Human Parainfluenza Virus Type 2 V Protein Modulates Iron Homeostasis

hPIV-2 V protein interferes with interaction between FTH1 and NCOA4 and inhibits NCOA4-mediated ferritin degradation, leading to the inhibition of iron release to the cytoplasm. This iron homeostasis modulation allows infected cells to avoid apoptotic cell death, resulting in effective growth of hPIV-2.

ABSTRACT

Intracellular iron concentration is tightly controlled for cell viability. It is known to affect the growth of several viruses, but the molecular mechanisms are not well understood. We found that iron chelators inhibit growth of human parainfluenza virus type 2 (hPIV-2). Furthermore, infection with hPIV-2 alters ferritin localization from granules to a homogenous distribution within cytoplasm of iron-stimulated cells. The V protein of hPIV-2 interacts with ferritin heavy chain 1 (FTH1), a ferritin subunit. It also binds to nuclear receptor coactivator 4 (NCOA4), which mediates autophagic degradation of ferritin, so-called ferritinophagy. V protein consequently interferes with interaction between FTH1 and NCOA4. hPIV-2 growth is inhibited in FTH1 knockdown cell line where severe hPIV-2-induced apoptosis is shown. In contrast, NCOA4 knockdown results in the promotion of hPIV-2 growth and limited apoptosis. Our data collectively suggest that hPIV-2 V protein inhibits FTH1-NCOA4 interaction and subsequent ferritinophagy. This iron homeostasis modulation allows infected cells to avoid apoptotic cell death, resulting in effective growth of hPIV-2.

IMPORTANCE hPIV-2 V protein interferes with interaction between FTH1 and NCOA4 and inhibits NCOA4-mediated ferritin degradation, leading to the inhibition of iron release to the cytoplasm. This iron homeostasis modulation allows infected cells to avoid apoptotic cell death, resulting in effective growth of hPIV-2.

INTRODUCTION

Iron is an essential element in several cellular activities, including oxygen binding/transport/release, ATP production, and a variety of enzymatic reactions (1, 2). Excessive amounts of iron, however, can cause the generation of cytotoxic reactive oxygen species (ROS) through Fenton reactions (3). Therefore, the concentration of intracellular free iron needs to be tightly controlled. Cellular iron is mainly stored in ferritin, a ubiquitous iron-binding protein consisting of a mixture of 24 subunits of ferritin heavy chain (FTH1) and ferritin light chain (FTL). When intracellular iron levels are high, ferritin synthesis is induced at both transcriptional and posttranscriptional levels (4). To capture and store the excess intracellular free iron, FTH1 binds to iron and converts Fe(II) to Fe(III) by ferroxidase activity. The ferritin-bound form of iron is nontoxic but is unavailable for intracellular use. FTL does not have ferroxidase activity but is important for the long-term storage of iron (5). When intracellular iron levels are low, ferritin is degraded through the autophagy pathway by nuclear receptor coactivator 4 (NCOA4), resulting in the release of iron from lysosome to cytosol (6). This autophagic degradation of ferritin is called ferritinophagy. During this process, NCOA4 binds to ferritin and recruits it to lysosome to degrade it. NCOA4 recognizes FTH1 via its C-terminal region, and FTH1 Arg23 is essential for binding with NCOA4 (7). Mutant NCOA4 that lacks binding sites with FTH1 showed defective ferritinophagy, indicating that interaction between FTH1 and NCOA4 is essential for functional ferritinophagy.

Iron concentration affects not only cell viability but also growth of several viruses. High concentrations of iron inhibit replication of hepatitis C virus (HCV) (8), herpes simplex virus 1 (HSV-1), and bovine viral diarrhea virus (BVDV) (9). In contrast, iron removal by chelators inhibits the replication of human immunodeficiency virus type 1 (HIV-1) (10), West Nile virus (WNV) (11), and human cytomegalovirus (HCMV) (12). These viruses are taxonomically diverse, so iron seems to have multiple effects on viral growth; however, the molecular mechanisms behind iron involvement in virus replication have not been elucidated.

