HNRNPA2B1

UniProt ID: P22626
Organism: Homo sapiens
Review Status: COMPLETE
πŸ“ Provide Detailed Feedback

Gene Description

Heterogeneous nuclear ribonucleoprotein A2/B1 is an abundant, broadly expressed RNA-binding protein of the hnRNP A/B family. It has two N-terminal RNA recognition motifs and a C-terminal glycine-rich low-complexity domain, and the B1 and A2 isoforms differ by a 12-residue N-terminal insert. The tandem RRMs bind RNA in a sequence-specific way, with RRM1 reading an AGG motif and RRM2 a UAG motif; the protein associates with nascent pre-mRNA and packages it into hnRNP particles, is a component of the spliceosome including the catalytic step 2 complex, and influences splice-site selection. Through the same sequence preference it recognises the A2 response element, a 21-nucleotide motif in the 3' untranslated region of myelin basic protein mRNA and other localised transcripts, and escorts those mRNAs from the nucleus to the cytoplasm in oligodendrocytes and neurons; HIV-1 genomic RNA carries the same element and is trafficked by the same route. It also binds primary microRNAs and promotes their processing by enhancing DGCR8 recruitment, and, when sumoylated, sorts specific microRNAs into exosomes. A second nucleic-acid-binding activity recognises single-stranded G-rich telomeric DNA, whose repeat matches the same consensus as the A2RE. The low-complexity domain drives liquid-liquid phase separation and assembly of RNA-processing granules; it is disordered both as a monomer and within the condensed phase, is regulated by arginine methylation, and contains a steric-zipper motif. Mutations that strengthen that motif convert condensation into self-seeding fibril formation and cause multisystem proteinopathy - inclusion body myopathy with Paget disease of bone and frontotemporal dementia, and amyotrophic lateral sclerosis - while frameshift variants in the same region cause oculopharyngeal muscular dystrophy. Two further roles are unsettled. The protein binds N6-methyladenosine-containing transcripts and mediates m6A-dependent effects on splicing and pri-miRNA processing, but direct affinity measurements find binding to methylated and unmethylated GGACU essentially equal, with binding energy supplied by the flanking AGG and UAG motifs, so whether it recognises the methyl mark itself or responds to an m6A-induced change in RNA structure is disputed. Separately, nuclear HNRNPA2B1 has been reported to act as a sensor of viral DNA that homodimerises, is demethylated at Arg226 by JMJD6, translocates to the cytoplasm and activates TBK1-IRF3 signalling to induce type I interferon; this role has been built upon by subsequent work but has not been independently re-tested.

