IFIH1

UniProt ID: Q9BYX4
Organism: Homo sapiens
Review Status: COMPLETE
πŸ“ Provide Detailed Feedback

Gene Description

Interferon-induced helicase C domain-containing protein 1 (IFIH1/MDA5) is a cytosolic RIG-I-like receptor that couples recognition of duplex RNA to type I interferon induction. It has two N-terminal caspase activation and recruitment domains (CARDs), a central DExD/H-box helicase core with an inserted domain, and a zinc-binding C-terminal regulatory domain. Unlike RIG-I, which reads the end of a duplex, MDA5 binds the internal stem of long double-stranded RNA and polymerises head-to-tail along it; ATP binding and hydrolysis drive assembly and disassembly of these filaments, giving a proofreading mechanism that makes signalling depend on duplex length rather than on a chemical mark at the RNA end. Filament formation clusters the CARDs, which nucleate MAVS prion-like polymers on the mitochondrial outer membrane and activate TBK1/IKBKE and IRF3/IRF7, inducing interferon-alpha/beta and pro-inflammatory cytokines; MDA5 itself relocalises from cytosol to mitochondria upon activation and ISGylation. MDA5 is required for the response to picornaviruses such as encephalomyocarditis virus, and contributes to responses to coronaviruses, flaviviruses and some DNA viruses. Which RNA actually engages MDA5 during infection is unresolved. Long viral replicative duplexes and viral mRNA lacking the host 2'-O-methyl cap mark activate it, but crosslinking of endogenous MDA5 in SARS-CoV-2- and encephalomyocarditis-virus-infected cells recovers almost exclusively cellular RNA, concentrated in introns and near inverted Alu repeats, and infection-induced splicing defects are sufficient to trigger it; on that evidence MDA5 has been proposed to monitor the fidelity of host RNA processing rather than to detect a pathogen-specific molecular pattern. The same ambiguity is visible in disease: gain-of-function IFIH1 variants cause Aicardi-Goutieres syndrome and Singleton-Merten-spectrum type I interferonopathies through sterile activation, and adenosine-to-inosine editing by ADAR1 exists in part to prevent MDA5 from responding to endogenous duplex RNA. Common IFIH1 variants are associated with type 1 diabetes and other autoimmune conditions.

Proposed New Ontology Terms

aberrant cellular RNA sensor activity

Definition: Binding to improperly processed or mislocalised cellular RNA, such as intron-retaining or unspliced transcripts accumulating in the cytoplasm, and thereby initiating an innate immune response. Unlike pattern recognition receptor activity, this activity does not require the bound ligand to be of pathogen origin; the discriminating feature is the processing state of the RNA rather than its provenance.

Justification: GO:0038187 pattern recognition receptor activity is defined as combining with a pathogen-associated molecular pattern. MDA5 is annotated to it, yet the PAMP has never been identified - it is described in the 2026 literature as one of the few remaining 'orphan' PRRs - and crosslinking of endogenous MDA5 in virus-infected cells recovers host rather than viral RNA. Sterile activation in ADAR1 deficiency and in IFIH1 gain-of-function interferonopathies points the same way. There is currently no molecular function term for a receptor that initiates innate immunity by detecting the processing state of self RNA, so curators must either use a term whose definition does not fit or leave the activity unannotated. The same gap affects ZBP1 and OAS1, for which comparable host-RNA and transcription-termination sensing has been reported.

Parent term: double-stranded RNA binding

Supporting Evidence:

