Hu Antigen R Regulates Antiviral Innate Immune Responses through the Stabilization of mRNA for Polo-like Kinase 2

Innate immune responses to virus infection are initiated upon the sensing of viral nucleic acid species by host pattern-recognition receptors, such as membrane-bound TLR3, TLR7, and TLR9 and cytosolic proteins retinoic acid–inducible gene I (RIG-I)–like receptors (RLRs) and cyclic GMP-AMP synthase (cGAS). TLR3 and TLR7 sense dsRNA and ssRNA, respectively, whereas TLR9 senses DNA. The RLRs RIG-I and melanoma differentiation-associated gene 5 (MDA5) are cytoplasmic RNA helicases that sense viral RNA, and cGAS is a cytoplasmic sensor for DNA (1, 2). Upon ligand ligation, they activate downstream signaling pathways, culminating in the induction of inflammatory cytokines and type I IFNs. TLR7 and TLR9 are known to play central roles in plasmacytoid dendritic cells (DCs) via the recruitment of the adapter MyD88, which eventually activates the transcription factors NF-κB and IRF7. TLR3 is expressed in various cell types, including conventional DCs, macrophages, and nonimmune cells, and it uses another adapter, TRIF, to activate NF-κB and IFN regulatory factor 3 (IRF3). RLRs and cGAS use the mitochondrial protein IFN-β promoter stimulator 1 (IPS-1) (also called MAVS) and the ER protein STING as an adapter, respectively, which likewise culminates in the activation of NF-κB and IRF3 (3, 4). NF-κB largely regulates the expression of inflammatory cytokine genes, whereas IRF3 and IRF7 regulate the expression of type I IFNs.

In the unstimulated condition, IRF3 is expressed in the cytoplasm. After viral infection or other simulation, IRF3 is phosphorylated by the kinase TBK1 and/or its related kinase IKKi (also known as IKKε) (5). This phosphorylation induces conformational changes in IRF3, which result in the formation of an IRF3 homodimer and its subsequent translocation into the nucleus, where it binds to target DNA and upregulates the transcription of type I IFN genes. The activation of IRF3 is tightly regulated by multiple mechanisms. IRF3 binding to the lipid phosphatidylinositol 5-phosphate, which is increased upon viral infection, causes IRF3 to be phosphorylated by TBK1/IKKi (6). The conjugation of the ubiquitin-like protein ISG15 by HERC5 is involved in sustained IRF3 activation (7). Additionally, Polo-like kinase (PLK) 2 is associated with IRF3 nuclear translocation (8). In contrast, multiple proteins, including PIN1, YAP, RBCK1, RAUL, PTEN, PP2A, MAPK phosphatase 5, SENP2, TRIM21 (Ro52), TRIM26, FoxO1, c-cbl, Rubicon, ERRα, MST1, and AGO2, have been reported to negatively regulate IRF3 activation via distinct mechanisms such as proteasome-dependent degradation, dephosphorylation, de-SUMOylation, and/or prevention of protein–protein interactions (9–18).

Posttranscriptional modifications, including mRNA decay and stabilization, are critical for the regulation of antiviral innate immune responses. Stabilization and degradation of mRNA for cytokines or signaling molecules contribute to maintaining the proper innate immune responses. RNA-binding proteins (RBPs), such as Regnase-1 (also known as Zc3h12a or Mcpip1), Roquin, and Arid5a, were found to regulate inflammatory responses by binding to their target mRNAs, which encode inflammatory cytokines like IL-6. Regnase-1 and Roquin recognize the stem-loop structure of the 3′ untranslated region (UTR) and promote the degradation of inflammation-related mRNAs (19). Regnase-1–deficient mice showed severe autoimmune inflammatory diseases (20). In contrast, Arid5a increases mRNA stability through binding to the 3′UTR of target mRNA (21).

HuR (also called ELAV-like protein 1[Elavl1]), an RBP that has three RNA-recognition motifs (RRMs), belongs to the Hu protein family, which is composed of HuR, HuB, HuC, and HuD. HuR is ubiquitously expressed, whereas HuB, HuC, and HuD are specifically expressed in neuronal tissues (22). Normally, HuR is localized in the nuclei, but stimulation with UV radiation, alterations in the microenvironment, or pathogen infections cause HuR relocalization to cytoplasmic dotlike structures known as stress granules (SGs). SGs contain numerous messenger ribonucleoprotein and RBPs, such as TIA1 and TIAR, and are formed under stresses such as changes in the microenvironment, chemical compounds, and bacterial or viral infection. SGs are thought to stall translation during stress exposure (23). HuR was shown to interact with U- and AU-rich mRNA within the 3′UTR, which are termed AU-rich RNA elements, and HuR positively regulates their mRNA stability and translation. Additionally, HuR increases the stability of target mRNAs by protecting the mRNAs from other RBPs or microRNAs (miRNAs) that promote mRNA degradation (22, 24). Numerous mRNAs, such as those for VEGF-A, COX-2, IL-8, CCL2, CCL8, TNF-α, and cyclins, have been characterized as HuR targets; in this manner, HuR contributes to many aspects of biological processes, such as development, inflammation, and cancer progression (22, 25–27). However, function of HuR in antiviral responses remains unclear.