Existing Annotations Review

GO Term Evidence Action Reason
GO:0000398 mRNA splicing, via spliceosome
IBA
GO_REF:0000033
ACCEPT
Summary: mRNA splicing via the spliceosome. HNRNPA2B1 binds intronic sequence and modulates splice-site choice.
Reason: Core biological process, supported by spliceosome composition, curator inference from those studies, and a loss-of-function experiment. Note that the IMP source (PMID:26321680) framed the splicing changes as m6A-dependent; the splicing role itself is independent of that framing and stands regardless of how the m6A-reader question resolves.
GO:0003730 mRNA 3'-UTR binding
IBA
GO_REF:0000033
ACCEPT
Summary: Binding to mRNA 3'-untranslated regions, notably the A2 response element (A2RE) of myelin basic protein mRNA.
Reason: Core and mechanistically specific: the A2RE is a 21-nucleotide 3' UTR element, point mutations that abolish hnRNP A2 binding also abolish transport, and this is the basis of the mRNA-trafficking function.
Supporting Evidence:
PMID:10567417
Point mutations of the A2RE11 that eliminated binding to hnRNP A2 also markedly reduced the ability of these oligoribonucleotides to support RNA transport.
GO:0043047 single-stranded telomeric DNA binding
IBA
GO_REF:0000033
KEEP AS NON CORE
Summary: Single-stranded telomeric DNA binding.
Reason: A genuine second nucleic-acid-binding activity - the ssDNA telomere repeat matches the same consensus as the A2RE RNA element, so one sequence-specific site accommodates both - but peripheral to the hnRNP function that defines this protein, and demonstrated largely in rat brain protein rather than in a human cellular setting.
Supporting Evidence:
PMID:15659580
Both the hnRNP A2-binding cis-acting element for the cytoplasmic RNA trafficking element, A2RE, and the ssDNA telomere repeat match a consensus sequence for binding to a second sequence-specific site identified by mutational analysis.
GO:0051028 mRNA transport
IBA
GO_REF:0000033
ACCEPT
Summary: mRNA transport.
Reason: Core biological process. The A2RE/A2RE11 element in myelin basic protein mRNA is necessary and sufficient for oligodendrocyte transport, and knockdown of hnRNP A2 disrupts it.
Supporting Evidence:
PMID:10567417
Oligodendrocytes treated with antisense oligonucleotides directed against the translation start site of hnRNP A2 had reduced levels of this protein and disrupted transport of microinjected myelin basic protein RNA.
GO:0071013 catalytic step 2 spliceosome
IBA
GO_REF:0000033
ACCEPT
Summary: Component of the catalytic step 2 spliceosome.
Reason: Correct and more specific; identified in purified C complex.
GO:0003676 nucleic acid binding
IEA
GO_REF:0000002
MODIFY
Summary: Bare nucleic acid binding from the InterPro RRM signature.
Reason: Correct but one level too general; RNA binding is the substantiated claim and is already annotated with experimental evidence.
Proposed replacements: RNA binding
GO:0003723 RNA binding
IEA
GO_REF:0000120
ACCEPT
Summary: RNA binding. HNRNPA2B1 is a core hnRNP: two RRMs plus a glycine-rich low-complexity domain, binding nascent pre-mRNA and packaging it into hnRNP particles.
Reason: Core molecular function, supported by focused and high-throughput work alike. General, but for an hnRNP it is the right general statement; the specific children this gene carries (pre-mRNA intronic binding, mRNA 3'-UTR binding, miRNA binding) refine it.
GO:0005576 extracellular region
IEA
GO_REF:0000044
KEEP AS NON CORE
Summary: Extracellular region, transferred from the UniProt subcellular-location vocabulary.
Reason: A less specific restatement of the exosome localisation; kept on the same footing. HNRNPA2B1 is not a secreted protein in the conventional sense.
GO:0005634 nucleus
IEA
GO_REF:0000120
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005654 nucleoplasm
IEA
GO_REF:0000044
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005737 cytoplasm
IEA
GO_REF:0000120
ACCEPT
Summary: Cytoplasmic localisation.
Reason: Correct. HNRNPA2B1 shuttles: it escorts A2RE-containing mRNAs to the cytoplasm, is found in cytoplasmic mRNP granules and stress granules, and relocalises on stimulation. As with the nucleus annotation, PMID:31320558 is cited here for localisation only, not for the innate immune function it reports.
GO:0005515 protein binding
IPI
PMID:18519039
The p53 target protein Wig-1 binds hnRNP A2/B1 and RNA Helic...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:23976881
The HILDA complex coordinates a conditional switch in the 3'...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:25702638
The L1TD1 protein interactome reveals the importance of post...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:26496610
A human interactome in three quantitative dimensions organiz...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:30021884
Histone Interaction Landscapes Visualized by Crosslinking Ma...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:35271311
OpenCell: Endogenous tagging for the cartography of human ce...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:36950384
Protein interaction studies in human induced neurons indicat...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:37207277
Using brain cell-type-specific protein interactomes to inter...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0097157 pre-mRNA intronic binding
IEA
GO_REF:0000107
ACCEPT
Summary: Binding to intronic sequence of pre-mRNA.
Reason: Core: this is the binding activity that underlies the splicing-regulatory role, and it fits the measured sequence preference (AGG/UAG-containing motifs, abundant in introns).
GO:0005654 nucleoplasm
IDA
GO_REF:0000052
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005634 nucleus
EXP
PMID:10772824
The RGG domain in hnRNP A2 affects subcellular localization.
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005634 nucleus
EXP
PMID:17289661
Molecular composition of IMP1 ribonucleoprotein granules.
