DDIT3 Targets Innate Immunity via the DDIT3-OTUD1-MAVS Pathway To Promote Bovine Viral Diarrhea Virus Replication

Extensive studies have demonstrated roles of DDIT3 in apoptosis and autophagy during viral infection. However, the role of DDIT3 in innate immunity remains largely unknown.

ABSTRACT

DNA damage-inducible transcript 3 (DDIT3) plays important roles in endoplasmic reticulum (ER) stress-induced apoptosis and autophagy, but its role in innate immunity is not clear. Here, we report that DDIT3 inhibits the antiviral immune response during bovine viral diarrhea virus (BVDV) infection by targeting mitochondrial antiviral signaling (MAVS) in Madin-Darby bovine kidney (MDBK) cells and in mice. BVDV infection induced high DDIT3 mRNA and protein expression. DDIT3 overexpression inhibited type I interferon (IFN-I) and IFN-stimulated gene production, thereby promoting BVDV replication, while DDIT3 knockdown promoted the antiviral innate immune response to suppress viral replication. DDIT3 promoted NF-κB-dependent ovarian tumor (OTU) deubiquitinase 1 (OTUD1) expression. Furthermore, OTUD1 induced upregulation of the E3 ubiquitin ligase Smurf1 by deubiquitinating Smurf1, and Smurf1 degraded MAVS in MDBK cells in a ubiquitination-dependent manner, ultimately inhibiting IFN-I production. Moreover, knocking out DDIT3 promoted the antiviral innate immune response to reduce BVDV replication and pathological changes in mice. These findings provide direct insights into the molecular mechanisms by which DDIT3 inhibits IFN-I production by regulating MAVS degradation.

IMPORTANCE Extensive studies have demonstrated roles of DDIT3 in apoptosis and autophagy during viral infection. However, the role of DDIT3 in innate immunity remains largely unknown. Here, we show that DDIT3 is positively regulated in bovine viral diarrhea virus (BVDV)-infected Madin-Darby bovine kidney (MDBK) cells and could significantly enhance BVDV replication. Importantly, DDIT3 induced OTU deubiquitinase 1 (OTUD1) expression by activating the NF-κB signaling pathway, thus increasing intracellular Smurf1 protein levels to degrade MAVS and inhibit IFN-I production during BVDV infection. Together, these results indicate that DDIT3 plays critical roles in host innate immunity repression and viral infection facilitation.

INTRODUCTION

Innate immune responses, the first line of defense against viral infection, are activated through the recognition of pathogen‐associated molecular patterns (PAMPs) by pattern recognition receptors (PRRs) (1–3). The RIG-I-like receptor (RLR) family is the major PRR that recognizes viral RNAs and initiates antiviral responses (4–6). Upon RNA viral infection, RIG‐I is activated after binding to cytoplasmic viral RNA, resulting in a conformational change that subsequently leads to its interaction with mitochondrial antiviral signaling protein (MAVS, also known as VISA, Cardiff, or IPS-1) (4, 7–11). Then, MAVS forms virus-like functional aggregates and serves as a platform for recruiting downstream proteins to initiate a series of signaling cascades (12). Subsequently, the MAVS signalosome activates the IRF3/7 and NF‐κB pathway, leading to the production of alpha/beta interferon (IFN-α/β) and cytokines, thereby inducing the expression of multiple interferon-stimulated genes (ISGs) that ultimately restrict viral replication (8, 12, 13).

As the sole adaptor protein required downstream of RLRs, MAVS plays an extremely important role in innate immunity against RNA viruses (14, 15). Therefore, some viruses have evolved mechanisms to regulate MAVS activity and availability to evade protective immune responses. For example, the human hepatitis A virus (HAV) 3C protease can directly degrade MAVS to escape innate immune responses after the precursor protein 3ABC is localized to the mitochondria (16). Sendai virus (SeV) and vesicular stomatitis virus (VSV) promote the expression of ovarian tumor (OTU) deubiquitinase 1 (OTUD1) to promote the ubiquitination and proteasomal degradation of MAVS (17). The Newcastle disease virus (NDV) V protein degrades MAVS through the ubiquitin-proteasome pathway via the E3 ubiquitin ligase RNF5 (18). Flaviviridae viruses have similar strategies to interfere with RLR signaling and IFN-I responses by regulating MAVS. Hepatitis C virus (HCV) NS3/4A specifically recognizes the cysteine at position 508 located in the transmembrane domain of MAVS to cleave MAVS (19). The HCV NS5A protein can interact with the leucine-rich PPR motif-containing (LRPPRC) protein, inhibiting MAVS binding to TRAF3 and TRAF6, which inhibits signal transmission by downstream signaling pathways (20). Both the dengue virus (DENV) and Zika virus NS4A proteins disrupt the RIG-I–MAVS interaction in mitochondrion-associated endoplasmic reticulum membranes (MAMs) by binding MAVS (21, 22).

Bovine viral diarrhea virus (BVDV), a member of the family Flaviviridae, is an enveloped, positive-sense, single-stranded RNA virus that causes numerous diseases in cattle, resulting in great economic losses in the cattle industry (23, 24). BVDV has a positive-sense, single-stranded RNA genome with a size of 12.3 kb that encodes a single open reading frame (ORF) including 4 structural proteins (Erns, envelope proteins E1 and E2, and a capsid protein) and 7 or 8 nonstructural proteins (N^pro, p7, NS2-3 or NS2 and NS3, NS4A, NS4B, NS5A, and NS5B) and is cleaved by viral or host cell proteases (25). N^pro, the first protein coded in the ORF, interacts with interferon regulatory factor 3 (IRF-3), resulting in the polyubiquitination and subsequent proteasomal degradation of IRF-3, thereby inhibiting IFN production (26, 27). The E^rns glycoprotein, which has double-stranded-RNA (dsRNA)-binding activity, can compete with PRRs to bind dsRNA and block the IFN-I signaling pathway. While further exploring the relationship between BVDV and innate immunity by transcriptomic sequencing, we found that DNA damage-inducible transcript 3 (DDIT3) expression was significantly upregulated, which means that DDIT3 may be involved in the BVDV replication process. However, the effect of BVDV infection on the key component MAVS in the antiviral response remains to be investigated.

