UFSP2

Summary: Phylogenetic inference that UFSP2 acts at the endoplasmic reticulum, where it deUFMylates RPL26 on ER-associated 60S ribosomes. Consistent with direct experimental evidence.

Reason: UFSP2 is an ER-membrane-tethered deUFMylase; the IBA localization is corroborated by direct IDA evidence. ORGANELLES check (2026-11-06): the PAINT IBD for GO:0005783 sits on PTN008489070 (Eumetazoa) and is seeded by mouse Ufsp2, human UFSP2 itself and a worm ortholog, so the target in its own sources reflects its own IDA. UFSP2 has no TM segment or signal peptide; it is anchored on the cytosolic face by the ER-membrane protein ODR4. Under the tethering route of rule 7 (2026-11-11) that stable, partner-dependent anchoring is recorded by a NEW GO:0098554 cytoplasmic side of endoplasmic reticulum membrane row beside this root row, which is kept; no ER lumen row is possible. The is_active_in qualifier is justified: ODR4 knockout, which removes UFSP2 from the ER, phenocopies UFSP2 loss for RPL26 deUFMylation (PMID:35926457), so the activity is shown at the ER.

Summary: Phylogenetic inference of UFM1-specific protease (deUFMylase) activity, the core molecular function of UFSP2, directly supported by extensive biochemistry.

Reason: deUFMylase activity is the central, repeatedly demonstrated function of UFSP2.

Summary: Phylogenetic inference of nuclear localization. Nuclear localization for human UFSP2 is supported only by sequence similarity to the mouse ortholog, not by direct human data; the dominant functional pool is at the ER/cytosol.

Reason: Possible minor nuclear pool (e.g. TRIP4-related nuclear-receptor regulation) but not the core site of UFSP2's deUFMylase function; retained as non-core.

Summary: Generic proteolysis term reflecting the peptidase activity of UFSP2. The specific activity is UFM1-specific isopeptide/peptide cleavage, better captured by deUFMylase activity and protein maturation.

Reason: Technically correct (UFSP2 is a protease) but uninformative compared with the specific deUFMylase activity; retained as non-core parent process.

Summary: Electronic localization to nucleus, inherited from the mouse ortholog; see the IBA/ISS nucleus annotations.

Reason: Not the core site of action; supported only by orthology, retained as non-core.

Summary: Electronic localization to cytoplasm, consistent with the cytoplasmic/ER-surface site where UFSP2 acts on ribosomal and other substrates.

Reason: Cytoplasm (including the cytoplasmic face of the ER) is a documented compartment for UFSP2 activity.

Summary: Electronic localization to ER, consistent with the experimentally established ER-membrane-tethered localization of UFSP2.

Reason: ER is the principal functional site of UFSP2 (RPL26 deUFMylation); agrees with direct evidence. ORGANELLES check (2026-11-06): the IEA rests on the UniProt location "Endoplasmic reticulum {ECO:0000269|PubMed:35926457}", curated from a paper on human UFSP2, and on mouse Ufsp2. UFSP2 has no TM segment or signal peptide; it is anchored on the cytosolic face by the ER-membrane protein ODR4. Under the tethering route of rule 7 (2026-11-11) that stable, partner-dependent anchoring is recorded by a NEW GO:0098554 cytoplasmic side of endoplasmic reticulum membrane row beside this root row, which is kept; no ER lumen row is possible.

Summary: High-throughput binary interactome screen recording UFSP2 interactions with APOBEC1 and LMO1; the bare protein binding term is uninformative and these partners do not inform UFSP2's deUFMylase function.

Reason: Bare GO:0005515 protein binding carries no functional information, so under the generic protein-binding policy this row is REMOVE. Removal does not mean the reported interaction is false; the summary and any supporting text keep it on record.

Summary: Ortholog-based (Ensembl) electronic annotation transferring an interferon-gamma production phenotype from the mouse gene. No direct human evidence; not part of the characterized UFM1-protease function.

Reason: Phenotype-based orthology transfer with no mechanistic link to UFSP2's deUFMylase activity in human; likely an indirect downstream consequence rather than a core function. The curated UniProt function describes a deUFMylase with no direct role in interferon production.

