The Microtubule-Associated Innate Immune Sensor GEF-H1 Does Not Influence Mouse Norovirus Replication in Murine Macrophages

Abstract

Norovirus is an acute infection of the gastrointestinal tract causing rapid induction of vomiting and diarrhoea. The infection is sensed and controlled by the innate immune system, particularly by the RNA helicase MDA-5 and type I and III interferons (IFNs). We have observed that intracellular replication of murine norovirus (MNV) occurs in membranous clusters proximal to the microtubule organising centre, a localisation dependent on intact microtubules. Recently, it was shown that the host protein guanine nucleotide exchange factor-H1 (GEF-H1) is a microtubule-associated innate immune sensor that activates interferon Regulatory Factor 3 to induce the production of type I IFNs. Thus, we interrogated the potential role of GEF-H1 in controlling MNV infections. We observed that GEF-H1 was recruited to the MNV replication complex; however RNAi-mediated suppression of GEF-H1 did not outwardly affect replication. We furthered our studies to investigate the impact of GEF-H1 on MNV innate detection and observed that GEF-H1 did not contribute to type I IFN induction during MNV infection or influenza virus infection but did result in a small reduction of interferon⁻β (IFNβ) during West Nile virus infection. Intriguingly, we discovered an interaction of GEF-H1 with the viral MNV non-structural protein 3 (NS3), an interaction that altered the location of GEF-H1 within the cell and prevented the formation of GEF-H1-induced microtubule fibres. Thus, our results indicate that GEF-H1 does not contribute significantly to the innate immune sensing of MNV, although its function may be modulated via interaction with the viral NS3 protein.

Keywords: GEF-H1; innate immunity; microtubules; mouse norovirus.

Figure 1

The replication complex (RC) of murine norovirus (MNV) co-localises with the endogenous guanine nucleotide exchange factor-H1 (GEF-H1) in the perinuclear area. (A) RAW264.7 cells were infected with MNV (multiplicity of infection (MOI) 5) and fixed at 12 h and 18 h after infection. Panels a, e, and i show the staining for the anti-GEF-H1 antibody (green); panels b, f, and j display the staining with an anti-non-structural protein 4 (NS4) antibody (red); panel c, g, and k represent the nuclear staining with DAPI (blue); and panels d, h, and l show the merged pictured of all channels. Stained cells were analysed via confocal microscopy and the co-localisation was quantified with the Pearson’s coefficient. (B) Quantitation of the Pearson’s coefficients. MNV 12 h = 0.57 ± 0.08 (n = 16) and MNV 18 h = 0.50 ± 0.16 (n = 14). Bars represent average ± standard error of the mean (SEM). Images were collected over triplicate experiments and analysed in GraphPad Prism (www.graphpad.com/scientific-software/prism/).

Figure 2

The expression of GEF-H1 WT induces the dispersion of the MNV RC. Murine macrophages (RAW264.7) were transfected with GFP-tagged GEF-H1 forms (WT, C53R: unable to associate with microtubules, ΔDH: no GEF activity, S885A: constantly active GEF) and at 12 h post-transfection (h.p.t.) were subsequently infected with MNV (MOI 5). Cells were fixed at 12 h.p.i. and immune-labelled with an anti-NS4 antibody for confocal microscopy analysis. The GFP signal of the different GEF-H1 forms is displayed in panels a, e, i, and m (green), while panels b, f, j, and n show the staining with the anti-NS4 antibody (red). Panels c, g, k, and o show the merged image including the nuclear stain (DAPI; blue). Panels d, h, l, and p represent the merged image of the GEF-H1-GFP transfected but uninfected cells. Macrophages expressing the microtubule-associated forms of GEF-H1 (WT, ΔDH, S885A) all showed a dispersion of the MNV RC (white arrows).