DDIT3, also known as C/EBP homologous protein (CHOP) or growth arrest and DNA damage-inducible gene 153 (GADD153), is a member of the CCAAT/enhancer binding protein (C/EBP) family and is involved in the regulation of genes that encode proteins involved in proliferation, differentiation, expression, and energy metabolism (28–31). DDIT3 contains two functional domains, an N-terminal transcriptional activation domain and a C-terminal basic-leucine zipper (bZIP) domain (30). DDIT3, as a transcription factor, can induce apoptosis by regulating the transcription of antiapoptotic and proapoptotic genes during viral infection. Infectious bronchitis virus (IBV) infection can upregulate the expression of DDIT3 to further inhibit the expression of the antiapoptotic protein Mcl-1, thereby promoting cell apoptosis and the release of the virus (32). DDIT3 not only regulates virus infection through apoptosis but also participates in virus regulation through autophagy. During HCV infection, DDIT3 upregulates the transcription of LC3B by directly binding to the promoter of the autophagy gene LC3B, which promotes the development of autophagy (33).

In this study, we report a novel function of DDIT3 that supports BVDV proliferation by promoting the ubiquitin-proteasome pathway-mediated degradation of MAVS, resulting in the inhibition of antiviral signaling. Initially, we found that DDIT3 was positively regulated in MDBK cells infected with BVDV and could enhance BVDV replication. We further found that DDIT3 positively regulated Smurf1 protein levels by facilitating NF-κB-dependent expression of OTUD1. Moreover, BVDV infection promoted the interaction between Smurf1 and MAVS, resulting in the proteasome-dependent degradation of MAVS. Interestingly, increased protein expression of MAVS and minor pathological damage were detected in DDIT3-deficient mice, which triggered a relatively strong innate immune response against BVDV infection by increasing IFN-β production. Our findings indicate that DDIT3 can promote viral replication and play an essential role in innate immune responses against BVDV infection.

RESULTS

BVDV infection upregulates the expression of DDIT3 in MDBK cells.

The effect of BVDV infection on gene expression in Madin-Darby bovine kidney (MDBK) cells was initially studied by transcriptomic analysis. Differentially expressed genes (DEGs) were verified by quantitative reverse transcription-PCR (qRT-PCR), and we observed that BVDV infection induced an 8- to 9-fold upregulation of DDIT3 mRNA levels, while the expression of other genes did not change as much, indicating that BVDV infection greatly promoted the transcriptional activation of DDIT3 (Fig. 1A and B). To confirm the effect of BVDV infection on the expression of DDIT3, we measured the expression of DDIT3 between 12 and 48 h postinfection (hpi) by qRT-PCR and Western blotting. The results showed that compared to mock infection, BVDV infection upregulated the expression of DDIT3 (Fig. 1C and D). These results clearly demonstrate that BVDV infection triggers the expression of DDIT3 in MDBK cells and suggest that DDIT3 plays a role in viral infection.

FIG 1.

BVDV infection upregulates the expression of DDIT3. (A) Heat map of mRNA expression in MDBK cells infected with BVDV at 24 h. Colors indicate the log₂ ratios of virus to mock infection. MDBK cells were mock infected or infected with BVDV (MOI of 5) for 24 h. (B) Differentially expressed genes in MDBK cells infected with BVDV (MOI of 5) were identified by quantitative real-time PCR (qPCR). (C and D) qPCR and Western blotting were used to determine the mRNA and protein levels, respectively, of DDIT3 between 12 and 48 h after BVDV (MOI of 1) infection of MDBK cells. Means and SD from three independent experiments are shown. **, P ≤ 0.01; ns, not significant.

The DDIT3 gene promotes BVDV replication in MDBK cells.

To determine whether elevated levels of DDIT3 are associated with an increased BVDV load, MDBK cells were infected with recombinant lentiviruses that overexpress DDIT3 or empty vectors. Stable cell lines were constructed, the DDIT3-overexpressing stable cell lines were evaluated at 24 h after BVDV infection by Western blotting (Fig. 2A), and DDIT3 promoted BVDV replication, as determined by measuring relative viral RNA expression and viral titers (Fig. 2B and C). The growth curve of the virus showed that the viral titer in the virus-infected DDIT3-overexpressing stable cell lines was significantly higher than that in control cells at 12 to 24 hpi (Fig. 2D). Since DDIT3 overexpression facilitates BVDV replication, we further hypothesized that DDIT3 knockdown has a negative impact on BVDV replication. To test this hypothesis, the lentiviruses shDDIT3-1, shDDIT3-2, and shNC were generated and used to infect MDBK cells. We verified the knockdown efficiencies of shDDIT3-1 and shDDIT3-2 in MDBK cell lines upon BVDV infection at 24 h by Western blotting (Fig. 2E). Moreover, we found that downregulation of DDIT3 levels significantly reduced BVDV replication in DDIT3 knockdown MDBK cells by quantitative real-time PCR (qPCR), measuring viral titers and calculating the growth curve of the virus (Fig. 2F to H). These results suggest that the DDIT3 gene supports the proliferation of BVDV.

FIG 2.

DDIT3 gene promotes BVDV replication in MDBK cells. (A) Establishment of an MDBK cell line stably overexpressing DDIT3. The stable overexpression of DDIT3 was confirmed by Western blotting with a rabbit anti-DDIT3 polyclonal antibody (PAb) and comparing the DDIT3-overexpressing cell line with a negative-control (NC) MDBK cell line transduced with an empty retrovirus. (B) qPCR analysis of BVDV RNA levels in DDIT3-overexpressing stable cell lines and MDBK control cell lines at 24 and 48 h after BVDV infection (MOI of 1). (C and D) Virus replication in DDIT3-overexpressing or empty retrovirus-transduced MDBK cells infected at an MOI of 1. (C) Supernatants were collected at 24 and 48 hpi, and viral titers were determined by plaque assays with DDIT3-overexpressing cells or control MDBK cells. (D) The viral titers of DDIT3 cell lines and control cell lines at 24 to 48 hpi were determined by plaque assays. (E) Establishment of stable MDBK cell lines with knocked-down DDIT3 expression (shDDIT3). The stable knockdown of DDIT3 was confirmed by Western blotting with a rabbit anti-DDIT3 PAb and comparing the DDIT3 knockdown cell lines with a negative-control (shNC) MDBK cell line transduced with an empty retrovirus. (F) qPCR analysis of BVDV RNA levels in DDIT3 knockdown stable cell lines and control cell lines at 24 and 48 h after BVDV infection (MOI of 1). (G and H) Viral replication in DDIT3 knockdown or empty-retrovirus-transduced MDBK cells infected at an MOI of 1. (G) Supernatants were collected at 24 and 48 hpi, and viral titers were determined by plaque assays with DDIT3 knockdown cells or control cells. (H) The viral titers of DDIT3-knockdown cells and control cell lines at 24 to 48 hpi were determined by plaque assays. (I and J) MDBK cell lines stably overexpressing DDIT3 were treated with Z-VAD (5 μM) and then infected with BVDV at an MOI of 1 for 24 h. The expression of cleaved caspase 3 was examined by Western blotting (I), and the viral titers of DDIT3 cell lines and control cell lines at 24 hpi were determined by plaque assays (J). (K and L) MDBK cell lines stably overexpressing DDIT3 were treated with 3-MA (5 mM) and then infected with BVDV at an MOI of 1 for 24 h. The expression of LC3-I and -II was examined by Western blotting (K), and the viral titers of DDIT3 cell lines and control cell lines at 24 hpi were determined by plaque assays (L). Mean and SD from three independent experiments are shown. *, P ≤ 0.05; **, P ≤ 0.01.

