjugated anti-goat IgG antibody (Li-Cor), respectively.

Statistical analysis.

All statistical analyses were performed using Prism, version 6 (GraphPad). The P values were calculated using unpaired t tests for pairwise data comparisons (Fig. 1B, 2B, and 6I, J, and K) or using a one-way analysis of variance (ANOVA) test for multiple data set comparisons (Fig. 3B and 8A).

FIG 1.

Overexpression of HDAC6 inhibits IAV infection. (A and B) Effect of HDAC6 overexpression on IAV infection. A549 cells (4 × 10⁵) were transfected with empty pcDNA3 (pc) or pcDNA3 containing the human HDAC6 (HD) gene for 48 h. Cells were then infected with PR8 at an MOI of 1.0 for 24 h. The virion yield in the culture medium was measured by WB of viral NP (A) and by microplaque assay on MDCK cells (B). Data are presented as means ± standard errors of the means (n = 3). (C) HDAC6 overexpression and its effect on acetylated α-tubulin. Total lysate of the infected cells was prepared, and HDAC6, acetylated α-tubulin, viral NP, and protein disulfide isomerase (PDI) were detected by WB. NP and PDI were detected as the infection marker and loading control, respectively. (D) Transfection efficiency of A549 cells. Cells were transfected with a plasmid expressing green fluorescent protein as described above. After 48 h, the image was acquired on an Olympus inverted fluorescence microscope under a magnification of ×10.

FIG 2.

Knockdown of HDAC6 expression promotes IAV infection. (A and B) Effect of HDAC6 depletion on IAV infection. A549 cells (2 × 10⁵) were transfected with a nontargeting control (Cont) siRNA or HDAC6-targeting (HD) siRNA for 72 h. Cells were then infected with PR8 at an MOI of 1.0 for 24 h. The virion yield in the culture medium was measured by WB (A) and by microplaque assay (B). Data are presented as means ± standard errors of the means (n = 3). (C) HDAC6 depletion and its effect on acetylated α-tubulin. Total lysate of the infected cells was prepared, and HDAC6, acetylated α-tubulin, viral NP, and PDI were detected by WB.

FIG 6.

Acetylated microtubules are involved in the trafficking of viral HA to the plasma membrane. (A to F) Effect of tubacin treatment on the intracellular and cell surface distribution of HA. A549 cells were infected with PR8 at an MOI of 1.0 and either mock treated (−Tubacin) or treated with 10 μM tubacin (+Tubacin). After 24 h, cells were fixed, permeabilized, and stained with mouse anti-acetylated α-tubulin followed by Alexa 594-conjugated rabbit anti-mouse IgG antibody. Cells were then stained with goat anti-HA followed by Alexa 488-conjugated rabbit anti-goat IgG antibody (left and middle columns). Cells shown in the right column were stained without permeabilization. (G and H) Mock-infected cells were stained as described above. Finally, cells were stained with the DNA-binding dye Hoechst and analyzed by confocal microscopy. Scale bar, 10 μm. (I to K) Quantification of the cell-surface HA by flow cytometry. A549 cells were treated as follows: infected with PR8 at an MOI of 1.0 and mock treated or treated with 10 μM tubacin (Tub) (I); transfected with a nontargeting control (Cont) or HDAC6-targeting (HD) siRNA for 72 h and then infected with PR8 at an MOI of 1.0 for 24 h (J); transfected with empty pcDNA3 (pc) or pcDNA3 containing the human HDAC6 (HD) gene for 48 h and then infected with PR8 at an MOI of 1.0 for 24 h (K). For each treatment, single-cell populations were prepared and stained with goat anti-HA followed by Alexa 488-conjugated rabbit anti-goat IgG antibody. Ten thousand cells were analyzed by flow cytometry. Data are presented as means ± standard errors of the means (n = 3). Mock, mock infected cells.

FIG 3.

Inhibition of HDAC6 activity promotes IAV infection. (A and B) Effect of the inhibition of HDAC6 activity on IAV infection. A549 cells (4 × 10⁵) were infected with PR8 at an MOI of 1.0 and subsequently treated with the indicated concentrations of tubacin for 24 h. The virion yield in the culture medium was measured by WB (A) and by microplaque assay (B). Data are presented as means ± standard errors of the means (n = 3). (C) Effect of the inhibition of HDAC6 activity on acetylated α-tubulin. Total lysate of the infected cells was prepared, and acetylated α-tubulin, viral NP, and PDI were detected by WB.