Human parainfluenza virus type 2 (hPIV-2) is a member of the Orthorubulavirus genus of the Paramyxoviridae (https://talk.ictvonline.org/taxonomy). It is an enveloped virus, and its genome is nonsegmented, negative-sense RNA. The genome contains six genes encoding the nucleocapsid (NP), phospho (P), V, matrix (M), fusion (F), receptor-binding (RBP), and large RNA polymerase (L) proteins (13). The unedited P gene of the Orthorubulavirus genus encodes the V open reading frame (ORF), whereas the insertion of two pseudotemplated G nucleotides shifts the mRNA to the P ORF (14). Therefore, the N-terminal region of V and P proteins has common amino acid sequences. The unique C-terminal region of V protein contains three Trp residues and seven Cys residues that are highly conserved among the Paramyxoviridae (13). The V protein interacts with several host proteins, such as MDA-5 (15), STATs (16), AIP1/Alix (17), LGP2 (18), TRAF6 (19), tetherin (20), Graf1 (21), caspase1 (22), inactive RhoA (23), profilin2 (24), and Cavin3 (25). Most of these host proteins interact with the C-terminal region of V protein, except for Graf1, which interacts with the N-terminal V/P common region (21).

In this study, we investigated whether iron is involved in hPIV-2 growth. We tried to identify iron-related proteins that affect hPIV-2 growth and analyzed interaction between these proteins and hPIV-2 proteins. Furthermore, we demonstrated the molecular mechanisms behind how iron affects hPIV-2 growth and how hPIV-2 modulates iron homeostasis in infected cells.

RESULTS

Iron concentration affects the growth of hPIV-2.

To investigate the effects of iron concentration on hPIV-2 growth, HeLa cells were treated with Tiron or deferoxamine mesylate (DFO), both of which are iron chelators, followed by infection with hPIV-2 at a multiplicity of infection (MOI) of 0.01. The virus titer was measured by plaque assay. Virus growth was significantly decreased at 48 and 72 h postinfection (hpi) by treatment of not only Tiron (Fig. 1A) but also DFO (Fig. 1B), indicating that iron contributes to the growth of hPIV-2.

FIG 1.

Effects of iron on hPIV-2 growth. HeLa cells were treated with 2.5 mM Tiron (A) or 100 μM deferoxamine mesylate (DFO) (B) for 3 h and were infected with hPIV-2 at an MOI of 0.01 for the indicated hours in the presence of each iron chelator. The amount of virus in the supernatants was measured by plaque assay. PFU/ml values are shown as the means from three independent experiments. *, P < 0.05 compared to values of nontreated cells. Error bars indicate standard deviations.

hPIV-2 inhibits punctate ferritin formation.

Ferritin is a key protein in iron homeostasis (4). To investigate whether hPIV-2 infection affects ferritinophagy, we analyzed subcellular localization of ferritin in hPIV-2-infected cells. HeLa cells were infected with hPIV-2 for 24 h. The cells were cultured for a further 24 h in the presence of ferric ammonium citrate (FAC). In mock-infected cells, FAC treatment induced punctate ferritin formation (Fig. 2, mock, +FAC). In contrast, hPIV-2 infection altered ferritin localization from granules to homogenous distribution throughout the cytoplasm of iron-stimulated cells (Fig. 2, hPIV-2, +FAC). These results suggest that hPIV-2 infection inhibits lysosomal degradation of ferritin because FAC-induced punctate ferritin localizes in lysosomes (6). Ferritin was hardly visible in hPIV-2-infected cells as well as mock-infected cells in the absence of FAC (Fig. 2, −FAC), indicating that hPIV-2 infection itself did not induce ferritin expression.

FIG 2.

Effects of V protein on subcellular localization of ferritin. At 24 h infected with or without hPIV-2, HeLa cells were treated with 540 μM FAC. After 24 h, cells were fixed, permeabilized, and stained with anti-ferritin (red) and anti-P (green). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). Scale bar, 10 μm.

FTH1 suppresses hPIV-2 growth.

FTH1 is a component of ferritin and has ferroxidase activity, which is essential for iron storage (4). To examine the effects of FTH1 on hPIV-2 growth, we generated a knockdown cell line of FTH1 (HeLa/FTH1 KD) (Fig. 3A). This cell line and its control cell line (HeLa/ctrl KD) were infected with hPIV-2 at an MOI of 0.01, and the virus titer was measured by plaque assay at different times postinfection. The virus growth in HeLa/FTH1 KD was 2-, 4-, and 5-fold lower than that in HeLa/ctrl KD at 24, 48, and 72 hpi, respectively (Fig. 3B).