Considering that the roles of DDIT3 in apoptosis and autophagy during viral infection, we separately tested the effect of DDIT3 on apoptosis or autophagy in BVDV infection and its effect on virus replication. Our results showed that DDIT3 induces higher levels of apoptosis during BVDV infection, and inhibiting apoptosis with the apoptosis inhibitor (Z-VAD) does not affect BVDV replication in DDIT3-overexpressing cell lines (Fig. 2I and J). These results also suggested that apoptosis has no effect on BVDV replication, consistently with previous studies (34, 35). It has been reported that autophagy induced by BVDV infection could promote BVDV replication at early stages (36–38). Moreover, we found that BVDV infection in DDIT3-overexpressing cells could induce higher levels of autophagy (Fig. 2K). After inhibition of autophagy in DDIT3-overexpressing cells, it was still about 6-fold higher than that of the control group (Fig. 2K and L). Taken together, these results suggest that DDIT3 could promote BVDV replication in a manner independent of apoptosis.

DDIT3 suppresses the expression of IFN-β, MX1, and ISG56 during BVDV infection.

We next investigated the molecular mechanisms that are responsible for the roles of DDIT3 in the innate immune response to BVDV. Given previous studies showing that IFN-Is have critical roles in antiviral innate immunity, we determined the expression of IFN-β, MX1, and ISG56 when DDIT3-overexpressing stable cell lines were infected with BVDV. The mRNA levels of IFN-β, MX1, and ISG56 were analyzed by qRT-PCR at 24 hpi. The results showed that DDIT3 overexpression reduced the levels of BVDV-induced IFN-β, MX1, and ISG56, and the inhibitory effect of DDIT3 on these genes was more significant, at approximately 50% (Fig. 3A and B). To confirm the role of DDIT3 in the regulation of the production of IFN-Is, endogenous DDIT3 expression was knocked down, and IFN-β, MX1, and ISG56 mRNA levels after BVDV infection were analyzed. The results showed that knocking down DDIT3 expression promoted the mRNA expression of IFN-β, MX1, and ISG56 (Fig. 3C and D). To further explore whether DDIT3 promotes BVDV replication primarily by inhibiting interferon antiviral responses, the IFN-independent role of DDIT3 in BVDV replication was examined in type I IFN-deficient cells. The knockdown efficiency of siRNA targeting IFNAR1 was confirmed by qRT-PCR (Fig. 3E). We silenced IFNAR1 in the DDIT3-overexpressing stable cell lines; then the cells were infected with BVDV, and the viral titers were measured at 24 hpi. After silencing of IFNAR1 in DDIT3-overexpressing stable cell lines or control MDBK cells, there was no significant difference in the viral titer of BVDV (Fig. 3F), indicating that DDIT3 promotes BVDV replication mainly by inhibiting the interferon antiviral response. Next, we explored whether the expression of DDIT3 was linked to the induction of IFN-I and ISG production. DDIT3 could not be induced by IFN-α in MDBK cells (Fig. 3G), indicating that the induction of DDIT3 by BVDV infection is independent of IFN response. Collectively, these results indicate that DDIT3 plays important roles in IFN-I induction in host defense against BVDV.

FIG 3.

DDIT3 suppresses the expression of IFN-β, MX1, and ISG56 during BVDV infection. (A and B) qPCR analysis of IFN-β, MX1, and ISG56 mRNAs in DDIT3-overexpressing or empty-retrovirus-transduced MDBK cells infected with BVDV at an MOI of 1. (C and D) qPCR analysis of IFN-β, MX1, and ISG56 mRNAs in DDIT3-knockdown or control MDBK cell lines infected with BVDV (MOI of 1). (E) The silencing efficiency of IFNAR1 was measured by qPCR in MDBK cells at 24 h posttransfection. (F) At 48 h after transfection with si-NC or si-IFNAR1, MDBK cells were infected with BVDV at an MOI of 1 and harvested at 24 hpi for virus titration analyses. (G) The MDBK cells was treated with cow IFN-α (200 U/ml) for 4, 8, and 12 h. The indicated protein was detected by Western blotting. Means and SD from three independent experiments are shown. **, P ≤ 0.01; ns, not significant.

DDIT3 promotes the degradation of MAVS to inhibit the innate immune response.