Existing Annotations Review

GO Term Evidence Action Reason
GO:0140374 antiviral innate immune response
IBA
GO_REF:0000033
ACCEPT
Summary: MDA5 drives the antiviral innate immune response to picornaviruses, coronaviruses and other agents.
Reason: Core biological process, and unaffected by the ligand question: whether MDA5 is triggered by viral RNA or by infection-induced aberrant host RNA, the process it participates in is antiviral innate immunity.
GO:0005737 cytoplasm
IBA
GO_REF:0000033
ACCEPT
Summary: MDA5 is a cytoplasmic protein.
Reason: Correct and consistently observed; this is where the sensor resides and acts.
GO:0039530 MDA-5 signaling pathway
IBA
GO_REF:0000033
ACCEPT
Summary: The MDA-5 signalling pathway itself: ligand-bound MDA5 filaments nucleate CARD oligomers that activate MAVS, TBK1/IKBKE and IRF3/IRF7.
Reason: Core biological process, named for this gene and supported by independent experimental work.
GO:0003725 double-stranded RNA binding
IBA
GO_REF:0000033
ACCEPT
Summary: MDA5 binds long double-stranded RNA and cooperatively assembles head-to-tail filaments along the duplex; this is its core ligand-binding activity.
Reason: Core molecular function, established structurally and biochemically and not in dispute. What is disputed is the origin of the duplex RNA that engages MDA5 in infected cells, not the binding activity itself. Sampaio et al. 2026 (PMID:42581186) applied iCLIP to endogenous MDA5 in SARS-CoV-2- and EMCV-infected cells and recovered essentially no viral RNA: the bound population was host RNA, enriched in introns, Alu-proximal regions and predicted duplexes, and MDA5 activation was abolished by SRSF3 overexpression, which restores splicing fidelity. The authors frame MDA5 as surveying RNA processing fidelity ('guarding') rather than detecting a pathogen-encoded pattern, and state plainly that they cannot exclude direct viral dsRNA detection in other settings. This reframes which RNA activates MDA5; it does not challenge the biochemistry of duplex-RNA binding. The host ligands it identifies - inverted Alu repeats and structured intronic RNA - are themselves double-stranded, so this term remains correct under either model.
Supporting Evidence:
PMID:23273991
MDA5 recognizes the internal duplex structure, whereas RIG-I recognizes the terminus of dsRNA
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
PMID:42581186
upon infection with SARS-CoV-2 or encephalomyocarditis virus, MDA5 bound overwhelmingly to cellular RNAs
GO:0003727 single-stranded RNA binding
IBA
GO_REF:0000033
KEEP AS NON CORE
Summary: Single-stranded RNA binding reported alongside the duplex-binding activity.
Reason: Real but peripheral. The signalling-competent ligand for MDA5 is a long duplex bound by a cooperative filament; single-stranded segments are bound within higher-order structures rather than as the activating species. Retained as a genuine binding property, demoted from core.
Supporting Evidence:
PMID:19656871
Activation of MDA5 requires higher-order RNA structures generated during virus infection.
GO:0008270 zinc ion binding
IBA
GO_REF:0000033
ACCEPT
Summary: Zinc coordinated by the C-terminal regulatory domain.
Reason: A structural metal site, directly observed, required to fold the domain that caps the RNA duplex. Correct, though architectural rather than a function in its own right.
GO:0003677 DNA binding
IEA
GO_REF:0000002
REMOVE
Summary: DNA binding assigned electronically from an InterPro signature.
Reason: This is an InterPro2GO artefact, not a claim about MDA5. GO_REF:0000002 maps IPR006935 (Helicase/UvrB, N-terminal; Pfam PF04851 ResIII) to GO:0003677, GO:0005524 and GO:0016787, because that domain family is dominated by DNA-acting enzymes such as UvrB and type III restriction endonucleases. MDA5 is a cytosolic RNA sensor: every characterised ligand is RNA, the C-terminal regulatory domain and the helicase core are structurally resolved in complex with dsRNA, and no study reports direct DNA binding by MDA5. MDA5 is activated by some DNA viruses, but the accepted mechanism is indirect, via RNA produced during those infections. Removing an electronic inference that can be argued against on biological grounds; no experimental annotation is being overruled.
GO:0003724 RNA helicase activity
IEA
GO_REF:0000003
MODIFY
Summary: RNA helicase activity, defined as ATP-driven unwinding of an RNA helix. MDA5 has a bona fide DExD/H ATPase that is stimulated by duplex RNA, but processive duplex unwinding is not what the ATPase does.
Reason: The cited IMP (PMID:19211564) mutated the conserved helicase motifs and measured ATP hydrolysis and signalling, not duplex unwinding, and concluded that interferon induction is independent of helicase catalytic activity. In MDA5 the ATPase powers filament dynamics and length-dependent proofreading. The ATP-hydrolysis term the gene already carries (GO:0016887) states what was measured; 'ATP-dependent activity, acting on RNA' (GO:0008186) captures the RNA-dependence without asserting unwinding.
Supporting Evidence:
PMID:19211564
type I interferon production mediated by full-length MDA5 and RIG-I is independent of the helicase domain catalytic activity
PMID:19211564
Results indicate that all targeted motifs are required for ATP hydrolysis, but a subset is involved in RNA binding.
GO:0005524 ATP binding
IEA
GO_REF:0000002
ACCEPT
Summary: ATP binding by the P-loop of the helicase core.
Reason: Correct and necessary for the ATPase cycle, though subsidiary to the hydrolysis term.
GO:0005634 nucleus
IEA
GO_REF:0000044
MARK AS OVER ANNOTATED
Summary: Nuclear localisation transferred electronically from the UniProt subcellular-location vocabulary.
Reason: UniProt's own nucleus assignment for IFIH1 carries ECO:0000305 (curator inference) and is qualified as 'May be found in the nucleus, during apoptosis'. Propagating that into a plain located_in nucleus annotation overstates it for a protein whose entire characterised biology is cytosolic.
GO:0005737 cytoplasm
IEA
GO_REF:0000120
ACCEPT
Summary: MDA5 is a cytoplasmic protein.
Reason: Correct and consistently observed; this is where the sensor resides and acts.
GO:0005739 mitochondrion
IEA
GO_REF:0000044
KEEP AS NON CORE
Summary: Mitochondrial localisation observed after activation.
Reason: Real but conditional: MDA5 is not a resident mitochondrial protein. Upon ligand binding and ISGylation it translocates to the mitochondrial outer membrane to engage MAVS. Meaningful for signalling, but not a primary location.
Supporting Evidence:
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized by the SARS-CoV-2 papain-like protease to evade host innate immunity.
GO:0007165 signal transduction
IEA
GO_REF:0000117
MODIFY
Summary: Bare signal transduction from an ARBA machine-learning model.
Reason: Correct but far too general; the specific pathway this protein initiates is already annotated.
Proposed replacements: MDA-5 signaling pathway
GO:0016787 hydrolase activity
IEA
GO_REF:0000002