In this study, we found that HuR plays an important role in Ifnb1 induction during RLR signaling. HuR binds to and stabilizes mRNA for PLK2, a kinase regulating IRF3 nuclear translocation.

HEK293T, RAW264.7, and mouse embryonic fibroblast (MEF) cells were cultured in DMEM (Nacalai Tesque) supplemented with 10% heat-inactivated FBS in a 5% CO₂ incubator. THP-1 cells were cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS in a 5% CO₂ incubator. MEF cells were derived from wild type (WT) C57BL/6 mice. Bone marrow–derived macrophages (BMMs) were obtained from mouse bone marrow cells cultured in RPMI 1640 supplemented with 10% FBS, 100 μM 2-ME, and 2 ng/ml mecombinant mouse M-CSF (BioLegend). High m.w. (HMW) polyinosinic-polycytidylic acid [poly(I:C)], low m.w. (LMW) poly(I:C), and LPS were purchased from InvivoGen. Sense and anti-sense IFN stimulatory DNA (ISD) sequences were synthesized (Fasmac) and annealed manually (sense, 5′-TACAGATCTACTAGTGATCTATGACTGATCTGTACATGATCTACA-3′). HMW poly(I:C), LMW poly(I:C), and ISD were each mixed with Lipofectamine 2000 (Life Technologies) at a ratio of 1:1 (microgram/microliter) in Opti-MEM (Life Technologies) for intracellular stimulation. The transcriptional inhibitor actinomycin D was purchased from Sigma-Aldrich. Newcastle disease virus (NDV) was prepared as described previously (3).

Full-length mouse HuR, HuB, HuC, HuD, and PLK2 coding sequence (CDS) were amplified from murine brain, lung, and bone marrow cDNAs and inserted into a pFlag-CMV-2 expression vector (Sigma-Aldrich). A series of mutants for HuR and PLK2 expression plasmids were generated by PCR from the full-length HuR pFlag-CMV-2 expression vector. For constructing the pGL3-Promoter Vector (pGL3) harboring mouse Plk2 3′UTR (pGL3-mPlk2 3′UTR), the Plk2 3′UTR sequence was amplified by PCR from murine bone marrow cDNA and inserted into XbaI-digested pGL3. Each of the deletion mutant vectors was generated by PCR from pGL3-mPlk2 3′UTR. The reporter plasmids for IFN-β and NF-κB and the expression plasmids for IPS-1, TRIF, and STING were constructed as described previously (3, 28).

Single-guide RNA (gRNA) targeting murine HuR exon 4 (gRNA no. 1, 5′-GAAGACATGTTTTCTCGGTT-3′; gRNA no. 2, 5′-GACCATGACACAGAAGGATG-3′) and mouse PLK2 exon 1 (gRNA, 5′-GCGGACTATCACCTACCAGC-3′) were inserted into pX330-U6-Chimeric_BB-CBh-hSpCas9 (Addgene). Partial fractions of the murine HuR and mouse PLK2 CDS, including the single-guide RNA–targeted site, were inserted into pCAG-EGxxFP (29). The pCAG-EGxxFP plasmid was provided by M. Ikawa (Osaka University). These plasmids were electroporated into RAW264.7 cells, and EGFP-positive cells were sorted and plated onto 96-well plates using FACSAria (BD Biosciences). After cells were grown, HuR and PLK2 deficiency was confirmed by sequence analysis and/or Western blotting (W.B.).

Mouse anti-HuR mAb (3A2; Santa Cruz Biotechnology), rabbit anti-TOM20 polyclonal Ab (FL-145; Santa Cruz Biotechnology), rabbit anti-IRF3 mAb (D83B9; Cell Signaling Technology), rabbit anti–phospho-IRF3 (Ser396) mAb (4D4G; Cell Signaling Technology), rabbit anti–NF-κB p65 mAb (D14E12; Cell Signaling Technology), rabbit anti-IRF3 polyclonal Ab (FL-425; Santa Cruz Biotechnology), rabbit anti–phospho–NF-κB p65 (Ser536) mAb (93H1; Cell Signaling Technology), rabbit anti-G3BP polyclonal Ab (G6046; Sigma-Aldrich), goat anti-actin polyclonal Ab (I-19; Santa Cruz Biotechnology), goat anti–lamin B Ab (C-20; Santa Cruz Biotechnology), and mouse anti–FLAG M2 mAb (Sigma-Aldrich) were purchased as commercially available products.

