The helicase DDX41 recognizes the bacterial secondary messengers cyclic di-GMP and cyclic di-AMP to activate a type I interferon immune response

Abstract

The induction of type I interferons by the bacterial secondary messengers cyclic di-GMP (c-di-GMP) or cyclic di-AMP (c-di-AMP) is dependent on a signaling axis that involves the adaptor STING, the kinase TBK1 and the transcription factor IRF3. Here we identified the heliase DDX41 as a pattern-recognition receptor (PRR) that sensed both c-di-GMP and c-di-AMP. DDX41 specifically and directly interacted with c-di-GMP. Knockdown of DDX41 via short hairpin RNA in mouse or human cells inhibited the induction of genes encoding molecules involved in the innate immune response and resulted in defective activation of STING, TBK1 and IRF3 in response to c-di-GMP or c-di-AMP. Our results suggest a mechanism whereby c-di-GMP and c-di-AMP are detected by DDX41, which forms a complex with STING to signal to TBK1-IRF3 and activate the interferon response.

Figure 1

c-di-GMP and c-di-AMP mediated induction of the innate immune response in murine DCs and human monocytes requires DDX41. (a) Immunoblot analysis of DDX41 and STING in D2SC mDCs transfected with non-targeting scrambled shRNA, or shRNAs targeting DDX41 or STING. (b-d) Expression of IFN-β mRNA measured via qPCR in control-shRNA D2SC mDCs left unstimulated (US) or in control-shRNA, DDX41-shRNA or STING-shRNA mDCs stimulated with L. monocytogenes (b), c-di-GMP (c) and c-di-AMP (d) for 6 h. (e-h) Expression of Mx1 mRNA (e,f) or IL-6 mRNA (g,h) measured by qPCR in D2SC cells treated as in c-d and stimulated with c-di-GMP (e,g) and c-di-AMP (f,h). (i) Immunoblot analysis of DDX41and STING in THP-1 monocytes treated with non-targeting scrambled shRNA control, or shRNAs targeting either DDX41 (two shRNAs: DDX41-a and DDX41-b) or STING. (j,k) ELISA of IFN-β cytokine production in control-shRNA, DDX41-shRNA or STING-shRNA THP-1 monocytes 16 h after stimulation with c-di-GMP (j) or c-di-AMP (k). Error bars indicate standard error. Data are representative of at least three independent experiments.

Figure 2

Cyclic dinucleotides activate IFN via DDX41 in primary cells. (a) Immunoblot analysis of DDX41 in primary mouse BMDCs treated with non-targeting scrambled shRNA, or shRNAs targeting DDX41 (three shRNAs: DDX41-a, DDX41-b and DDX41-c). (b-d) ELISA of IFN-β cytokine production in control-shRNA or DDX41-shRNA BMDCs treated with L. monocytogenes (b), c-di-GMP (c) or c-di-AMP (d) for 16 h. (e-g) Quantification of DDX41 mRNA (e) or IFN-β mRNA induction (f,g) by qPCR in primary mouse peritoneal macrophages transfected with either control siRNA or siRNA targeting DDX41 (e), then stimulated with L. monocytogenes (f) or c-di-GMP (g) for 6 h. (h-j) Quantification of DDX41 mRNA (h) and IFN-β mRNA induction (i,j) via qPCR performed as in e-g using primary human peripheral blood monocytes electroporated with either control siRNA or siRNA targeting DDX41 (h), then stimulated with L. monocytogenes (i) or c-di-GMP (j) for 6 h.. Error bars indicate standard error. Data are representative of at least two independent experiments.

Figure 3

DDX41 is a direct sensor of c-di-GMP. (a) Pulldown and immunoblot analysis of biotinylated c-di-GMP interactions with Myc-DDX41 from 293T cell lyasates. (b) Confocal imaging of c-di-GMP-DDX41 colocalization in D2SC cells co-transfected with Myc-DDX41 and biotinylated c-di-GMP. (c) Pulldown and immunoblot analysis between biotinylated c-di-GMP and GST-DDX41. (d) Pulldown and immunoblot analysis of biotinylated c-di-GMP (2.0 μM) interactions with GST-DDX41 alone or with increasing amounts of unlabeled ligands as indicated, (3× = 6.0 μM, 10× = 20 μM). Immunoanalysis was performed as in c. I, input GST-DDX41; N, no competitor. Data are representative of at least two independent experiments.

