LGP2 is a positive regulator of RIG-I– and MDA5-mediated antiviral responses

To investigate the physiological role of LGP2 in vivo, we established Lgp2^−/− mice (Fig. S1 A and B). As reported previously, the expression of Lgp2 mRNA was highly induced in response to IFN-β stimulation in MEFs (Fig. S1C) (17). Expression of Lgp2 mRNA was not detected in Lgp2^−/− MEFs, whereas the expression levels of Rig-I and Mda5 mRNAs were comparable between WT and Lgp2^−/− cells (Fig. S1C). The Lgp2^−/− progenies obtained from Lgp2^+/− intercrosses were lower than the expected Mendelian ratio (Fig. S2A), indicating that homozygous mutations of the Lgp2 gene cause embryonic lethality at a high frequency. In addition, adult female Lgp2^−/− mice showed an enlarged uterus filled with fluid resulting from vaginal atresia (Fig. S2 B and C).

First, we examined the production of IFN-β in cDCs derived from bone marrow (BM) in the presence of GM-CSF after infection with a variety of RNA viruses (Fig. 1). The productions of IFN-β in response to picornaviridae, EMCV, and mengovirus were severely impaired in Lgp2^−/− cDCs compared with WT cells (Fig. 1A). IL-6 production induced by EMCV infection was also severely impaired in Lgp2^−/− cells (Fig. 1B). Furthermore, LGP2 was involved in the productions of IFN-β in response to several RNA viruses recognized by RIG-I, such as VSV, SeV, and JEV (Fig. 1A). IFN-β production in response to reovirus, a dsRNA virus, was also impaired in Lgp2^−/− cells (Fig. 1A). In contrast, the IFN-β productions in response to infection with influenza virus were comparable between WT and Lgp2^−/− cDCs (Fig. 1A). Stimulation with CpG-DNA, a TLR9 ligand, induced comparable amounts of IFN-β in WT and Lgp2^−/− cells (Fig. 1A).

Next, we examined whether the defect in type I IFN production in response to EMCV was controlled at the mRNA level. The expressions of the genes encoding IFN-β, CXCL10 and IL-6 after infection with EMCV were severely impaired in Lgp2^−/− macrophages (Fig. 2A). However, the influenza virus-induced expressions of IFN-β and CXCL10 mRNAs were comparable between WT and Lgp2^−/− MEFs throughout the whole time course (Fig. 2B). Therefore, it is unlikely that LGP2 negatively regulates RIG-I–mediated responses, even during the later period of infection. These results indicate that LGP2 is involved in positive, but not negative, regulation of virus recognition by MDA5 and RIG-I, with the exception of influenza virus.

We previously showed that RIG-I– and MDA5-dependent RNA virus recognition occurs in cDCs but not in plasmacytoid dendritic cells (pDCs) (8). To determine whether LGP2 functions in a cell type–specific fashion, B220⁻CD11c⁺ cDCs and CD11c⁺B220⁺ pDCs were purified from WT and Lgp2^−/− splenocytes. EMCV-induced IFN-β production was severely impaired in Lgp2^−/− splenic cDCs compared with WT cDCs, whereas splenic pDCs from WT and Lgp2^−/− mice produced comparable amounts of IFN-β (Fig. S3). These data indicate that LGP2 functions in cDCs but not in pDCs.

To investigate whether LGP2 regulates the primary responses to RNA virus infections, we examined the activation of intracellular signaling pathways. Electrophoretic mobility shift assays (EMSAs) revealed that the activations of NF-κB and IFN-stimulated regulatory elements (ISREs) in response to EMCV infection were severely impaired in Lgp2^−/− cells (Fig. 2 C and D). Furthermore, the phosphorylation of STAT1 was abrogated in Lgp2^−/− cells (Fig. 2E). Nevertheless, the expressions of Lgp2 in response to IFN-β treatment were not altered in Rig-I^−/− and Mda5^−/− cells (Fig. S4). These results suggest that LGP2 is required for the initial recognition of EMCV, leading to activation of transcription factors involved in the expression of type I IFNs.

Next, we examined whether the expression of Lgp2 could rescue the virus-mediated IFN responses. IFN-β–dependent reporter gene expression was induced in response to EMCV infection in WT, but not in Lgp2^−/−, MEFs (Fig. 3A). Expression of exogenous LGP2 in Lgp2^−/− cells restored the EMCV-induced IFN-β promoter activity as well as IFN-β production (Fig. 3 A and B). Although overexpression of either LGP2 or MDA5 alone in Lgp2^−/−Mda5^−/− MEFs failed to confer EMCV-induced IFN-β promoter activity, coexpression of both LGP2 and MDA5 restored EMCV responsiveness (Fig. 3C). Overexpression of the CARDs from RIG-I or MDA5 in Lgp2^−/− MEFs activated the IFN-β promoter (Fig. 3D), suggesting that LGP2 functions upstream of RIG-I and MDA5.