The small interfering RNA (siRNA) sequences for murine HuR, human HuR, and firefly luciferase are as indicated below: murine HuR, 5′-AGGUUGAAUCUGCAAAGCUUAUUTT-3′ (sense); human HuR, 5′-GCUCAGAGGUGAUCAAAGATT-3′ (sense); firefly luciferase, 5′-CGUACGCGGAAUACUUCGATT-3′ (sense). siRNAs were synthesized and annealed by Fasmac. siRNAs were electroporated into BMM and THP-1 cells using Neon. Before electroporation of siRNA, THP-1 cells were differentiated in the presence of 100 ng/ml PMA for 24 h. Cells were then washed with PBS and cultured with 10% FBS RPMI 1640 for 48 h before electroporation. BMM and THP-1 cells were subjected to experiments 2 d after electroporation. For the construction of short hairpin RNA (shRNA)–expressing retroviral vectors, the oligo DNA was inserted into the BglII and HindIII sites of pSUPER.retro.puro (OligoEngine). The oligo DNA sequences used are as follows: scrambled shRNA, 5′-CCTAAGGCTATGAAGAGATACTTCAAGAGAGTATCTCTTCATAGCCTTATTTTT-3′; HuR shRNA, 5′-GAGAACGAATTTAATTGTCAACTTTCAAGAGAAGTTGACAATTAAATTCGTTCTC-3′. These vectors were transfected into Platinum-E cells using Lipofectamine 2000. The produced supernatant was filtered through a 0.22-μm filter 48 h after transfection and infected into MEFs. Following 48 h of infection, the cells were treated with 2 μg/ml puromycin and cultured for 48 h. Surviving MEFs were used for experiments.

For the construction of HuR-expressing retroviral vectors, murine HuR CDS was inserted into pMXs-IRES-puro (Cell Biolabs) that contained Flag-tag and CMV promoter sequence. The retroviral vectors were transfected into Platinum-E cells using Lipofectamine 2000. The produced supernatant was filtered with a 0.22-μm filter 48 h after transfection and infected into RAW264.7 cells. After 48 h of infection, the cells were treated with 4 μg/ml puromycin for selecting virus-infected cells. Living cells were subjected to following experiments.

RAW264.7 cells were cultured in six-well plates and stimulated with 1 μg/ml HMW poly(I:C) for 1 or 3 h. Whole-cell lysates were prepared by lysing cells in 50 mM Tris-HCl (pH 8), 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS. After centrifugation at 800 × g for 10 min at 4°C, supernatants were collected and used as whole-cell lysates. Cytoplasmic and nuclear fractions were prepared as follows. Cells were lysed for 5 min on ice using hypotonic buffer (10 mM HEPES-KOH [pH 7.8], 10 mM KCl, 0.1 mM EDTA [pH 8], 0.1% Nonidet P-40, 1 mM DTT, and protease inhibitor mixture [Roche]). After centrifugation of the samples at 2300 × g for 5 min at 4°C, supernatants were collected and used as cytoplasmic fractions. The remaining pellets were then lysed in hypertonic buffer (50 mM HEPES-KOH [pH 7.8], 420 mM KCl, 0.1 mM EDTA [pH 8], 5 mM MgCl₂, 1 mM DTT, and protease inhibitor mixture). The pellets were incubated for 30 min on ice and mixed by vortex every 10 min. After centrifugation of the samples at 17,800 × g for 10 min at 4°C, supernatants were collected and used as nuclear fractions. Whole-cell lysates and cytoplasmic and nuclear fractions were all subjected to SDS-PAGE and transferred to an Immun-Blot PVDF membrane (Bio-Rad). The membrane was then immunoblotted with the indicated Abs. The bound Abs were visualized with HRP-conjugated Abs against mouse, rabbit, or goat IgG (Sigma-Aldrich) using Western Lightning Plus-ECL (PerkinElmer). HRP activity was detected by a LAS 4000 (Fujitsu Life Sciences).