To determine the molecular function of DDIT3 in the BVDV-triggered signaling pathway, we examined the effects of DDIT3 on the transcription of IFN-β mediated by components of the virus-triggered pathway. DDIT3 inhibited the transcription of IFN-β induced by overexpression of RIG-I–caspase recruitment domain (CARD) and MAVS but did not inhibit IFN-β transcription induced by TBK1, IKKε, IRF3-5D, or IRF7 (Fig. 4A). Consistent with this finding, data generated by a reporter gene assay indicated that overexpression of DDIT3 significantly inhibited RIG-I–CARD- and MAVS-mediated activation of the IFN-stimulated response element (ISRE) promoter, whereas ISRE promoter activation by TBK1, IKKε, IRF3-5D, or IRF7 was not inhibited by DDIT3 coexpression (Fig. 4B). These results suggest that DDIT3 inhibits the IFN-mediated antiviral response downstream of MAVS and upstream of TBK1, which means that DDIT3 blocks the transmission of antiviral signals by inhibiting MAVS. Because MAVS activation is a critical event in antiviral signaling, we next determined whether the expression level of MAVS is affected by DDIT3. As shown in Fig. 4C, the total amount of MAVS in DDIT3-overexpressing cell lines was lower than that in negative-control (NC) cell lines at 12, 24, 36, and 48 hpi. Furthermore, DDIT3-overexpressing cells and control cells were infected with BVDV at a multiplicity of infection (MOI) of 5, and the degradation of MAVS in DDIT3-overexpressing cells was completely blocked by the proteasome inhibitor MG132 but not the autophagy inhibitor chloroquine (CQ) (Fig. 4D). The relatively high levels of endogenous MAVS ubiquitination in DDIT3-overexpressing cell lines also indicated that DDIT3 promoted the ubiquitination and degradation of MAVS (Fig. 4E). These collective data suggest that DDIT3 promotes the degradation of MAVS via the proteasome pathway during BVDV infection.

FIG 4.

DDIT3 promotes the degradation of MAVS to inhibit the innate immune response. (A) HEK293T cells were transfected with the indicated expression plasmids expressing key proteins in the antiviral response together with control or DDIT3 plasmids for 24 h. After 24 h, qPCR was performed to detect IFN-β mRNA levels. (B) Luciferase reporter assays were performed with HEK293T cells cotransfected with ISRE promoter reporter plasmids and pRL-TK plasmids plus the indicated expression plasmids together with control or DDIT3 plasmids for 24 h. Means and SDs from three independent experiments are shown. **, P ≤ 0.01. (C) Immunoblot analysis of MAVS in negative-control MDBK cells and DDIT3-overexpressing MDBK cells infected with BVDV for the indicated lengths of time was performed. (D) DDIT3-overexpressing MDBK cells and negative-control MDBK cells were pretreated with the proteasome inhibitor MG132 (0.5 μM) or autophagy inhibitor CQ (10 μM) for 4 h and then infected with BVDV (MOI of 5) for 24 h. The cells were harvested to detect the expression of MAVS by Western blotting. (E) The ubiquitination levels of MAVS in DDIT3-overexpressing or negative-control MDBK cells were analyzed as indicated. Data are representative of three independent experiments.

OTUD1 promotes BVDV replication by inhibiting interferon antiviral responses.

To elucidate how DDIT3 inhibits MAVS expression, we analyzed transcriptomic high-throughput sequencing data and found that the protein expression of the deubiquitinating enzyme OTUD1 was upregulated in BVDV-infected DDIT3-overexpressing cells (Fig. 5A). The increased expression of OTUD1 in DDIT3-overexpressing cells was confirmed by RT-qPCR at 12 and 24 hpi (Fig. 5B). The Western blotting also confirmed the positive regulatory effect of DDIT3 on OTUD1 expression (Fig. 5C). It has been reported that OTUD1 could promote the degradation of the MAVS/TRAF3/TRAF6 signalosome during VSV or SeV infection (17). We speculate that DDIT3 promotes the downregulation of the MAVS ubiquitin-proteasome pathway by upregulating the expression of OTUD1 in MDBK cells. Therefore, we constructed OTUD1-overexpressing cells to explore whether OTUD1 promotes the degradation of MAVS in MDBK cells. OTUD1-overexpressing stable cell lines were evaluated by Western blotting (Fig. 5D). We further found that the MAVS protein levels but not the mRNA levels in OTUD1-overexpressing cells were lower than those in negative-control cells after BVDV infection (Fig. 5E). OTUD1 inhibited the expression of IFN-β, Mx1, and ISG56 in MDBK cells (Fig. 5F), and the BVDV titer was significantly increased by approximately 10-fold at 24 hpi after OTUD1 overexpression (Fig. 5G). These data indicate that OTUD1 inhibits antiviral responses to promote BVDV replication by reducing MAVS protein levels.

FIG 5.

OTUD1 promotes BVDV replication by inhibiting interferon antiviral responses. (A) Genes differentially expressed in DDIT3-overexpressing MDBK cell lines after BVDV infection (MOI of 5). A heat map of mRNA expression in DDIT3-overexpressing MDBK cell lines infected with BVDV for 24 h is shown; colors indicate the log₂ ratios of virus to mock infection. (B) qPCR analysis of OTUD1 mRNA expression in DDIT3-overexpressing MDBK cells 0 to 24 h after BVDV infection (MOI of 5). The data are shown as fold changes and were normalized to control cell data. (C) Immunoblot analysis of OTUD1 protein levels in DDIT3-overexpressing MDBK cells 24 h after BVDV infection (MOI of 5). (D) Stable overexpression of OTUD1 confirmed by Western blotting with a rabbit anti-OTUD1 PAb and comparing OTUD1-overexpressing cells with a negative-control (NC) MDBK cell line transduced with an empty retrovirus. (E) qPCR and immunoblot analysis of MAVS expression in OTUD1-overexpressing cells and negative-control cells during BVDV infection (MOI of 5). (F) qPCR analysis of IFN-β, MX1, and ISG56 mRNA expression in OTUD1-overexpressing cells and negative-control cells during BVDV infection (MOI of 5). (G) Viral titers of OTUD1 cell lines and control cell lines at 24 and 48 hpi determined by plaque assays. Means and SDs from three independent experiments are shown. *, P ≤ 0.05; **, P ≤ 0.01; ns, not significant.

OTUD1 negatively regulates MAVS protein levels via upregulation of Smurf1 expression.

