Summary: IBA annotation for mannosyl-oligosaccharide 1,2-alpha-mannosidase activity. Drosophila EDEM2 belongs to glycosyl hydrolase family 47 and contains the catalytic domain and active site residues characteristic of alpha-1,2-mannosidases (UniProt features show four active site residues). While EDEMs were initially considered lectins, recent studies have revealed that some EDEMs can function as mannosidases (PMID:25716426: "recent studies have revealed that EDEMs can function as mannosidases"). The IBA annotation is phylogenetically appropriate given the conserved catalytic residues. However, whether Drosophila EDEM2 has actual catalytic mannosidase activity versus acting primarily as a lectin/sensor has not been directly demonstrated. Falcon deep research adds direct Drosophila genetic evidence bearing on the catalytic claim: overexpression of wild-type dEDEM2 reduces steady-state NHK (a glycoprotein ERAD substrate), whereas the catalytic mutant E144Q (EF-hand motif, abolishing alpha-1,2-mannosidase activity) instead increases NHK, indicating that mannosidase catalytic activity contributes to clearance of at least some glycoprotein ERAD substrates (PMID:28633019). Mammalian EDEM1/EDEM2 show bona fide but folding-state-dependent mannosidase activity, higher on unfolded/denatured glycoproteins (M8 to M5). Thus the catalytic mannosidase MF annotation is supported, not merely a domain-based lectin inference. (Note: protective effects on non-glycoprotein substrates such as Abeta42 are mannosidase-independent, reflecting a separate substrate-engagement role rather than refuting the catalytic MF.)

Reason: EDEM2 belongs to GH47 family with conserved active site residues. The IBA annotation is phylogenetically sound across EDEM family members. Drosophila catalytic-mutant evidence (E144Q increases NHK; wild-type reduces NHK) supports that mannosidase catalytic activity contributes to glycoprotein ERAD substrate clearance (PMID:28633019), and mammalian EDEM1/EDEM2 show bona fide folding-state-dependent mannosidase activity. This is a core molecular function annotation.

Summary: IBA annotation for ER unfolded protein response. This is well supported by experimental data in Drosophila showing that EDEM2 expression is transcriptionally induced by ER stress (PMID:25716426: "Drosophila EDEM2 expression was transcriptionally induced" upon DTT treatment). Furthermore, EDEM2 knockdown in combination with EDEM1 leads to increased Hsc3 levels under ER stress conditions (PMID:25716426: "this increase was even more pronounced when both EDEM1 and EDEM2 were knocked down"). The IBA annotation is consistent with the IMP and IGI annotations from the same gene.

Reason: EDEM2 is transcriptionally induced by ER stress and plays a protective role during the UPR by promoting degradation of misfolded proteins. The IBA is consistent with IMP (PMID:25716426) and IGI (PMID:19805114) evidence for this same term.

Summary: IBA annotation for ubiquitin-dependent glycoprotein ERAD pathway. This is strongly supported by experimental evidence. PMID:25716426 demonstrated that Drosophila EDEMs promote ubiquitination and degradation of misfolded A1AT glycoprotein variants, and knockdown leads to accumulation of glycosylated ERAD substrates. The IBA annotation is consistent with the IMP annotation from PMID:25716426.

Reason: ERAD is the core biological process function of EDEM2. The IBA annotation is well supported by direct experimental evidence showing EDEM2 promotes ubiquitin-dependent degradation of misfolded glycoproteins (PMID:25716426).

Summary: IBA annotation for ER localization. Directly confirmed by IDA evidence from PMID:25716426 showing co-localization with Hsc3 (Drosophila BiP ortholog) in the ER. UniProt also lists ER as the subcellular location. The IBA annotation is consistent with the IDA and IEA annotations for the same term.

Reason: ER localization is the primary constitutive location of EDEM2, directly demonstrated by immunolocalization (PMID:25716426). The IBA is well supported.

Summary: IEA annotation for ER quality control compartment, inferred from involvement in ER mannose trimming (GO:1904380). EDEM2 functions in the ERAD quality control pathway, recognizing misfolded glycoproteins for degradation. The ER quality control compartment is functionally appropriate for an ERAD component.

Reason: EDEM2 is an ER quality control factor involved in ERAD substrate recognition. The ER quality control compartment localization is consistent with its function in glycoprotein quality control and degradation.

Summary: IEA annotation for mannosyl-oligosaccharide 1,2-alpha-mannosidase activity from InterPro domain mapping (GH47 family). Consistent with the IBA annotation for the same term. EDEM2 contains the GH47 catalytic domain.

Reason: Consistent with IBA annotation and GH47 family membership. The InterPro mapping is appropriate for this domain-based molecular function.