Cells were stimulated with 1 μg/ml HMW or LMW poly(I:C) for 8 h, with 1 μg/ml ISD for 6 h, or with 1 μg/ml LPS for 2 h. To measure mRNA stability, transcriptional activity was terminated by the addition of actinomycin D (2.5 or 5 μg/ml). Total RNA was extracted with TRIzol reagent (Invitrogen), and total RNA was used for cDNA synthesis using ReverTra Ace (Toyobo) according to the manufacturer’s instructions. Real-time quantitative PCR (RT-qPCR) was performed using Power SYBR Green PCR Master Mix (Applied Biosystems) with a LightCycler 96 system (Roche Diagnostics). The primers used for RT-qPCR analysis are described in Supplemental Table I.

Cells were seeded on 96-well plates and stimulated with 1 μg/ml HMW poly(I:C) or 1 μg/ml ISD for 24 h or 1 μg/ml LPS for 12 h. The cytokine levels of IFN-β, CXCL10, and IL-10 in the culture supernatant were measured by using a Lumikine mIFN-β (InvivoGen), Mouse CXCL10 DuoSet ELISA (R&D Systems), and Mouse IL-10 Uncoated ELISA (Invitrogen) according to the manufacturer’s instructions.

HEK293T cells were harvested in 24-well plates and transiently transfected with 100 ng of reporter plasmid for IFN-β, NF-κB, Plk2 3′UTR, or one of the Plk2 3ʹUTR deletion mutants and 500 ng of expression plasmid or empty plasmid. As an internal control, 10 ng of pRL-TK (Promega) was transfected simultaneously. The medium was replaced at 6 h posttransfection. After 24 or 48 h of transfection, luciferase activity was measured with a TriStar² LB 942 Multidetection Microplate Reader (Berthold) using the Dual-Glo Luciferase System (Promega) according to the manufacturer’s instructions.

Cells were cultured on poly-l-lysine–coated coverslips in 24-well plates. Cells were stimulated with 1 μg/ml HMW poly(I:C) for 6 h and fixed with 4% paraformaldehyde for 20 min. Cells were washed three times with 0.02% Triton(R) X-100 in PBS and permeabilized with PBS containing 100 mM glycine and 0.02% Triton(R) X-100 for 30 min. Cells were then incubated overnight at 4°C in PBS with 10% FBS and 0.02% Triton(R) X-100, after which they were reacted overnight at 4°C with the indicated Ab in PBS with 10% FBS and 0.02% Triton(R) X-100. Cells were washed and incubated for 2 h with anti-mouse and/or anti-rabbit secondary Ab conjugated to Alexa Fluor 488 or 568. Nuclei were stained with Hoechst 33342. Stained cells were mounted with Fluoro-KEEPER Antifade Reagent (Nacalai Tesque). Fluorescence images were obtained by LSM700 (Carl Zeiss), and the percentages of cells showing nuclear localization of IRF3 or NF-κB p65 were determined by counting 200 cells per coverslip.

Cells were stimulated with 1 μg/ml HMW poly(I:C) for 8 h. Total RNA was then extracted using TRIzol reagent and purified with an RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. Total RNA was amplified and biotinylated using the Ovation RNA Amplification System V2 and Encore Biotin Module (both from NuGEN) according to the manufacturer’s instructions. The biotinylated cDNA was hybridized with a GeneChip Mouse Gene 2.0 ST Array (Affymetrix) using an Affymetrix GeneChip Fluidics Station 450, and the array was scanned with a GeneChip Scanner 3000 7G (Affymetrix). The microarray dataset has been deposited in the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE103459.

For preparing the Ab-conjugated beads, Protein-A Sepharose beads were washed by NT2 buffer (50 mM Tris-Hcl [pH 7.4], 150 mM NaCl, 1 mM MgCl₂, and 0.05% Nonidet P-40) and incubated with NT2 buffer supplemented with 5% BSA at 4°C for 2 h. After washing with NT2 buffer, IgG or anti-HuR Ab was added to the bead slurry, and samples were incubated overnight at 4°C, followed by washing with NT2 buffer. To prepare the messenger ribonucleoprotein lysate, RAW264.7 cells were stimulated with or without 1 μg/ml HMW poly(I:C) using Lipofectamine 2000. Following 8 h of stimulation, the cells were suspended in polysome lysis buffer (100 mM KCl, 5 mM MgCl₂, 10 mM HEPES [pH 7], 0.5% Nonidet P-40, 1 mM DTT, RNaseOUT, and protease inhibitor mixture) and lysed by pumping using a 26G syringe needle (Terumo). After centrifugation at 15,300 × g for 15 min at 4°C, the resulting supernatants were rotated with the Ab-conjugated beads for 2 h at room temperature. Beads were then washed five times with NT2 buffer and treated with TRIzol reagent to extract the RNA.

Statistical significance was determined by Student unpaired t test and ANOVA with Tukey test. A p value <0.05 was considered significant.