Figure 3

Co-expression with non-structural protein 3 (NS3) changes the GEF-H1 morphology. (A) Vero cells were co-transfected with GEF-H1-GFP WT and the HIS-tagged viral proteins. Cells expressing GEF-H1-GFP is shown in panels a, e, and i (green). Panels b, f, and j display the staining with an anti-α–tubulin antibody (blue), while panels c, g, and k show the antibody staining for anti-6× HIS (red). Panels d, h, and l represent the merged image of all channels including the nuclear stain (grey). Co-transfected cells displayed the typical bundle-like structures of GEF-H1 in the cell periphery, except for cells co-transfected with NS3. NS3 and GEF-H1 WT appeared to co-localise and a rather reticular distribution of GEF-H1 could be observed. Additionally, GEF-H1 WT appeared to change the distribution of NS3 as well, compared to cells transfected with NS3 only. The co-localisation was quantified with the Pearson’s coefficient and is indicated in the merged image. (B) Quantitation of the Pearson’s coefficients. NS1–2 = 0.34 ± 0.20 (n = 9), NS3 = 0.70 ± 0.11 (n = 17), and NS4 = 0.45 ± 0.13 (n = 9). Bars represent average ± SEM and * = p < 0.05 and ** = p < 0.01. Images were collected over triplicate experiments and analysed in GraphPad Prism. (C) The 293T cells were co-transfected with cDNA expression plasmids encoding for 6× HIS-tagged NS3 and GEF-H1-GFP proteins, and cell lysates were analysed via immune precipitation. Lysates were incubated with a cOmplete^TM His-Tag Purification Resin to pull down the 6× HIS-tagged NS3 and possible interaction partners. Lysates were then transferred to a nitrocellulose membrane for immunoblotting with anti-GFP and anti-6× HIS antibodies and visualised by chemiluminescence. A representative image from three independent experiments is presented.

Figure 4

Expression of NS3 changes the location and morphology of the different GEF-H1 forms. (A) Vero cells were co-transfected with the HIS-tagged viral protein NS3 and the four different GEF-H1-GFP forms (WT, C53R, ∆DH, S885A). The GFP signal of the different GEF-H1 forms is displayed in panels a, e, I, and m, while panels b, f, j, and n show the staining with the anti-6× HIS antibody. Panels c, g, k, and o show the merged image including the nuclear stain (DAPI). Panels d, h, l, and p represent the merged image of GEF-H1-GFP transfected but uninfected cells. Co-transfected cells displayed a dispersed morphology of GEF-H1 compared to cells transfected with the GEF-H1 forms only (panels d, h, l, and p). The co-localisation was quantified with the Pearson’s coefficient and is indicated in the merged image (panels c, g, k, and o). (B) Quantitation of the Pearson’s coefficients: C53R = 0.42 ± 0.19 (n = 12), ∆DH = 0.66 ± 0.18 (n = 15), S885A = 0.60 ± 0.14 (n = 17), and GEF-H1 = 0.65 ± 0.01 (n = 19). Bars represent average ± SEM. Images were collected over triplicate experiments and analysed in GraphPad Prism.

Figure 5

Suppression of the GEF-H1 expression does not influence MNV replication or the protein production of the infectious virus release. RAW264.7 cells were treated twice with GEF-H1 siRNA (siRNA3) or control siRNA (siRNA-ve), before cells were infected with MNV (MOI 5) for 12 h. (A) The immunoblot analysis of whole cell lysates stained with antibodies against GEF-H1, NS7, and actin. (B) Panels a, d, and g show the staining for the anti-GEF-H1 antibody; panels b, e, and h display the staining with an anti-dsRNA antibody; and panels c, f, and i represent the merged pictured of all the channels including the nuclear stain (DAPI). Stained cells were analysed via confocal microscopy. (C) Cells were lysed, and RNA was extracted. The relative fold expression of MNV mRNA compared to mock cells was analysed via RT-qPCR. (D) The supernatant of the infected cells was collected and used to determine viral titres via plaque assay. (n = 3, average ± SEM, not significant [ns]: p > 0.05; one-way ANOVA).

Figure 6

Knockdown of GEF-H1 in murine macrophage cells does not lead to significant changes in the mRNA expression of IFNβ and TNFα upon MNV infection. RAW264.7 cells were treated twice with GEF-H1 siRNA (siRNA3) or the control siRNA (siRNA-ve) before cells were either infected with MNV (MOI 5), treated with poly(I:C), or left untreated for 12 h. Cells were lysed, and the RNA was extracted. Relative fold expression of IFNβ (panel A) and TNFα mRNA (panel B) compared to mock cells was analysed via RT-qPCR (n = 3, average +/− SEM, not significant [ns]: p > 0.05; one-way ANOVA). In panels C and D, RAW264.7 cells were treated twice with GEF-H1 siRNA, the control siRNA, or left untreated before cells were either infected with MNV (MOI 5), X31 (MOI 5), or West Nile Virus (WNV) (MOI 1). Cells were lysed, and the RNA was extracted. Relative fold expression of IFNβ (A,C), TNFα (B,D), and viral mRNA (E) compared to infected but untreated cells was analysed via RT-qPCR (n = 4, average ± SEM, not significant [ns]: p > 0.05, * p < 0.05; one-way ANOVA).

Figure 7

Proposed for the interplay between GEF-H1, MNV replication, microtubules, the MNV NS3 protein, and the innate immune response during the MNV infection of macrophages. Question marks (?) indicate unanswered understanding between the MNV NS3 protein and GEF-H1.