FIG 3.

Effects of FTH1 on hPIV-2 growth. (A) Lysates of the indicated cell lines were subjected to immunoblotting using anti-FTH1 pAb. Actin was used as a loading control. (B) HeLa/ctrl KD and HeLa/FTH1 KD were infected with hPIV-2 at an MOI of 0.01 for the indicated number of hours. The amount of virus in the supernatants was measured as described in the legend to Fig. 1. Relative PFU/ml values are also shown. *, P < 0.05 compared to values of HeLa/ctrl KD. Error bars indicate standard deviations.

hPIV-2 V protein binds to FTH1.

Next, we investigated the interaction between hPIV-2 proteins and FTH1 using immunoprecipitation. COS cells were transfected with FLAG-tagged FTH1-encoded pCMV-3-Tag8 together with SRα carrying the hPIV-2 V or P gene. FTH1 could precipitate V but not P protein (Fig. 4A, lanes 1 to 4). A deletion mutant consisting of only the V/P common region (V/P) was not precipitated with FTH1 (Fig. 4A, lanes 5 and 6). Thus, the C-terminal region of V protein may be important for binding with FTH1. To determine the amino acids essential for binding with FTH1, one Trp mutant (V_{W178H/W182E/W192A}) and three Cys mutants (V_C193/197A, V_{C209/211/214A}, and V_C218/221A) were used for immunoprecipitation with FTH1. As shown in our previous studies (16, 26), V_{W178H/W182E/W192A} cannot bind to most of V's interacting partners, while none of the seven Cys residues are involved in the binding capacity to them (see the introduction). A similar case was shown in V-FTH1 interaction, meaning Trp mutation lost the binding capacity to FTH1, while all Cys mutants bound to FTH1 (Fig. 4A, lanes 7 to 14). These results indicate that Trp residues in the C-terminal region of V protein are necessary for its binding with FTH1.

FIG 4.

Interaction between FTH1 and hPIV-2 proteins. (A and C) COS cells were transfected with various combinations of the indicated plasmids. V/P indicates a deletion mutant consisting of only common regions of V and P proteins. After 48 h, cell lysates were analyzed directly by immunoblotting (input). Immunoprecipitates with anti-FLAG MAb were probed by anti-V/P or anti-FLAG MAb. Double and single asterisks indicate immunoglobulin heavy chain and light chain, respectively. All experiments were performed at least three times independently. (B) Schematic diagram of full-length FTH1 (full), its C-terminal mutant (ΔC), and N-terminal deletion mutant (ΔN) with C-terminal FLAG tag (F). Deleted regions are indicated by dotted lines.

To identify the region important for binding with the V protein, we used C-terminally truncated (FTH1 ΔC) and N-terminally truncated (FTH1 ΔN) mutants of FTH1 for immunoprecipitation (Fig. 4B). FTH1 ΔC, but not FTH1 ΔN, could precipitate V protein (Fig. 4C), indicating that the N-terminal region of FTH1 is important for binding with V protein.

NCOA4 enhances hPIV-2 growth.

hPIV-2 infection may inhibit ferritinophagy, so we analyzed the effects of NCOA4 on hPIV-2 growth using an NCOA4 knockdown cell line (HeLa/NCOA4 KD) (Fig. 5A). HeLa/NCOA4 KD and HeLa/ctrl KD cells were infected with hPIV-2 at an MOI of 0.01, and the virus titer was measured by plaque assay. hPIV-2 growth in HeLa/NCOA4 KD was approximately 2- and 3-fold higher than that in HeLa/ctrl at 48 and 72 hpi, respectively, while these cell lines showed similar growth patterns at 24 hpi (Fig. 5B).

FIG 5.