To understand the involvement of Hu family proteins in antiviral innate immunity, we constructed expression plasmids for Hu family proteins HuR, HuB, HuC, and HuD (Fig. 1A) and cotransfected HEK293T cells with each individual plasmid along with a luciferase reporter plasmid driven by the IFN-β promoter. Overexpression of HuR, but not of HuB, HuC, or HuD, markedly increased the IFN-β promoter activity (Fig. 1A). Furthermore, HuR overexpression activated the IFN-β promoter in a dose-dependent manner, whereas it failed to activate the NF-κB promoter (Fig. 1B). We then constructed a series of deletion mutants of the HuR expression plasmid in which individual RRM domains were removed. The results showed that the deletion of each RRM domain abrogated the IFN-β promoter activity, suggesting that HuR-mediated IFN-β promoter activity requires its RNA-binding properties (Fig. 1C).

To examine the cellular localization of HuR, we stained MEF cells with anti-HuR Ab and found that HuR localizes exclusively in the nuclei (Fig. 1D). Following stimulation with transfection of HMW poly(I:C) (an MDA5 ligand), HuR was observed in dotlike structures in the cytoplasm that were partially merged with Ras-GAP SH3 domain binding protein (G3BP), a marker of SGs (Fig. 1D). These results suggest that HuR localizes in a portion of SGs in poly(I:C)-stimulated cells.

To understand the roles of HuR in antiviral innate immunity, we knocked down HuR expression in BMMs and the human monocyte cell line THP-1. Knockdown of HuR mRNA in BMMs was confirmed by RT-qPCR and Western blot (Fig. 2A, 2B). We found that expression of Ifnb1 and Cxcl10 (encoding IP-10 chemokine) after HWM poly(I:C) stimulation was decreased in HuR knockdown BMMs (Fig. 2C). Similarly, HuR knockdown in THP-1 cells also inhibited induction of IFNB1 and CXCL10 mRNA in response to poly(I:C) (Fig. 2D–F). Next, we examined the role of HuR in MEF cells. We prepared a shRNA-expressing retrovirus vector targeting HuR and transferred it into MEF cells. The efficiency of HuR knockdown was verified by RT-qPCR (Fig. 2G) and W.B. (Fig. 2H). The expression of Ifnb1 and Cxcl10 after HMW poly(I:C) stimulation was reduced in HuR knocked-down cells compared with that in control cells (Fig. 2I). In contrast, the inductions of Ifnb1 and Cxcl10 were comparable between control and HuR knocked-down cells following stimulation with bacterial LPS (a TLR4 agonist) or ISD (a cGAS agonist) (Fig. 2I). These results suggest that HuR is required for expression of IFN-β following poly(I:C) stimulation.

Using the CRISPR/Cas9 systems, we established a macrophage cell line RAW264.7 that lacks HuR (knockout [KO]1, KO2). The HuR deficiency of these cells was verified by DNA sequencing and Western blot analyses (Fig. 3A, 3B, Supplemental Fig. 1A). We measured the expression levels of Ifnb1, Cxcl10, and Il10 following stimulation with HMW poly(I:C), LMW poly(I:C) (a RIG-I agonist), ISD, or LPS (Fig. 3C–F). After stimulation with HMW or LMW poly(I:C), the mRNA expression levels of Ifnb1 and Cxcl10 were lower in HuR-deficient cells compared with WT cells (Fig. 3C, 3D). In contrast, Il10 expression was comparable between WT and HuR-deficient cells. Moreover, the mRNA expression of Ifnb1, Cxcl10, and Il10 was unimpaired in response to ISD or LPS in HuR-deficient cells (Fig. 3E, 3F). Consistent with mRNA expression, IFN-β and CXCL10 protein production induced by HMW poly(I:C) was also reduced in HuR-deficient cells (Fig. 3G). In contrast, protein levels of IFN-β and CXCL10 after stimulation with ISD and LPS were comparable between WT and HuR-deficient cells (Fig. 3H, 3I). We also generated HuR-deficient RAW264.7 cells using another sequence of gRNA (Supplemental Fig. 1B–E), which also showed reduced Ifnb1 and Cxcl10 expression after poly(I:C) stimulation (Supplemental Fig 1C). We then evaluated the responses of HuR-deficient cells against virus infection. Ifnb1 and Cxcl10 mRNA expression levels following infection with NDV, which is sensed by RIG-I, were markedly reduced in HuR-deficient cells compared with those in WT cells (Fig. 3J). Additionally, the amount of NDV transcript was higher in HuR-deficient cells than in WT cells, as measured by RT-qPCR (Fig. 3K). We also confirmed that complementation of Flag-tagged HuR into HuR-deficient cells restored the mRNA expression of Ifnb1 and Cxcl10 in response to HMW poly(I:C) (Fig. 3L, 3M). These results suggest that HuR regulates RLR-mediated antiviral innate immune responses.

