Fig 1.
STAT1- and type I IFN-independent RIG-I expression in response to dsRNA.
2fTGH, U3A, and U5A cells were transfected with poly I:C (100 ng) and incubated for up to 8 h. The expression levels of RIG-I were determined by quantitative RT-PCR. The means (±SD) of three experiments are shown.
Fig 2.
STAT1- and type I IFN-independent RIG-I expression requires IRF3.
HeLa (A), and 2fTGH, U3A, and U5A cells (B) were transfected with siRNA against IRF-1 or IRF-3 or control siRNA for 48 h and then transfected with poly I:C (100 ng) for 4 h. The expression levels of RIG-I were determined by quantitative RT-PCR. The means (±SD) of three experiments are shown; †P < 0.05, *P < 0.01.
Fig 3.
Intracellular localization of IRF-1 and IRF-3.
(A) HeLa cells were subjected to the following treatments: control (a-c, j-l), transfected with poly I:C (100 ng, for 4 h) (d-f, m-o), or treated with IFN-γ (2 ng/mL, for 4 h) (g-i, p-r). The cells were then fixed with 4% paraformaldehyde and co-stained for IRF-1 (red, left panels) or IRF-3 (red, right panels). Nuclei were counterstained with DAPI (blue). A representative result for five random fields is shown. (B) HeLa cells were transfected with poly I:C (100 ng, for 4 h) or treated with IFN-γ (2 ng/mL, for 4 h). The cells were then harvested in lysis buffer and fractionated. The cell extracts were subsequently subjected to SDS-PAGE, blotted, and probed with anti-IRF-1, anti-IRF-3, anti-HSP90, or anti-histone H1 antibodies. The results are representative of three independent experiments.
Fig 4.
Promoter analysis of the human RIG-I gene.
(A) Putative consensus sequences of STAT1, ISRE, c-Rel and IRF-E on the RIG-I promoter are shown. (B) and (C) HeLa cells were co-transfected with pGL4.11 (empty) or serial human RIG-I luciferase reporter constructs and Renilla luciferase expression vector (pGL4.74) for 24 h. The cells were further transfected with poly I:C (100 ng) for 4 h. The reporter activities are shown as relative values, specifically ratios of the firefly luciferase activities driven by the RIG-I promoters to the Renilla luciferase activities. The means (±SD) of three experiments are shown; †P < 0.05, *P < 0.01.
Fig 5.
IRF-E on the RIG-I promoter regulates the transcriptional activity of RIG-I.
(A) Putative consensus sequences of ISRE, GAS and IRF-E within the proximal region of the RIG-I promoter are shown. (B) A series of single (double or triple) deletion constructs on the RIG-I promoter is shown. (C) U3A and U5A cells were co-transfected with the RIG-I deletion constructs and a renilla luciferase expression vector as shown in (B); †P < 0.05, *P < 0.01.
Fig 6.
Critical role of IRF-E on the IRF3-mediated activation of the RIG-I promoter.
(A) The nucleotide sequence of the RIG-I promoter (-22 to +8) is shown, and putative IRF-3-binding sites are underlined. The numbers show the positions relative to the transcription start site (+1). (B) The luciferase activities driven by vectors with serial truncations of the IRF-E in the RIG-I promoters (left panel) were analyzed (center panel). The right panel shows the effect of IRF-3 knockdown on the promoter activities of RIG-I promoters with different lengths; †P< 0.05, *P < 0.01.
Fig 7.
IRF-3 binding to the RIG-I promoter in vivo.
HeLa cells were transfected with poly I:C (100 ng) for 4 h and fixed; the DNA was then fragmented. Chromatin immunoprecipitation was performed using an anti-IRF-3 antibody or control IgG. The means (±SD) of three experiments are shown. †P < 0.05
Fig 8.
IRF-3 binds to IRF-E in the RIG-I promoter.
(A) EMSA was performed using a DIG-labelled RIG-I IRF-E wild-type (WT) probe. Nuclear extracts were prepared from poly I:C-transfected HeLa cells. For the supershift assay, rabbit anti-IRF-3 antibody was pre-incubated with the reaction mixture. (B) The labeled RIG-I IRF-E-WT probe was combined with 500 nM (r)GST or (r)GST-IRF-3(5D) fusion protein at various concentrations (500 nM, 125 nM, and 50 nM; wedges). Arrows indicate the IRF-3-oligonucleotide probe complex.
Fig 9.
Characterization of RIG-I-IRF-E for IRF-3 binding.
(A) Description of oligonucleotide probes of wild-type (WT) (-22 to +8) and mutant (MT) RIG-I IRF-E are shown. Putative IRF-3-binding sites are indicated in lowercase and underlined. (B) The RIG-I IRF-E-WT and RIG-I IRF-E-MT probes were mixed with the nuclear extract as described in Fig 6, and an EMSA was then performed. A competition assay was performed based on the addition of a 50-fold molar excess of unlabeled WT or MT probes. (C) The labeled WT or MT probes were combined with recombinant (r)GST or (r)GST-IRF-3(5D) fusion protein at various concentrations (500 nM). Arrows indicate the IRF-3-oligonucleotide probe complex. N.S.: non-specific signal.
Fig 10.
Proposal model for the direct role of IRF-3 in both constitutive and induced RIG-I expression.
Upon viral infection, a low level of constitutively expressed RIG-I recognizes viral RNA, inducing the cells to reach an antiviral state. In the antiviral state, IRF-3 is phosphorylated in response to RLR signaling and translocates to the nucleus to induce type I IFNs. Our findings indicate that activated IRF-3 is also able to directly enhance the expression of RIG-I to enhance antiviral signaling. Infected cells produce IFNs, which subsequently activate STAT1, leading to the robust expression of RIG-I in neighboring cells.