AP5B1 encodes beta-5 adaptin, the large scaffolding subunit of AP-5, the fifth and most recently identified heterotetrameric adaptor protein complex. It is an 878-residue protein built on the beta-adaptin plan, an N-terminal alpha-helical solenoid trunk followed by a C-terminal appendage that retains both the beta-sandwich and the alpha/beta platform subdomains. Its N-terminal quarter binds the medium subunit mu-5 in the same arrangement that beta subunits use in AP-1 and AP-2, and mu-5 protein is destabilised when beta-5 is depleted, so beta-5 holds the complex together. Together with zeta-5, mu-5 and sigma-5 it forms the AP-5 core, and in cells that core is found in a larger, stoichiometric assembly with spatacsin (SPG11) and spastizin (ZFYVE26/SPG15), in which spastizin docks the assembly onto membranes through a FYVE domain that reads phosphatidylinositol 3-phosphate, and spatacsin is thought to provide a clathrin-like scaffold. Recruitment to late endosomes and lysosomes is by coincidence detection of PI3P and Rag GTPases and is enhanced when cells are starved, tying the complex to nutrient signalling. Unlike the beta subunits of AP-1 and AP-2, beta-5 carries no usable clathrin-recruitment apparatus: it lacks the type-II LLDLL box and the tryptophan-based W-box, its single clathrin-box consensus match lies buried inside a folded C-terminal domain rather than in an exposed linker, and the trunk-to-appendage linker that carries the box in AP1B1 and AP2B1 is roughly a third as long. AP-5 correspondingly does not associate with clathrin and is insensitive to brefeldin A. Functionally, AP-5 acts in a retrieval step out of late endosomes back to the Golgi, in parallel with and as a backup to retromer. Depleting beta-5 traps the cation-independent mannose 6-phosphate receptor and the retromer subunit Vps26 in enlarged perinuclear compartments and produces swollen multivesicular bodies with emanating tubules; loss of AP-5 more broadly causes retrieval defects for GOLIM4 and GOLM1 as well, and leads to the accumulation of aberrant multilamellar endolysosomes. The assembled AP-5/SPG11/SPG15 complex binds PI3P, senses membrane curvature and remodels membranes in vitro, and is required to initiate autolysosome tubulation. Bi-allelic loss-of-function and recurrent missense alleles of AP5B1 cause recessive macular dystrophy, and the protein is present in human retinal pigment epithelium in puncta that overlap Rab7-positive late endosomes and, partly, the trans-Golgi network.
Definition: A protein-containing complex consisting of the four subunits of the AP-5 adaptor complex (beta-5, zeta-5, mu-5 and sigma-5) together with spatacsin and spastizin, in approximately 1:1:1:1:1:1 stoichiometry. It is found in the cytosol and on late endosomal and lysosomal membranes, onto which it is recruited by coincidence detection of phosphatidylinositol 3-phosphate, bound by the FYVE domain of spastizin, and of Rag GTPases. It functions as a coat-like assembly in retrieval of cargo out of late endosomes, and remodels endolysosomal membranes.
Justification: This is the form in which AP-5 acts in cells, and GO cannot express it. GO:0044599 covers only the tetrameric core and has no children, while ComplexPortal curates the hexamer as CPX-20045. The absence forces genuine co-membership evidence to be recorded as uninformative GO:0005515 protein binding: on AP5B1 alone, four rows for spatacsin and one for spastizin, each traceable to an experiment showing stoichiometric co-purification rather than a pairwise interaction. The same gap affects AP5Z1, AP5M1, AP5S1, SPG11 and ZFYVE26, and prevents GO-CAM modelling of the retrieval step.
