Essential Role of the N-terminal Domain in the Regulation of RIG-I ATPase Activity

Abstract

Retinoic acid-inducible gene I (RIG-I) is a cytosolic receptor that recognizes viral RNA and activates the interferon-mediated innate antiviral response. To understand the mechanism of signal activation at the receptor level, we cloned, expressed, and purified human RIG-I containing the two caspase activation and recruitment domains (CARDs) followed by the C-terminal helicase domain. We found that recombinant RIG-I is a functional protein that interacts with double-stranded RNA with substantially higher affinity as compared with single-stranded RNA structures unless they contain a 5′-triphosphate group. Viral RNA binding to RIG-I stimulates the velocity of ATP hydrolysis by 33-fold, which at the cellular level translates into a 43-fold increase of interferon-β expression. In contrast, the isolated ATPase/helicase domain is constitutively activated while also retaining its RNA ligand binding properties. These results support the recent model by which RIG-I signaling is autoinhibited in the absence of RNA by intra-molecular interactions between the CARDs and the C terminus. Based on pH profile and metal ion dependence experiments, we propose that the active site of RIG-I cannot efficiently accommodate divalent cations under the RNA-free repressed conformation. Overall, these results show a direct correlation between RNA binding and ATPase enzymatic function leading to signal transduction and suggest that a tight control of ATPase activity by the CARDs prevents RIG-I signaling in the absence of viral RNA. Retinoic acid-inducible gene I (RIG-I) is a cytosolic receptor that recognizes viral RNA and activates the interferon-mediated innate antiviral response. To understand the mechanism of signal activation at the receptor level, we cloned, expressed, and purified human RIG-I containing the two caspase activation and recruitment domains (CARDs) followed by the C-terminal helicase domain. We found that recombinant RIG-I is a functional protein that interacts with double-stranded RNA with substantially higher affinity as compared with single-stranded RNA structures unless they contain a 5′-triphosphate group. Viral RNA binding to RIG-I stimulates the velocity of ATP hydrolysis by 33-fold, which at the cellular level translates into a 43-fold increase of interferon-β expression. In contrast, the isolated ATPase/helicase domain is constitutively activated while also retaining its RNA ligand binding properties. These results support the recent model by which RIG-I signaling is autoinhibited in the absence of RNA by intra-molecular interactions between the CARDs and the C terminus. Based on pH profile and metal ion dependence experiments, we propose that the active site of RIG-I cannot efficiently accommodate divalent cations under the RNA-free repressed conformation. Overall, these results show a direct correlation between RNA binding and ATPase enzymatic function leading to signal transduction and suggest that a tight control of ATPase activity by the CARDs prevents RIG-I signaling in the absence of viral RNA. Pathogen recognition receptors constitute one of the first lines of human host cell defense against viral infections. Among them, retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA-5) helicases are specialized sensors of viral nucleic acid. RIG-I was shown to be involved in the recognition of Sendai virus, Hepatitis C virus (HCV), 2The abbreviations used are: HCVhepatitis C virusCARDcaspase activation and recruitment domainIFNinterferoncIREScomplementary internal ribosome entry sitedsRNAdouble-stranded RNAssRNAsingle-stranded RNA3P-ssRNAtriphosphorylated single-stranded RNAEMSAelectromobility shift assay. 2The abbreviations used are: HCVhepatitis C virusCARDcaspase activation and recruitment domainIFNinterferoncIREScomplementary internal ribosome entry sitedsRNAdouble-stranded RNAssRNAsingle-stranded RNA3P-ssRNAtriphosphorylated single-stranded RNAEMSAelectromobility shift assay. and vesicular stomatitis virus, whereas MDA-5 mediates the antiviral response to picornaviruses (1Hornung V. Ellegast J. Kim S. Brzozka K. Jung A. Kato H. Poeck H. Akira S. Conzelmann K.K. Schlee M. Endres S. Hartmann G. Science. 