To examine the involvement of HuR in SG formation, SGs induced by poly(I:C) stimulation were examined in WT and HuR-deficient RAW264.7 cells using anti-G3BP Ab (Supplemental Fig. 2A). G3BP was recruited into cytosolic dotlike structures in both WT and HuR-deficient cells, suggesting that HuR deficiency does not affect the induction of SG formation.

We conducted microarray analyses of poly(I:C)-stimulated WT and HuR-deficient RAW264.7 cells (Supplemental Fig. 2B, 2C). Compared with WT cells, HuR-deficient cells showed lower expression levels of various chemokine genes, such as Ccl2, Ccl7, and Ccl12. Additionally, the transcript amount of Plk2 was also lower in HuR-deficient cells. We then performed RT-qPCR to measure the expression level of signaling molecules that are involved in innate immune responses. The mRNA expression level of Plk2 was lower in HuR-deficient cells, whereas that of other PLK family members (Plk1, Plk3, and Plk4) was not different from those in WT cells (Fig. 4A, Supplemental Fig. 2D). Expression levels of RLR signaling molecules, such as Mavs, Tmem173 (encoding STING), Ticam1 (TRIF), Tbk1, Traf2, Traf6, Mapk14 (p38α), and Map3k7 (TAK1), were not different in HuR-deficient cells (Fig. 4B). In addition, restoration of HuR expression into HuR-deficient cells increased Plk2 mRNA expression (Fig. 4C). These results suggest that HuR affects the abundance of Plk2 mRNA.

To understand the roles of PLK2 in the induction of antiviral innate immune responses, we examined whether PLK2 enhanced IFN-β promoter activity. Overexpression of PLK2 in HEK293T cells increased the IFN-β promoter activity (Fig. 5A). By contrast, overexpression of a kinase-negative mutant PLK2 (K108M) failed to enhance the IFN-β promoter activity (Fig. 5B). These results suggest that PLK2 promotes IFN-β induction via its kinase activity. Moreover, the expression level of Plk2 was upregulated in response to poly(I:C) stimulation in RAW264.7 cells, whereas the expression of Plk1, Plk3, and Plk4 was not altered (Fig. 5C).

We then generated PLK2-deficient RAW264.7 cells using CRISPR/Cas9 systems. We obtained PLK2-deficient cells with deletion of 1 and 2 bp in the PLK2 exon1 sequence, creating frameshift mutations (Supplemental Fig. 3A, 3B). Expression of Plk2 mRNA was decreased in PLK2-deficient cells compared with WT cells, as verified by RT-qPCR (Supplemental Fig. 3C). We measured the expression levels of Ifnb1 and Cxcl10 in PLK2-deficient cells stimulated with poly(I:C) or ISD. The induction of Ifnb1 and Cxcl10 mRNA after poly(I:C) stimulation was reduced in PLK2-deficient cells when compared with WT cells, whereas expression of these cytokines in mRNA after ISD stimulation was comparable between WT and PLK2-deficient cells (Fig. 5D, 5E). These results suggest that PLK2 contributes to induction of antiviral innate immune responses to poly(I:C).

A previous study indicated that PLK2 is involved in the production of antiviral cytokines via mediating the nuclear translocation of IRF3 (8). Therefore, we next investigated if HuR affects IRF3 translocation to the nucleus. We stimulated WT and HuR-deficient RAW264.7 cells with HMW poly(I:C), and then whole-cell lysates were prepared and blotted with anti-IRF3 or anti-phosphorylated IRF3 (p-IRF3) Ab (Fig. 6A). HuR deficiency did not influence IRF3 phosphorylation. We then examined IRF3 nuclear translocation by separating the cell lysates into nuclear and cytoplasmic fractions. The amount of phosphorylated IRF3 and total IRF3 in the nuclei were both increased 3 h after poly(I:C) stimulation in WT cells, and these were significantly reduced in HuR-deficient cells (Fig. 6A, 6B). IRF3 phosphorylation in the cytoplasmic fraction was unaffected in HuR-deficient cells (Fig. 6A, 6B). To confirm further that HuR deficiency abrogates IRF3 nuclear translocation, we verified cellular localization of IRF3 by immunofluorescence. In WT cells, IRF3 localized to the cytoplasm in unstimulated condition. After poly(I:C) stimulation, ∼80% of cells showed IRF3 nuclear localization. However, cells displaying IRF3 nuclear localization were significantly reduced in HuR-deficient cells (Fig. 6C). In contrast, nuclear translocation of NF-κB p65 subunit after poly(I:C) stimulation was comparable between WT and HuR-deficient cells (Fig. 6D). We next examined IRF3 and NF-κB p65 nuclear translocation in PLK2-deficient cells. Whereas IRF3 nuclear translocation was significantly reduced in PLK2-deficient cells (Fig. 6E), NF-κB p65 nuclear translocation was unimpaired in these cells (Fig. 6F). These results strongly suggest that the HuR–PLK2 axis mediates antiviral innate immune responses via regulating the nuclear translocation of IRF3.