Previous studies have demonstrated that OTUD1 upregulates the intracellular Smurf1 (E3 ubiquitin ligase) protein levels by removing the ubiquitination of Smurf1, and Smurf1 could enhance proteasome-mediated degradation of MAVS (17, 39). Subsequently, we tested whether OTUD1 induced by DDIT3 upregulated the expression of Smurf1 in MDBK cells. Both DDIT3 and OTUD1 upregulated the protein level of Smurf1 during BVDV infection (Fig. 6A); moreover, the interaction between OTUD1 and Smurf1 amplified from bovine cDNA was confirmed by coimmunoprecipitation (Co-IP) in HEK293T cells (Fig. 6B). Similarly, we observed that BVDV infection obviously promoted the protein expression and accumulation of both endogenous OTUD1 and Smurf1 in MDBK cells and enhanced the colocalization of OTUD1 and Smurf1, as determined by a confocal immunofluorescence assay (Fig. 6C). The OTUD1 amplified from bovine cDNA significantly removed the Smurf1 ubiquitination in HEK293T cells (Fig. 6D). Next, we asked whether BVDV infection is capable of inducing the interaction between Smurf1 and MAVS to promote the degradation of MAVS. Our results clearly showed that Co-IP analysis supported the interaction of Smurf1 and MAVS (Fig. 6E), and colocalization of the Smurf1 protein and MAVS was increased in MDBK cells infected with BVDV for 24 h (Fig. 6F). The data showed that the colocalization of Smurf1 and MAVS mainly occurred in mitochondria (Fig. 6G), and Smurf1 promoted the MAVS ubiquitination in HEK293T cells (Fig. 6H).

FIG 6.

OTUD1 negatively regulates MAVS protein levels via upregulation of Smurf1 expression. (A) Immunoblot analysis of Smurf1 protein expression in DDIT3- or OTUD1-overexpressing cells 24 h after BVDV infection (MOI of 5). (B) HA-Smurf1 immunoprecipitation performed in HEK293T cells expressing an empty vector or Flag-OTUD1. (C) MDBK cells infected with BVDV (MOI of 5) for 24 h. The cells were stained with anti-Smurf1 and anti-OTUD1 antibodies. Cell nuclei were stained with Hoechst. The fluorescent images were captured with a Leica SP8 confocal microscope. (D) Flag-Smurf1 immunoprecipitation was performed in HEK293T cells expressing an empty vector or HA-OTUD1, and ubiquitination levels of Smurf1 were detected by a specific ubiquitin (Ub) antibody. (E) Flag-MAVS immunoprecipitation performed in HEK293T cells expressing an empty vector or HA-Smurf1. (F and G) MDBK cells infected with BVDV (MOI of 5) for 24 h. (F) The cells were stained with anti-Smurf1 and anti-MAVS antibodies. (G) The cells were incubated with MitoTracker Red CMXRos for 15 min and then stained with anti-Smurf1 antibody. Cell nuclei were stained with Hoechst. The fluorescent images were captured with a Leica SP8 confocal microscope. (H) Flag-MAVS immunoprecipitation was performed in HEK293T cells expressing an empty vector or HA-Smurf1, and ubiquitination levels of MAVS were detected by a specific Ub antibody. (I) Immunoblot analysis of the silencing effect of a lentivirus on Smurf1 using anti-Smurf1 antibodies. (J and K) An OTUD1-overexpressing cell line infected with a Smurf1-silencing lentivirus for 36 h and then infected with BVDV (MOI of 5) for 24 h. (J) Immunoblot analysis of the OTUD1, Smurf1, and MAVS proteins in whole-cell lysates of these cells is shown. (K) Relative IFN-β mRNA expression was determined by qPCR. The data are fold changes and were normalized to data from the infected control lentivirus group. (L) Immunoblot analysis of OTUD1 expression in OTUD1 mutant-overexpressing cell lines. (M) Immunoblot analysis of the OTUD1, Smurf1, and MAVS proteins in whole-cell lysates after OTUD1- and OTUD1 mutant-overexpressing cell lines were infected with BVDV (MOI of 5) for 24 h. (N) qPCR analysis of IFN-β mRNA expression in OTUD1- and OTUD1 mutant-overexpressing cell lines after BVDV infection (MOI of 5) for 24 h. The data are shown as fold changes and were normalized to control cell data. **, P ≤ 0.01; ns, not significant.

We further found that the protein level of MAVS was rescued and IFN-β inhibition was also ameliorated when Smurf1 expression was knocked down in OTUD1-overexpressing cell lines (Fig. 6I to K), indicating that Smurf1 is required for the OTUD1-mediated inhibition of the IFN-mediated antiviral response. We continued to verify whether the deubiquitinase activity of OTUD1 in MDBK cells is critical for removing ubiquitin and upregulating Smurf1 protein levels. According to previous research (40), we constructed OTUD1 mutants with mutations in two conservative sites (C320A and H431Q) in bovine-derived OTUD1; both mutations were added in the mutants, resulting in the loss of OTUD1 deubiquitinase activity (Fig. 6L). Compared with OTUD1 wild-type cell lines, OTUD1 mutant cell lines had lost the ability to upregulate Smurf1 protein expression and downregulate MAVS expression (Fig. 6M). In addition, we observed that compared with negative-control cell lines, OTUD1 mutant-overexpressing cell lines also failed to inhibit the production of IFN-β (Fig. 6N). Collectively, these data suggest that the deubiquitinase activity of OTUD1 is required for Smurf1 upregulation and MAVS downregulation to inhibit innate immunity.

DDIT3 promotes the expression of OTUD1 by upregulating NF-κB expression in BVDV infection.

We next investigated how DDIT3 affects the levels of OTUD1 during BVDV infection. We found that BVDV infection upregulated NF-κB p65 phosphorylation in DDIT3-overexpressing cells (Fig. 7A) and that the levels of NF-κB p65 phosphorylation were lower in DDIT3-knockdown MDBK cells than in negative-control cells (Fig. 7B). Subsequently, we used the NF-κB inhibitor pyrrolidine dithiocarbamate (PDTC) and found that the upregulation of P-P65 expression was inhibited in DDIT3-overexpressing cells during BVDV infection (Fig. 7C). We found that the BVDV-induced upregulation of OTUD1 mRNA expression was significantly inhibited in DDIT3-overexpressing cells treated with the NF-κB inhibitor PDTC (Fig. 7D). Furthermore, the declines in IFN-β, MX1, and ISG56 mRNA levels and the increase in BVDV virus titers were inhibited in DDIT3-overexpressing cells treated with PDTC (Fig. 7D and E). Thus, these data demonstrated that DDIT3 promoted the activation of NF-κB signaling and the expression of OTUD1 during BVDV infection.

FIG 7.

