SPG21

Existing Annotations Review

GO Term	Evidence	Action	Reason
GO:0005829 cytosol	IBA GO_REF:0000033	ACCEPT	Summary: Maspardin has no transmembrane segment and genuinely partitions between a soluble cytosolic pool and a peripherally membrane-bound pool; the cytosolic assertion is correct. Reason: Correct, and consistent across every study from the original cloning paper to the 2026 work. Uninformative on its own, though - the endosome membrane annotation this gene also carries says where maspardin acts.
GO:0030140 trans-Golgi network transport vesicle	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Propagated from this gene's own 2001 immunofluorescence localisation. Later work places maspardin chiefly on late endosomes and lysosomes. Reason: Not contradicted - a TGN/late-endosomal distribution was reported repeatedly, and the compartments are contiguous - but the site at which maspardin's now-established RAB7A effector function is exercised is the late endosome, not the TGN-derived transport vesicle. Retained as a secondary localisation.
GO:0042609 CD4 receptor binding	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Phylogenetic projection of a real but vertebrate-restricted interaction. The WITH/FROM for this row is PANTHER:PTN002698091 plus UniProtKB:Q9NZD8, i.e. the node was seeded by this gene's own IPI. Reason: The underlying interaction is genuine and I am not overruling it. What is questionable is its scope as a family-level function: CD4 is vertebrate-specific, whereas maspardin is conserved to insects and beyond, so a node assertion of CD4 receptor binding cannot describe the ancestral function of the family. Maspardin is also ubiquitously expressed and its disease is neurological, not immunological, and no follow-up in twenty-five years has built on the CD4 link. Demoted to non-core rather than removed.
GO:0005794 Golgi apparatus	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: Electronic transfer of the UniProt subcellular-location statement. Reason: A coarse parent of the trans-Golgi network transport vesicle annotation and subject to the same caveat: real, but not where the RAB7A effector function operates.
GO:0005829 cytosol	IEA GO_REF:0000044	ACCEPT	Summary: Electronic transfer of the cytosolic localisation. Reason: Correct; maspardin has a substantial soluble cytosolic pool.
GO:0007166 cell surface receptor signaling pathway	IEA GO_REF:0000117	MARK AS OVER ANNOTATED	Summary: An ARBA machine-learning inference that generalises the 2001 CD4 story into a broad signalling process term. Reason: Maspardin is not a component of a cell-surface receptor pathway. Its only tie to receptor signalling is a single unreplicated two-hybrid interaction with the CD4 cytoplasmic tail, and the process term derived from it is both unsupported and too broad to be informative. An electronic generalisation of a narrow, dated observation.
GO:0010008 endosome membrane	IEA GO_REF:0000044	ACCEPT	Summary: The most informative location GOA currently carries for maspardin, and the one the 2025-2026 work supports. Reason: Core location. Both independent 2025-2026 studies place maspardin on endolysosomal membranes, the Bonifacino group via RAB7A-dependent recruitment and the Boonen group via fractionation and colocalisation with late endosomal markers. Supporting Evidence: PMID:40833810 Herein, we report that the SPG21 protein localizes to endolysosomes through interaction with the GTP-bound form of RAB7A. PMID:41400694 It is caused by mutations in the SPG21 gene, which encodes maspardin, a cytosolic protein of unknown function that associates with the late endosomal/lysosomal membrane.
GO:0016020 membrane	IEA GO_REF:0000044	MODIFY	Summary: Correct but maximally general; the specific membrane is now known. Reason: Maspardin is not an integral membrane protein and has no transmembrane segment; it associates peripherally with a specific membrane, the late endosomal/lysosomal one. The bare membrane term should be replaced by the endosome membrane term the gene already carries. Proposed replacements: endosome membrane