We next examined the interaction between HuR and Plk2 mRNA in RAW264.7 cells by RNA immunoprecipitation. The cell lysates prepared from RAW264.7 cells stimulated with HMW poly(I:C) were subjected to immunoprecipitation using control IgG or anti-HuR Ab. Following RNA extraction from these immunoprecipitates, we measured the abundance of Plk2 mRNA and found that the amount of Plk2 mRNA was enriched in anti-HuR immunoprecipitates (Fig. 7A). In contrast, the amounts of Plk1, Plk3, Plk4, and Mavs mRNA were comparable between IgG and anti-HuR Ab immunoprecipitates (Fig. 7A). To determine if HuR affects the mRNA stability of Plk2, we then measured the t_1/2 of Plk2 mRNA. WT and HuR-deficient RAW264.7 cells were treated with poly(I:C) and actinomycin D, which is a transcriptional inhibitor, and the t_1/2 of Plk2 mRNA was quantified by RT-qPCR. In HuR-deficient cells, Plk2 mRNA was destabilized compared with WT cells (Fig. 7B). Conversely, HEK293T cells overexpressing HuR showed sustained PLK2 mRNA expression compared with control cells (Fig. 7C).

To understand the involvement of the Plk2 3′UTR in Plk2 mRNA stability, we constructed a luciferase plasmid harboring the mouse Plk2 3′UTR, which consists of the sequence 2183–2802 of mouse Plk2 mRNA (pGL3-mPlk2 3′UTR). We transfected HEK293T cells with the control pGL3 plasmid or the pGL3-mPlk2 3′UTR plasmid, together with a mock or HuR expression plasmid. HuR overexpression increased the luciferase activity in cells transfected with pGL3-mPlk2 3′UTR but not in those transfected with control pGL3 (Fig. 7D). It has been reported that HuR preferentially binds to AU- and U-rich sequences (30, 31), and the Plk2 3′UTR contains four AU- and U-rich consecutive sequences. To clarify the involvement of these sequences in the stability of Plk2 mRNA, we constructed a series of deletion mutant plasmids lacking the Plk2 3′UTR sequences 2183–2214 (Δ1), 2626–2655 (Δ2), 2687–2711 (Δ3), and 2744–2760 (Δ4) (Fig. 7E). We then transfected these reporter plasmids into HEK293T cells and measured the resulting luciferase activity. The luciferase activity was decreased in cells transfected with the Δ4 mutant (Fig. 7F) but not in those transfected with any of the other mutants. Taken together, these results suggest that HuR regulates Plk2 mRNA stability by interacting with the Plk2 3′UTR.

In this study, we identified HuR, an RBP, as a critical regulator in antiviral innate immunity. HuR positively mediates antiviral immune responses through stabilizing Plk2 mRNA. It has been reported that HuR binds to numerous target mRNAs, including Ifnb1 mRNA (32). However, our results demonstrate that HuR overexpression induced IFN-β promoter activity and HuR deficiency suppressed the expression of Ifnb1 as well as that of other cytokines and chemokines in response to poly(I:C) stimulation or RNA virus infection. Moreover, IRF3 nuclear translocation was also suppressed by HuR deficiency. These findings suggest that HuR regulates antiviral innate immune responses by stabilizing mRNA encoding RLR signaling molecules. Among the innate immune molecules tested in this study, we found that Plk2 mRNA was strikingly reduced in HuR-deficient cells. Our RNA immunoprecipitation experiment using anti-HuR Ab showed that HuR bound to Plk2 mRNA and that the overexpression of HuR stabilized Plk2 mRNA. Although PLK2 and PLK4 have been shown to regulate TLR-mediated IRF3 nuclear translocation in DCs (8), our results suggest that the HuR-mediated stabilization of Plk2 mRNA is required for IRF3 nuclear translocation during RLR signaling. Intriguingly, the mRNA expression of Plk2, but not that of Plk4, was increased after poly(I:C) stimulation in RAW264.7 cells. Furthermore, PLK2-deficient RAW264.7 cells showed reduced expression of Ifnb1 in response to poly(I:C). These findings suggest that PLK2 plays a major role in antiviral responses in macrophages. HuR-deficient cells exhibited impaired IRF3 nuclear translocation during RLR signaling, whereas the phosphorylation of cytoplasmic IRF3 was unimpaired in these cells. Therefore, it may be possible that PLK2 catalyzes the phosphorylation of the substrate responsible for IRF3 nuclear trafficking. IRF3 translocation was shown to be mediated by PTEN, Rubicon, and IPO5/Importin-β3 (13, 14, 33), which may be regulated by PLK2.