Effects of NCOA4 on hPIV-2 growth. (A) Lysates of the indicated cell lines were subjected to immunoblotting using anti-NCOA4 pAb. Actin was used as a loading control. (B) HeLa/ctrl KD and HeLa/NCOA4 KD were infected with hPIV-2 at an MOI of 0.01 for the indicated hours. The amount of virus in the supernatants was measured as described in the legend to Fig. 1. Relative PFU/ml values are also shown. *, P < 0.05 compared to values of HeLa/ctrl KD. Error bars indicate standard deviations.

NCOA4 binds to V protein.

We next investigated the interaction between NCOA4 and V protein. COS cells were transfected with FLAG-tagged NCOA4-encoded pCMV-3-Tag8 together with SRα encoding the V or P gene, followed by immunoprecipitation using anti-FLAG antibody (Ab). NCOA4 precipitated V but not P and V/P proteins (Fig. 6A, lanes 1 to 6), indicating the importance of the C-terminal region of V protein for binding with NCOA4. V_{W178H/W182E/W192A} was not immunoprecipitated with NCOA4, while V_C193/197A, V_{C209/211/214A}, and V_C218/221A retained the binding capacity (Fig. 6A, lanes 7 to 14), indicating the importance of Trp residues within the C-terminal region of V protein.

FIG 6.

Interaction between NCOA4 and hPIV-2 proteins. (A and C) COS cells were transfected with various combinations of the indicated plasmids. Immunoblotting and immunoprecipitation were performed as described for Fig. 4A. (B) Schematic diagram of full-length NCOA4 (full), its C-terminal mutant (ΔC), and N-terminal deletion mutant (ΔN) with C-terminal FLAG tag (F). Deleted regions are indicated by the dotted lines.

We determined the region of NCOA4 required for binding with V protein. C-terminally truncated (NCOA4 ΔC) and N-terminally truncated mutants of NCOA4 (NCOA4 ΔN) (Fig. 6B) were subjected to immunoprecipitation with V protein. V protein was coimmunoprecipitated with NCOA4 ΔC but not with NCOA4 ΔN (Fig. 6C), indicating that the N-terminal region of NCOA4 is important for binding with the V protein.

V protein interferes with binding of FTH1 and NCOA4.

FTH1-NCOA4 interaction is required for functional ferritinophagy. To investigate whether V protein affects binding of NCOA4 with FTH1, binding between FTH1 and NCOA4 in the presence of V protein was analyzed using immunoprecipitation. V protein interacts with both FTH1 and NCOA4 at their N-terminal regions (Fig. 7A), so V protein itself might mediate FTH1-NCOA4 binding. To exclude this possibility, we used the deletion mutant of NCOA4 (NCOA4 ΔN) (Fig. 6B). COS cells were independently transfected with myc-tagged FTH1, FLAG-tagged NCOA4 ΔN, or V protein, and the cell lysates were combined, followed by immunoprecipitation using anti-FLAG polyclonal Ab (pAb). FTH1 was precipitated with NCOA4 ΔN (Fig. 7B, lane 2). In the presence of wild-type (wt) V protein, the amount of precipitated FTH1 was decreased to approximately 50% (Fig. 7B, lane 3, and C). Expression of V_{W178H/W182E/W192A} did not affect the amount of precipitated FTH1 (Fig. 7B, lane 4, and C). These results indicate that V protein interferes with FTH1-NCOA4 interaction.

FIG 7.

Effects of V protein on interaction between FTH1 and NCOA4. (A) Summary of interactions between NCOA4, FTH1, and V. Arrows indicate the interaction. Interaction of V with FTH1 or NCOA4 are the results from Fig. 4 and 6. (B) COS cells were independently transfected with each plasmid carrying NCOA4 ΔN, FTH1, or V protein. After 48 h, cell lysates were combined and directly analyzed by immunoblotting (input). Immunoprecipitates with anti-FLAG pAb were probed by anti-myc, anti-FLAG, or anti-V/P MAb. All experiments were performed three times independently. (C) The quantitative densitometry of precipitated FTH1 in panel B was performed using ImageJ software (http://rsb.info.nih.gov/ij). Data are means from three independent experiments and are presented as the relative value (NCOA4 ΔN + FTH1 = 1). *, P < 0.05 compared to values of NCOA4 ΔN + FTH1.

FTH1 and NCOA4 affect hPIV-2-induced apoptosis.