MODIFY
Summary: Bare hydrolase activity carried over from the InterPro signature IPR006935 (Helicase/UvrB, N-terminal).
Reason: Correct in kind but uninformative at this level. The hydrolase activity of MDA5 is ATP hydrolysis, already annotated with experimental evidence.
Proposed replacements: ATP hydrolysis activity
GO:0016887 ATP hydrolysis activity
IEA
GO_REF:0000116
ACCEPT
Summary: Duplex-RNA-stimulated ATP hydrolysis by the DExD/H helicase core.
Reason: Core molecular function. Directly measured, and mechanistically central: ATP turnover drives the assembly/disassembly cycle that discriminates long duplexes from short ones.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
PMID:19211564
Results indicate that all targeted motifs are required for ATP hydrolysis, but a subset is involved in RNA binding.
GO:0051259 protein complex oligomerization
IEA
GO_REF:0000117
ACCEPT
Summary: Protein complex oligomerization: MDA5 filament assembly and CARD oligomerisation.
Reason: Directly observed and mechanistically central - the oligomerisation step is how duplex length is converted into a signalling threshold.
GO:0140374 antiviral innate immune response
IEA
GO_REF:0000117
ACCEPT
Summary: MDA5 drives the antiviral innate immune response to picornaviruses, coronaviruses and other agents.
Reason: Core biological process, and unaffected by the ligand question: whether MDA5 is triggered by viral RNA or by infection-induced aberrant host RNA, the process it participates in is antiviral innate immunity.
GO:0005515 protein binding
IPI
PMID:17460044
Negative regulation of the RIG-I signaling by the ubiquitin ...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:17600090
Negative regulation of MDA5- but not RIG-I-mediated innate a...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:19403670
A shared interface mediates paramyxovirus interference with ...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:19416887
ISG56 is a negative-feedback regulator of virus-triggered si...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:21903422
Mapping a dynamic innate immunity protein interaction networ...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:22301138
Herpes simplex virus 1 tegument protein US11 downmodulates t...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:22623778
Human respiratory syncytial virus nucleoprotein and inclusio...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:23328395
Paramyxovirus V proteins disrupt the fold of the RNA sensor ...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:25865883
RIOK3-mediated phosphorylation of MDA5 interferes with its a...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:28591694
The nucleocapsid proteins of mouse hepatitis virus and sever...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:31015422
Physical and functional interaction between A20 and ATG16L1-...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:32296183
A reference map of the human binary protein interactome.
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:33372174
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:34084167
SARS-CoV-2 Membrane Protein Inhibits Type I Interferon Produ...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0042802 identical protein binding
IPI
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filamen...
ACCEPT
Summary: MDA5 self-association. Filament formation along duplex RNA and CARD-CARD oligomerisation are the mechanism by which ligand binding is converted into a signal.
Reason: Unlike bare protein binding, homotypic self-association is mechanistically load-bearing for this protein and is directly observed.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
GO:0042802 identical protein binding
IPI
PMID:25865883
RIOK3-mediated phosphorylation of MDA5 interferes with its a...
ACCEPT
Summary: MDA5 self-association. Filament formation along duplex RNA and CARD-CARD oligomerisation are the mechanism by which ligand binding is converted into a signal.
Reason: Unlike bare protein binding, homotypic self-association is mechanistically load-bearing for this protein and is directly observed.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
GO:0009615 response to virus
IEA
GO_REF:0000107
ACCEPT
Summary: Response to virus.
Reason: Correct but general; the gene carries more specific terms (GO:0140374, GO:0051607, GO:0039530) that say the same thing better.
GO:0045087 innate immune response
IEA
GO_REF:0000107
ACCEPT
Summary: Innate immune response.
Reason: Correct high-level process term.
GO:0140374 antiviral innate immune response
IDA
PMID:40184173
MDA5 ISGylation is crucial for immune signaling to control v...
ACCEPT
Summary: MDA5 drives the antiviral innate immune response to picornaviruses, coronaviruses and other agents.
Reason: Core biological process, and unaffected by the ligand question: whether MDA5 is triggered by viral RNA or by infection-induced aberrant host RNA, the process it participates in is antiviral innate immunity.
GO:0005737 cytoplasm
EXP
PMID:32169843
NOD1 Promotes Antiviral Signaling by Binding Viral RNA and R...
ACCEPT
Summary: MDA5 is a cytoplasmic protein.
Reason: Correct and consistently observed; this is where the sensor resides and acts.
GO:0016887 ATP hydrolysis activity
EXP
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filamen...
ACCEPT
Summary: Duplex-RNA-stimulated ATP hydrolysis by the DExD/H helicase core.
Reason: Core molecular function. Directly measured, and mechanistically central: ATP turnover drives the assembly/disassembly cycle that discriminates long duplexes from short ones.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
PMID:19211564
Results indicate that all targeted motifs are required for ATP hydrolysis, but a subset is involved in RNA binding.
GO:0005737 cytoplasm
IDA
PMID:11805321
mda-5: An interferon-inducible putative RNA helicase with do...
ACCEPT
Summary: MDA5 is a cytoplasmic protein.
Reason: Correct and consistently observed; this is where the sensor resides and acts.
GO:0038187 pattern recognition receptor activity
IDA
PMID:23273991
Structural basis for dsRNA recognition, filament formation, ...
ACCEPT
Summary: Pattern recognition receptor activity: MDA5 binds an RNA ligand and, through its tandem CARDs, initiates MAVS-dependent innate immune signalling. The GO definition requires the ligand to be a pathogen-associated molecular pattern, and that requirement is exactly what the 2026 iCLIP data call into question.
