SAMD9 senses cytosolic double-stranded nucleic acids in epithelial and mesenchymal cells to induce antiviral immunity

Abstract

Sensing of cytosolic, double-stranded (ds) DNA or dsRNA molecules derived from microbial or endogenous sources triggers cell-intrinsic innate immunity, but sensors recognizing both cytosolic dsDNA and dsRNA are sparsely reported. Here we find that full-length human SAMD9 protein directly binds to synthetic or viral dsDNA or dsRNA. Overexpression of SAMD9 from various vertebrate species leads to robust production of interferons and pro-inflammatory cytokines. By contrast, loss of endogenous SAMD9 impairs the interferon responses to cytosolic dsDNA and dsRNA stimulation in multiple cell types and enhances the infectivity of pathogenic dsDNA and dsRNA viruses. Mice lacking Samd9l, the human SAMD9 homolog, show increased viral load and severe clinical manifestations of rotavirus and reovirus infections. Rotavirus-encoded non-structural protein 1 targets SAMD9 for proteasomal degradation. Collectively, our data demonstrate that SAMD9 may serve as a pattern-recognition receptor for cytosolic dsDNA and dsRNA across different domains of life and represents a potential target of viral innate immune evasion.

Fig. 1. SAMD9 overexpression activates IFN signaling in a MAVS-TBK1-IRF3 dependent manner.

A A volcano plot of RNA sequencing results indicating top gene transcripts upregulated (red) and downregulated (blue) by SAMD9 overexpression as compared to empty vector (Ev) transfection in A549 cells. B–D A549 cells transfected with Ev or SAMD9-Myc-DDK plasmid for 24 h and the mRNA levels of IFNL3, CCL5, and CXCL10 measured by qRT-PCR (B), and levels of secreted IFN-λ3 (C) or CCL5 (D) quantified by ELISA. E Validation of wild-type (WT), single clonal MAVS KO, TBK1 KO A549 cells by western blot with indicated antibodies. F, G WT, single clonal MAVS KO and TBK1 KO A549 cells transfected with Ev or SAMD9-Myc-DDK plasmid for 24 h and the mRNA levels of IFNL3 (F) and CCL5 (G) measured by qRT-PCR. H A549 cells transfected with Ev or SAMD9-Myc-DDK plasmid for 24 h, and cell lysates harvested for SDS-PAGE electrophoresis with indicated antibodies. I, J A549 cells incubated with ruxolitinib (10 μM) for 1 h, transfected with Ev or SAMD9-Myc-DDK for 24 h and the mRNA levels of IFNL3 (I) and CCL5 (J) measured by qRT-PCR. The RNA sequencing experiment in (A) was performed once in duplicate, other experiments were repeated at least three times. Individual points in (B–D, F, G, I, and J) represent independent experiments. Data are represented as mean ± SEM. Statistical significance calculated by two-way ANOVA with Sidak’s multiple-comparisons tests (B, F, G, I, and J), or unpaired two-tailed Student’s t-test (C, D): n.s., not significant, *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file. Detailed P-values are provided in the Source Data file.

Fig. 2. SAMD9 senses cytosolic dsDNA and dsRNA.