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005634 nucleus
EXP
PMID:24098712
Arginine methylation of hnRNP A2 does not directly govern it...
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005634 nucleus
EXP
PMID:31320558
Nuclear hnRNPA2B1 initiates and amplifies the innate immune ...
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005634 nucleus
EXP
PMID:37643469
FAM76B regulates NF-ΞΊB-mediated inflammatory pathway by infl...
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005737 cytoplasm
EXP
PMID:10772824
The RGG domain in hnRNP A2 affects subcellular localization.
ACCEPT
Summary: Cytoplasmic localisation.
Reason: Correct. HNRNPA2B1 shuttles: it escorts A2RE-containing mRNAs to the cytoplasm, is found in cytoplasmic mRNP granules and stress granules, and relocalises on stimulation. As with the nucleus annotation, PMID:31320558 is cited here for localisation only, not for the innate immune function it reports.
GO:0005737 cytoplasm
EXP
PMID:17289661
Molecular composition of IMP1 ribonucleoprotein granules.
ACCEPT
Summary: Cytoplasmic localisation.
Reason: Correct. HNRNPA2B1 shuttles: it escorts A2RE-containing mRNAs to the cytoplasm, is found in cytoplasmic mRNP granules and stress granules, and relocalises on stimulation. As with the nucleus annotation, PMID:31320558 is cited here for localisation only, not for the innate immune function it reports.
GO:0005737 cytoplasm
EXP
PMID:31320558
Nuclear hnRNPA2B1 initiates and amplifies the innate immune ...
ACCEPT
Summary: Cytoplasmic localisation.
Reason: Correct. HNRNPA2B1 shuttles: it escorts A2RE-containing mRNAs to the cytoplasm, is found in cytoplasmic mRNP granules and stress granules, and relocalises on stimulation. As with the nucleus annotation, PMID:31320558 is cited here for localisation only, not for the innate immune function it reports.
GO:0005737 cytoplasm
EXP
PMID:37643469
FAM76B regulates NF-ΞΊB-mediated inflammatory pathway by infl...
ACCEPT
Summary: Cytoplasmic localisation.
Reason: Correct. HNRNPA2B1 shuttles: it escorts A2RE-containing mRNAs to the cytoplasm, is found in cytoplasmic mRNP granules and stress granules, and relocalises on stimulation. As with the nucleus annotation, PMID:31320558 is cited here for localisation only, not for the innate immune function it reports.
GO:0032392 DNA geometric change
ISS
GO_REF:0000024
MARK AS OVER ANNOTATED
Summary: DNA geometric change, from a sequence-similarity transfer.
Reason: The term asserts an induced change in the twist or writhe of a DNA double helix. What the underlying observations show is that hnRNP A2 binds single-stranded DNA and protects the telomeric repeat from nuclease digestion. Single-stranded-nucleic-acid-binding proteins destabilise duplexes as a consequence of binding one strand, but that is not the same as carrying out a topological transformation, and no study assigns HNRNPA2B1 such an activity. The binding terms this gene already carries state the evidence accurately.
GO:0070182 DNA polymerase binding
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: DNA polymerase binding, from a sequence-similarity transfer (GO_REF:0000024).
Reason: Retained rather than removed - a curator made this transfer on explicit sequence-similarity judgment and I have not examined the source annotation - but it is the weakest annotation on this gene: no human experimental support, no partner named, and no mechanistic role in any of the established hnRNP functions. It should not be treated as part of the core function.
GO:0098505 G-rich strand telomeric DNA binding
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: G-rich strand telomeric DNA binding, from a sequence-similarity transfer.
Reason: A more specific restatement of the single-stranded telomeric DNA binding activity; kept on the same footing, peripheral.
GO:0005515 protein binding
IPI
PMID:37643469
FAM76B regulates NF-ΞΊB-mediated inflammatory pathway by infl...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0140693 molecular condensate scaffold activity
IDA
PMID:29358076
Mechanistic View of hnRNPA2 Low-Complexity Domain Structure,...
ACCEPT
Summary: Molecular condensate scaffold activity: the glycine-rich low-complexity domain drives liquid-liquid phase separation and assembly of RNA-processing granules.
Reason: Core and uncontested. The LC domain is disordered as a monomer and stays largely disordered in the phase-separated state; disease mutations in it extend the aggregation-prone region and convert condensation into fibril formation, which is the molecular basis of multisystem proteinopathy. Arginine methylation of the same domain tunes phase separation.
Supporting Evidence:
PMID:29358076
The hnRNPA2 low-complexity (LC) domain is compact and intrinsically disordered as a monomer, retaining predominant disorder in a liquid-liquid phase-separated form.
GO:0005515 protein binding
IPI
PMID:25678563
Peptidylprolyl isomerase A governs TARDBP function and assem...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0070062 extracellular exosome
TAS
Reactome:R-HSA-8950292
KEEP AS NON CORE
Summary: Extracellular exosome localisation.
Reason: Real - sumoylated HNRNPA2B1 is present in exosomes and directs miRNA loading there - but a minor, conditional destination relative to the nuclear and cytoplasmic pools.
GO:0000781 chromosome, telomeric region
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: Chromosome, telomeric region, from a sequence-similarity transfer.
Reason: Consistent with the single-stranded telomeric DNA binding activity and with colocalisation with telomeric chromatin in ALT-cell PML bodies, but peripheral to the hnRNP function.
GO:0015030 Cajal body
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: Cajal body localisation, from a sequence-similarity transfer.
Reason: Plausible for an RNA-processing factor and consistent with the condensate-scaffold activity, but not supported by human experimental evidence here and not part of the core function.