Reason: Retained, but this is the annotation on which the contested question turns, and both sides are recorded here rather than resolved. FOR the term: GOA's own supporting experiments are real - Zuest et al. 2011 (PMID:21217758) show MDA5-dependent interferon induction by viral mRNA lacking the host 2'-O-methyl cap mark (self/non-self discrimination), and Wu et al. 2013 (PMID:23273991) show recognition of the internal duplex of long dsRNA, a canonical microbial pattern. AGAINST the term as defined: Sampaio et al. 2026 (PMID:42581186) applied iCLIP to endogenous MDA5 in SARS-CoV-2- and EMCV-infected cells and recovered essentially no viral RNA: the bound population was host RNA, enriched in introns, Alu-proximal regions and predicted duplexes, and MDA5 activation was abolished by SRSF3 overexpression, which restores splicing fidelity. The authors frame MDA5 as surveying RNA processing fidelity ('guarding') rather than detecting a pathogen-encoded pattern, and state plainly that they cannot exclude direct viral dsRNA detection in other settings. This reframes which RNA activates MDA5; it does not challenge the biochemistry of duplex-RNA binding. The same paper calls MDA5 one of the few remaining 'orphan' pattern recognition receptors because no PAMP has been pinned down. Independently, IFIH1 gain-of-function variants cause Aicardi-Goutieres syndrome and related interferonopathies through excessive signalling in the absence of infection (PMID:24686847), and ADAR1 editing exists to stop MDA5 sensing endogenous duplex RNA (PMID:26275108) - both are evidence that the physiological ligand pool includes self RNA distinguished by processing state rather than by pathogen origin. ACCEPT rather than MODIFY or REMOVE because (i) these are IDA annotations whose full experimental basis stands, and (ii) the mismatch is with the wording of GO:0038187, not with this gene: no existing MF term describes a receptor that senses aberrantly processed cellular RNA. A replacement term is proposed under proposed_new_terms and the question is raised in suggested_questions rather than being settled by an edit here.
Supporting Evidence:
PMID:21217758
the induction of type I interferon by viruses deficient in 2'-O-methyltransferase was dependent on the cytoplasmic RNA sensor Mda5
PMID:42581186
We did not detect MDA5 binding sites in viral RNA; instead, MDA5 bound only host RNA.
PMID:42581186
MDA5's RNA agonists are not well defined.
PMID:24686847
these mutations confer gain of function such that mutant IFIH1 binds RNA more avidly, leading to increased baseline and ligand-induced interferon signaling
GO:0005737 cytoplasm
IDA
PMID:28031478
TRIM65-catalized ubiquitination is essential for MDA5-mediat...
ACCEPT
Summary: MDA5 is a cytoplasmic protein.
Reason: Correct and consistently observed; this is where the sensor resides and acts.
GO:0019904 protein domain specific binding
IPI
PMID:28031478
TRIM65-catalized ubiquitination is essential for MDA5-mediat...
MARK AS OVER ANNOTATED
Summary: Protein domain specific binding from a two-hybrid/interaction study.
Reason: As uninformative as bare protein binding; it names no partner and no function.
GO:0051259 protein complex oligomerization
IDA
PMID:28031478
TRIM65-catalized ubiquitination is essential for MDA5-mediat...
ACCEPT
Summary: Protein complex oligomerization: MDA5 filament assembly and CARD oligomerisation.
Reason: Directly observed and mechanistically central - the oligomerisation step is how duplex length is converted into a signalling threshold.
GO:0009615 response to virus
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
ACCEPT
Summary: Response to virus.
Reason: Correct but general; the gene carries more specific terms (GO:0140374, GO:0051607, GO:0039530) that say the same thing better.
GO:0003724 RNA helicase activity
IMP
PMID:19211564
Regulation of signal transduction by enzymatically inactive ...
MODIFY
Summary: RNA helicase activity, defined as ATP-driven unwinding of an RNA helix. MDA5 has a bona fide DExD/H ATPase that is stimulated by duplex RNA, but processive duplex unwinding is not what the ATPase does.
Reason: The cited IMP (PMID:19211564) mutated the conserved helicase motifs and measured ATP hydrolysis and signalling, not duplex unwinding, and concluded that interferon induction is independent of helicase catalytic activity. In MDA5 the ATPase powers filament dynamics and length-dependent proofreading. The ATP-hydrolysis term the gene already carries (GO:0016887) states what was measured; 'ATP-dependent activity, acting on RNA' (GO:0008186) captures the RNA-dependence without asserting unwinding.
Supporting Evidence:
PMID:19211564
type I interferon production mediated by full-length MDA5 and RIG-I is independent of the helicase domain catalytic activity
PMID:19211564
Results indicate that all targeted motifs are required for ATP hydrolysis, but a subset is involved in RNA binding.
GO:0016887 ATP hydrolysis activity
IMP
PMID:19211564
Regulation of signal transduction by enzymatically inactive ...
ACCEPT
Summary: Duplex-RNA-stimulated ATP hydrolysis by the DExD/H helicase core.
Reason: Core molecular function. Directly measured, and mechanistically central: ATP turnover drives the assembly/disassembly cycle that discriminates long duplexes from short ones.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
PMID:19211564
Results indicate that all targeted motifs are required for ATP hydrolysis, but a subset is involved in RNA binding.
GO:0039530 MDA-5 signaling pathway
IMP
PMID:19211564
Regulation of signal transduction by enzymatically inactive ...
ACCEPT
Summary: The MDA-5 signalling pathway itself: ligand-bound MDA5 filaments nucleate CARD oligomers that activate MAVS, TBK1/IKBKE and IRF3/IRF7.
Reason: Core biological process, named for this gene and supported by independent experimental work.
GO:0003723 RNA binding
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
MODIFY
Summary: Bare RNA binding.
Reason: Correct but too general for a protein whose RNA-binding specificity is the whole point. The duplex-binding term is the informative one and the gene already carries it with independent experimental support.
Proposed replacements: double-stranded RNA binding
GO:0005737 cytoplasm
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
ACCEPT
Summary: MDA5 is a cytoplasmic protein.
Reason: Correct and consistently observed; this is where the sensor resides and acts.
GO:0005739 mitochondrion
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
KEEP AS NON CORE
Summary: Mitochondrial localisation observed after activation.
Reason: Real but conditional: MDA5 is not a resident mitochondrial protein. Upon ligand binding and ISGylation it translocates to the mitochondrial outer membrane to engage MAVS. Meaningful for signalling, but not a primary location.
Supporting Evidence:
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized by the SARS-CoV-2 papain-like protease to evade host innate immunity.
GO:0032728 positive regulation of interferon-beta production
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
ACCEPT
Summary: Positive regulation of interferon-beta production.
Reason: Core biological output of the pathway, with multiple independent experimental annotations.
GO:0039530 MDA-5 signaling pathway
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
ACCEPT
Summary: The MDA-5 signalling pathway itself: ligand-bound MDA5 filaments nucleate CARD oligomers that activate MAVS, TBK1/IKBKE and IRF3/IRF7.