The excess amount of free iron produces ROS, which causes the induction of apoptosis (27). We investigated the role of FTH1 and NCOA4 in hPIV-2-induced apoptosis. HeLa/ctrl KD, HeLa/FTH1 KD, and HeLa/NCOA4 KD were infected with hPIV-2 at an MOI of 0.01 for various times, and DNA ladder experiments were performed to measure apoptosis. Obvious DNA ladders were observed in hPIV-2-infected HeLa/ctrl KD (Fig. 8A, lanes 1 to 4), indicating hPIV-2-induced apoptosis. Formation of DNA ladders was promoted in HeLa/FTH1 KD (Fig. 8A, lanes 5 to 8). In contrast, DNA fragmentation was suppressed by NCOA4 knockdown (Fig. 8A, lanes 9 to 12). Apoptosis was also analyzed by caspase-3/7 activity assay. Caspase-3/7 activity was significantly higher in hPIV-2-infected HeLa/FTH1 KD than in HeLa/ctrl KD (Fig. 8B). In contrast, NCOA4 knockdown resulted in slightly lower caspase-3/7 activity (Fig. 8B). These results suggest that hPIV-2-induced apoptosis is promoted by NCOA4 and is inhibited by FTH1.

FIG 8.

Effects of FTH1 and NCOA4 on hPIV-2-induced apoptosis. HeLa/ctrl KD, HeLa/FTH1 KD, and HeLa/NCOA4 KD were infected with hPIV-2 at an MOI of 0.01 for the indicated number of hours. (A) DNA was extracted as described in Materials and Methods and subjected to electrophoresis. The graph shows the quantitative densitometry of the DNA ladder at 72 hpi in panel A performed as described for Fig. 7C. Data are means from three independent experiments and are presented as the relative values (HeLa/ctrl KD = 1). *, P < 0.05 compared to values of HeLa/ctrl KD. (B) Cell suspension was subjected to Caspase-Glo 3/7 assay. Data are means from three independent experiments and are presented as the relative value (0 hpi = 1). *, P < 0.05 compared to values of HeLa/ctrl KD.

DISCUSSION

Iron is essential for almost all living organisms, but excess iron is harmful, since it catalyzes ROS formation. Therefore, iron modulation is important not only for host cells but also for infectious microorganisms. In this study, we investigated the effects of iron on hPIV-2 growth. hPIV-2 was found to suppress NCOA4-mediated ferritinophagy, and hPIV-2 V protein bound to both FTH1 and NCOA4 (Fig. 4 and 6). NCOA4 binds to FTH1 to deliver it to lysosome, where ferritin is degraded via ferritinophagy. Therefore, we hypothesized that V protein affects FTH1-NCOA4 interaction. As expected, V protein interfered with the interaction between FTH1 and NCOA4 (Fig. 7B, lane 3, and C), but this inhibition was not observed in the presence of Trp-mutated V protein (Fig. 7B, lane 4, and C). Furthermore, rPIV-2 carrying Trp-mutated V protein could not inhibit FAC-induced punctate ferritin formation (data not shown). Thus, the binding of V protein to FTH1 and/or NCOA4 is suggested to be essential to inhibit FTH1-NCOA4 interaction and subsequent ferritinophagy. Both V protein and NCOA4 interact with the N-terminal region of FTH1 (Fig. 7A), so V protein might competitively inhibit the binding of NCOA4 with FTH1. Despite the inability of V protein to bind to NCOA4 ΔN (Fig. 6C, lane 4), V protein was coimmunoprecipitated by NCOA4 ΔN in the presence of FTH1 (Fig. 7B, lane 3), probably because FTH1 mediates the interaction between V protein and NCOA4 ΔN. V protein may indirectly bind to FTH1 and NCOA4 through an unidentified cellular protein(s).

Ferritin captures and sequesters iron to avoid ROS generation by excess iron, while NCOA4 promotes iron release by degradation of ferritin in lysosome (4, 6). FAC treatment was reported to cause granule formation of ferritin in lysosome (6), leading to ferritin degradation. hPIV-2 infection was found to inhibit FAC-induced punctate ferritin formation (Fig. 2), suggesting that hPIV-2 suppresses ferritinophagy. Inhibition of ferritinophagy reduces the amount of active iron and ROS, resulting in the suppression of apoptotic cell death (27).