HuR-deficient cells showed an impairment of IFN-β induction against cytosolic RNA, whereas they exhibited normal IFN-β induction responses against LPS and ISD. After poly(I:C) stimulation, HuR localizes to cytoplasmic dotlike structures known as SGs. SGs have been shown to play a role in protecting mRNAs from harmful conditions, and they contain a number of RBPs. Notably, RIG-I is also retained in a subset of SGs termed antiviral SGs (avSGs), and the inhibition of avSG formation abrogates RIG-I signaling, thus indicating that avSGs are a platform for viral RNA detection and the initiation of antiviral innate immune responses (34, 35). In HuR-deficient cells, the formation of SGs was not abrogated, which suggests that the observed reduction of IFN-β induction in HuR-deficient cells was not due to impaired SG formation (Supplemental Fig. 2A). Therefore, it is unlikely that HuR regulates RIG-I-mediated RNA sensing and the subsequent activation of downstream signaling. Rather, it may indirectly regulate antiviral innate immune responses through the stabilization of signaling molecules, such as PLK2, and the modulation of IRF3 translocation. However, it remains unclear how PLK2 regulates IRF3 nuclear translocation during RLR signaling. PLK2 may be recruited to SGs or mitochondria for interaction with RIG-I or IPS-1, which triggers RLR-mediated IRF3 nuclear translocation. Alternatively, PLK2 may be activated in response to RNA virus infection. We previously reported that the intracellular level of phosphatidylinositol 5-phosphate (PI5P) is increased during RLR signaling, which acts as the second messenger that directly binds and activates IRF3, and it is produced by the lipid kinase PIKfyve (6). Thus, it is speculated that PLK2 is directly activated by PIKfyve-dependent phosphorylation. This should be clarified in the future.

Previous studies reported that HuR profoundly interacts with the AU- and U-rich sequences such as AUUUUA, AUUUUUA, and UUUUUUU (30, 31). The Plk2 3′UTR contains four AU- and U-rich consecutive sequences. Our analysis using a series of deletion mutants of the Plk2 3′UTR demonstrated that the sequence from 2744 to 2760 in the Plk2 3′UTR is involved in HuR-mediated Plk2 mRNA stability. This sequence is also highly conserved in the Homo sapiens PLK2 3′UTR, suggesting that the sequence from 2744 to 2760 in the Plk2 3′UTR plays an important role in controlling the stability of Plk2 mRNA. Although it remains unclear how HuR increases the stability of Plk2 mRNA, it is possible that HuR prevents the binding of Plk2 mRNA by other RBPs and miRNA responsible for mRNA decay. For example, HuR has been shown to compete with TTP, KSRP, and AUF1, which promote mRNA degradation, for binding to mRNAs encoding TNF, iNOS, COX-2, and p16 (27, 36–38). Notably, HuR and TTP share more than 1000 genes as their targets (39). HuR also protects target mRNAs from degradation by miRNA, and it competes with let-7, miR-16, miR-200b, and miR-331-3p for antagonizing the function of each miRNA to degrade the mRNA encoding CAT-1, COX-2, VEGF-A, and ERBB-2, both in vitro and in vivo (24, 40–42). Therefore, Plk2 mRNA degradation in HuR-deficient cells is induced by competing with other RBPs or miRNAs that mediate Plk2 mRNA decay.

Type I IFNs and RLRs play essential roles in protection against infectious viruses and are also involved in exacerbating autoimmunity and bacterial infection (43, 44). Control of type I IFN production is crucial for achieving an appropriate host defensive response. Recently, small molecules that could inhibit an interaction between HuR and target mRNAs have been developed (45, 46). Therefore, these molecules may be useful to prevent autoimmunity in which type I IFNs and RLRs are involved.

We thank K. Abe and C. Suzuki for secretarial assistances and K. Oakley from Edanz Group (www.edanzediting.com/ac) for editing a draft of this article.

The authors have no financial conflicts of interest.