Summary: IEA annotation for calcium ion binding from InterPro domain mapping. UniProt lists a Ca2+ binding site at position 487, and GH47 mannosidases are known to require calcium as a cofactor (UniProt COFACTOR section: "Ca(2+)"). This is consistent with the enzyme's catalytic requirements.

Reason: Calcium binding is documented in the UniProt record as a cofactor for the GH47 mannosidase domain. The InterPro mapping is appropriate.

Summary: IEA annotation for ER localization from UniProt subcellular location mapping. Consistent with the IBA and IDA annotations for the same term. UniProt lists ER as the subcellular location.

Reason: Consistent with IBA and IDA evidence for ER localization. Redundant but acceptable.

Summary: IEA annotation for carbohydrate metabolic process from InterPro domain mapping. This is a very broad parent term. The more specific terms for ER mannose trimming (GO:1904380) and ERAD pathway (GO:0097466, GO:0036503) are already annotated and are much more informative. While technically correct as a parent, this adds little value.

Reason: Too broad to be informative. EDEM2 is involved in ER mannose trimming and ERAD, not general carbohydrate metabolism. The more specific terms GO:1904380, GO:0097466, and GO:0036503 are already annotated and are much more informative.

Summary: IEA annotation for response to unfolded protein from UniProt keyword mapping (Unfolded protein response keyword). EDEM2 is transcriptionally induced by ER stress (PMID:25716426) and participates in the ERAD response to misfolded proteins. This is a broader term than the more specific GO:0030968 "ER unfolded protein response" which is already annotated. Acceptable as a parent term.

Reason: EDEM2 is part of the cellular response to unfolded proteins in the ER. While more specific terms are annotated (GO:0030968), this broader IEA term is acceptable and consistent with the gene's role.

Summary: IEA annotation for membrane from InterPro domain mapping. EDEM2 is an ER-resident protein. While the UniProt entry does not explicitly annotate a transmembrane domain (it has a signal peptide), the ER localization and potential membrane association are plausible. This is a very generic CC term.

Reason: Acceptable as a broad localization term. EDEM2 is associated with the ER membrane system. The more specific ER annotation is more informative.

Summary: IEA annotation for hydrolase activity from UniProt keyword mapping. EDEM2 belongs to the GH47 glycosyl hydrolase family. This is a broad parent term; the more specific GO:0004571 is already annotated. Acceptable as a parent term.

Reason: Correct as a broad parent term for the GH47 mannosidase activity. The more specific GO:0004571 is already annotated.

Summary: IEA annotation for hydrolase activity acting on glycosyl bonds from UniProt keyword mapping. EDEM2 belongs to GH47 family. This is an intermediate-specificity term between GO:0016787 (hydrolase activity) and GO:0004571 (mannosyl-oligosaccharide 1,2-alpha-mannosidase activity). Acceptable.

Reason: Correct intermediate-level MF term for GH47 family membership. Consistent with the more specific GO:0004571 annotation.

Summary: IEA annotation for metal ion binding from UniProt keyword mapping. EDEM2 binds calcium as a cofactor (UniProt COFACTOR section). This is a generic term; GO:0005509 "calcium ion binding" is already annotated and is more specific.

Reason: Correct as a broad parent of GO:0005509 calcium ion binding. Redundant with the more specific term but acceptable.

Summary: IEA annotation for ER mannose trimming from InterPro domain mapping. EDEM2 is part of the mannose trimming machinery that processes misfolded glycoproteins for ERAD. The annotation is consistent with the mannosidase activity (GO:0004571) and ERAD pathway (GO:0097466) annotations.

Reason: ER mannose trimming is a core biological process for EDEM2, directly related to its role in ERAD substrate recognition via mannose trimming of N-glycans. The InterPro mapping is appropriate. Falcon deep research supports EDEM2 acting as an alpha-1,2-mannose trimming enzyme on high-mannose N-glycans of misfolded glycoproteins, with conserved framing placing EDEM2 at the first mannose-trimming step initiating gpERAD.

Summary: IDA annotation for ER localization from PMID:25716426. This study directly demonstrated by immunofluorescence that epitope-tagged Drosophila EDEM2 co-localizes with Hsc3 (Drosophila BiP) in the ER when expressed in embryo amnioserosa cells (PMID:25716426: "Immunolabeling with anti-myc antibody revealed that the Drosophila EDEMs co-localized with Hsc3").

Reason: Direct IDA evidence from immunofluorescence co-localization with the ER marker Hsc3 (PMID:25716426). This is the strongest evidence for EDEM2 ER localization.

Summary: IMP annotation for ER unfolded protein response from PMID:25716426. The study showed that knockdown of both EDEM1 and EDEM2 led to increased Hsc3 (BiP) levels under ER stress conditions, and that EDEM2 expression is transcriptionally induced by ER stress (DTT treatment). This demonstrates that EDEM2 plays a protective role in the UPR by promoting degradation of misfolded proteins.