DDIT3 promotes the expression of OTUD1 by upregulating NF-κB activity in BVDV infection. (A) Immunoblot analysis of P-P65 and P65 protein expression in DDIT3-overexpressing cell lines after BVDV infection (MOI of 5). (B) Immunoblot analysis of the P-P65 and P65 proteins in DDIT3 knockdown cell lines. (C to E) DDIT3-overexpressing cells were pretreated with the NF-κB inhibitor PDTC (100 μM) for 1 h and then infected with BVDV (MOI of 5) as indicated. (C) Immunoblotting was used to analyze P-P65 and P65 proteins in whole-cell lysates. (D) Relative OTUD1, IFN-β, MX1, and ISG56 mRNA expression was determined by qPCR. The data are fold changes and were normalized to mock-infected-cell data. (E) Viral titers of PDTC- or dimethyl sulfoxide (DMSO)-treated DDIT3-overexpressing or control cells determined by plaque assays at 24 hpi. **, P ≤ 0.01; ns, not significant.

DDIT3 deficiency inhibits BVDV replication, reducing pathological damage in mice.

To investigate the role and functional importance of DDIT3 in the host antiviral response in vivo, we challenged wild-type and DDIT3^−/− mice with BVDV by oral inoculation. At 24 h after BVDV infection, the DDIT3^−/− mice had much lower viral RNA levels (Fig. 8A), and immunofluorescence results also showed that BVDV replication was inhibited in the DDIT3^−/− mice (Fig. 8B). Furthermore, we found that the lung and liver lesions in DDIT3^−/− mice were less severe than those in wild-type mice by hematoxylin and eosin (H&E) staining, which indicated reductions in symptoms of balloon-like degeneration in the liver and alveolar walls (Fig. 8C and D). Collectively, our data suggest that DDIT3^−/− mice exhibit more potent host defense against BVDV than wild-type mice via promotion of the induction of IFN-I.

FIG 8.

DDIT3 deficiency inhibits BVDV replication, reducing pathological damage in mice. (A) qPCR analysis of BVDV mRNA expression in the liver and lungs of wild-type and DDIT3^−/− mice (n = 7 per group) 24 h after intraperitoneal injection of BVDV. The data are shown as fold changes and were normalized to wild-type mouse data. (B) Immunofluorescence assays of the BVDV p125/p80 protein in the liver of wild-type or DDIT3^−/− mice 3 days after BVDV infection. Bar, 50 μm. (C and D) Hematoxylin-and-eosin staining of liver (C) and lung (D) sections from wild-type and DDIT3^−/− mice. Bar, 20 μm. Means and SDs from three independent experiments are shown. **, P ≤ 0.01.

Knocking out DDIT3 promotes the innate immune response against BVDV in mice.

To further study the effect of DDIT3 on the innate immune response in vivo, IFN-β, Mx1, and ISG56 mRNA levels in the liver and lungs (BVDV-susceptible organs) after BVDV infection were analyzed. The mRNA expression levels of IFN-β, Mx1, and ISG56 were increased significantly in the DDIT3^−/− mice, indicating that the antiviral response was enhanced (Fig. 9A). Additionally, the DDIT3^−/− mice produced much higher expression of IFN-β in the serum than did the wild-type mice (Fig. 9B). The IFN-β protein levels in the supernatants of DDIT3^−/− peritoneal macrophages infected with BVDV or stimulated with poly(I·C) were higher than those in the supernatants of wild-type peritoneal macrophages (Fig. 9C), which also showed that the antiviral innate immune response was enhanced. Next, we examined the expression of OTUD1, MAVS, and genes encoding molecules in IFN-β signaling pathways in the liver and lungs, which are BVDV-susceptible organs. We found that the expression of OTUD1 in DDIT3^−/− mice was lower than that in wild-type mice at both the protein and mRNA levels during BVDV infection (Fig. 9D and E). Additionally, after DDIT3 knockout, MAVS expression increased only at the protein level; the mRNA level did not change (Fig. 9D and E). Overall, our data suggest that DDIT3-deficient mice promote the induction of IFN-Is and ISGs through reduced expression of OTUD1 to defend against BVDV infection.

FIG 9.

Knocking out DDIT3 promotes the innate immune response against BVDV in mice. (A) qPCR analysis of IFN-β, MX1, and ISG56 mRNA expression in the organs. (B) Enzyme-linked immunosorbent assay (ELISA) analysis of serum IFN-β levels in wild-type or DDIT3^−/− mice 12 h after oral BVDV or medium (Med) inoculation. (C) ELISA analysis of IFN-β in the supernatants of wild-type and DDIT3^−/− peritoneal macrophages infected for 12 h with BVDV or stimulated with poly(I·C). (D) qPCR analysis of OTUD1 and MAVS mRNA expression in the liver and lungs 24 h after wild-type or DDIT3^−/−mice were inoculated with BVDV. (E) Immunoblot analysis of MAVS and OTUD1 protein expression in the livers of wild-type or DDIT3^−/− mice at 24 h after oral BVDV inoculation. Means and SDs from three independent experiments are shown. *, P ≤ 0.05; **, P ≤ 0.01; ns, not significant.

DISCUSSION

It has been widely reported that DDIT3 affects viral replication through apoptosis or autophagy. Recombinant NDV promotes autophagy by inducing upregulated expression of DDIT3, promoting the expression of beclin-1 and the conversion of LC3-I to LC3-II (41). More recently, it has been reported that the induction of eIF2α-DDIT3-BCL-2/JNK promotes apoptosis and cytokine secretion to support NDV proliferation (42). DDIT3 is involved in the replication process of many viruses, but its relationship with BVDV replication is unclear. In this study, transcriptomic analysis revealed that DDIT3 expression was upregulated after BVDV infection (Fig. 1). After DDIT3 overexpression, the viral titer of BVDV was increased approximately 10-fold at 24 hpi, and after silencing DDIT3, the viral titer was reduced 3-fold at 24 hpi (Fig. 2). This study examined the natural immune response of DDIT3-overexpressing cell lines infected with BVDV. Compared with control cells, DDIT3-overexpressing cells exhibited inhibited expression of IFN-Β and the ISGs MX1 and ISG56 (Fig. 3). This is the first report on DDIT3 inhibiting the innate immune response to promote viral replication.