FIG 8.

IAV vRNP interacts with acetylated α-tubulin. (A) Quantification of the colocalization of vRNP constituents NP and PB2 with acetylated α-tubulin (AcTub). Percent pixel overlap was calculated using Manders' overlap coefficient determined from the analysis of individual images. Data are presented as means ± standard errors of the means (n = 3 for M-βCOP; n = 15 for NP–acetylated α-tubulin and PB2–acetylated α-tubulin). (B) Coimmunoprecipitation of viral NP with acetylated α-tubulin. A549 cells (8 × 10⁵) were infected with PR8 at an MOI of 1.0 for 24 h. Cell lysates were then prepared, and immunoprecipitation (IP) was carried out with mouse IgG, rabbit anti-PDI, mouse anti-NP, and mouse anti-acetylated α-tubulin antibody. The total and coprecipitated NP was then detected by WB using goat anti-NP (left panel). Total acetylated α-tubulin in the lysate was detected by IP with mouse anti-acetylated α-tubulin and WB with the same antibody using a Quick Western kit reagent (right panel).

RESULTS

Ectopically expressed HDAC6 inhibits IAV infection.

To test our hypothesis that HDAC6 is an anti-IAV host factor, we analyzed IAV infection in cells overexpressing human HDAC6. Human lung epithelial A549 cells were transfected with an HDAC6-expressing plasmid or the empty vector. Cells were then infected with the influenza virus A/PR/8/34 (H1N1) strain (here to referred as PR8), and the culture medium and the infected cells were harvested separately. The culture medium was divided into two parts; one part was analyzed by Western blotting (WB), and the other part was titrated by plaque assay to measure the amount of total and infectious IAV progeny released, respectively. In addition, the lysates of infected cells were subjected to WB to analyze HDAC6 overexpression and the subsequent effect of overexpressing HDAC6 on its substrate, acetylated α-tubulin. We found that the overexpression of HDAC6 in A549 cells caused a reduction in IAV infection. The HDAC6-overexpressing cells released less total viral progeny (Fig. 1A) and about 3-fold fewer infectious virions (Fig. 1B) in the culture medium than the cells transfected with the empty vector. The overexpression of HDAC6 in the cells was confirmed by WB (Fig. 1C). A noticeable deacetylation of the acetylated α-tubulin in HDAC6-overexpressing cells compared to the vector-only control confirmed that ectopically expressed HDAC6 was enzymatically active (Fig. 1C). The transfection efficiency of A549 cells is shown in Fig. 1D.

Endogenous HDAC6 possesses anti-IAV properties.

After discovering that A549 cells with ectopically expressed HDAC6 produce less IAV progeny, we wanted to find out whether depletion of the endogenous HDAC6 would have the opposite effect. To test this, we knocked down the expression of HDAC6 in A549 cells using RNA interference. A549 cells were transfected with an HDAC6-targeting siRNA and a nontargeting control siRNA and subsequently infected with PR8. The culture medium and the cell lysates were then analyzed as described in the legend of Fig. 1. Indeed, in contrast to the results obtained with HDAC6 overexpression, the knockdown of HDAC6 expression made A549 cells produce more IAV progeny. The cells transfected with an HDAC6-targeting siRNA released more total virions (Fig. 2A) and about six times more infectious virions (Fig. 2B) than the cells transfected with the control siRNA. WB analysis of the infected cell lysates confirmed the depletion of HDAC6 expression and the resulting increase in the level of acetylated α-tubulin (Fig. 2C). Interestingly, a more prominent double band of viral NP was detected after depletion of HDAC6 (Fig. 2A). Influenza NP is known to be cleaved by caspase 3 in infected cells (20). HDAC6 is also cleaved by caspase 3 in IAV-infected cells (12); however, a potential link between HDAC6 and caspase pathways remains to be investigated.

Deacetylase activity plays a role in the antiviral property of HDAC6.