2006; 314: 994-997Crossref PubMed Scopus (1861) Google Scholar, 2Kato H. Takeuchi O. Sato S. Yoneyama M. Yamamoto M. Matsui K. Uematsu S. Jung A. Kawai T. Ishii K.J. Yamaguchi O. Otsu K. Tsujimura T. Koh C.S. Reis e Sousa C. Matsuura Y. Fujita T. Akira S. Nature. 2006; 441: 101-105Crossref PubMed Scopus (2862) Google Scholar, 3Pichlmair A. Schulz O. Tan C.P. Naslund T.I. Liljestrom P. Weber F. Reis e Sousa C. Science. 2006; 314: 997-1001Crossref PubMed Scopus (1743) Google Scholar). Both RIG-I and MDA-5 are modular proteins of ∼920 residues that contain an ATPase/helicase domain for RNA binding and two caspase activation and recruitment domains (CARDs) located in tandem at the N-terminal end of the protein (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google Scholar). The CARDs, by way of homotypic interactions, are required for the signaling function between RIG-I and the newly discovered interferon-β promoter stimulator-1 (IPS-1), also called Cardif, MAVS, and VISA (for review see Ref. 5Hiscott J. Lin R. Nakhaei P. Paz S. Trends Mol. Med. 2006; 12: 53-56Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar). The latter serves as an intermediate to the cascade leading to the activation of NF-κB and IRF-3/7, which results in the production of antiviral cytokines such as type-1 interferons (6Seth R.B. Sun L. Ea C.K. Chen Z.J. Cell. 2005; 122: 669-682Abstract Full Text Full Text PDF PubMed Scopus (2433) Google Scholar, 7Kawai T. Takahashi K. Sato S. Coban C. Kumar H. Kato H. Ishii K.J. Takeuchi O. Akira S. Nat. Immunol. 2005; 6: 981-988Crossref PubMed Scopus (1990) Google Scholar, 8Xu L.G. Wang Y.Y. Han K.J. Li L.Y. Zhai Z. Shu H.B. Mol. Cell. 2005; 19: 727-740Abstract Full Text Full Text PDF PubMed Scopus (1492) Google Scholar, 9Meylan E. Curran J. Hofmann K. Moradpour D. Binder M. Bartenschlager R. Tschopp J. Nature. 2005; 437: 1167-1172Crossref PubMed Scopus (1937) Google Scholar). The receptor function of RIG-I is non-redundant and therefore essential to coordinate the signaling cascade leading to antiviral response (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google Scholar), as confirmed by knock-out studies (10Kato H. Sato S. Yoneyama M. Yamamoto M. Uematsu S. Matsui K. Tsujimura T. Takeda K. Fujita T. Takeuchi O. Akira S. Immunity. 2005; 23: 19-28Abstract Full Text Full Text PDF PubMed Scopus (1100) Google Scholar). Notably, the Huh7.5 cell line of hepatocytes is particularly permissive to HCV and vesicular stomatitis virus infection as the result of a defective RIG-I protein bearing a single mutation in its first CARD (11Sumpter Jr., R. Loo Y.M. Foy E. Li K. Yoneyama M. Fujita T. Lemon S.M. Gale Jr., M. J. Virol. 2005; 79: 2689-2699Crossref PubMed Scopus (721) Google Scholar); the antiviral signaling pathway can be restored by overexpressing wild-type RIG-I (12Cheng G. Zhong J. Chung J. Chisari F.V. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 9035-9040Crossref PubMed Scopus (131) Google Scholar). Viruses have found ways to avoid the RIG-I-mediated antiviral response. For instance, the NS3/4A protease of HCV cleaves IPS-1 to prevent its interaction with RIG-I (9Meylan E. Curran J. Hofmann K. Moradpour D. Binder M. Bartenschlager R. Tschopp J. Nature. 2005; 437: 1167-1172Crossref PubMed Scopus (1937) Google Scholar, 13Li X.D. Sun L. Seth R.B. Pineda G. Chen Z.J. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 17717-17722Crossref PubMed Scopus (649) Google Scholar, 14Lin R. Lacoste J. Nakhaei P. Sun Q. Yang L. Paz S. Wilkinson P. Julkunen I. Vitour D. Meurs E. Hiscott J. J. Virol. 2006; 80: 6072-6083Crossref PubMed Scopus (191) Google Scholar, 15Loo Y.M. Owen D.M. Li K. Erickson A.K. Johnson C.L. Fish P.M. Carney D.S. Wang T. Ishida H. Yoneyama M. Fujita T. Saito T. Lee W.M. Hagedorn C.H. Lau D.T. Weinman S.A. Lemon S.M. Gale Jr., M. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 6001-6006Crossref PubMed Scopus (349) Google Scholar). Negative regulation of RIG-I signaling can also be mediated by cellular factors as a way to regulate uncontrolled production of interferon (IFN). Among them, LGP2 prevents the activation of RIG-I by sequestering its double-stranded RNA ligand (16Rothenfusser S. Goutagny N. DiPerna G. Gong M. Monks B.G. Schoenemeyer A. Yamamoto M. Akira S. Fitzgerald K.A. J. Immunol. 2005; 175: 5260-5268Crossref PubMed Scopus (483) Google Scholar), whereas RNF125 induces the ubiquitination of RIG-I (17Arimoto K. Takahashi H. Hishiki T. Konishi H. Fujita T. Shimotohno K. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 7500-7505Crossref PubMed Scopus (350) Google Scholar). hepatitis C virus caspase activation and recruitment domain interferon complementary internal ribosome entry site double-stranded RNA single-stranded RNA triphosphorylated single-stranded RNA electromobility shift assay. hepatitis C virus caspase activation and recruitment domain interferon complementary internal ribosome entry site double-stranded RNA single-stranded RNA triphosphorylated single-stranded RNA electromobility shift assay. The molecular mechanism of activation of RIG-I in response to RNA binding is still poorly understood. Although overexpression of the isolated CARDs induces a constitutive signaling independent from viral infection, mutagenesis experiments demonstrated that full-length RIG-I requires a functional ATPase to activate the response leading to IFN production (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google Scholar). More recently, it has been proposed that RIG-I signaling is intrinsically inhibited by intramolecular interactions between the CARDs and the last 190 amino acids from the C terminus, named the repressor domain (18Saito T. Hirai R. Loo Y.M. Owen D. Johnson C.L. Sinha S.C. Akira S. Fujita T. Gale Jr., M. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 582-587Crossref PubMed Scopus (572) Google Scholar). According to this new model, latent RIG-I must undergo a conformational change and self-associate in order to activate IPS-1-mediated response. The present study was conducted to reconcile these independent observations and to establish a link between RNA binding, ATPase activity, and CARD-mediated signal transduction leading to IFN expression. Using enzyme kinetics combined with signal transduction assays, we discovered that the ATPase function of RIG-I is negatively regulated by the CARDs in the absence of RNA, which reveals the pivotal role of the ATPase/helicase domain in the modulation of RIG-I signaling. RNA Methods—Short from by the as by the used in double-stranded was to was in from the with the of in RNA was with RNA and while was used for of RNA in the of and was purified with The RNA was with the as K. V. S. H. H. C. M. C. K. D. J. N. R. I. J. 2006; Full Text Full Text PDF PubMed Scopus Google Scholar). and from and wild-type human RIG-I was from and of the gene to a containing at the of the and by mutagenesis with and on RIG-I was in with with for at a at in pH and and by at for The was with with at for on with a of to RIG-I was purified with affinity and against pH and at ATPase ATPase in the helicase pH in the of used the domain full-length RIG-I in a at for from 5 to by in with and for and the was at the RNA activation under as and RNA RIG-I was at with RNA for at in helicase the on a a at and by The of with a binding model the as L. C. L. P. 2006; PubMed Scopus Google Scholar), is the of RNA, is the is the of RNA, is the of RIG-I is the and is the and in with and cell line are with a HCV and are a of the Huh7.5 cell The was from the cell in the cell line To the of the in was from the The RIG-I was the and with of RIG-I as a into the was with in RNA and cellular RNA was RNA was isolated RNA to the was in a of cellular RNA To the level of and RIG-I gene of in was in a of with the of and with with the The human RIG-I and was from to