Figure 4

DDX41 DEAD box domain is required for c-di-GMP and c-di-AMP mediated induction of IFN-β. (a) Schematic of deletion constructs of DDX41. (b) Input immunoblot of HA-tagged DDX41 deletion mutants used in co-immunoprecipitation experiments (for c and d). (c,d) Co-immunoprecipitation and immunoblot analysis of biotinylated c-di-GMP (c) or biotinylated c-di-AMP (d) incubated with lysates from 293T cells transfected with HA-DDX41-full length or deletion constructs labeled A-E as shown. (e) Immunoblot of THP-1 monocytes treated with control shRNA or shRNA targeting the 3’UTR of DDX41, then transfected with HA-DDX41 full length (X41A) or HA-DDX41 lacking the DEAD box domain (X41C). (f,g) ELISA for IFN-β cytokine production from THP-1 cells treated with control shRNA or shRNA targeting DDX41 that were reconstituted as in e following treatment with c-di-GMP (f) or c-di-AMP (g) for 16h. Error bars indicate standard error. Data are representative of at least two independent experiments.

Figure 5

c-di-GMP and c-di-AMP require DDX41 for STING dependent signaling. (a) Immunoprecipitation and immunoblot analysis of DDX41-STING interactions is D2SC cells transfected with c-di-GMP or c-di-AMP for 4 h. (b) Immunoprecipitation and immunoblot analysis of STING-TBK1 interactions in control-shRNA or DDX41-shRNA D2SC cells transfected as in a. (c) Immunoblot analysis of TBK1, IRF3, p65 and STAT1 phosphorylations in control-shRNA, DDX41-shRNA or STING-shRNA D2SC mDCs transfected with c-di-GMP, c-di-AMP, Poly (I:C) or B-DNA for 4 h. Data are representative of at least two independent experiments.

Figure 6

c-di-GMP binds DDX41with a greater affinity than STING (a) Pulldown and immunoblot analysis of biotinylated c-di-GMP interactions with Myc-DDX41 or Myc-STING from 293T cell lyasates. (b) Pulldown and immunoblot analysis of biotinylated c-di-GMP or biotinylated c-di-AMP interactions with endogenous DDX41 and STING from D2SC cell lysates. (c) Confocal microscopy indicating c-di-GMP interactions with DDX41 or STING in 293T cells co-transfected with Myc-DDX41 and biotinylated c-di-GMP (top) or Myc-STING and biotinylated c-di-GMP (bottom). (d) Hill plots of Affinity Capillary Electrophoresis analysis showing binding affinities between recombinant DDX41 and c-di-GMP (left panel) or recombinant STING CTD (139-379) and c-di-GMP (right panel). DDX41-c-di-GMP K_d =5.65 μM, R² = 0.99992. STING CTD-c-di-GMP K_d = 14.54 μM, R² = 0.98342. K_d, dissociation constant. (e) Pulldown and immunoblot analysis of biotinylated c-di-GMP interactions with bacterially purified DDX41 and STING CTD. Data are representative of at least two independent experiments.

Figure 7

DDX41 is required for c-di-GMP downstream association with STING. (a) Pulldown and immunoblot analysis of biotinylated c-di-GMP interactions with HA-STING from lysates of 293T cells transfected with control siRNA or siRNA targeting DDX41. (b) Confocal analysis showing c-di-GMP colocalizations with DDX41 or STING in control-shRNA or DDX41-shRNA (upper right panel) and STING-shRNA (lower right panel) D2SC mDCs co-transfected with biotinylated c-di-GMP and Myc-STING (upper panels) or biotinylated c-di-GMP and Myc-DDX41 (lower panels). (c) Quantification of c-di-GMP colocalizations with DDX41 or STING in control-shRNA (upper panel), c-di-GMP colocalization with DDX41 in STING-shRNA (middle panel) or c-di-GMP colocalization with STING in DDX41-shRNA (lower panel) D2SC cells from b. Error bars indicate standard error. Data are representative of at least two independent experiments.