Functional analysis of the RNF114 psoriasis susceptibility gene implicates innate immune responses to double-stranded RNA in disease pathogenesis

Abstract

Psoriasis is an immune-mediated skin disease, the aetiology of which remains poorly understood. In recent years, genome-wide association studies (GWAS) have helped to illuminate the molecular basis of this condition, by demonstrating the pathogenic involvement of multiple genes from the IL-23 and NF-κB pathways. A GWAS carried out by our group also identified RNF114, a gene encoding a novel ubiquitin binding protein, as a determinant for psoriasis susceptibility. Although the function of RNF114 is unknown, its paralogue RNF125 has been shown to regulate the RIG-I/MDA5 innate antiviral response. This signalling cascade, which is activated by the presence of double-stranded RNA (dsRNA) within the cytoplasm, induces the production of type I interferon (IFN) through the activation of the IRF3 and NF-κB transcription factors. Here, we explore the hypothesis that RNF114 may also modulate RIG-I/MDA5 signalling. We show that RNF114 associates with ubiquitinated proteins and that it is a soluble cytosolic protein that can be induced by interferons and synthetic dsRNA. Moreover, we demonstrate that RNF114 over-expression enhances NF-κb and IRF3 reporter activity and increases type I and type III IFN mRNA levels. These results indicate that RNF114 regulates a positive feedback loop that enhances dsRNA induced production of type I IFN. Thus, our data point to a novel pathogenic pathway, where dysregulation of RIG-I/MDA5 signalling leads to the over-production of type I IFN, a key early mediator of epithelial inflammation.

Figure 1.

RNF114 localizes to the cytoplasm. (A) Nuclear (N) and cytoplasmic (C) fractions from the HaCaT keratinocyte cell line were analysed by western blot using an anti-RNF114 antibody. Antibodies detecting known nuclear (Lamin B1) and cytoplasmic proteins (α-tubulin) were used as controls. (B) Primary keratinocytes obtained from healthy donors were analysed by immunofluorescence, using an anti-RNF114 antibody (green). Nuclei were stained with DAPI (blue). (C) HeLa cells were homogenized in hypotonic buffer and fractionated on discontinuous 70/65/10% sucrose gradients. Fractions were collected from the top (fraction 1) and aliquots were analysed by western blot. Antibodies against known soluble (Hsp90) and membrane-associated proteins (transferrin receptor, TfR) were used as controls.

Figure 2.

RNF114 binds ubiquitinated proteins and has ubiquitin ligase activity in vitro. (A) HEK293T cells were transfected with myc-tagged RNF114 and HA-ubiquitin. Ubiquitinated proteins were isolated from whole-cell lysates by incubation with GST–UBA beads and detected with anti-HA by western blotting (central panel). Pulldowns with a mutant form of GST–UBA (GST–UBA-m) that is unable to bind ubiquitin were performed as negative controls. Western blotting with anti-myc was performed to detect the association of RNF114 with ubiquitinated proteins (right, top panel). Pulldown samples were stained using Coomassie to demonstrate the presence of comparable amounts of GST–UBA and GST–UBAm (right, bottom panel). (B) RNF114-myc and RNF125-myc were immunoprecipitated from transfected HEK293T cells and the sample was split into two equal parts. One half was incubated for 90 min at 30°C in reaction buffer only (−), while the other half was incubated in reaction buffer supplemented with purified E1, the E2 UbcH5a, biotinylated ubiquitin, DTT and ATP (+). Ubiquitination products were revealed by western blots, using a streptavidin–HRP (SA–HRP) conjugate (top), and immunoprecipitated RNF114 and RNF125 were detected with anti-myc (bottom). The difference in migration between RNF114-myc in the (−) and (+) lane results from the presence of DTT in the (+) sample. An anti-myc immunoprecipitate from untransfected HEK293T cells was used as a negative control in this experiment. (C) Bacterially expressed RNF114 and RNF125 GST-fusion proteins were captured on glutathione–sepharose beads and used in in vitro ubiquitination reactions as described in (B). The RNF114 C29A,C32A mutant (RING mut) contains two cysteine-to-alanine substitutions at conserved residues in the RING domain. The reactions were analysed by western blotting using a streptavidin–HRP conjugate to detect modification of the GST fusion proteins with biotinylated ubiquitin (top panel), and with anti-GST to detect the GST-fusion proteins (bottom panel). The RNF114 and RNF125 samples were analysed on the same nitrocellulose blot (for SA–HRP or anti-GST) and the images are of the same X-ray autoradiography exposure of these blots.