Because HDAC6 is an enzyme, we next wanted to determine whether the deacetylase activity of HDAC6 plays a role in its anti-IAV function. To accomplish this we employed tubacin, a specific and cell-permeable small-molecule inhibitor of HDAC6 (21). To inhibit the HDAC6 activity, tubacin binds to one of the two HDAC6 catalytic domains (a feature unique to HDAC6) possessing α-tubulin deacetylase function (21). A549 cells were infected with PR8 and subsequently treated with various concentrations of tubacin, and the medium and the cells were then analyzed as described above. We found that, like HDAC6 depletion, tubacin treatment increased the release of total and infectious viral progeny in the medium in a dose-dependent manner as measured by WB (Fig. 3A) and plaque assay (Fig. 3B), respectively. Compared to untreated cells, infected cells treated with 5, 10, and 15 μM tubacin released about 3.8-, 4.9-, and 5.2-fold more infectious virions, respectively (Fig. 3B). The inhibition of HDAC6 activity by tubacin was confirmed by detecting the acetylated α-tubulin level in infected cell lysates by WB. As expected, the level of acetylated α-tubulin in tubacin-treated cells increased in a dose-dependent manner compared to that in mock-treated cells (Fig. 3C). The above-described experiments were done using an MOI of 1 PFU per cell. To determine whether inhibition of HDAC6 function would have a similar effect on virion release from the cells infected at a higher MOI, cells were infected with PR8 at an MOI of 5 PFU per cell and treated with 10 μM tubacin, and virion release was measured by plaque assay. About a 3.5-fold increase in the virion release was observed from tubacin-treated cells compared to untreated cells (data not shown), indicating that HDAC6's role is mostly independent of the MOI used.

To further analyze the effect of HDAC6 inhibition on IAV infection, PR8-infected cells were examined by thin-section electron microscopy (EM). Consistent with the increased virion release observed after HDAC6 depletion and tubacin treatment, about twice the number of budding viral progeny were visualized on the plasma membranes of tubacin-treated cells (Fig. 4B and C) as on those of mock-treated cells (Fig. 4A and C). No obvious defects in the IAV morphology were observed after tubacin treatment. As with untreated cells, spherical virions, which were either already pinched off or ready to pinch off from the plasma membrane, can be visualized on the surface of tubacin-treated cells (Fig. 4B).

FIG 4.

HDAC6 inhibition increases IAV budding. MDCK cells were infected with PR8 at an MOI of 1.0 and mock treated (A) or treated with 10 μM tubacin (B). After 12 h, cells were processed and visualized by thin-section electron microscopy. (C) Quantification of the number of virions per μm of the plasma membrane. Data are presented as mean virion counts from 25 images. Tub, tubacin.

HDAC6 regulates the trafficking of viral hemagglutinin to the plasma membrane via its substrate, acetylated microtubules.

To understand the anti-IAV mechanism of HDAC6, we next determined the stage of the IAV life cycle targeted by HDAC6. IAV enters the host cell by receptor-mediated endocytosis, replicates its RNA genome in the nucleus, and assembles its progeny on the plasma membrane (22). For virion assembly, all viral components need to be transported to the plasma membrane. The EM data shown in Fig. 4 indicated that HDAC6-deficient cells assemble more virions on the plasma membrane, and to assemble more viral progeny a greater quantity of the viral components is needed at the plasma membrane. The acetylation of microtubules has been shown to promote the vesicular transport (23, 24). Therefore, to inhibit IAV infection, HDAC6 potentially impedes the trafficking of viral components to the plasma membrane by deacetylating the acetylated α-tubulin and consequently reducing the level of acetylated microtubules in the infected cells. We previously reported that IAV downregulates the activity of HDAC6 and increases the level of acetylated α-tubulin in epithelial cells (11). A time course experiment revealed that such an increase occurs during 6 to 12 h of infection (Fig. 5), which is the growth period of the IAV life cycle in cultured cells. Potentially, IAV is downregulating the HDAC6 activity to sustain or increase the level of acetylated microtubules in infected cells to maintain the normal trafficking of viral components. The data presented here so far support this hypothesis as the decrease in virion release and in the intracellular level of acetylated α-tubulin consequent to HDAC6 overexpression coincides with an increase in the intracellular level of viral NP (Fig. 1C). Conversely, an increase in virion release and intracellular acetylated α-tubulin level due to HDAC6 depletion and tubacin treatment corresponds with the decrease in intracellular NP level (Fig. 2C and 3C). Together, these observations indicate the regulation of the trafficking of viral components by HDAC6 via acetylated microtubules.

FIG 5.

Increase in acetylated α-tubulin level in IAV-infected cells occurs between 6 and 12 h of infection. MDCK cells were infected with PR8 at an MOI of 0.5 and harvested after 0, 6, 12, and 24 h of infection. Total cell lysates were prepared, and acetylated α-tubulin, total α-tubulin, viral NP, and actin (as a loading control) were detected by WB. The arrow points to the proteolytically processed form of viral NP.