GO:0016363 nuclear matrix
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: Nuclear matrix localisation, from a sequence-similarity transfer.
Reason: A classical description of hnRNP fractionation behaviour rather than a distinct functional site; retained as non-core.
GO:1904358 positive regulation of telomere maintenance via telomere lengthening
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: Positive regulation of telomere maintenance via telomere lengthening, from a sequence-similarity transfer.
Reason: Follows the telomeric ssDNA binding activity and is kept on the same footing: real enough to retain, too peripheral and too indirectly supported to count as core.
GO:0005515 protein binding
IPI
PMID:24965446
Host factors that interact with the pestivirus N-terminal pr...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0000398 mRNA splicing, via spliceosome
IMP
PMID:26321680
HNRNPA2B1 Is a Mediator of m(6)A-Dependent Nuclear RNA Proce...
ACCEPT
Summary: mRNA splicing via the spliceosome. HNRNPA2B1 binds intronic sequence and modulates splice-site choice.
Reason: Core biological process, supported by spliceosome composition, curator inference from those studies, and a loss-of-function experiment. Note that the IMP source (PMID:26321680) framed the splicing changes as m6A-dependent; the splicing role itself is independent of that framing and stands regardless of how the m6A-reader question resolves.
GO:0003730 mRNA 3'-UTR binding
IDA
PMID:10567417
Mutational analysis of a heterogeneous nuclear ribonucleopro...
ACCEPT
Summary: Binding to mRNA 3'-untranslated regions, notably the A2 response element (A2RE) of myelin basic protein mRNA.
Reason: Core and mechanistically specific: the A2RE is a 21-nucleotide 3' UTR element, point mutations that abolish hnRNP A2 binding also abolish transport, and this is the basis of the mRNA-trafficking function.
Supporting Evidence:
PMID:10567417
Point mutations of the A2RE11 that eliminated binding to hnRNP A2 also markedly reduced the ability of these oligoribonucleotides to support RNA transport.
GO:0005515 protein binding
IPI
PMID:26321680
HNRNPA2B1 Is a Mediator of m(6)A-Dependent Nuclear RNA Proce...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0005737 cytoplasm
IDA
PMID:24356509
Sumoylated hnRNPA2B1 controls the sorting of miRNAs into exo...
ACCEPT
Summary: Cytoplasmic localisation.
Reason: Correct. HNRNPA2B1 shuttles: it escorts A2RE-containing mRNAs to the cytoplasm, is found in cytoplasmic mRNP granules and stress granules, and relocalises on stimulation. As with the nucleus annotation, PMID:31320558 is cited here for localisation only, not for the innate immune function it reports.
GO:0006406 mRNA export from nucleus
IDA
PMID:10567417
Mutational analysis of a heterogeneous nuclear ribonucleopro...
ACCEPT
Summary: mRNA export from the nucleus.
Reason: Core: HNRNPA2B1 binds A2RE-containing transcripts and promotes their passage to the cytoplasm.
GO:0031053 primary miRNA processing
IDA
PMID:26321680
HNRNPA2B1 Is a Mediator of m(6)A-Dependent Nuclear RNA Proce...
ACCEPT
Summary: Primary miRNA processing: HNRNPA2B1 binds pri-miRNAs, interacts with DGCR8 and promotes Microprocessor cleavage.
Reason: The pri-miRNA processing role is directly demonstrated and is retained as correct. It is separable from the m6A-reader claim made in the same paper: HNRNPA2B1 can promote Microprocessor recruitment to pri-miRNAs it binds through its AGG/UAG sequence preference without recognizing the methyl mark itself, which is the 'm6A switch' model. The process annotation therefore survives the demotion of GO:1990247.
Supporting Evidence:
PMID:26321680
HNRNPA2B1 binds to m(6)A marks in a subset of primary miRNA transcripts, interacts with the microRNA Microprocessor complex protein DGCR8, and promotes primary miRNA processing
GO:0035198 miRNA binding
IDA
PMID:24356509
Sumoylated hnRNPA2B1 controls the sorting of miRNAs into exo...
ACCEPT
Summary: miRNA binding, via sequence motifs that direct miRNAs into exosomes.
Reason: Directly demonstrated, and the binding specificity is sequence-based, consistent with the AGG/UAG preference established structurally.
Supporting Evidence:
PMID:24356509
The protein heterogeneous nuclear ribonucleoprotein A2B1 (hnRNPA2B1) specifically binds exosomal miRNAs through the recognition of these motifs and controls their loading into exosomes.
GO:0070062 extracellular exosome
IDA
PMID:24356509
Sumoylated hnRNPA2B1 controls the sorting of miRNAs into exo...
KEEP AS NON CORE
Summary: Extracellular exosome localisation.
Reason: Real - sumoylated HNRNPA2B1 is present in exosomes and directs miRNA loading there - but a minor, conditional destination relative to the nuclear and cytoplasmic pools.
GO:1990247 N6-methyladenosine-containing RNA reader activity
IDA
PMID:26321680
HNRNPA2B1 Is a Mediator of m(6)A-Dependent Nuclear RNA Proce...
MARK AS OVER ANNOTATED
Summary: m6A reader activity, from Alarcon et al. 2015 (PMID:26321680). The GO definition requires a protein adaptor that 'recognizes and binds' m6A-modified RNA; two independent studies find that HNRNPA2B1 binds such RNA without recognizing the methyl mark.