gene in gene with wild-type RIG-I in the of RNA as a to the in as The ATPase of RIG-I and purified the full-length human RIG-I containing the two CARDs for signal transduction and the domain as a RNA ligand receptor The ATPase activity of RIG-I was by the of with of The was also conducted in the of RNA a The double-stranded was for this it the the of the HCV these we that the of the from to in the of activation was by a higher to in ATP We RNA structures can the activation of The ATPase activity of RIG-I was also in the of with that the ATPase domain is activated binding to double-stranded RNA structures of in We also the of In the absence of RNA, RIG-I in between In the a for the as by the and increase in of ATP and these results show that purified RIG-I is a functional RNA receptor with an RNA In the absence of the ATPase activity of RIG-I is of RNA RIG-I to RNA to establish a link between ligand binding and ATPase We the RNA binding of RIG-I by an electromobility shift of protein with RNA for at and the to a these RIG-I was to to to of RNA at and The of molecular one of was with a binding model a We also the of the to binding, as it was to an role in RIG-I activation (1Hornung V. Ellegast J. Kim S. Brzozka K. Jung A. Kato H. Poeck H. Akira S. Conzelmann K.K. Schlee M. Endres S. Hartmann G. Science. 2006; 314: 994-997Crossref PubMed Scopus (1861) Google Scholar, 3Pichlmair A. Schulz O. Tan C.P. Naslund T.I. Liljestrom P. Weber F. Reis e Sousa C. Science. 2006; 314: 997-1001Crossref PubMed Scopus (1743) Google Scholar). in by RNA was with the helicase to a the of the 5′-triphosphate on the of the as compared with Overall, we discovered that RIG-I a profile of binding with a for RNA as binding is with the profile for of these on ATPase activity, with a increase of with also observations that RIG-I with (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google Scholar, G. Zhong J. Chung J. Chisari F.V. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 9035-9040Crossref PubMed Scopus (131) Google Scholar). The the of the helicase domain of RIG-I by its amino against the one of LGP2 The is and to be it therefore to the functional of RIG-I the regulation of its signaling. To these we to and of the helicase domain containing of the between and We that the regulate the enzymatic function of the helicase domain and to the the of this to as We also at the of RNA on ATPase To of the was to the activity of the isolated helicase domain To that this in regulation was by an RNA binding we the of the by The isolated domain the RNA binding as the full-length enzyme the and profile of in the absence of RNA to the of the activated full-length RIG-I these results that the helicase domain is constitutively activated in the absence of the ATPase for of and in helicases are a of We to by mutagenesis the of of these to the ATPase activity of RIG-I The proteins containing a mutation in in still to RNA enzymatic activity compared with wild-type RIG-I under RNA activation The results with the proteins which the of of the two residues for ATPase To the of at the cellular level, we gene in hepatocytes as a for RIG-I of RNA a increase in gene in The of was also confirmed by of gene gene that in that The absence of response to RNA in containing a defective RIG-I that is a of the signaling We with wild-type RIG-I to of recombinant RIG-I followed by RNA these wild-type RIG-I restored the response to with RNA, in a 43-fold increase in gene compared with In the and in RIG-I production of and antiviral activation of the activity Overall, these results with the enzymatic of the purified proteins and show that the ATPase activity of RIG-I is required for the production of IFN under RNA The ATPase of RIG-I of helicases metal to coordinate the and of ATP the by a are a of with ATPase activity G. R. PubMed Scopus Google Scholar, S. J. Mol. 2007; PubMed Scopus Google Scholar, J. Full Text Full Text PDF PubMed Scopus Google Scholar). ATPase conducted with to it is to RNA we discovered that full-length RIG-I for its ATPase activity the metal ion in the in the divalent metal such as the of of by which the of metal The is to RIG-I as the to is confirmed by the that such as can the ATPase activity of RIG-I and domain We also at the of pH on has been shown with helicases that the binding affinity of the enzyme to PubMed Scopus Google Scholar), in a of activity Johnson R.B. Sun K.A. J. Wang J. PubMed Google Scholar, G. J. J. Virol. 