The IAV particle consists of three major components, the envelope, viral ribonucleoprotein (vRNP) core, and M1 (22). The envelope consists of the viral membrane proteins hemagglutinin (HA), neuraminidase (NA), and M2, whereas vRNP mainly comprises nonmembrane viral proteins NP, PB1, PB2, PA, and genomic RNA (22). To examine whether the acetylated form of the microtubules is involved in the trafficking of IAV components to the plasma membrane, we visualized the intracellular and cell surface distribution of HA as the representative of viral envelope components in mock-treated and tubacin-treated cells. PR8-infected cells were stained for HA and acetylated microtubules and then analyzed by confocal microscopy. Before staining, cells were either permeabilized to visualize the intracellular distribution of HA or were not permeabilized to visualize the cell surface distribution of HA. We found that in mock-treated cells, HA was almost uniformly distributed in the cytoplasm, with some staining concentrated in the juxtanuclear region (Fig. 6B). A basal level of acetylated microtubules was visualized in mock-treated cells (Fig. 6A). However, in tubacin-treated cells, concomitant to an increase in the basal level of acetylated microtubules due to inhibition of HDAC6 activity (Fig. 6D), viral HA was redistributed from the cytoplasm to the plasma membrane (Fig. 6E). The redistribution of HA to the plasma membrane after tubacin treatment was more clearly visualized on the surface of nonpermeabilized cells (Fig. 6F). This indicated that the acetylated microtubules mediate the trafficking of viral HA and, potentially, of all envelope components to the plasma membrane. To further validate this, we quantified the amount of HA present on the surface of mock-treated and tubacin-treated cells. A single-cell population of PR8-infected cells was prepared, stained for HA, and analyzed by flow cytometry. Consistent with confocal microscopy data, about 2.5-fold more HA was detected on the surfaces of tubacin-treated cells than of mock-treated cells (Fig. 6I). Similarly, the amount of HA present on the surfaces of HDAC6-depleted and HDAC6-overexpressing cells was quantified by flow cytometry. As with tubacin-treated cells, about 2.0-fold more HA was present on the surfaces of the cells transfected with an HDAC6-targeting siRNA than with a nontargeting control siRNA (Fig. 6J). Conversely, there was about a 2.0-fold decrease in the amount of HA present on the surfaces of the cells transfected with an HDAC6-expressing plasmid than with an empty pcDNA3 plasmid (Fig. 6K). The Fig. 6I, J, and K data were consistent with the virion release data shown in Fig. 3, 2, and 1, respectively.

Acetylated microtubules are involved in the trafficking of IAV vRNP.

To analyze the role of acetylated microtubules in the trafficking of nonmembrane IAV components, we selected the vRNP complex as a representative component. PR8-infected cells were stained for vRNP complex and acetylated microtubules and were subsequently analyzed by confocal microscopy. The vRNP complex was visualized by staining for vRNP constituents NP (Fig. 7E and K) and PB2 (Fig. 7H). We found that vRNPs were more or less evenly distributed in the cytoplasm of infected cells as punctate structures (Fig. 7E and H), and many of these punctate structures were colocalized with acetylated microtubules (Fig. 7F and I). The colocalization of vRNP with acetylated microtubule filaments is visualized more clearly in a further magnified portion of the Fig. 7F image (Fig. 7C). To quantify the colocalization of vRNP with acetylated microtubules, Manders' overlap coefficient was determined from 15 independent images to calculate the percent pixel overlap between NP or PB2 and acetylated microtubules. Three images of plasmid-expressed SARS M protein colocalizing with the Golgi complex were used as a positive control. As shown in Fig. 8A, about 62 and 57% pixel overlap was observed between NP and acetylated microtubules and between PB2 and acetylated microtubules, respectively. A 93% pixel overlap between SARS M protein and the Golgi marker βCOP indicates the validity of the method used to quantitate the colocalization of vRNP with acetylated microtubules (Fig. 8A). SARS M protein is known to colocalize with the Golgi complex (25). Further, similar to the redistribution of HA, cytoplasmic vRNPs were redistributed toward the cell periphery in tubacin-treated cells (Fig. 7K, arrows). We counted the cells in 35 images (acquired under ×63 magnification) to quantitate the vRNP redistribution and found that about 36% of the cells showed a vRNP shift toward the cell periphery after tubacin treatment. These data indicate that acetylated microtubules mediate the trafficking of vRNPs, too. Finally, we carried out a coimmunoprecipitation assay to identify the molecular interactions between NP and acetylated α-tubulin. Lysates prepared from PR8-infected cells were incubated with Dynabeads protein G conjugated to normal mouse IgG, PDI, NP, or acetylated α-tubulin antibody. Coprecipitated NP was then detected by WB. Indeed, consistent with confocal microscopy data, viral NP coprecipitated with acetylated α-tubulin antibody but not with irrelevant mouse IgG or PDI antibody (Fig. 8B). These data were also consistent with an earlier observation where nonacetylated α-tubulin was coprecipitated with reconstituted vRNP (26). Collectively, these data indicate that acetylated microtubules mediate the trafficking of nonmembrane IAV components.