Reason: Core biological process, named for this gene and supported by independent experimental work.
GO:0042802 identical protein binding
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
ACCEPT
Summary: MDA5 self-association. Filament formation along duplex RNA and CARD-CARD oligomerisation are the mechanism by which ligand binding is converted into a signal.
Reason: Unlike bare protein binding, homotypic self-association is mechanistically load-bearing for this protein and is directly observed.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
GO:0045071 negative regulation of viral genome replication
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
KEEP AS NON CORE
Summary: Negative regulation of viral genome replication.
Reason: A true consequence of MDA5-driven interferon induction, but effected by interferon-stimulated genes rather than by MDA5 itself.
GO:0051607 defense response to virus
IDA
PMID:33727702
ISG15-dependent activation of the sensor MDA5 is antagonized...
ACCEPT
Summary: Defense response to virus.
Reason: Correct and experimentally supported.
GO:0098586 cellular response to virus
IEP
PMID:28146100
Echovirus 6 Infects Human Exocrine and Endocrine Pancreatic ...
ACCEPT
Summary: Cellular response to virus, from expression evidence.
Reason: Correct; IFIH1 is itself an interferon-stimulated gene and is induced by infection.
GO:0042802 identical protein binding
IDA
PMID:23090998
MDA5 assembles into a polar helical filament on dsRNA.
ACCEPT
Summary: MDA5 self-association. Filament formation along duplex RNA and CARD-CARD oligomerisation are the mechanism by which ligand binding is converted into a signal.
Reason: Unlike bare protein binding, homotypic self-association is mechanistically load-bearing for this protein and is directly observed.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
GO:0042802 identical protein binding
IDA
PMID:30449722
Cryo-EM Structures of MDA5-dsRNA Filaments at Different Stag...
ACCEPT
Summary: MDA5 self-association. Filament formation along duplex RNA and CARD-CARD oligomerisation are the mechanism by which ligand binding is converted into a signal.
Reason: Unlike bare protein binding, homotypic self-association is mechanistically load-bearing for this protein and is directly observed.
Supporting Evidence:
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
GO:0038187 pattern recognition receptor activity
IDA
PMID:21217758
Ribose 2'-O-methylation provides a molecular signature for t...
ACCEPT
Summary: Pattern recognition receptor activity: MDA5 binds an RNA ligand and, through its tandem CARDs, initiates MAVS-dependent innate immune signalling. The GO definition requires the ligand to be a pathogen-associated molecular pattern, and that requirement is exactly what the 2026 iCLIP data call into question.
Reason: Retained, but this is the annotation on which the contested question turns, and both sides are recorded here rather than resolved. FOR the term: GOA's own supporting experiments are real - Zuest et al. 2011 (PMID:21217758) show MDA5-dependent interferon induction by viral mRNA lacking the host 2'-O-methyl cap mark (self/non-self discrimination), and Wu et al. 2013 (PMID:23273991) show recognition of the internal duplex of long dsRNA, a canonical microbial pattern. AGAINST the term as defined: Sampaio et al. 2026 (PMID:42581186) applied iCLIP to endogenous MDA5 in SARS-CoV-2- and EMCV-infected cells and recovered essentially no viral RNA: the bound population was host RNA, enriched in introns, Alu-proximal regions and predicted duplexes, and MDA5 activation was abolished by SRSF3 overexpression, which restores splicing fidelity. The authors frame MDA5 as surveying RNA processing fidelity ('guarding') rather than detecting a pathogen-encoded pattern, and state plainly that they cannot exclude direct viral dsRNA detection in other settings. This reframes which RNA activates MDA5; it does not challenge the biochemistry of duplex-RNA binding. The same paper calls MDA5 one of the few remaining 'orphan' pattern recognition receptors because no PAMP has been pinned down. Independently, IFIH1 gain-of-function variants cause Aicardi-Goutieres syndrome and related interferonopathies through excessive signalling in the absence of infection (PMID:24686847), and ADAR1 editing exists to stop MDA5 sensing endogenous duplex RNA (PMID:26275108) - both are evidence that the physiological ligand pool includes self RNA distinguished by processing state rather than by pathogen origin. ACCEPT rather than MODIFY or REMOVE because (i) these are IDA annotations whose full experimental basis stands, and (ii) the mismatch is with the wording of GO:0038187, not with this gene: no existing MF term describes a receptor that senses aberrantly processed cellular RNA. A replacement term is proposed under proposed_new_terms and the question is raised in suggested_questions rather than being settled by an edit here.
Supporting Evidence:
PMID:21217758
the induction of type I interferon by viruses deficient in 2'-O-methyltransferase was dependent on the cytoplasmic RNA sensor Mda5
PMID:42581186
We did not detect MDA5 binding sites in viral RNA; instead, MDA5 bound only host RNA.
PMID:42581186
MDA5's RNA agonists are not well defined.
PMID:24686847
these mutations confer gain of function such that mutant IFIH1 binds RNA more avidly, leading to increased baseline and ligand-induced interferon signaling
GO:0140374 antiviral innate immune response
IDA
PMID:15563593
The V proteins of paramyxoviruses bind the IFN-inducible RNA...
ACCEPT
Summary: MDA5 drives the antiviral innate immune response to picornaviruses, coronaviruses and other agents.
Reason: Core biological process, and unaffected by the ligand question: whether MDA5 is triggered by viral RNA or by infection-induced aberrant host RNA, the process it participates in is antiviral innate immunity.
GO:0005515 protein binding
IPI
PMID:30193849
The Zinc-Finger Protein ZCCHC3 Binds RNA and Facilitates Vir...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0051607 defense response to virus
IDA
PMID:21478870
A diverse range of gene products are effectors of the type I...
ACCEPT
Summary: Defense response to virus.
Reason: Correct and experimentally supported.
GO:0032727 positive regulation of interferon-alpha production
IMP
PMID:21957149
DHX9 pairs with IPS-1 to sense double-stranded RNA in myeloi...
ACCEPT
Summary: Positive regulation of interferon-alpha production.
Reason: Direct downstream output of MDA5 signalling, supported by loss-of-function experiments.
GO:0032728 positive regulation of interferon-beta production
IMP
PMID:21957149
DHX9 pairs with IPS-1 to sense double-stranded RNA in myeloi...
ACCEPT
Summary: Positive regulation of interferon-beta production.
Reason: Core biological output of the pathway, with multiple independent experimental annotations.
GO:0032755 positive regulation of interleukin-6 production
IMP
PMID:21957149