To screen for the DDIT3 target protein that regulates antiviral innate immunity, we performed a double-luciferase reporter test and found that MAVS was a candidate gene (Fig. 4A and B). MAVS is an important adaptor receptor protein in antiviral responses. Many viruses have evolved strategies to degrade MAVS to inhibit innate immune responses. For example, Golgi protein 73 (GP73) activated by HCV infection promotes MAVS/TRAF6 complex degradation through a proteasome-dependent pathway to facilitate viral replication (43), and NLK interacts with and phosphorylates MAVS, inducing the degradation of MAVS during SeV or VSV infection (44). In this study, we did not find that overexpression of DDIT3 regulated GP73 or NLK protein expression to downregulate MAVS expression (data not shown). Therefore, we first determined whether MAVS degradation occurs after BVDV infection via the ubiquitin-proteasome pathway or autophagy-lysosomal degradation pathway. By treating cells with MG132 or CQ, it was found that MAVS recurred only in DDIT3 cells after MG132 treatment, and MAVS had a relatively high level of ubiquitination in DDIT3-overexpressing cell lines, indicating that DDIT3 promotes MAVS ubiquitin-proteasome pathway-dependent degradation (Fig. 4C to E). Currently, there are no reports on BVDV targeting MAVS to regulate the host antiviral response.

In this study, transcriptomic sequencing revealed that DDIT3 upregulated the expression of OTUD1 (Fig. 5A and B). Overexpression of OTUD1 inhibited the natural immune response and promoted BVDV replication (Fig. 5F and G). The upregulation of OTUD1 expression induced by VSV is NF-κB dependent (17). In this study, DDIT3 induced NF-κB pathway activation, thereby promoting the upregulation of OTUD1 expression (Fig. 7). It has been reported that E3 ubiquitin ligases such as AIP4, Smurf1, and MARCH5 mediate the MAVS ubiquitination and degradation induced by RNA virus infection (39, 45, 46). Among these ligases, OTUD1 negatively regulates the innate immune response by enhancing the E3 ligase Smurf1-mediated degradation of MAVS (17). OTUD1 also upregulated the expression of Smurf1 in MDBK cells to inhibit natural immunity by promoting the degradation of MAVS (Fig. 6).

We also performed experiments with DDIT3 knockout mice infected with BVDV. Compared with that of wild-type mice, the BVDV load of DDIT3^−/− mice was reduced, resulting in weaker pathological damage (Fig. 8). Further testing found that the decrease in the expression of OTUD1 in DDIT3 knockout mice resulted in higher levels of antiviral responses than those in wild-type mice, which particularly manifested as increased MAVS protein levels and IFN-I and ISGs mRNA levels in susceptible organs. In addition, the content of IFN-β both in the serum and in the supernatant of peritoneal macrophages infected with BVDV or stimulated with poly(I·C) was much higher in knockout mice than in wild-type mice (Fig. 9). These data suggest that knocking out DDIT3 expression in mice induces an enhanced antiviral response against BVDV infection.

In summary, we report the mechanism whereby DDIT3 gene expression upregulation by BVDV infection promotes MAVS protein degradation to antagonize antiviral innate immunity and facilitate viral replication. Our data showed that DDIT3 induced the expression of OTUD1 by enhancing the NF-κB signaling pathway, thus increasing Smurf1 protein levels to degrade MAVS, thereby inhibiting the expression of IFN-Is and ISGs and ultimately blocking the transmission of antiviral response signals during BVDV infection (Fig. 10). The findings of this study reveal for the first time that DDIT3 plays critical roles in the repression of host innate immunity and facilitation of viral infection.

FIG 10.

Model depicting viral infection promotion by DDIT3-OTUD1-MAVS pathway activation during BVDV infection. The upregulated expression of DDIT3 caused by BVDV infection induced increased levels of OTUD1 expression by activating NF-κB, thereby promoting the expression of the E3 ubiquitin ligase Smurf1. Smurf1 promoted the ubiquitination of MAVS, thereby blocking signal transmission by antiviral signaling pathways and inhibiting the expression of IFN-I and ISGs.

MATERIALS AND METHODS

Cell lines and virus.

HEK293T (GDC0067) were provided by the China Center for Type Culture Collection (CCTCC), and MDBK cells were provided by the American Type Culture Collection (ATCC CCL-22). MDBK cells and HEK293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco) supplemented with 10% horse serum (Gibco) and antibiotics (100 μg/ml streptomycin and 100 U/ml penicillin) in a humidified incubator at 37°C and 5% CO₂.

The BVDV reference strain NADL, obtained from the China Veterinary Culture Collection Center (CVCC), was propagated and titrated in MDBK cells, and the 50% tissue culture infectious dose (TCID₅₀) of BVDV determined by the Reed-Muench method was 10⁷ TCID₅₀/ml viral titers.

RNA sequencing and analysis.

MDBK cells and DDIT3-overexpressing cells infected with or without BVDV at a multiplicity of infection (MOI) of 5 for 24 h were harvested, and total RNA was extracted from each group of cells using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. cDNA library construction and the quality analysis of the cDNA library were performed with an Agilent 2100. cDNA libraries were sequenced using an Illumina HiSeq2000 platform, and adaptor sequences and low-quality sequences were removed to obtain clean reads. The remaining high-quality reads were mapped to a reference genome with TopHat2. The expression level of each gene was normalized and calculated as the value of fragments per transcript kilobase per million fragments mapped (FPKM), which eliminated the influences of different gene lengths and sequencing discrepancies. DEGs were selected with a threshold false discovery rate (FDR) of <0.05 and an absolute value of log₂ fold change of >1. The DEGs were annotated with gene ontology (GO) functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis.

RT and real-time PCR.

mRNA levels were determined by RT-qPCR as described by Hou et al. (49). Total RNA was extracted from cultured cells with TRIzol (Invitrogen) according to the manufacturer’s instructions. Reverse transcription was performed using the PrimeScript RT reagent kit (TaKaRa, Japan). SYBR green-based real-time qRT-PCR was carried out using a Premix Ex Taq kit according to the manufacturer’s protocols (TaKaRa, Japan). The Roche LightCycler 480 real-time PCR system (Roche Applied Science, Germany) was used to detect mRNA expression. All of the gene transcripts were quantified by real-time PCR with SYBR green qPCR master mix and a LightCycler 480 real-time quantitative PCR system (Roche Applied Science). Relative fold induction was calculated with the 2^−ΔΔCT method. For the primers used for real-time PCR, refer to previous studies (47, 48).

Luciferase reporter assay.