We examined the responses of Lgp2^−/− cells to synthetic RNAs recognized by RIG-I and MDA5. Unexpectedly, WT and Lgp2^−/− cDCs produced comparable amounts of IFN-β in response to poly I:C, in vitro–transcribed dsRNA and RNA with a 5′ triphosphate end (Tri-P) (Fig. 4A). Similarly, Lgp2-deficiency did not affect the IFN-β productions in response to the synthesized RNAs in fibroblasts (Fig. 4A). In addition, no differences were observed in the responses to the various concentrations of poly I:C examined (Fig. 4B). These data suggest that LGP2 is dispensable for the recognition of synthesized dsRNA and 5′ triphosphate RNA.

The recognition of dsRNA and RNA viruses by RIG-I or MDA5 requires ATPase activity (17, 25). To examine the role of the LGP2 ATPase activity in LGP2-mediated virus recognition, we reconstituted Lgp2^−/− MEFs with WT LGP2 or LGP2-K30A harboring a Lys-to-Ala point mutation in the Walker ATP-binding motif using a retrovirus system. Expression of WT LGP2 in Lgp2^−/− MEFs conferred IFN-β promoter activity as well as IFN-β production in response to EMCV, whereas expression of LGP2-K30A failed to confer these responses to EMCV infection (Fig. 5 A and B).

To further examine the role of the LGP2 ATPase activity in vivo, we generated mice harboring the LGP2 K30A point mutation (Fig. S5 A and B). Expression of Lgp2 mRNA was comparably induced in response to IFN-β stimulation in WT and Lgp2^K30A/K30A MEFs (Fig. S5C). We confirmed the insertion of the point mutation by sequencing analysis (Fig. S5D). The Lgp2^K30A/K30A mice were born at the expected Mendelian ratio, and did not show any developmental defects. We examined the IFN-β productions in cDCs in response to infections with RNA viruses. The IFN-β productions in response to infections with EMCV, mengovirus, VSV, SeV, and reovirus, but not with influenza virus, were severely impaired in Lgp2^K30A/K30A cDCs compared with WT cells (Fig. 5C). The IL-6 production in response to EMCV infection was also impaired in Lgp2^K30A/K30A cDCs (Fig. 5D). The defects observed in Lgp2^K30A/K30A cDCs were as severe as those observed in Lgp2^−/− cDCs, suggesting that the ATPase activity of LGP2 is essential for the recognition of viruses. The productions of IFN-β in response to transfections of synthetic RNAs and poly I:C were comparable between WT and Lgp2^K30A/K30A cells (Fig. 5E), further confirming that LGP2 is not involved in the responses to the transfection of synthetic RNAs. Taken together, these results indicate that the ATPase activity of LGP2 is essential for LGP2 to function as a positive regulator in MEFs. This finding is in marked contrast to in vitro experiments in which overexpression of WT LGP2 and LGP2-K30A in HEK293 cells suppressed RIG-I–mediated IFN-β promoter activity (Fig. S6), suggesting that overexpression of LGP2 in HEK293 cells nonspecifically inhibits the RIG-I–mediated pathway.

Finally, we assessed the role of LGP2 and its ATPase activity in antiviral responses in vivo. When Lgp2^−/− mice were challenged with EMCV, IFN-β production was not detected in their sera (Fig. 6A). Furthermore, Lgp2^−/− mice were highly susceptible to EMCV infection compared with their littermate controls (Fig. 6C). Consistent with the increased susceptibility to EMCV, the viral titer in the heart was remarkably higher in Lgp2^−/− mice than in control mice (Fig. 6E). Similar to the results for Lgp2^−/− mice, Lgp2^K30A/K30A mice showed severe defects in IFN-β production in response to EMCV infection (Fig. 6B). Lgp2^K30A/K30A mice were highly susceptible to infection with EMCV, with highly increased viral titers in their hearts (Fig. 6 D and F). These data indicate that LGP2 also plays a key role in the host defenses against RNA viruses recognized by MDA5 in vivo.