DHX9 pairs with IPS-1 to sense double-stranded RNA in myeloi...
KEEP AS NON CORE
Summary: Positive regulation of interleukin-6 production.
Reason: A downstream inflammatory consequence of MDA5-MAVS signalling rather than a distinct activity of the protein.
GO:0032760 positive regulation of tumor necrosis factor production
IMP
PMID:21957149
DHX9 pairs with IPS-1 to sense double-stranded RNA in myeloi...
KEEP AS NON CORE
Summary: Positive regulation of tumor necrosis factor production.
Reason: A downstream inflammatory consequence of MDA5-MAVS signalling rather than a distinct activity of the protein.
GO:0060760 positive regulation of response to cytokine stimulus
IMP
PMID:21957149
DHX9 pairs with IPS-1 to sense double-stranded RNA in myeloi...
KEEP AS NON CORE
Summary: Positive regulation of response to cytokine stimulus.
Reason: Distal and non-specific; a consequence of the interferon amplification loop.
GO:0071360 cellular response to exogenous dsRNA
IMP
PMID:21957149
DHX9 pairs with IPS-1 to sense double-stranded RNA in myeloi...
ACCEPT
Summary: Cellular response to exogenous double-stranded RNA.
Reason: Correct and directly tested: transfected dsRNA activates MDA5-dependent signalling.
GO:0005829 cytosol
TAS
Reactome:R-HSA-5696600
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0032728 positive regulation of interferon-beta production
IDA
PMID:17600090
Negative regulation of MDA5- but not RIG-I-mediated innate a...
ACCEPT
Summary: Positive regulation of interferon-beta production.
Reason: Core biological output of the pathway, with multiple independent experimental annotations.
GO:0039530 MDA-5 signaling pathway
IDA
PMID:17600090
Negative regulation of MDA5- but not RIG-I-mediated innate a...
ACCEPT
Summary: The MDA-5 signalling pathway itself: ligand-bound MDA5 filaments nucleate CARD oligomers that activate MAVS, TBK1/IKBKE and IRF3/IRF7.
Reason: Core biological process, named for this gene and supported by independent experimental work.
GO:0005515 protein binding
IPI
PMID:22328336
Ankrd17 positively regulates RIG-I-like receptor (RLR)-media...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:22908223
Tetraspanin 6 (TSPAN6) negatively regulates retinoic acid-in...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005829 cytosol
TAS
Reactome:R-HSA-168909
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-168934
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-913725
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-918225
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-918227
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-918229
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-918230
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-918232
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933523
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933525
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933526
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933527
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933530
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933532
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933537
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933538
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-933539
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-936381
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-936475
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-937343
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9705137
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9705145
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9705320
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9705323
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9754745
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9755244
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9833393
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-990526
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-990528
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9920772
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-HSA-9924831
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005829 cytosol
TAS
Reactome:R-NUL-936401
ACCEPT
Summary: Cytosolic localisation, asserted by Reactome pathway curation.
Reason: Correct and more specific than bare cytoplasm.
GO:0005515 protein binding
IPI
PMID:21791617
DDX60, a DEXD/H box helicase, is a novel antiviral factor pr...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0002753 cytoplasmic pattern recognition receptor signaling pathway
TAS
PMID:21616437
Immune signaling by RIG-I-like receptors.
ACCEPT
Summary: Cytoplasmic pattern recognition receptor signalling pathway, the parent of the MDA-5 pathway.
Reason: Correct as the general pathway class. Its definition inherits the same PAMP wording as GO:0038187 and inherits the same caveat, recorded there; the specific child term GO:0039530 is the informative annotation.
GO:0003725 double-stranded RNA binding
IDA
PMID:19656871
Activation of MDA5 requires higher-order RNA structures gene...
ACCEPT
Summary: MDA5 binds long double-stranded RNA and cooperatively assembles head-to-tail filaments along the duplex; this is its core ligand-binding activity.
Reason: Core molecular function, established structurally and biochemically and not in dispute. What is disputed is the origin of the duplex RNA that engages MDA5 in infected cells, not the binding activity itself. Sampaio et al. 2026 (PMID:42581186) applied iCLIP to endogenous MDA5 in SARS-CoV-2- and EMCV-infected cells and recovered essentially no viral RNA: the bound population was host RNA, enriched in introns, Alu-proximal regions and predicted duplexes, and MDA5 activation was abolished by SRSF3 overexpression, which restores splicing fidelity. The authors frame MDA5 as surveying RNA processing fidelity ('guarding') rather than detecting a pathogen-encoded pattern, and state plainly that they cannot exclude direct viral dsRNA detection in other settings. This reframes which RNA activates MDA5; it does not challenge the biochemistry of duplex-RNA binding. The host ligands it identifies - inverted Alu repeats and structured intronic RNA - are themselves double-stranded, so this term remains correct under either model.
Supporting Evidence:
PMID:23273991
MDA5 recognizes the internal duplex structure, whereas RIG-I recognizes the terminus of dsRNA
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
PMID:42581186
upon infection with SARS-CoV-2 or encephalomyocarditis virus, MDA5 bound overwhelmingly to cellular RNAs
GO:0003727 single-stranded RNA binding
IDA
PMID:19656871
Activation of MDA5 requires higher-order RNA structures gene...
KEEP AS NON CORE
Summary: Single-stranded RNA binding reported alongside the duplex-binding activity.
Reason: Real but peripheral. The signalling-competent ligand for MDA5 is a long duplex bound by a cooperative filament; single-stranded segments are bound within higher-order structures rather than as the activating species. Retained as a genuine binding property, demoted from core.