Reason: GOA does carry an m6A-reader term for this gene, so there is a real annotation to adjudicate. The underlying observation is sound - HNRNPA2B1 binds m6A-bearing transcripts in vivo and in vitro, and its CLIP footprint overlaps the m6A consensus - but the inference from overlap to 'reader' does not survive direct affinity measurement, and this was already flagged in 2018, not only in 2026. Wu et al. 2018 (PMID:29379020) solved crystal structures of the tandem RRMs on several RNAs, found sequence-specific recognition of AGG by RRM1 and UAG by RRM2, and concluded the protein acts through an 'm6A switch' rather than as a direct reader. Park et al. 2026 (PMID:41662154) combined MD simulation with microscale thermophoresis on the RRM1-RRM2 construct: with m6A placed in a GGACU motif between the two domains, affinity is nearly identical to and marginally worse than the unmodified sequence, and binding energy is dominated by the flanking AGG/UAG motifs rather than by the methylated adenine. The GGAC consensus written by METTL3 overlaps the AGG/UAG motifs HNRNPA2B1 prefers, which is a sufficient explanation for the CLIP overlap without any methyl-specific recognition. MARK_AS_OVER_ANNOTATED rather than REMOVE: this is an IDA whose full experimental basis I have not read, HNRNPA2B1 genuinely does bind m6A-containing RNA and does mediate m6A-dependent effects on splicing and pri-miRNA processing, and the 2026 measurements used an isolated RRM1-RRM2 fusion rather than full-length protein with its low-complexity domain. What is over-stated is the selectivity that the word 'reader' asserts.
Supporting Evidence:
PMID:41662154
the binding affinity is nearly identical and slightly less favourable to the unmodified sequence
PMID:41662154
Our study suggests that HNRNPA2B1 is not a (selective) reader but has high affinity for m6A in a sequence-dependent manner.
PMID:41662154
This can be attributed to the strong interactions conferred by AGG and UAG motifs rather than adenine or m6A in the GGACU motif
PMID:29379020
elucidating specific recognitions of AGG and UAG motifs by RRM1 and RRM2 domains, respectively
PMID:26321680
We find that the RNA-binding protein HNRNPA2B1 binds m(6)A-bearing RNAs in vivo and in vitro and its biochemical footprint matches the m(6)A consensus motif.
GO:1990428 miRNA transport
IDA
PMID:24356509
Sumoylated hnRNPA2B1 controls the sorting of miRNAs into exo...
KEEP AS NON CORE
Summary: miRNA transport into exosomes.
Reason: Directly demonstrated and mechanistically specific (sumoylation-dependent, motif-driven loading), but a specialised secondary role rather than part of the central hnRNP function.
GO:0016020 membrane
HDA
PMID:19946888
Defining the membrane proteome of NK cells.
MARK AS OVER ANNOTATED
Summary: Membrane localisation from a high-throughput proteomics dataset.
Reason: HNRNPA2B1 has no transmembrane segment, lipid anchor or signal peptide. Recovery in a membrane fraction is expected for an abundant, aggregation-prone RNA-binding protein and does not support a membrane location.
GO:0005515 protein binding
IPI
PMID:22720776
PHF6 interacts with the nucleosome remodeling and deacetylat...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0005634 nucleus
IDA
PMID:22720776
PHF6 interacts with the nucleosome remodeling and deacetylat...
ACCEPT
Summary: Nuclear localisation, the principal site of hnRNP particle assembly and splicing.
Reason: Correct and heavily supported. Worth flagging what GOA has and has not taken from one of these sources: PMID:31320558 (Wang, Wen & Cao, Science 2019) is the paper proposing hnRNPA2B1 as a nuclear DNA sensor that detects viral DNA, homodimerises, is demethylated by JMJD6 and translocates to activate TBK1-IRF3. GOA cites it only for nucleus and cytoplasm localisation. There is no molecular function term for DNA sensing and no biological process term for innate immune response or interferon production on this gene, even though UniProt states the innate-immunity role in its FUNCTION comment. That asymmetry is itself the finding: the DNA-sensor claim is a standing, widely cited claim that GO curators have so far declined to turn into a functional annotation.
Supporting Evidence:
PMID:31320558
Upon DNA virus infection, nuclear-localized hnRNPA2B1 senses viral DNA, homodimerizes
GO:0005515 protein binding
IPI
PMID:24549040
C9ORF72, implicated in amytrophic lateral sclerosis and fron...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0003723 RNA binding
HDA
PMID:22658674
Insights into RNA biology from an atlas of mammalian mRNA-bi...
ACCEPT
Summary: RNA binding. HNRNPA2B1 is a core hnRNP: two RRMs plus a glycine-rich low-complexity domain, binding nascent pre-mRNA and packaging it into hnRNP particles.
Reason: Core molecular function, supported by focused and high-throughput work alike. General, but for an hnRNP it is the right general statement; the specific children this gene carries (pre-mRNA intronic binding, mRNA 3'-UTR binding, miRNA binding) refine it.
GO:0003723 RNA binding
HDA
PMID:22681889
The mRNA-bound proteome and its global occupancy profile on ...
ACCEPT
Summary: RNA binding. HNRNPA2B1 is a core hnRNP: two RRMs plus a glycine-rich low-complexity domain, binding nascent pre-mRNA and packaging it into hnRNP particles.
Reason: Core molecular function, supported by focused and high-throughput work alike. General, but for an hnRNP it is the right general statement; the specific children this gene carries (pre-mRNA intronic binding, mRNA 3'-UTR binding, miRNA binding) refine it.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-72103
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770119
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770129
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770131
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770132
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770141
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770142
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770145
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770236
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9770847
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9794542
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9921507
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970141
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970179
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970189
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970190
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970191