GO:0005515 protein binding	IPI PMID:16189514 Towards a proteome-scale map of the human protein-protein in...	REMOVE	Summary: Bare protein binding from a yeast two-hybrid interactome screen. Reason: Carries no functional information. This term, repeated across ten interactome papers, is currently almost the whole of maspardin's molecular-function record in GOA, which is precisely the pattern that makes a protein look annotated while remaining functionally dark. The informative replacement is small GTPase binding (RAB7A). 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:21516116 Next-generation sequencing to generate interactome datasets.	REMOVE	Summary: Bare protein binding from a high-throughput interaction dataset. Reason: Carries no functional information; see the other protein binding rows. 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:22458338 Host-pathogen interactome mapping for HTLV-1 and -2 retrovir...	REMOVE	Summary: Bare protein binding from a host-pathogen interactome screen. Reason: Carries no functional information; see the other protein binding rows. 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:25416956 A proteome-scale map of the human interactome network.	REMOVE	Summary: Bare protein binding from a proteome-scale interactome map. Reason: Carries no functional information; see the other protein binding rows. 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:25910212 Widespread macromolecular interaction perturbations in human...	REMOVE	Summary: Bare protein binding from an interaction-perturbation study. Reason: Carries no functional information; see the other protein binding rows. 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:28514442 Architecture of the human interactome defines protein commun...	REMOVE	Summary: Bare protein binding from an interactome-architecture study. Reason: Carries no functional information; see the other protein binding rows. 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:31515488 Extensive disruption of protein interactions by genetic vari...	REMOVE	Summary: Bare protein binding from a variant-interaction screen. Reason: Carries no functional information; see the other protein binding rows. 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 the HuRI binary interactome reference map. Reason: Carries no functional information; see the other protein binding rows. 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:33961781 Dual proteome-scale networks reveal cell-specific remodeling...	REMOVE	Summary: Bare protein binding from a cell-type-resolved affinity-purification network. Reason: Carries no functional information; see the other protein binding rows. 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:40205054 Multimodal cell maps as a foundation for structural and func...	REMOVE	Summary: Bare protein binding from a multimodal cell-map study. Reason: Carries no functional information; see the other protein binding rows. 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	IDA GO_REF:0000052	ACCEPT	Summary: Immunofluorescence localisation to the cytosol. Reason: Correct; consistent with every other study of maspardin's distribution.
GO:0005829 cytosol	IDA PMID:11113139 Cloning of ACP33 as a novel intracellular ligand of CD4.	ACCEPT	Summary: Experimental cytosolic localisation in the original ACP33 cloning paper. Reason: Correct; the soluble pool is real and substantial.
GO:0030140 trans-Golgi network transport vesicle	IDA PMID:11113139 Cloning of ACP33 as a novel intracellular ligand of CD4.	KEEP AS NON CORE	Summary: Experimental localisation to TGN-derived transport vesicles, reported in 2001. Reason: Experimental and therefore retained; I have not read the full text and am not overruling the curator. But the compartment where maspardin's established RAB7A effector function operates is the late endosome/lysosome, so this is a secondary location rather than the core one.
GO:0042609 CD4 receptor binding	IPI PMID:11113139 Cloning of ACP33 as a novel intracellular ligand of CD4.	KEEP AS NON CORE	Summary: The interaction that named the protein: ACP33 was recovered in a two-hybrid screen with the CD4 cytoplasmic tail, and the interaction maps to CD4's C-terminal hydrophobic residues. Reason: A genuine experimental interaction that has never been retracted, and the only informative molecular-function term maspardin carries in GOA - so it is worth keeping. It is nonetheless not the core function: CD4 is vertebrate-restricted while the protein is conserved far more widely, maspardin is ubiquitously expressed, the disease phenotype is neurological, and no independent replication or mechanistic follow-up has appeared since 2001. The conserved molecular function that has now been established, in two independent laboratories, is binding of GTP-loaded RAB7A. Notably, the same 2001 paper already identified maspardin's hydrolase fold as non-catalytic and repurposed for protein-protein interaction, which is the framing that has held up. Supporting Evidence: PMID:11113139 Furthermore, we demonstrate that interaction with CD4 is mediated by the noncatalytic alpha/beta hydrolase fold domain of ACP33. PMID:11113139 we propose that ACP33 modulates the stimulatory activity of CD4