Supporting Evidence:
PMID:19656871
Activation of MDA5 requires higher-order RNA structures generated during virus infection.
GO:0005515 protein binding
IPI
PMID:21156324
MDA5 is SUMOylated by PIAS2Ξ² in the upregulation of type I i...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0009615 response to virus
TAS
PMID:21616437
Immune signaling by RIG-I-like receptors.
ACCEPT
Summary: Response to virus.
Reason: Correct but general; the gene carries more specific terms (GO:0140374, GO:0051607, GO:0039530) that say the same thing better.
GO:0016925 protein sumoylation
IDA
PMID:21156324
MDA5 is SUMOylated by PIAS2Ξ² in the upregulation of type I i...
MARK AS OVER ANNOTATED
Summary: Protein sumoylation, annotated involved_in.
Reason: MDA5 is the substrate of SUMO conjugation, not a participant in the sumoylation machinery. An involved_in annotation on a modified substrate misstates the direction of the relationship.
GO:0032727 positive regulation of interferon-alpha production
IMP
PMID:19656871
Activation of MDA5 requires higher-order RNA structures gene...
ACCEPT
Summary: Positive regulation of interferon-alpha production.
Reason: Direct downstream output of MDA5 signalling, supported by loss-of-function experiments.
GO:0032728 positive regulation of interferon-beta production
IMP
PMID:19656871
Activation of MDA5 requires higher-order RNA structures gene...
ACCEPT
Summary: Positive regulation of interferon-beta production.
Reason: Core biological output of the pathway, with multiple independent experimental annotations.
GO:0034344 regulation of type III interferon production
TAS
PMID:21616437
Immune signaling by RIG-I-like receptors.
KEEP AS NON CORE
Summary: Regulation of type III interferon production.
Reason: Real but secondary to the type I interferon output that defines the pathway.
GO:0045087 innate immune response
TAS
PMID:21616437
Immune signaling by RIG-I-like receptors.
ACCEPT
Summary: Innate immune response.
Reason: Correct high-level process term.
GO:0003725 double-stranded RNA binding
TAS
PMID:17079289
Inhibition of retinoic acid-inducible gene I-mediated induct...
ACCEPT
Summary: MDA5 binds long double-stranded RNA and cooperatively assembles head-to-tail filaments along the duplex; this is its core ligand-binding activity.
Reason: Core molecular function, established structurally and biochemically and not in dispute. What is disputed is the origin of the duplex RNA that engages MDA5 in infected cells, not the binding activity itself. Sampaio et al. 2026 (PMID:42581186) applied iCLIP to endogenous MDA5 in SARS-CoV-2- and EMCV-infected cells and recovered essentially no viral RNA: the bound population was host RNA, enriched in introns, Alu-proximal regions and predicted duplexes, and MDA5 activation was abolished by SRSF3 overexpression, which restores splicing fidelity. The authors frame MDA5 as surveying RNA processing fidelity ('guarding') rather than detecting a pathogen-encoded pattern, and state plainly that they cannot exclude direct viral dsRNA detection in other settings. This reframes which RNA activates MDA5; it does not challenge the biochemistry of duplex-RNA binding. The host ligands it identifies - inverted Alu repeats and structured intronic RNA - are themselves double-stranded, so this term remains correct under either model.
Supporting Evidence:
PMID:23273991
MDA5 recognizes the internal duplex structure, whereas RIG-I recognizes the terminus of dsRNA
PMID:22160685
Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
PMID:42581186
upon infection with SARS-CoV-2 or encephalomyocarditis virus, MDA5 bound overwhelmingly to cellular RNAs
GO:0009597 detection of virus
TAS
PMID:17079289
Inhibition of retinoic acid-inducible gene I-mediated induct...
KEEP AS NON CORE
Summary: Detection of virus, from a review article.
Reason: Retained as a fair description of the physiological outcome, but this is the BP term most exposed to the ligand question: the 2026 iCLIP data indicate that in infected cells MDA5 occupies host rather than viral RNA, so what is detected may be the infection's effect on RNA processing rather than the virus itself. Demoted from core; GO:0140374 and GO:0039530 carry the same biology without the mechanistic commitment.
Supporting Evidence:
PMID:42581186
we propose that MDA5 surveys RNA processing fidelity and can detect infections by sensing perturbations of post-transcriptional events such as splicing.
GO:0032727 positive regulation of interferon-alpha production
TAS
PMID:17079289
Inhibition of retinoic acid-inducible gene I-mediated induct...
ACCEPT
Summary: Positive regulation of interferon-alpha production.
Reason: Direct downstream output of MDA5 signalling, supported by loss-of-function experiments.
GO:0032728 positive regulation of interferon-beta production
TAS
PMID:17079289
Inhibition of retinoic acid-inducible gene I-mediated induct...
ACCEPT
Summary: Positive regulation of interferon-beta production.
Reason: Core biological output of the pathway, with multiple independent experimental annotations.
GO:0005515 protein binding
IPI
PMID:16127453
IPS-1, an adaptor triggering RIG-I- and Mda5-mediated type I...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0005515 protein binding
IPI
PMID:20434986
NLRC5 negatively regulates the NF-kappaB and type I interfer...
REMOVE
Summary: Bare protein binding from interaction screens.
Reason: Carries no functional information. Per project guidance, a specific molecular function term should be used instead where one is available. No more informative molecular function is supportable from this evidence, so the row is removed rather than rewritten. Removal does not imply the reported interaction is false.
GO:0043021 ribonucleoprotein complex binding
IPI
PMID:19881509
PCBP2 mediates degradation of the adaptor MAVS via the HECT ...
KEEP AS NON CORE
Summary: Binding to a ribonucleoprotein complex.
Reason: Plausible and consistent with MDA5 engaging RNA that is already protein-bound - a point the 2026 iCLIP work makes explicitly, proposing that host RNA-binding proteins relocalised to the cytoplasm may deliver unspliced RNA to MDA5 - but peripheral to the core function.
GO:0008270 zinc ion binding
IDA
PMID:19380577
Solution structures of cytosolic RNA sensor MDA5 and LGP2 C-...
ACCEPT
Summary: Zinc coordinated by the C-terminal regulatory domain.
Reason: A structural metal site, directly observed, required to fold the domain that caps the RNA duplex. Correct, though architectural rather than a function in its own right.
GO:0008270 zinc ion binding
IDA
PMID:19531363
Structural basis of double-stranded RNA recognition by the R...
ACCEPT
Summary: Zinc coordinated by the C-terminal regulatory domain.
Reason: A structural metal site, directly observed, required to fold the domain that caps the RNA duplex. Correct, though architectural rather than a function in its own right.