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970193
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970282
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970294
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970318
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970320
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970428
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970429
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:0005654 nucleoplasm
TAS
Reactome:R-HSA-9970431
ACCEPT
Summary: Nucleoplasmic localisation.
Reason: Correct and more specific than bare nucleus; consistent with exclusion from the nucleolus.
GO:1990904 ribonucleoprotein complex
IDA
PMID:2557628
Primary structures of the heterogeneous nuclear ribonucleopr...
ACCEPT
Summary: Part of a ribonucleoprotein complex - the hnRNP particle.
Reason: Core: hnRNP particle assembly on nascent pre-mRNA is what this protein is for.
GO:0005515 protein binding
IPI
PMID:21784977
Zinc finger protein tristetraprolin interacts with CCL3 mRNA...
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:0000398 mRNA splicing, via spliceosome
IC
PMID:11991638
Purification and characterization of native spliceosomes sui...
ACCEPT
Summary: mRNA splicing via the spliceosome. HNRNPA2B1 binds intronic sequence and modulates splice-site choice.
Reason: Core biological process, supported by spliceosome composition, curator inference from those studies, and a loss-of-function experiment. Note that the IMP source (PMID:26321680) framed the splicing changes as m6A-dependent; the splicing role itself is independent of that framing and stands regardless of how the m6A-reader question resolves.
GO:0071013 catalytic step 2 spliceosome
IDA
PMID:11991638
Purification and characterization of native spliceosomes sui...
ACCEPT
Summary: Component of the catalytic step 2 spliceosome.
Reason: Correct and more specific; identified in purified C complex.
GO:0005515 protein binding
IPI
PMID:17289661
Molecular composition of IMP1 ribonucleoprotein granules.
REMOVE
Summary: Bare protein binding from large-scale interaction screens.
Reason: Carries no functional information. HNRNPA2B1 is abundant, sticky and phase- separating, which makes it a frequent hit in affinity proteomics; the resulting IPI annotations say nothing about its molecular function. 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:1990904 ribonucleoprotein complex
IDA
PMID:17289661
Molecular composition of IMP1 ribonucleoprotein granules.
ACCEPT
Summary: Part of a ribonucleoprotein complex - the hnRNP particle.
Reason: Core: hnRNP particle assembly on nascent pre-mRNA is what this protein is for.
GO:0003723 RNA binding
IDA
PMID:15659580
hnRNP A2, a potential ssDNA/RNA molecular adapter at the tel...
ACCEPT
Summary: RNA binding. HNRNPA2B1 is a core hnRNP: two RRMs plus a glycine-rich low-complexity domain, binding nascent pre-mRNA and packaging it into hnRNP particles.
Reason: Core molecular function, supported by focused and high-throughput work alike. General, but for an hnRNP it is the right general statement; the specific children this gene carries (pre-mRNA intronic binding, mRNA 3'-UTR binding, miRNA binding) refine it.
GO:0006397 mRNA processing
IDA
PMID:2557628
Primary structures of the heterogeneous nuclear ribonucleopr...
ACCEPT
Summary: mRNA processing.
Reason: Correct parent process; the gene carries the more specific splicing and export terms.
GO:0043047 single-stranded telomeric DNA binding
IDA
PMID:15659580
hnRNP A2, a potential ssDNA/RNA molecular adapter at the tel...
KEEP AS NON CORE
Summary: Single-stranded telomeric DNA binding.
Reason: A genuine second nucleic-acid-binding activity - the ssDNA telomere repeat matches the same consensus as the A2RE RNA element, so one sequence-specific site accommodates both - but peripheral to the hnRNP function that defines this protein, and demonstrated largely in rat brain protein rather than in a human cellular setting.
Supporting Evidence:
PMID:15659580
Both the hnRNP A2-binding cis-acting element for the cytoplasmic RNA trafficking element, A2RE, and the ssDNA telomere repeat match a consensus sequence for binding to a second sequence-specific site identified by mutational analysis.
GO:0050658 RNA transport
IDA
PMID:17004321
Trafficking of HIV-1 RNA is mediated by heterogeneous nuclea...
ACCEPT
Summary: RNA transport.
Reason: Correct; demonstrated for HIV-1 genomic RNA, which carries A2RE motifs, and mechanistically the same activity as the cellular mRNA trafficking role.
GO:0000398 mRNA splicing, via spliceosome
IC
PMID:9731529
Mass spectrometry and EST-database searching allows characte...
ACCEPT
Summary: mRNA splicing via the spliceosome. HNRNPA2B1 binds intronic sequence and modulates splice-site choice.
Reason: Core biological process, supported by spliceosome composition, curator inference from those studies, and a loss-of-function experiment. Note that the IMP source (PMID:26321680) framed the splicing changes as m6A-dependent; the splicing role itself is independent of that framing and stands regardless of how the m6A-reader question resolves.
GO:0005681 spliceosomal complex
IDA
PMID:9731529
Mass spectrometry and EST-database searching allows characte...
ACCEPT
Summary: Component of the spliceosomal complex.
Reason: Directly demonstrated by spliceosome purification; consistent with the splicing role.
GO:0003723 RNA binding
TAS
PMID:7789969
Structure and expression of the gene (HNRPA2B1) encoding the...
ACCEPT
Summary: RNA binding. HNRNPA2B1 is a core hnRNP: two RRMs plus a glycine-rich low-complexity domain, binding nascent pre-mRNA and packaging it into hnRNP particles.
Reason: Core molecular function, supported by focused and high-throughput work alike. General, but for an hnRNP it is the right general statement; the specific children this gene carries (pre-mRNA intronic binding, mRNA 3'-UTR binding, miRNA binding) refine it.