A Validation of WT, SAMD9 KO, SAMD9-mGFP add-back SAMD9 KO, STING KO, STING SAMD9 double KO HT-29 cells by western blot with indicated antibodies. B WT, SAMD9 KO, SAMD9 KO complemented with SAMD9-mGFP, STING KO, and STING SAMD9 double KO HT-29 cells transfected with poly(dA:dT) (1 μg/ml) for 24 h and the IFNL3 mRNA level measured by qRT-PCR. C Validation of WT, SAMD9 KO, SAMD9-mGFP add-back SAMD9 KO, RIG-I KO, RIG-I SAMD9 double KO, MDA5 KO, MDA5 SAMD9 double KO, RIG-I MDA5 double KO, RIG-I MDA5 SAMD9 triple KO, MAVS KO, MAVS SAMD9 double KO HT-29 cells by western blot with indicated antibodies. D WT, SAMD9 KO, SAMD9 KO complemented with SAMD9-mGFP, RIG-I KO, RIG-I SAMD9 double KO, MDA5 KO, MDA5 SAMD9 double KO, RIG-I MDA5 double KO, RIG-I MDA5 SAMD9 triple KO, MAVS KO, and MAVS SAMD9 double KO HT-29 cells transfected with LMW poly(I:C) (1 μg/ml) for 24 h and the IFNL3 mRNA level measured by qRT-PCR. E, F Validation of WT and two individual single clonal SAMD9 KO human intestinal organoids by western blot (E) and immunofluorescence staining F. G WT and two individual SAMD9 KO clones of human ileal organoids transfected with poly(dA:dT) or LMW poly(I:C) (1 μg/ml) for 24 h and the IFNL3 mRNA level measured by qRT-PCR. Human ileal organoids fixed and stained for SAMD9 (green) and nuclei (blue) by microscopy. H Validation of WT and SAMD9 KO primary human hepatocytes by western blot. I IFNL3 mRNA levels in poly(dA:dT) or LMW poly(I:C) (1 μg/ml) transfected WT and SAMD9 KO primary human hepatocytes measured by qRT-PCR. For all figures, experiments were repeated at least three times. Individual points in (B, D, G, and I) represent independent experiments. Data are represented as mean ± SEM. Statistical significance calculated by two-way ANOVA with Tukey’s multiple-comparisons tests (B, D, G, and I): n.s., not significant, *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file. Detailed P-values are provided in the Source Data file.

Fig. 3. SAMD9 binds to dsDNA and dsRNA.

A Coomassie stain of His-SAMD9 protein analyzed by native-PAGE (left) or SDS-PAGE (right) electrophoresis. B Electrophoretic mobility shift assay (EMSA) of circular plasmid dsDNA with an increasing amount of SAMD9. C Measurement of the dissociation constant (K_D) of circular dsDNA with His-SAMD9, calculated based on a half-maximum binding on a curve fit. D EMSA of SAMD9 with linear dsDNA fragments of indicated lengths. E EMSA of His-SAMD9 with 80-bp dsDNA or 80-bp ssDNA. F EMSA of SAMD9 with linear dsRNA fragments of indicated lengths. G EMSA of His-SAMD9 with HMW poly(I:C) or single-stranded RNA polyuridine (ssRNA polyU). H EMSA of His-SAMD9, provided in 2-fold serial dilutions, with 512-bp dsRNA. I K_D measurements of 512 bp dsRNA with His-SAMD9, calculated based on a half-maximum binding on a curve fit. For all figures, experiments were repeated at least three times.

Fig. 4. SAMD9 OB-fold domain is essential for IFN induction.

A Schematic representation of SAMD9. FL full-length, OB oligonucleotide binding-fold domain. B Validation of WT HT-29 and SAMD9 KO HT-29 cells complemented with Ev, SAMD9-mGFP (FL), SAMD9-∆OB-mGFP (∆OB) by western blot with indicated antibodies. C WT HT-29 and SAMD9 KO HT-29 cells complemented with Ev, FL, or ∆OB transfected with poly(dA:dT) (1 μg/ml) for 24 h and the IFNL3 mRNA level measured by qRT-PCR. D WT HT-29 and SAMD9 KO HT-29 cells complemented with Ev, FL, or ∆OB transfected with LMW poly(I:C) (1 μg/ml) for 24 h and the IFNL3 mRNA level measured by qRT-PCR. E EMSA of His-MBP-tagged SAMD9 FL or ∆OB proteins with circular dsDNA. F Immunoblotting analysis of streptavidin immunoprecipitates of cell lysates from WT HT-29 and SAMD9 KO HT-29 cells complemented with FL or ∆OB incubated with biotinylated 48,502-bp dsDNA. G Immunoblotting analysis of anti-Flag antibody immunoprecipitates of cell lysates from HEK293 cells transfected with Myc-DDK-tagged SAMD9 FL or ∆OB. H WT HT-29 and SAMD9 KO HT-29 cells complemented with FL or ∆OB transfected with poly(dA:dT) (10 μg/ml) for 8 h, and cell lysates harvested for native-PAGE or SDS-PAGE electrophoresis with indicated antibodies. For all figures, experiments were repeated at least three times. Individual points in (C and D) represent independent experiments. Data are represented as mean ± SEM. Statistical significance in (C and D) calculated by one-way ANOVA with Tukey’s post-test: ***P < 0.001. Source data are provided as a Source Data file. Detailed P-values are provided in the Source Data file.