Figure 3.

RNF114 expression is up-regulated by IFNα, IFNγ and synthetic dsRNA. (A) HaCaT cells were treated with 100 ng/ml IFNα (5 h) or 50 ng/ml IFNγ (24 h). RNF114 induction was measured by real-time PCR, using PPIA as an internal control; ** P < 0.005. (B) Primary keratinocytes obtained from healthy donors (n = 2) were treated and analysed as in (A); * P < 0.05, ** P < 0.005. (C) Primary keratinocytes were treated with 100 ng/ml IFNα (24 h) or 50 ng/ml IFNγ (48 h) and RNF114 induction was measured by western blot, using a monoclonal anti-RNF114 antibody. (D and E) Primary keratinocytes were treated in duplicate with the indicated concentrations of cytokine for 5 (IFNα) or 24 h (IFNγ). RNF114 induction was measured by real-time PCR, using PPIA as an internal control. (F and G) Primary keratinocytes were treated in duplicate with 100 ng/ml IFNα or 50 ng/ml IFNγ. Cells were harvested at the indicated time points and RNF114 induction was measured by real-time PCR, using PPIA as an internal control. (H) HEK293T cells were transfected with 2 μg/ml poly(I:C) for the indicated times. IFNB1 (left) and RNF114 (right) induction was measured by real-time PCR, using PPIA as internal control.

Figure 4.

RNF114 positively regulates poly(I:C) induced IFNβ production. (A and B) HEK293T cells were transfected with pcDNA3.1-myc/His plasmids containing either RNF114 cDNA (RNF114) or no insert (EV: empty vector) together with pJ7lacZ and a luciferase reporter plasmid for IRF3 (pISG54-luc in A) or NFκB (pPRDII-luc in B). After 30 h, cells were transfected with 1 μg/ml poly(I:C) or mock transfected with PBS.  Beta-galactosidase and luciferase activities were measured after a further 18 h, using a dual-light reporter assay. The plots refer to normalized luciferase ratios, measured in one of three independent experiments, each carried out in triplicate; *P < 0.01, **P < 0.001. (C) HEK293T cells were transfected with pJ7lacZ and the indicated plasmids, together with the IFNβ reporter IFNb-luc. After 30 h, cells were transfected with either PBS or 1 μg/ml poly(I:C).  Beta-galactosidase and luciferase activities were measured after a further 18 h, using a dual-light reporter assay. The plots refer to normalized luciferase ratios, measured in one of three independent experiments, each carried out in triplicate (left). RING mut: mutated RNF114 construct, bearing two Ala substitutions at Cys residues 29 and 32 in the RING domain; UIM mut: RNF114 construct, bearing a Pro substitution at the conserved Ser 224 in the UIM domain; *P < 0.05, **P < 0.001. The expression of all RNF114 constructs was verified by western blot (right).

Figure 5.

RNF114 over-expression results in the induction of IFNβ, IL-29, CCL5 and CXCL10. (A–D) HEK293T cells were transfected with pcDNA3.1-myc/His plasmids containing either RNF114 cDNA (RNF114) or no insert (EV: empty vector) or were left untransfected (UT). Forty-eight hours post-transfection, cells were harvested and the expression of IFNB1 (A), IL-29 (B), CCL5 (C) and CXCL10 (D) was measured by real-time PCR, using PPIA as internal control. * P < 0.005 relative to the EV control. (E) THP1 cells were cultured for 24 h in cell-culture supernatants of HEK293T cells that had been transfected for 48 h with either an empty pcDNA3.1-myc/His vector (EV) or one containing RNF114 cDNA (RNF114). THP1 cells were harvested and IFNB1 mRNA induction was measured by real-time PCR.