Summary: Automated electronic annotation of deUFMylase activity, redundant with and consistent with extensive experimental support.

Reason: Correct core molecular function.

Summary: Sequence-similarity-based nuclear localization transferred from the mouse ortholog (Q99K23). See the IBA/IEA nucleus annotations.

Reason: Orthology-only support; not the core functional compartment.

Summary: Sequence-similarity-based cytoplasmic localization, consistent with the cytoplasmic/ER-surface site of UFSP2 action.

Reason: Cytoplasmic localization is well supported, including by direct human data.

Summary: Direct evidence for UFSP2 deUFMylase activity in the context of HRD1 UFMylation and ER homeostasis.

Reason: Direct experimental support for the core deUFMylase activity.

Summary: Direct evidence for UFSP2 deUFMylase activity.

Reason: Direct experimental support for the core deUFMylase activity.

Summary: UFSP2 deUFMylates CYB5R3; because CYB5R3 UFMylation signals ER-phagy (degradation), UFSP2 activity opposes this catabolic route. The negative regulation of a degradation process is a downstream consequence of its deUFMylase activity.

Reason: Plausible process outcome of UFSP2 reversing CYB5R3 UFMylation, but a downstream/context-specific effect rather than the core molecular function.

Summary: Direct evidence for UFSP2 deUFMylase activity on a target substrate.

Reason: Direct experimental support for the core deUFMylase activity.

Summary: Direct demonstration that wild-type UFSP2, but not the catalytic Cys302Ala mutant, removes UFM1 from CYB5R3, establishing UFSP2 deUFMylase activity on this ER substrate.

Reason: Active-site-dependent removal of UFM1 from a defined substrate; strong direct support for the core function.

Summary: UFSP2 deUFMylation of RPL26/uL24 is required to release the UFM1 E3 ligase from UFMylated 60S subunits, enabling 60S recycling after RQC-mediated splitting of ribosomes stalled during co-translational ER translocation.

Reason: Directly supported role in the ribosome-rescue/recycling arm of ER-associated RQC; this is a key biological process for UFSP2.

Summary: Direct evidence that UFSP2 acts at the ER (in contrast to cytosolic UFSP1), where it removes UFM1 from the ribosomal subunit RPL26.

Reason: Establishes the ER as the functional compartment for UFSP2's RPL26 deUFMylation. ORGANELLES check (2026-11-06): membrane, cytosol and nuclear fractions and ODR4 knockout cells show ER localization of UFSP2 depends on the ER-resident TM protein ODR4, and ODR4 loss phenocopies UFSP2 loss for RPL26 UFMylation, so the is_active_in qualifier is justified. UFSP2 has no TM segment or signal peptide; it is anchored on the cytosolic face by the ER-membrane protein ODR4. Under the tethering route of rule 7 (2026-11-11) that stable, partner-dependent anchoring is recorded by a NEW GO:0098554 cytoplasmic side of endoplasmic reticulum membrane row beside this root row, which is kept; no ER lumen row is possible.

Summary: UFSP2-mediated deUFMylation of RPL26 releases the UFM1 E3 ligase clamp from the 60S subunit, contributing to disassembly/recycling of post-RQC 60S ribosomal subunits.

Reason: Directly supported role in releasing the E3-UFM1 clamp to allow 60S recycling.

Summary: Mutational/phenotypic evidence supporting UFSP2 as the active deUFMylase mediating deconjugation of UFM1 from target proteins.

Reason: Supports the core deUFMylase activity; UFSP2 mainly acts as a protein deUFMylase.

Summary: Direct evidence that UFSP2 removes UFM1 from RPL26, confirming deUFMylase activity.

Reason: Direct experimental support for the core deUFMylase activity.

Summary: UFSP2 deUFMylates RPL26 on 60S subunits, the activity underlying 60S recycling.

Reason: Direct support for the core deUFMylase activity in the ribosome-recycling context.