GO:0050851 antigen receptor-mediated signaling pathway	IC PMID:11113139 Cloning of ACP33 as a novel intracellular ligand of CD4.	KEEP AS NON CORE	Summary: Curator inference from the CD4 interaction: if maspardin modulates CD4's costimulatory activity, it is downstream of the T-cell receptor. Reason: Follows logically from the CD4 result and is not contradicted, but it inherits all of that result's limitations - one screen, one cell context, no replication - and cannot be the function of a ubiquitously expressed, deeply conserved protein whose loss causes spastic paraplegia. Retained as non-core, tied to the CD4 annotation it derives from.
GO:0031267 small GTPase binding	IPI PMID:40833810 The hereditary spastic paraplegia type 21 (SPG21) protein is...	NEW	Summary: Proposed addition, and the informative replacement for the bare protein binding rows: maspardin binds RAB7A, preferentially in its GTP-loaded form, and this is what recruits it to endolysosomes. Reason: Established independently by two laboratories in 2025-2026 - the Bonifacino group by nucleotide-state binding and localisation analysis, the Boonen group by co-immunoprecipitation in control cells - and anticipated by a 2010 interaction report. GO merged the former 'Rab GTPase binding' into GO:0031267, so this is the most specific molecular-function term available. It is a binding term and therefore still leaves the effector output unstated; that residual gap is recorded in knowledge_gaps rather than papered over with a more confident term. Supporting Evidence: PMID:40833810 Herein, we report that the SPG21 protein localizes to endolysosomes through interaction with the GTP-bound form of RAB7A. PMID:40833810 We conclude that SPG21 acts as a RAB7A effector that promotes noncanonical mTORC1-catalyzed phosphorylation of TFEB, thereby suppressing its nuclear localization and transcriptional activity. PMID:41400694 In the present study, we found that maspardin binds to RAB7 in control cells
GO:0005770 late endosome	IDA PMID:41400694 Maspardin/SPG21 controls lysosome motility and TFEB phosphor...	NEW	Summary: Proposed addition. The compartment at which maspardin holds RAB7A, and the one GOA currently represents only through the generic endosome membrane term. Reason: Both 2025-2026 studies localise maspardin to late endosomes and lysosomes, and the mechanism they describe depends on this specific placement: maspardin loss moves RAB7A from retromer- and TBC1D5-positive late endosomes onto lysosomes. Supporting Evidence: PMID:41400694 It is caused by mutations in the SPG21 gene, which encodes maspardin, a cytosolic protein of unknown function that associates with the late endosomal/lysosomal membrane. PMID:41400694 This redistribution decreases the interaction between RAB7 and its GTPase-activating protein (GAP), TBC1D5.
GO:1900181 negative regulation of protein localization to nucleus	IMP PMID:40833810 The hereditary spastic paraplegia type 21 (SPG21) protein is...	NEW	Summary: Proposed addition. Maspardin keeps TFEB phosphorylated and cytoplasmic; its loss drives TFEB into the nucleus. Reason: Demonstrated independently in two laboratories, by CRISPR knockout and by siRNA knockdown, with rescue by re-expression of wild-type but not mutant maspardin, and with the specificity control that canonical mTORC1 substrates are unaffected. This is the clearest process-level consequence of the RAB7A effector function. Supporting Evidence: PMID:40833810 Biochemical studies reveal that SPG21 depletion does not affect phosphorylation of canonical mTORC1 substrates such as ULK1, S6K1, 4E-BP1, but reduces phosphorylation of the noncanonical mTORC1 substrate TFEB. PMID:40833810 We conclude that SPG21 acts as a RAB7A effector that promotes noncanonical mTORC1-catalyzed phosphorylation of TFEB, thereby suppressing its nuclear localization and transcriptional activity.
GO:0032418 lysosome localization	IMP PMID:41400694 Maspardin/SPG21 controls lysosome motility and TFEB phosphor...	NEW	Summary: Proposed addition. Loss of maspardin shifts lysosomes toward anterograde, plus-end-directed movement through excess FYCO1 recruitment by mislocalised GTP-RAB7. Reason: A direct, mechanistically explained consequence of maspardin's control of RAB7 positioning, measured by live-cell tracking in SPG21 knockout cells. Currently unrepresented in GOA. Supporting Evidence: PMID:41400694 Consequently, RAB7 remains primarily GTP-bound, recruiting more FYCO1 to lysosomes and promoting the anterograde movement of these organelles along microtubules.

Gene Description

Existing Annotations Review

Core Functions

References

Suggested Questions for Experts

Suggested Experiments

📚 Additional Documentation

Notes

📄 View Raw YAML