FIG 7.

Acetylated microtubules are involved in the trafficking of IAV vRNP. (A to I) Influenza virus vRNPs colocalize with acetylated microtubules. MDCK cells were either mock infected (A and B) or infected with PR8 at an MOI of 0.5 for 12 h (D to I). Cells shown in panels A to E were stained with mouse anti-acetylated α-tubulin antibody followed by Alexa 594-conjugated rabbit anti-mouse IgG antibody and then stained with FITC-conjugated mouse anti-NP antibody. Cells shown in panels G and H were stained with mouse anti-PB2 antibody followed by Alexa 594-conjugated rabbit anti-mouse IgG antibody and then stained with Alexa 488-conjugated mouse anti-acetylated α-tubulin antibody. Stained cells were then analyzed by confocal microscopy. Panels F and I show merged images of panels D and E and panels G and H, respectively. Panel C shows a higher magnification of the boxed portion of panel F. (J to L) Tubacin treatment causes the redistribution of cytoplasmic vRNP toward the cell periphery. MDCK cells were infected with PR8 at an MOI of 0.5 and treated with 10 μM tubacin. After 12 h, cells were stained for acetylated microtubules (J) and viral NP (K) as described for panels D and E, respectively, and analyzed by confocal microscopy. Panel L shows the merged image of panels J and K. Arrows in panel K point to the redistribution of vRNP toward the cell periphery. Scale bar, 10 μm.

DISCUSSION

We have demonstrated here that HDAC6 is an anti-IAV host factor by using A549, MDCK, and HeLa (data not shown) cell lines. HDAC6 is expressed in uninfected cells, and its endogenous level is easily detectable in A549 and HeLa cells as well as in primary normal human bronchial epithelial cells (11, 12). We were unable to detect the endogenous HDAC6 in MDCK cells because the antibodies available to human HDAC6 do not recognize canine HDAC6. Recently, several interferon-induced host restriction factors, including IFITM3, ISG15, MxA, and viperin, have been shown to restrict IAV infection (7–10). At the moment, we do not know whether HDAC6 is also inducible by interferon in IAV-infected cells and belongs to the above category of restriction factors. However, in response to IAV infection, HDAC6 polypeptide levels do increase in primary human bronchial epithelial cells, but the HDAC6 activity is downregulated (11, 12). This intriguing relationship between the increased HDAC6 polypeptide level and the downregulation of HDAC6 activity in IAV-infected cells requires more detailed analysis.

We have also shown here that HDAC6 inhibits IAV infection by downregulating the trafficking of viral components to the host cell plasma membrane through acetylated microtubules. Host restriction factors employ distinct mechanisms to block various steps of the viral life cycle to restrict virus infection (6). IFITM3, MxA, and viperin interfere with IAV entry, transcription, and release, respectively. IFITM3 inhibits the release of incoming vRNPs from the endosomes, thereby preventing vRNP entry into the nucleus (27), whereas human MxA targets viral NP, blocking the vRNP function and subsequent viral replication (9, 28). Viperin, on the other hand, inhibits the release of newly budded IAV particles by altering the plasma membrane fluidity by perturbing lipid rafts, the plasma membrane microdomains where IAV budding takes place (10, 29). Because IAV assembles at the host cell plasma membrane and because all of the viral components are required to be there for virus assembly, it is conceivable that some of the host restriction factors will interfere with viral component trafficking to restrict IAV infection. In fact, ARHGAP21, a Cdc42 GTPase-activating protein, has recently been shown to negatively regulate the trafficking of IAV NA to the plasma membrane (30). However, the extent of increase or decrease in the intracellular viral components in the event of HDAC6 overexpression or depletion needs further investigation.