Core Functions

MDA5 binds the internal stem of long double-stranded RNA and polymerises head-to-tail along it. Unlike RIG-I it reads duplex length rather than a terminal chemical mark, and it is the RNA-binding event that licenses everything downstream. The duplexes it engages in cells are not necessarily viral: iCLIP of endogenous MDA5 during SARS-CoV-2 and encephalomyocarditis virus infection recovers host transcripts, concentrated in introns and near inverted Alu repeats, which are themselves capable of base-pairing.

Supporting Evidence:
  • PMID:23273991
    MDA5 recognizes the internal duplex structure, whereas RIG-I recognizes the terminus of dsRNA
  • PMID:42581186
    upon infection with SARS-CoV-2 or encephalomyocarditis virus, MDA5 bound overwhelmingly to cellular RNAs

Duplex-RNA-stimulated ATP hydrolysis by the DExD/H helicase core. The ATPase does not act as a processive duplex unwindase; it drives the assembly and disassembly of MDA5 filaments, and this turnover is the proofreading step that makes signalling depend on duplex length. Helicase catalytic activity is dispensable for interferon induction by full-length MDA5, so the ATPase tunes rather than transmits the signal.

Molecular Function:
ATP hydrolysis activity
Directly Involved In:
Cellular Locations:
Supporting Evidence:
  • PMID:22160685
    Cooperative assembly and dynamic disassembly of MDA5 filaments for viral dsRNA recognition.
  • PMID:19211564
    type I interferon production mediated by full-length MDA5 and RIG-I is independent of the helicase domain catalytic activity

Ligand-activated innate immune receptor activity: ligand binding occurs in the cytosol, where filament assembly clusters the tandem CARDs; the clustered CARDs then nucleate MAVS polymers at the mitochondrial outer membrane and drive TBK1/IKBKE and IRF3/IRF7 activation, inducing interferon-alpha/beta. This is MDA5's defining role, and is recorded here with an explicit caveat: GO:0038187 is defined as combining with a pathogen-associated molecular pattern, and whether MDA5's activating ligand is pathogen-derived is unsettled. Viral mRNA lacking 2'-O-methylation and long viral duplexes activate it, but in infected cells the bound RNA is predominantly host, gain-of-function variants cause sterile interferonopathy, and ADAR1 editing exists to keep endogenous duplexes from triggering it.

Supporting Evidence:
  • PMID:21217758
    the induction of type I interferon by viruses deficient in 2'-O-methyltransferase was dependent on the cytoplasmic RNA sensor Mda5
  • PMID:42581186
    MDA5's RNA agonists are not well defined.
  • PMID:24686847
    heterozygous mutations in the cytosolic double-stranded RNA receptor gene IFIH1 (also called MDA5) cause a spectrum of neuroimmunological features consistently associated with an enhanced interferon state

References

Loading supporting content…

Download this section (compressed HTML)

Suggested Questions for Experts

Q: Should IFIH1 keep GO:0038187 pattern recognition receptor activity, whose definition requires the ligand to be a pathogen-associated molecular pattern, when the RNA occupied by endogenous MDA5 during SARS-CoV-2 and EMCV infection is host-derived, and when the activating signal can be removed simply by restoring splicing fidelity with SRSF3? Or does GO need a molecular function term for a receptor that reads the processing state of self RNA?

Suggested experts: GO innate immunity curators, Jan Rehwinkel, Sun Hur

Q: Are the viral-ligand results and the host-RNA results in conflict, or are they measuring different things? Purified MDA5 demonstrably binds long viral duplexes and unmethylated viral mRNA in vitro, while iCLIP measures which RNAs MDA5 is crosslinked to in an infected cell, where viral RNA may be sequestered in replication organelles and coated by viral proteins. A direct test would be whether MDA5 activation survives in cells where host splicing is protected but viral replication proceeds normally.

Suggested experts: Jan Rehwinkel, Sun Hur, Michaela Gack

Q: If MDA5 is a guard rather than a PAMP receptor, is the gain-of-function interferonopathy phenotype better described as loss of a threshold on normal self-RNA surveillance than as inappropriate self recognition? That distinction matters for how the Aicardi-Goutieres and Singleton-Merten alleles should be annotated with respect to this gene's normal function.

Suggested experts: Yanick Crow, Sun Hur

Q: Does MDA5 have any physiologically relevant duplex-unwinding activity? GO:0003724 RNA helicase activity asserts ATP-driven unwinding of an RNA helix, but the IMP behind it measured ATP hydrolysis and signalling, and concluded that helicase catalytic activity is dispensable for interferon induction.

Suggested experts: Sun Hur, Anna Marie Pyle

Suggested Experiments

Experiment: Uncouple viral replication from host RNA-processing failure. In MDA5-competent cells, infect with EMCV or SARS-CoV-2 while restoring splicing fidelity (SRSF3 overexpression, or complementation of the virus-encoded splicing antagonist) and measure IFNB1 induction, CARD oligomerisation and MDA5 filament formation. Conversely, induce intron retention pharmacologically in the absence of infection and ask whether MDA5 is activated to the same extent. A clean dissociation in either direction would settle whether the activating ligand is viral.

Hypothesis: MDA5 activation during infection requires accumulation of unspliced or aberrantly processed host transcripts in the cytoplasm, and is not triggered by viral RNA per se.

Type: infection plus splicing-modulation, with interferon and MDA5 oligomerisation readouts

Experiment: Apply the published endogenous-MDA5 iCLIP protocol to uninfected patient-derived cells carrying AGS/Singleton-Merten IFIH1 alleles and to ADAR1-deficient cells, and compare the bound RNA populations with those recovered from infected wild-type cells. Overlap in intronic and inverted-Alu binding sites would argue that infection and interferonopathy converge on one ligand class, which would in turn argue that the molecular function term should not presuppose pathogen origin.

Hypothesis: The RNAs bound by MDA5 in gain-of-function interferonopathy patients are the same aberrantly processed host transcripts identified during infection.

Type: iCLIP on patient-derived and ADAR1-deficient cells

Experiment: Run side-by-side single-molecule unwinding assays on purified full-length MDA5 and on a bona fide processive RNA helicase using identical duplex substrates and ATP concentrations, and compare with filament assembly/disassembly kinetics on the same substrates. A clear absence of unwinding would justify replacing GO:0003724 with a term that asserts ATP-dependent activity on RNA without asserting unwinding.

Hypothesis: MDA5 does not unwind RNA duplexes processively; its ATPase serves filament proofreading only.

Type: single-molecule helicase assay

πŸ“š Additional Documentation

Notes

(IFIH1-notes.md)

Loading supporting content…

Download this section (compressed HTML)

πŸ“„ View Raw YAML

Loading supporting content…

Download this section (compressed HTML)