As shown in Fig. 8, we observed apoptotic cell death by hPIV-2. hPIV-2-induced apoptosis was promoted by FTH1 knockdown and inhibited by NCOA4 knockdown, so hPIV-2 infection might trigger ROS generation. ROS production has been observed in several viruses under different mechanisms, including respiratory syncytial virus (RSV) (28), HIV-1 (29), and HCV (30), all of which lead to apoptotic cell death. hPIV-2 growth was inhibited in HeLa/FTH1 KD (Fig. 3B), where severe hPIV-2-induced apoptosis was observed (Fig. 8). In contrast, NCOA4 knockdown resulted in only limited apoptosis and promotion of hPIV-2 growth. These results suggest suppression of hPIV-2 growth by NCOA4-mediated ferritinophagy. NCOA4 knockdown significantly affected hPIV-2 growth at only 72 hpi (Fig. 5B), perhaps because NCOA4 knockdown slowly inhibited hPIV-2-induced apoptosis. This explanation is supported by the higher degree of hPIV-2-induced apoptosis in HeLa/NCOA4 KD than in HeLa/ctrl KD observed as late as 72 hpi (Fig. 8). In contrast, FTH1 knockdown affected hPIV-2-induced apoptosis during early infection (Fig. 8), suggesting that FTH1 knockdown affects apoptosis faster than NCOA4 knockdown.

We propose a model of modulation of iron homeostasis by hPIV-2 (Fig. 9). hPIV-2 V protein inhibits FTH1-NCOA4 interaction and subsequent delivery of ferritin to lysosome where ferritin is degraded, which leads to suppression of iron release to cytoplasm. This suppression allows infected cells to avoid apoptotic cell death, leading to effective growth of hPIV-2. A similar pattern was observed in HCMV. pUL38 of HCMV binds to ubiquitin-specific protease 24 (USP24) to reduce lysosomal degradation of ferritin (12). Binding of pUL38 with USP24 also contributes to NCOA4 stabilization. In contrast, hPIV-2 V protein directly targets ferritin and NCOA4 to avoid apoptosis. A viral protein directly modulating FTH1 and NCOA4, the main players of iron homeostasis, is reported here for the first time. hPIV-2 V protein was shown to contribute to inhibition of ferritinophagy and apoptosis by modulating iron homeostasis. Since Trp residues of V proteins are well conserved among the Paramyxoviridae, especially the Orthorubulavirus genus, their V proteins might have a similar role. Further study will aim to fully elucidate the mechanism behind iron regulation by hPIV-2.

FIG 9.

Models for iron homeostasis modulation by V protein. NCOA4 binds to and recruits FTH1 to lysosome to degrade it, resulting in the release of iron to cytoplasm. hPIV-2 V protein interferes with FTH1-NCOA4 interaction and suppresses NCOA4-mediated ferritin degradation in lysosome. This iron homeostasis modulation leads to a reduction of apoptosis.

MATERIALS AND METHODS

Cells and viruses.

Vero cells were grown in Eagle's minimal essential medium (MEM) supplemented with 10% fetal calf serum (FCS). COS, HEK293, and HeLa cells and their derivatives were grown in Dulbecco's modified Eagle's MEM (DMEM) containing 10% FCS. All cells were maintained in a humidified incubator at 37°C with 5% CO₂. hPIV-2 (Toshiba strain) was used in this study.

Antibodies and reagents.

Monoclonal antibodies (MAbs) against hPIV-2 V/P (315-1) and P proteins (335A) were previously described (31). MAb and polyclonal Ab (pAb) to FLAG were obtained from Sigma (St. Louis, MO, USA). Anti-actin and anti-myc MAbs were purchased from Wako (Osaka, Japan) and MBL (Nagoya, Japan), respectively. Anti-FTH1 (3998), anti-NCOA4 (ARA70 antibody; A302-272A), and anti-ferritin (200-401-090-0100) pAbs were obtained from Cell Signaling Technology (Danvers, MA, USA), Bethyl Laboratories (Montgomery, TX, USA), and Rockland (Limerick, PA, USA), respectively. Tiron monohydrate, deferoxamine mesylate salt, and FAC were purchased from Tokyo Chemical Industry (Tokyo, Japan), Sigma, and Nacalai Tesque (Kyoto, Japan), respectively.