Core Functions

Sequence-specific binding of nascent pre-mRNA through tandem RRM domains - RRM1 reading an AGG motif and RRM2 a UAG motif - packaging transcripts into hnRNP particles and modulating splice-site selection as a component of the spliceosome, including the catalytic step 2 complex. This is the activity from which the protein's name and most of its biology follow, and it is the same sequence preference that explains its apparent affinity for m6A consensus sites without requiring recognition of the methyl mark.

Supporting Evidence:
  • PMID:29379020
    elucidating specific recognitions of AGG and UAG motifs by RRM1 and RRM2 domains, respectively

Recognition of the A2 response element, a 21-nucleotide motif in the 3' untranslated region of myelin basic protein mRNA and other localised transcripts, and escort of those mRNAs from the nucleus to the cytoplasm in oligodendrocytes and neurons. Binding and transport are coupled: point mutations that abolish binding abolish transport, and depleting the protein disrupts trafficking of injected MBP mRNA. HIV-1 genomic RNA carries the same element and is trafficked by the same activity.

Molecular Function:
mRNA 3'-UTR binding
Cellular Locations:
Supporting Evidence:
  • PMID:10567417
    Point mutations of the A2RE11 that eliminated binding to hnRNP A2 also markedly reduced the ability of these oligoribonucleotides to support RNA transport.
  • PMID:10567417
    Oligodendrocytes treated with antisense oligonucleotides directed against the translation start site of hnRNP A2 had reduced levels of this protein and disrupted transport of microinjected myelin basic protein RNA.

Scaffolding of RNA-containing molecular condensates through the glycine-rich low-complexity domain. The domain is intrinsically disordered as a monomer and remains largely disordered within the phase-separated state, and arginine methylation tunes the arginine-mediated contacts that drive condensation. This activity is also where the protein's pathology lives: a steric-zipper motif within the domain, strengthened by the multisystem-proteinopathy mutations, converts reversible condensation into self-seeding fibril formation.

Supporting Evidence:
  • PMID:29358076
    The hnRNPA2 low-complexity (LC) domain is compact and intrinsically disordered as a monomer, retaining predominant disorder in a liquid-liquid phase-separated form.
  • PMID:23455423
    Wild-type hnRNPA2 (the most abundant isoform of hnRNPA2B1) and hnRNPA1 show an intrinsic tendency to assemble into self-seeding fibrils, which is exacerbated by the disease mutations.

Binding of primary microRNA transcripts and promotion of their Microprocessor-dependent cleavage through interaction with DGCR8. The binding is retained as a core activity; what is not retained is the claim that the protein selects those pri-miRNAs by reading an N6-methyladenosine mark, since direct affinity measurement shows methylated and unmethylated GGACU bind essentially equally and the binding energy comes from the flanking AGG and UAG motifs.