HEK293T (2 × 10⁵) cells were seeded in 24-well plates and transfected with plasmids encoding an interferon-sensitive response element (ISRE) luciferase reporter (firefly luciferase; 100 ng) and pRL-TK (Renilla luciferase plasmid; 10 ng), together with various amounts of the appropriate control or protein-expressing plasmid(s). An empty vector (pcDNA3.1) was used to maintain equal amounts of DNA among wells. The cells were collected at 24 h after transfection, and then firefly and Renilla luciferase activities were measured with a dual-luciferase reporter assay system kit (E1910; Promega Corporation, Madison, WI, USA) using a SpectraMax M5 microplate reader (Molecular Devices Instruments Inc., USA) following the manufacturer’s instructions. Firefly luciferase activities were normalized against Renilla luciferase activities as the relative fluorescence intensity.

Immunofluorescence assays.

Transfected cells were cultured on coverslips, washed twice with ice-cold phosphate-buffered saline (PBS), and fixed in 4% paraformaldehyde. After fixation, the cells were washed three times in PBS and permeabilized in PBS containing 0.2% Triton X-100, and the cell nucleus was counterstained with Hoechst stain, followed by visualization with a Leica SP8 confocal microscope.

Western blot analysis.

Extracted proteins were subjected to 10% SDS-gel electrophoresis and transferred onto polyvinylidene difluoride (PVDF) membranes. The blots were blocked using Tris-buffered saline–Tween (TBST) with 5% nonfat dry milk. The following primary antibodies were used: anti-DDIT3 (2895; Cell Signaling), anti-OTUD1 (ab182511; Abcam), anti-MAVS (24930; Cell Signaling), anti-Smurf1 (sc-100616; Santa Cruz), anti-Ub (sc-8017; Santa Cruz), anti-HA (3724; Cell Signaling), anti-Flag (14793; Cell Signaling), anti-phospho-NF-κB p65 (3033; Cell Signaling), anti-GAPDH (AB0036, Abways Technology), and anti-β-actin (CP06; Calbiochem). The membranes were washed as before and visualized using enhanced chemiluminescence reagents according to the manufacturer’s protocol. Differences in band intensity were quantified by densitometry using ImageJ.

Generation of stable DDIT3- and OTUD1-expressing cell lines.

To stably express DDIT3 or OTUD1 in MDBK cells, we used a four-plasmid-based lentiviral packaging system including the eukaryotic expression vectors pLVX-IRES-puro, pLP1, pLP2, and pLP/VSVG (Clontech Laboratories, Mountain View, CA). Primers were designed to clone DDIT3 (GenBank no. NM_001078163.1) or OTUD1 (GenBank no. XM_010811195.3) into the vector pLVX-IRES-puro with a Flag tag at the C terminus. Lentiviruses and stable cell lines were generated according to protocols described in previous studies (48). The expression of DDIT3 or OTUD1 was detected with anti-DDIT3 or anti-Flag rabbit antibodies by Western blotting.

Short hairpin RNAs for stable knockdown of DDIT3 or Smurf1.

To stably knock down endogenous DDIT3 expression in MDBK cells using a lentiviral system, a lentiviral vector for delivering short hairpin RNA (shRNA) targeting the DDIT3 gene was constructed based on an established method with modifications. Two small interfering RNA sequences targeting the DDIT3 or Smurf1 gene were designed, synthesized, and cloned into the pYr-Lvsh lentiviral vector. Subsequently, lentiviral packaging, infection, and cell line construction were performed as previously described (49). The sequences of shRNAs targeting DDIT3 and Smurf1 are shown in Table 1.

TABLE 1.

shRNA target sequences

Name	Sequence (5′-3′)
shDDIT3-1	GACTCAAACAGGAAATCGAGC
shDDIT3-2	GATTGACCGGATGGTTAATCT
shSmurf1-1	GCGTTTGGATCTGTGCAAACT
shSmurf1-2	GGTGGACCCTGAACTCCATAA

BVDV infection in vitro.

MDBK cells were prepared for BVDV infection to detect gene production and signaling molecule expression. MDBK cells in serum-free medium were infected with BVDV at an MOI of 5 for 1 h, the supernatant was removed, and the cells were returned to medium containing 2% fetal bovine serum (FBS). Total cells and culture medium were harvested for virus titration at 12, 24, 36, and 48 hpi to determine BVDV replication levels by the Reed-Muench endpoint method.

Ethics statement.

All animal experimental procedures were performed strictly in accordance with the Regulations of the People’s Republic of China on the Administration of Experimental Animal Affairs approved by the State Council (1 November 1988). The mice were treated under the animal use guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shandong Normal University.

BVDV infection in vivo.

Mice carrying the DDIT3-null allele were purchased from The Jackson Laboratory (stock code 005530), and C57BL/6J mice were used as the wild-type mouse. Eight-week-old mice and BVDV were prepared for in vivo BVDV infection by oral inoculation as previously described (50). Each mouse was infected with 100 μl of virus fluid (10⁷ TCID₅₀/mice) or the same volume of medium via the oral route. The expression of IFN-β, MX1, ISG56, OTUD1, MAVS, and BVDV was detected in the organs via qPCR. BVDV-infected mice were harvested 3 days after infection for H&E staining and immunohistochemistry assays with an anti-BVDV antibody (sc-101592). IFN-β ELISA kits (E-EL-M0033C; Elabscience) were used to test the concentrations of the IFN-β protein in the sera of BVDV-infected mice collected 12 hpi.

Wild-type and DDIT3^−/− peritoneal macrophage isolation.

Mice were euthanized, and resident peritoneal macrophages were isolated from the mice by peritoneal lavage using 10 ml of ice-cold DMEM. The lavage fluid was centrifuged at 500 × g for 5 min, and 2 × 10⁶ cells were cultured in 6-well cell culture plates in DMEM containing 10% fetal bovine serum (Gibco) and antibiotics (100 μg/ml streptomycin and 100 U/ml penicillin). Nonadherent cells were removed by extensive washing with DMEM as previously described (51, 52).

Statistical analyses.

GraphPad Prism software (version 8.0; GraphPad, San Diego, CA, USA) was used to perform statistical analyses. Two-way analysis of variance (ANOVA) was used to calculate statistical significance, and multiple samples were compared by one-way ANOVA. The results are presented as means and standard deviations (SD) from at least three independent experiments, and a P value below 0.05 was considered statistically significant.