Reason: Well-supported IMP evidence showing that EDEM2 loss of function exacerbates ER stress (increased Hsc3 levels) and that EDEM2 is transcriptionally induced by ER stress (PMID:25716426). This is a core biological process annotation.

Summary: IEP annotation for response to ER stress from PMID:25716426. Based on expression pattern evidence: EDEM2 mRNA is transcriptionally induced by ER stress agents (DTT, tunicamycin, thapsigargin). The IEP evidence is appropriate for the expression-based observation.

Reason: Appropriate IEP annotation based on transcriptional induction of EDEM2 by multiple ER stress agents (PMID:25716426). Consistent with the broader UPR role. Falcon deep research further notes that Edem2 overexpression suppresses an ER-stress reporter (xbp1-EGFP) driven by misfolded Rh-1G69D, consistent with Edem2 reducing misfolded-client burden during the ER stress response.

Summary: IMP annotation for ubiquitin-dependent glycoprotein ERAD pathway from PMID:25716426. The study demonstrated that EDEM2 knockdown leads to accumulation of glycosylated A1AT mutant variants, and EDEM2 overexpression promotes their ubiquitination and degradation. This is the core biological process function of EDEM2.

Reason: Strong IMP evidence showing EDEM2 promotes ubiquitin-dependent degradation of misfolded glycoproteins. Knockdown causes accumulation; overexpression accelerates degradation and increases ubiquitination (PMID:25716426). Core function annotation.

Summary: ISM annotation for mannosyl-oligosaccharide 1,2-alpha-mannosidase activity from PMID:19805114 (sequence model-based inference). EDEM2 contains the GH47 catalytic domain with conserved active site residues. The ISM evidence is based on sequence analysis. Consistent with IBA and IEA annotations for the same term.

Reason: Consistent with the IBA and IEA annotations. The ISM evidence based on conserved active site residues is appropriate for this enzyme family assignment.

Summary: IGI annotation for ER unfolded protein response from PMID:19805114. This study showed that loss-of-function of ERAD factors (including EDEM2) resulted in increased levels of misfolded Rh-1 in ninaE mutant flies, while co-expression of ERAD factors reduced Rh-1 levels and suppressed ER stress reporter activation (PMID:19805114: "co-expression of certain ERAD factors was sufficient to reduce Rh-1 protein levels and to completely suppress ER stress reporter activation"). The IGI evidence (genetic interaction with ninaE mutants) supports EDEM2 role in UPR.

Reason: The IGI evidence from genetic interaction with ninaE (Rh-1) mutants supports EDEM2 involvement in the ER unfolded protein response. ERAD factor overexpression suppressed ER stress in the Drosophila retinal degeneration model (PMID:19805114).

Summary: IGI annotation for ERAD pathway from PMID:19805114. The study demonstrated that loss-of-function of ERAD factors including EDEM2 increased levels of misfolded Rh-1, and that ERAD factor overexpression reduced misfolded Rh-1 levels. The genetic interaction evidence supports EDEM2 involvement in ERAD.

Reason: ERAD is the core biological process for EDEM2. The IGI evidence from the Drosophila retinal degeneration model (PMID:19805114) supports this annotation. Consistent with the more specific GO:0097466 annotation.

Summary: IPI annotation for unfolded protein binding from PMID:19805114. The IPI evidence is based on physical interaction between EDEM2 and the misfolded Rh-1 protein (the WITH/FROM column lists FB:FBgn0002940, which is ninaE/Rh1). However, EDEM2 is an ERAD substrate recognition factor (lectin/sensor) that recognizes misfolded glycoproteins for degradation, not a chaperone that prevents aggregation. The interaction with misfolded Rh-1 is in the context of ERAD substrate recognition and targeting for degradation, not chaperone holdase or foldase activity. Per UPB project decision rules, ER quality control sensors should have GO:0051082 removed or marked as over-annotated. EDEM2 recognizes misfolded glycoprotein substrates via their N-glycan structures, which is fundamentally different from chaperone-mediated binding to unfolded polypeptide segments.

Reason: EDEM2 is an ERAD substrate recognition factor (lectin/sensor), not a chaperone. Its interaction with misfolded Rh-1 (PMID:19805114) reflects ERAD substrate recognition for degradation, not chaperone activity. Per UPB project rules, ER quality control sensors should not be annotated with GO:0051082. EDEM2 recognizes misfolded glycoproteins via N-glycan mannose trimming patterns, targeting them for ubiquitin- dependent proteasomal degradation (PMID:25716426). This is a sensor/lectin function, not an unfolded protein binding/chaperone function.