Fig. 5. SAMD9 restricts the infection of vaccinia virus, rotavirus, and reovirus.

A EMSA of SAMD9 with rotavirus (RV) segmented dsRNA genome. B Nuclear, membrane, and cytoplasmic fractions from A549 cells prepared by homogenization and differential centrifugation. Pellet (P8000) and supernatant (S8000) from centrifugation of 8000×g immunoblotted with indicated antibodies. C A549 cells infected with mock or RV (MOI = 300) for 8 h, then fixed and stained for SAMD9 (green), dsRNA (red), nuclei (blue) and colocalization (yellow) by confocal microscopy. D A549 cells infected with mock or RV (MOI = 3) for 8 h and cell lysates harvested for native-PAGE or SDS-PAGE electrophoresis with indicated antibodies. E, F WT, SAMD9 KO, SAMD9 KO complemented with SAMD9-mGFP, RIG-I KO, RIG-I SAMD9 double KO, MDA5 KO, MDA5 SAMD9 double KO, RIG-I MDA5 double KO, RIG-I MDA5 SAMD9 triple KO, MAVS KO, and MAVS SAMD9 double KO HT-29 cells infected by RV (MOI = 0.01) for 24 h, the IFNL3 mRNA level (E) or CCL5 mRNA level (F) measured by qRT-PCR. G, H WT and SAMD9 KO complemented with Ev or SAMD9-mGFP (Rescue) HT-29 cells infected with RV (MOI = 0.01) for 24 h, the levels of secreted IFN-λ3 (G) and CCL5 (H) quantified by ELISA. I WT and SAMD9 KO Hela cells infected by indicated viruses (MOI = 0.1) for 24 h and the mRNA levels of viral genes measured by qRT-PCR. RRV rhesus rotavirus, UK bovine rotavirus, WI61 human rotavirus, Reo mammalian orthoreovirus, IAV influenza A virus, VSV vesicular stomatitis virus, CHIKV chikungunya virus, ZIKV Zika virus, HAstV human astrovirus, AdV adenovirus. For all figures, experiments were repeated at least three times. Individual points in (E–I) represent independent experiments. Data are represented as mean ± SEM. Statistical significance calculated by two-way ANOVA with Tukey’s multiple-comparisons tests (E–I): n.s., not significant, *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file. Detailed P-values are provided in the Source Data file.

Fig. 6. SAMD9 from different vertebrate species mediates IFN induction.

A Maximum likelihood tree of SAMD9 and SAMD9L from selected vertebrate species with bootstrap values indicated on the branches. B A549 cells transfected with Ev, human SAMD9 (hSAMD9), and human SAMD9L (hSAMD9L) for 24 h, the SAMD9 and SAMD9L protein levels examined by western blot with indicated antibodies. C, D A549 cells transfected with Ev, hSAMD9, and hSAMD9L for 24 h, the IFNL3 mRNA level (C) and CCL5 mRNA level (D) measured by qRT-PCR. E A549 cells transfected with Ev, hSAMD9, C57Bl/6J mSAMD9L, and 129S6/SvEv mSAMD9L for 24 h, the SAMD9 and SAMD9L protein levels examined by western blot with indicated antibodies. F, G A549 cells transfected with Ev, hSAMD9, mSAMD9L from C57Bl/6J or 129S6/SvEv strains for 24 h, the IFNL3 mRNA level (F) and CCL5 mRNA level (G) measured by qRT-PCR. H A549 cells transfected with Ev, hSAMD9, hamster SAMD9 (hmSAMD9), hamster SAMD9L (hmSAMD9L), or zebrafish SAMD9L (zSAMD9L) for 24 h, the SAMD9 and SAMD9L protein levels examined by western blot with indicated antibodies. I, J A549 cells transfected with Ev, hSAMD9, hmSAMD9, hmSAMD9L, or zSAMD9L for 24 h, the IFNL3 mRNA level (I) and CCL5 mRNA level (J) measured by qRT-PCR. For all figures, experiments were repeated at least three times. Individual points in (C, D, F, G, I, and J) represent independent experiments. Data are represented as mean ± SEM. Statistical significance calculated by one-way ANOVA with Dunnett’s post-test (C, D, F, G, I, and J): *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file. Detailed P-values are provided in the Source Data file.