Plasmids.

A pcDL-SRα296 vector carrying hPIV-2 V, P, or their mutants was previously described (31, 32). cDNAs of FTH1 and NCOA4 were obtained from HEK293 cell total RNA by reverse transcription-PCR (RT-PCR) as previously described (33). The cDNAs and their deletion mutants were cloned into a pCMV-3Tag-8 vector with 3× FLAG tags at their C termini (Stratagene, La Jolla, CA, USA). The cDNA of FTH1 was also cloned into a pEF4-Myc/His vector (Invitrogen, Carlsbad, CA, USA). These constructs were all confirmed by DNA sequencing.

Generation of knockdown cell lines of FTH1 or NCOA4.

DNA fragments encoding anti-FTH1 or anti-NCOA4 short hairpin RNA (shRNA) were cloned into pHygH1dTO (34). The shRNA target sequences of FTH1 and NCOA4 were 5′-CCTGTCCATGTCTTACTACTT-3′ (corresponding to nucleotides 105 to 125 of the FTH1 gene) and 5′-CCCAGGAAGTATTACTTAATT-3′ (corresponding to nucleotides 1694 to 1714 of the NCOA4 gene), respectively. HeLa cells were transfected with pHygh1dTO carrying anti-FTH1 or anti-NCOA4 shRNA using XtremeGENE HP (Roche, Basel, Switzerland) according to the manufacturer's instructions. Stable transfectants were selected with 100 μg/ml hygromycin (Invitrogen). Clones showing FTH1 or NCOA4 depletion with high efficiency were used as FTH1 (HeLa/FTH1 KD) and NOCA4 (HeLa/NCOA4 KD) knockdown cell lines. A HeLa cell line transduced with the empty vector pHygh1dTO was used as a control (HeLa/ctrl KD) (25, 33).

Plaque assay.

Vero cells grown in 12-well plates were infected with hPIV-2 diluted serially 10-fold in MEM without FCS and cultured in MEM containing 2% FCS and 0.8% SeaKem ME agarose (FMC Bioproducts, Rockland, ME, USA) for 5 days. The cells were then stained with 0.05% neutral red, and the number of plaques was counted.

Immunoblot and immunoprecipitation assays.

Cells were harvested and sonicated for 30 s three times in lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.6% NP-40. After centrifugation, the supernatants were separated by SDS-PAGE and then transferred to a nitrocellulose membrane, followed by Western blotting (WB). For immunoprecipitation, the supernatants were incubated with nProtein A Sepharose 4 Fast Flow (GE Healthcare Bio-Sciences, Piscataway, NJ, USA) preincubated with the appropriate Abs. Precipitated proteins were analyzed by WB.

Immunofluorescence assay.

Cells grown on cover glasses in 24-well plates to 50% confluence were fixed with 4% paraformaldehyde for 20 min and permeabilized with phosphate-buffered saline (PBS) containing 0.2% Triton X-100 for 15 min. The cells were then incubated for 60 min with the appropriate Abs. The secondary antibodies used were Alexa Fluor 488 goat anti-mouse IgG and Alexa Fluor 594 goat anti-rabbit IgG (Invitrogen). The cells were mounted with Fluoromount-G (SouthernBiotech, Birmingham, AL, USA) and analyzed by a fluorescence microscope (BZ-X810; Keyence Co., Osaka, Japan).

Detection of apoptosis.

For DNA ladder experiments, cells grown in six-well plates were lysed in lysis buffer containing 10 mM Tris-HCl (pH 8.0), 0.5% Triton X-100, 10 mM EDTA. After centrifugation, the supernatants were treated with 10 μg/ml RNase A (Nippongene, Toyama, Japan) and 50 μg/ml proteinase K (Roche) at 37°C for 60 min. DNA was extracted using phenol-chloroform-isoamyl alcohol and subjected to electrophoresis.

To measure the caspase-3/7 activity, the cell suspension was subjected to Caspase-Glo 3/7 assay (Promega, Madison, WI) according to the manufacturer’s protocol.