Microtubules are the main cellular organelles that mediate anterograde and retrograde cargo transport in the cell, and most viruses take advantage of the microtubule-mediated transport pathways to enter or exit the host cell (31). Constituents of the IAV envelope (HA, NA, and M2) utilize the exocytic vesicular transport pathway for targeting to the plasma membrane (22), and such vesicular trafficking is mediated by microtubules. The transport mechanisms of vRNP and M1, the other two IAV components, were not well understood until recently. It has been shown lately that vRNP traffics via the Rab11 GTPase-mediated vesicular transport pathway, which is also closely linked to the microtubules (32–35). Further, because M1 interacts with HA and NP and regulates the nuclear export of vRNP, it is also likely to follow the same transport pathway (22). Thus, it can be inferred that host microtubules mediate the targeting of all IAV viral components to the plasma membrane. Therefore, to delineate the anti-IAV mechanism of HDAC6, we first looked at the regulation of the trafficking of IAV components by acetylated microtubules. However, this may not be the only mechanism by which HDAC6 exerts its anti-IAV function.

Microtubules are the largest cytoskeletal components and are involved in a variety of cellular functions, including intracellular transport, organelle positioning, cell shape and motility, and centrosome and cilium formation (36). To perform such diverse functions, microtubules associate with motor and nonmotor proteins and undergo posttranslational modifications such as acetylation and detyrosination (36). Acetylation was the second posttranslational modification of α-tubulin to be discovered after detyrosination (37). Therefore, the existence of acetylated microtubules has been known for some time, but their specific role in the transport of cellular cargo has only been described recently. It has been shown that acetylation promotes vesicular transport by promoting the binding of kinesin and dynein motor proteins to microtubules (23, 24). Further, enhanced acetylation of microtubules rescued the vesicular transport defect in Huntington's disease (24). A role for acetylated microtubules during virus infection has also been investigated lately. Herpesvirus and adenovirus have been shown to induce microtubule acetylation to facilitate viral entry (38, 39). But a direct role of HDAC6 in viral infection was not investigated in these reports. However, a direct role of HDAC6 in the infection of HIV, which, like IAV, also buds from the plasma membrane, has been investigated (40). It has been shown that HIV-1 induces the acetylation of microtubules to promote infection, and the overexpression or depletion of HDAC6 inhibited or enhanced HIV-1 infection, respectively (40, 41). It will be interesting to investigate the role of HDAC6 in infection of other viruses, especially negative-strand RNA viruses that bud from the plasma membrane.

Acetylated α-tubulin was the first HDAC6 substrate to be identified, and since then many HDAC6 substrates have been identified (14, 16). Besides acetylated α-tubulin, acetylated Hsp90, acetylated cortactin, and acetylated β-catenin could be the other potential effectors of the anti-IAV function of HDAC6. Hsp90 has been previously shown to regulate the assembly of IAV polymerase complex and genomic RNA synthesis (42, 43). Cortactin is an actin-binding protein and promotes actin polymerization and branching, and acetylation regulates the actin-binding ability of cortactin (18). We along with others previously have shown the interaction of IAV NP and M1 with actin (44, 45). Further, cortical actin has been proposed to maintain the correct organization of lipid rafts for budding of filamentous IAV virions (46). Beta-catenin, on the other hand, is a coactivator of interferon regulatory factor 3, and acetylation of β-catenin has been shown to regulate the transcription of type I interferon in Sendai virus-infected cells (47). Nevertheless, a specific and direct role of the acetylated form of Hsp90, cortactin, or β-catenin in IAV infection remains to be investigated.

IAV seasonal epidemics and occasional pandemics have serious medical and economic implications for global public health. A universal IAV vaccine is not available, and current IAV vaccines induce only short-term immunity (48). Further, currently approved antiviral drugs target viral components, and IAV has mutated those components to acquire drug resistance (49). As an alternative approach, the host signaling pathways are increasingly being studied as targets for novel antiviral therapies (50, 51). An advantage of this strategy is that it is less likely to induce viral resistance (50, 51). HDAC6 has been discovered to play a significant role in the development of various cancers, neurodegenerative diseases, and inflammatory disorders (14). Consequently, efforts are being mobilized to develop HDAC6 modulators for the treatment of relevant diseases (14, 52, 53). As an alternative anti-IAV approach, HDAC6 stimulators may be employed to stimulate the anti-IAV potential of endogenous HDAC6 to restrict IAV infection.