Fig. 7. Samd9l^−/− mice are more susceptible to dsRNA virus infections.

A–E Five-day-old Samd9l^+/+ and Samd9l^−/− neonatal mice orally infected with simian RV (2.4 × 10⁷ FFU per mouse). RNA extracted from duodenum (A), jejunum (B), ileum (C), pancreas (D) and lung (E) collected at 3 dpi, and RV NSP5 mRNA levels measured by qRT-PCR. Dotted lines: limits of detection. F–H Three-day-old Samd9l^+/+ and Samd9l^−/− neonatal mice intraperitoneally injected with reovirus (1 × 10⁴ PFU per mouse) and the percentages of diarrhea development (F), biliary disease (G), or survival (H) recorded from days 1 to 15 post-infection. I–K Three-day-old Samd9l^+/+ and Samd9l^−/− neonatal mice intraperitoneally injected with reovirus (1 × 10⁴ PFU per mouse). RNA extracted from liver (I), lung (J) and heart (K) collected at 9 dpi, and reovirus mRNA levels measured by qRT-PCR. Dotted lines: limits of detection. For all figures, experiments were repeated at least three times. Individual points in (A–E and I–K) represent individual mouse. Data are represented as mean ± SEM. Statistical significance calculated by unpaired two-tailed Student’s t-test (A–E and I–K), or log-rank test (L): *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file. Detailed P-values are provided in the Source Data file.

Fig. 8. Rotavirus NSP1 is a viral antagonist of SAMD9.

A HEK293 cells transfected with GFP-ev, or GFP-tagged RV-encoded proteins (VP1-VP4, VP6, VP7, and NSP1-NSP6) for 48 h and cell lysates harvested for SDS-PAGE electrophoresis with indicated antibodies. B MA104 cells infected by indicated recombinant simian RV SA11 (MOI = 3) for 8 h and cell lysates harvested for SDS-PAGE electrophoresis with indicated antibodies. C Schematic representation of N-terminally GFP-tagged WT and truncated RV NSP1 constructs. D Immunoblotting analysis of anti-GFP antibody immunoprecipitates of cell lysates from HEK293 cells transfected with GFP-tagged WT or truncated NSP1s. E Schematic representation of N-terminally 3 × Flag-tagged WT and truncated SAMD9 constructs. F HEK293 cells co-transfected with Flag-tagged SAMD9 mutants and GFP-ev or GFP-NSP1 for 48 h and cell lysates harvested for SDS-PAGE electrophoresis with indicated antibodies. G Immunoblotting analysis of anti-GFP antibody immunoprecipitates of cell lysates from HEK293 cells co-transfected with GFP-tagged NSP1 and Flag-tagged SAMD9 mutants. H HEK293 cells co-transfected with SAMD9-Myc-DDK and GFP-ev or GFP-NSP1 for 36 h, treated with MG132 (10 μM), PYR-41 (50 μM), bortezomib (10 μM), or MLN4924 (10 μM) for another 12 h, and cells lysates harvested for SDS-PAGE electrophoresis with indicated antibodies. I HEK293 cells transfected with control or NEDD8-specific siRNA for 48 h, then co-transfected with SAMD9-Myc-DDK and GFP-ev or GFP-NSP1 for another 36 h, and cell lysates harvested for SDS-PAGE electrophoresis with indicated antibodies. J Relative abundance of SAMD9 and SAMD9L protein levels from indicated species in SAMD9 KO HEK293 co-transfected with indicated SAMD9 or SAMD9L and NSP1s derived from different RV strains. Bv bovine. Darker colors indicate a higher level of SAMD9/9L inhibition by NSP1. For all figures, experiments were repeated at least three times.