Summary: A pathogenic UFSP2 variant impairs UFMylation homeostasis, consistent with UFSP2 functioning as the active UFM1-specific peptidase; mutant phenotype supports the enzymatic role.

Reason: Disease-variant evidence supports UFSP2's deUFMylase/UFM1-protease function.

Summary: Mutational evidence supporting UFSP2 deUFMylase activity.

Reason: Supports the core deUFMylase activity.

Summary: UFSP2 deUFMylates the coactivator TRIP4/ASC-1, regulating nuclear-receptor (including estrogen receptor) transactivation. A context-specific downstream role of UFSP2's deUFMylase activity.

Reason: A genuine but context-specific regulatory process downstream of UFSP2 deUFMylating TRIP4; not the core molecular function.

Summary: Generic proteolysis term reflecting UFSP2's peptidase activity (here in the TRIP4 deUFMylation study). The specific activity is UFM1-specific cleavage.

Reason: Correct but uninformative parent term relative to deUFMylase activity.

Summary: UFSP2 deUFMylates TRIP4; mutational/functional evidence supports the deUFMylase activity on this substrate.

Reason: Supports the core deUFMylase activity with a defined substrate (TRIP4).

Summary: Records a UFSP2-TRIP4 (Q15650) interaction (the substrate it deUFMylates). The bare protein binding term is uninformative; the functionally meaningful relationship is enzyme-substrate.

Reason: Bare GO:0005515 protein binding carries no functional information, so under the generic protein-binding policy this row is REMOVE. Removal does not mean the reported interaction is false; the summary and any supporting text keep it on record.

Summary: Proposed sub-compartment. UFSP2 is a soluble protease with no transmembrane segment or signal peptide that is held on the cytosolic face of the ER membrane by its integral-membrane partner ODR4, and it acts there on ER-bound 60S ribosomes. The existing annotations stop at the organelle.

Reason: Added under the tethering route of ORGANELLES rule 7 (2026-11-11), beside the three root GO:0005783 rows, which are kept. GO:0098554 is defined to include proteins peripherally associated with the cytoplasmic leaflet and is part_of GO:0005789, so it is the more specific statement of where UFSP2 sits. The route's four conditions are met. (a) The endogenous protein is at the ER: the curated IDA row from PMID:35926457 (UniProt "Endoplasmic reticulum", ECO:0000269). (b) A named integral partner binds it: ODR4 is a conserved tail-anchored transmembrane protein, and ODR-4 and ODR-8/UfSP2 form a complex at the ER membrane, with human ODR4 and UfSP2 also co-immunoprecipitating (PMID:24603482, Chen et al. 2014, overexpressed proteins). (c) Dependence, shown two ways: worm ODR-8 is released from digitonin-permeabilised HeLa cells when expressed alone but partly retained with ODR-4 (PMID:24603482, Fig 8F), a partner-dependent release test that also places it on the cytosolic face, since digitonin opens only the plasma membrane; and endogenous human UFSP2 is markedly reduced in ODR4 knockout cells, which phenocopy UFSP2 knockout for RPL26 UFMylation (PMID:35926457). (d) UniProt annotates no transmembrane, signal-peptide or transit feature, and in the literature worm ODR-8 is diffusely cytosolic when expressed without ODR-4 (PMID:24603482, Fig 8E), so UFSP2 is made in the cytosol and the cytoplasmic leaflet is the only face it can reach, where its substrate, membrane-associated 60S uL24, sits (PMID:38383785). Attribution: the anchoring statements in PMID:35926457 ("thought to anchor") and PMID:38383785 ("membrane anchor ODR4", ref. 2) both rest on Chen et al., so they are not independent corroboration. Caveats - the release test used overexpressed worm proteins in HeLa cells, while the human result is destabilisation rather than release; the cached text of PMID:35926457 describes membrane, cytosol and nuclear fractions but no ER-marker control for UFSP2 itself; and UniProt also lists cytoplasm and nucleus by similarity to mouse. Hence IC rather than IDA, and located_in rather than is_active_in: the is_active_in claim is carried by the root IDA and IBA rows, and the face-level location is an inference from topology and the partner.