Molecular Function:
miRNA binding
Directly Involved In:
Cellular Locations:
Supporting Evidence:
  • PMID:26321680
    HNRNPA2B1 binds to m(6)A marks in a subset of primary miRNA transcripts, interacts with the microRNA Microprocessor complex protein DGCR8, and promotes primary miRNA processing
  • PMID:41662154
    either adenine or m6A at the fifth position, irrespective of their conformations with respect to the protein, do not contribute much to the binding

References

Loading supporting content…

Download this section (compressed HTML)

Suggested Questions for Experts

Q: Should HNRNPA2B1 retain GO:1990247 N6-methyladenosine-containing RNA reader activity? The term's definition requires a protein adaptor that recognizes m6A-modified RNA, but two independent studies - a 2018 crystallographic and biochemical analysis and a 2026 simulation-plus-thermophoresis study - find affinity for methylated GGACU that is equal to or marginally worse than for the unmodified sequence, with binding driven by the flanking AGG and UAG motifs. If the protein mediates m6A effects through an 'm6A switch' (a methylation-induced change in RNA structure that exposes its binding site) rather than by reading the mark, does GO need a distinct term for that indirect mode, so that 'reader' can be reserved for YTH-domain-style direct recognition?

Suggested experts: GO RNA-modification curators, Jiabin Ma, Samie Jaffrey

Q: Would the same conclusion hold for full-length HNRNPA2B1? The 2026 affinity measurements used an isolated RRM1-RRM2 fusion. The low-complexity domain drives phase separation and could plausibly contribute avidity or context effects that an isolated two-domain construct cannot show. A full-length comparison of methylated and unmethylated substrates, ideally in a condensate, would close the gap.

Suggested experts: Nicolas Fawzi, Lydia Contreras

Q: What should GO do about the nuclear DNA-sensor claim? UniProt states in its FUNCTION comment that HNRNPA2B1 senses viral DNA in the nucleus and activates TBK1-IRF3 signalling, citing PMID:31320558, and later papers build on that role. GOA cites the same paper only for nucleus and cytoplasm localisation and carries no DNA-binding, pattern-recognition or interferon-production annotation for this gene. Is that a deliberate curatorial hold pending independent replication, or an annotation gap? Either answer is worth recording explicitly, because the claim is widely cited.

Suggested experts: GO innate immunity curators, UniProt curators

Q: Are the telomeric DNA-binding annotations, which rest largely on rat brain hnRNP A2 and on sequence-similarity transfer, worth retaining for the human protein, and does GO:0032392 DNA geometric change overstate nuclease protection of a single-stranded repeat as a topological activity?

Suggested experts: GO telomere curators

Suggested Experiments

Experiment: Measure binding of full-length HNRNPA2B1 (not the RRM1-RRM2 fusion) to a matched panel of transcripts in which the m6A site and the flanking AGG/UAG motifs are varied independently: methylated and unmethylated GGACU in an AGG/UAG context, and the same methylation states with the flanking motifs mutated. Pair equilibrium binding with structure probing (SHAPE or DMS) on the same RNAs to test whether methylation changes accessibility rather than affinity. A binding preference that tracks motif accessibility rather than methylation would confirm the switch model and settle the term.

Hypothesis: HNRNPA2B1 does not discriminate m6A-modified from unmodified RNA; apparent m6A dependence in cells reflects an m6A-induced structural change that exposes AGG/UAG motifs.

Type: quantitative binding plus RNA structure probing

Experiment: Separate the two models in cells with RRM point mutants that abolish AGG or UAG binding without affecting folding (guided by the published crystal structures) and compare their effect on METTL3-dependent splicing and pri-miRNA processing with that of wild-type protein. If the motif-binding mutants phenocopy full loss of function while methylation is unchanged, the m6A dependence is indirect.

Hypothesis: The m6A-dependent splicing and pri-miRNA processing effects attributed to HNRNPA2B1 are mediated by RNA structural change, not by direct methyl recognition.

Type: structure-guided separation-of-function rescue

Experiment: Independently re-test the core claim in a second laboratory: HSV-1 or vaccinia infection in cells with HNRNPA2B1 knocked out and rescued with wild-type protein, with an Arg226 non-methylatable mutant, and with a dimerisation-defective mutant, reading out interferon induction, TBK1-IRF3 activation and nucleocytoplasmic translocation. Include a construct carrying RRM mutations that abolish RNA binding, to ask whether the immune phenotype can be uncoupled from the splicing and transport functions. Independent replication is the precondition for GO annotating a molecular function here.

Hypothesis: The reported nuclear DNA-sensing activity of HNRNPA2B1 is reproducible and separable from its hnRNP function.

Type: independent replication with separation-of-function mutants

πŸ“š Additional Documentation

Notes

(HNRNPA2B1-notes.md)

Loading supporting content…

Download this section (compressed HTML)

πŸ“„ View Raw YAML

Loading supporting content…

Download this section (compressed HTML)