The Lgp2 gene was isolated from genomic DNA extracted from embryonic stem (ES) cells (GSI-I) by PCR. The targeting vector was constructed by replacing a 4-kb fragment encoding the Lgp2 ORF (including the DExH/H box) with a neomycin-resistance gene cassette (neo), and inserting herpes simplex virus thymidine kinase (HSV-TK) driven by the PGK promoter into the genomic fragment for negative selection. After the targeting vector was transfected into ES cells, G418 and gancyclovir double-resistant colonies were selected and screened by PCR, and recombination was confirmed by Southern blotting. Homologous recombinants were microinjected into C57BL/6 female mice, and heterozygous F1 progenies were intercrossed to obtain Lgp2^−/− mice. Lgp2^−/− and littermate control mice were used for subsequent experiments.

A point mutation was inserted into a genomic fragment harboring the exon encoding Lys-30 of murine Lgp2 by site-directed mutagenesis (Clontech) to replace this residue with Ala. A targeting vector was constructed with this genomic fragment and electroporated into ES cells. Homologous recombinants were selected and microinjected into C57BL/6 female mice, and heterozygous F1 progenies were crossed with CAG-Cre transgenic mice to excise the neo cassette. Next, the CAG-Cre transgene was removed from Lgp2^+/K30A mice by crossing the mice with C57BL/6 mice. Lgp2^K30A/K30A and littermate control mice were used for subsequent experiments.

All animal experiments were carried out with the approval of the Animal Research Committee of the Research Institute for Microbial Diseases (Osaka University).

Rig-I^−/− and Lgp2^−/−Mda5^−/− MEFs were prepared from embryos on 129Sv and C57BL/6 backgrounds derived at 11.5 days post coitum. BM-derived DCs were generated in RPMI medium 1640 containing 10% FCS, 50 mM 2-mercaptoethanol, and 10 ng/mL GM-CSF (PeproTech). pDCs and cDCs were isolated from the spleen by MACS using anti-B220 and anti-CD11c microbeads (Miltenyi Biotech). Poly I:C was purchased from Amersham Biosciences. RNAs were complexed with the cationic lipid Lipofectamine 2000 (Invitrogen) and added to cells. A/D-type CpG-oligodeoxynucleotides (D35) were synthesized by Hokkaido System Science. In vitro–transcribed dsRNA and triphosphate RNA were described previously (5, 9). Antibodies against phospho-STAT1 and STAT1 (Cell Signaling) were used for Western blotting as described previously (8).

VSV, VSV lacking an M protein variant (NCP), influenza virus ΔNS1, JEV, EMCV, and mengovirus were described previously (9). SeV lacking V protein (V–) was kindly provided by Dr. A. Kato. Reovirus was kindly provided by Dr. T. Dermody.

Total RNA was extracted from peritoneal macrophages infected with EMCV or MEFs infected with influenza virus using TRIzol reagent (Invitrogen). The obtained RNA was electrophoresed, transferred to nylon membranes, and hybridized with various cDNA probes. To detect the expression of Lgp2 mRNA, a 326-bp fragment (772–1098) was used as a probe.

Peritoneal macrophages (3 × 10⁶) were infected with EMCV for various periods. Nuclear extracts were purified from the cells using lysis buffer (10 mM Hepes-KOH pH 7.8, 10 mM KCl, and 10 mM EDTA, pH 8.0), incubated with specific probes for NF-κB or ISRE DNA-binding sites, electrophoresed, and visualized by autoradiography.

MEFs were transiently transfected with a reporter construct containing the IFN-β promoter together with an empty vector (Mock) or expression constructs for several genes using Lipofectamine 2000. As an internal control, the cells were transfected with a Renilla luciferase construct. The transfected cells were left untreated (medium alone) or infected with EMCV for 8 h. The cells were then lysed and subjected to a luciferase assay using a dual-luciferase reporter assay system (Promega) according to the manufacturer's instructions.

Murine LGP2 and LGP2K30A cDNAs were individually cloned into the pLZR-IRES/GFP retroviral vector (30). Retroviruses were produced by transient transfection of the constructs into PlatE cells. MEFs were separately infected with the retroviruses expressing LGP2 and LGP2 (K30A). At 2 days after infection, the cells were exposed to EMCV for 24 h, and the IFN-β concentrations in the culture supernatants were measured by ELISA.

At 48 h after EMCV infection, hearts were prepared and homogenized in PBS. The virus titers in the hearts were determined by a standard plaque assay as described previously (9). After centrifugation, the supernatants were serially diluted and added to plates containing HeLa cells. Cells were overlaid with DMEM containing 1% low-melting point agarose and incubated for 48 h. The numbers of plaques were counted.

Culture supernatants were collected, and the cytokine concentrations were measured using ELISA kits for IFN-β (PBL Biomedical Laboratories) and IL-6 (R&D Systems) according to the manufacturers’ instructions.

The statistical significance of differences between groups was determined by Student's t test, and survival curves were analyzed by the log-rank test. Values of P < 0.05 were considered to indicate statistical significance.