80: PubMed Scopus Google Scholar). In experiments, the ATPase activity of the isolated domain was to pH in and between pH and the the activity of full-length RIG-I was by pH from to and to to RNA that the level of of residues from the active site the binding of to in the of the in the ATPase of this is that RNA binding induces a conformational change of the active site of RIG-I in order to the residues in a to the of the intrinsically activated helicase domain. We present the first enzymatic of human recombinant RIG-I and show that the CARDs an essential role in the regulation of the ATPase domain. Both full-length enzyme and the isolated helicase domain are functional proteins that can be purified in E. experiments these two show that RIG-I ATP the activity is helicases O. J. M. P. Mol. Cell. Full Text Full Text PDF PubMed Scopus Google Scholar, PubMed Scopus Google and can be inhibited by the and the and results show that recombinant RIG-I is also a Using a protein shift we the binding for of and demonstrated that is a ligand for RIG-I unless it a group. These results are in with new that RIG-I can also against that contain a on (1Hornung V. Ellegast J. Kim S. Brzozka K. Jung A. Kato H. Poeck H. Akira S. Conzelmann K.K. Schlee M. Endres S. Hartmann G. Science. 2006; 314: 994-997Crossref PubMed Scopus (1861) Google Scholar, 3Pichlmair A. Schulz O. Tan C.P. Naslund T.I. Liljestrom P. Weber F. Reis e Sousa C. Science. 2006; 314: 997-1001Crossref PubMed Scopus (1743) Google that a for the between and RNA. the RIG-I recognizes double-stranded a it is to that the affinity of RIG-I for is higher for of the of protein in the latter study also demonstrated that RNA binding to a conformational change that stimulates the ATPase activity of RIG-I in with a for it is that double-stranded can the of IFN S. A. R. Nat. PubMed Scopus Google Scholar, M. Nat. PubMed Scopus Google Scholar). recent study that RIG-I recognizes that contain T. Wang D. M. P. Hartmann R. Fujita T. Nat. 2006; PubMed Scopus Google Scholar), which is the of RNA used in was to activate IFN in to a of under In with we that the double-stranded is also a of RIG-I ATPase activity (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google Scholar, G. Zhong J. Chung J. Chisari F.V. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 9035-9040Crossref PubMed Scopus (131) Google Scholar, Y.M. J. N. G. O. L. Akira S. A. Gale Jr., M. J. Virol. PubMed Scopus Google Scholar). In contrast, two knock-out studies that activates MDA-5 in which in response between human and H. Takeuchi O. Sato S. Yoneyama M. Yamamoto M. Matsui K. Uematsu S. Jung A. Kawai T. Ishii K.J. Yamaguchi O. Otsu K. Tsujimura T. Koh C.S. Reis e Sousa C. Matsuura Y. Fujita T. Akira S. Nature. 2006; 441: 101-105Crossref PubMed Scopus (2862) Google Scholar, L. S. M. M. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: PubMed Scopus Google Scholar). We found a correlation between activation of the ATPase activity and of IFN an RNA was used as a viral of containing a the of by RIG-I to for ATP The enzymatic activation into a 43-fold increase of IFN in hepatocytes with In study reveals that in a molecular of activation for purified the molecular for binding of RNA to RIG-I in a of the structures such as the present in HCV RNA. To RNA to the ATPase activity of the isolated domain Although it is to that the ATPase/helicase its to an N-terminal we that the between the CARDs and the ATPase is for the enzymatic activity and RNA binding and that RIG-I is a modular protein with two functional domains by a we this of regulation of enzymatic activity as an activation of the helicase domain in the absence of the the activity and the of the helicase domain are with the activity of full-length RIG-I RNA and result is particularly for two it the link between two that show that ATPase activity is essential for the signaling function of RIG-I (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google Scholar), and that the C of RIG-I interactions with the IPS-1 signaling (18Saito T. Hirai R. Loo Y.M. Owen D. Johnson C.L. Sinha S.C. Akira S. Fujita T. Gale Jr., M. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 582-587Crossref PubMed Scopus (572) Google Scholar). results support the of an autoinhibited in the absence of ligand and on the role of the CARDs in of RIG-I the ATPase The of this is that leading to the signaling function of RIG-I must at two conformational In the first autoinhibited RIG-I to RNA and a to the one of the isolated helicase domain. In the autoinhibited to in the enzyme is to RNA, to that the from latent to activated a of the active RIG-I is hydrolysis of ATP the for a conformational change to the CARDs and with also by RIG-I We therefore propose a mechanism to control of ATPase activity by the CARDs is to the regulation of the signaling function of Although we direct of interaction by ATPase in we model that the CARDs are for signaling the ATPase has been is by the that a of ATPase activity in translates into a of signaling function the The was with which a of signal activation of IFN by RIG-I to has been (4Yoneyama M. Kikuchi M. Matsumoto K. Imaizumi T. Miyagishi M. Taira K. Foy E. Loo Y.M. Gale Jr., M. Akira S. Yonehara S. Kato A. Fujita T. J. Immunol. 2005; 175: 2851-2858Crossref PubMed Scopus (1269) Google and be to a between defective RIG-I and helicases ATP to the conformational to of nucleic acid. is by HCV which requires of ATP hydrolysis S. V. Jr., I. C. Nature. 2006; PubMed Scopus Google Scholar). the of the is to be and hydrolysis in order to the is is RIG-I also RNA to signal be by the activity of RIG-I T. Wang D. M. P. Hartmann R. Fujita T. Nat. 2006; PubMed Scopus Google and interferon production in the the of the repressor domain to the ATPase activity of RIG-I also be The model for the control of RIG-I signaling is of the regulation of nucleic the protein In this in a latent X.D. Sun L. Seth R.B. Pineda G. Chen Z.J. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 17717-17722Crossref PubMed Scopus (649) Google Scholar). RNA binding induces a conformational change that the the of this mechanism is still Trends Sci. 2007; Full Text Full Text PDF PubMed Scopus Google Scholar). have that of is required for signal transduction S.A. Kim I. G. J. 2007; Full Text Full Text PDF PubMed Scopus Google Scholar). of RIG-I is to an role in activation of the signaling (18Saito T. Hirai R. Loo Y.M. Owen D. Johnson C.L. Sinha S.C. Akira S. Fujita T. Gale Jr., M. Proc. Natl. Acad. Sci. U. S. A. 2007; 104: 582-587Crossref PubMed Scopus (572) Google Scholar). We in the of molecular one of RNA. These results in of an of RIG-I RNA binding, experiments suggest that RIG-I can also the for RNA binding with a single function a binding to be with full-length RIG-I with the isolated helicase domain as by the and the of a of the CARDs in interactions and The direct role of these domains in the of proteins such as and has been N. T. L. Y. C. Chen S. R. J. D. J. G. J. Full Text Full Text PDF PubMed Scopus Google Scholar, G. K. J. Full Text Full Text PDF PubMed Scopus Google Scholar, Y. L. D.M. G. J. Full Text Full Text PDF PubMed Scopus Google Scholar), and it be the level of of RIG-I also its to with which residues are involved in and of RIG-I to the molecular that the of this ligand binding, ATPase activity, and signal In we the first enzymatic of RIG-I and that RIG-I ATPase activity is by the We also show that structures of RNA can this and signal transduction to The of the the activation of the ATPase which RIG-I a of RNA We J. and for in the of the The was by and and the was from We also for the and for the and and