Parent term: protein-containing complex
Mappings:
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005515 protein binding | IPI PMID:20613862 A genome-scale DNA repair RNAi screen identifies SPG48 as a ... | MODIFY | Summary: AP5B1 co-purifies with AP5Z1 (zeta-5), the other large subunit of AP-5. The evidence is reciprocal BAC-tagged immunoprecipitation followed by shotgun LC-MS/MS in which AP5B1, then called DKFZp761E198, was itself one of the baits, so the association is established in both directions rather than as a one-way prey hit. This is one of the two studies that between them defined the four-subunit AP-5 core. Reason: Bare protein binding conveys nothing about what AP5B1 does. The informative content of this row is that AP5B1 is a stoichiometric structural subunit of AP-5: it is the beta-adaptin-fold trunk that holds the complex together, its N-terminal quarter makes the canonical beta-mu contact, and depleting it destabilises mu-5. GO:0005198 structural molecule activity is defined as the action of a molecule that contributes to the structural integrity of a complex, which is exactly this claim and is the most specific molecular function GO currently offers for an adaptor-complex subunit. The complex membership itself is separately captured by GO:0044599. Note for whoever implements this: nine IPI rows on this gene carry the same proposed replacement and should collapse into a single GO:0005198 annotation rather than nine. That collapse also implies a re-code, because IPI is an odd evidence code for structural molecule activity - what actually supports the term is the mapped beta-mu interface (IDA) and the loss of mu-5 protein when beta-5 is depleted (IMP), both from PMID:22022230. Proposed replacements: structural molecule activity Supporting Evidence: PMID:20613862 Glycine eluates from KIAA0415-LAP, KIAA0415-NFLAP, SPG11-LAP, and DKFZp761E198-LAP immunopurifications were used for in-solution digestion and analyzed by shotgun-LC-MS/MS PMID:20613862 Reciprocal immunoprecipitation experiments followed by mass spectrometry analyses of in-gel and in-solution digests confirmed the existence of a protein complex, which consists of at least five core proteins: KIAA0415, SPG11, SPG15, C20orf29, and DKFZp761E198 PMID:22022230 The mapping of the binding site on DKFZp761E198 for C14orf108 to residues 16β229 is also consistent with a Ξ²-ΞΌ-like interaction |
| GO:0005515 protein binding | IPI PMID:20613862 A genome-scale DNA repair RNAi screen identifies SPG48 as a ... | MODIFY | Summary: AP5B1 co-purifies with spastizin (ZFYVE26/SPG15), the sixth member of the native assembly. Spastizin carries the FYVE domain that reads PI3P and docks the coat onto late endosomal and lysosomal membranes, so the membrane-binding determinant of the assembly sits on this partner rather than on AP5B1 itself. Reason: As for the SPG11 rows, the informative annotation is hexamer membership and no GO term exists for it. What can be said about AP5B1 from this association is that it is a structural constituent of the assembly, which GO:0005198 expresses. Recording it this way also keeps the review from implying that AP5B1 binds membranes or PI3P: the PI3P contact is made by the spastizin FYVE domain. Proposed replacements: structural molecule activity Supporting Evidence: PMID:20613862 Glycine eluates from KIAA0415-LAP, KIAA0415-NFLAP, SPG11-LAP, and DKFZp761E198-LAP immunopurifications were used for in-solution digestion and analyzed by shotgun-LC-MS/MS PMID:20613862 Reciprocal immunoprecipitation experiments followed by mass spectrometry analyses of in-gel and in-solution digests confirmed the existence of a protein complex, which consists of at least five core proteins: KIAA0415, SPG11, SPG15, C20orf29, and DKFZp761E198 PMID:33464297 recruitment of AP-5/SPG11/SPG15 is enhanced in starved cells and occurs by coincidence detection, requiring both phosphatidylinositol 3-phosphate (PI3P) and Rag GTPases |
| GO:0005515 protein binding | IPI PMID:20613862 A genome-scale DNA repair RNAi screen identifies SPG48 as a ... | MODIFY | Summary: AP5B1 co-purifies with spatacsin (SPG11). SPG11 is not an AP-5 subunit; it is one of the two additional proteins that make the native assembly a hexamer rather than a tetramer, and it is proposed to be the coat scaffold that AP-5 docks onto. This 2010 study found SPG11 in immunoprecipitates in which AP5B1 was one of the baits. Reason: The annotation that would actually capture this row is membership of the AP-5/SPG11/SPG15 complex, and GO has no term for it: QuickGO and OLS return no GO class for spatacsin, spastizin or the hexamer, and GO:0044599 has no children, while ComplexPortal does curate the assembly as CPX-20045. In the absence of that CC term, the defensible molecular statement is the one shared by all of AP5B1's intra-assembly associations: it is a stoichiometric structural component of a 1:1:1:1:1:1 complex. A proposed term for the hexamer is recorded in proposed_new_terms. Proposed replacements: structural molecule activity Supporting Evidence: PMID:20613862 Glycine eluates from KIAA0415-LAP, KIAA0415-NFLAP, SPG11-LAP, and DKFZp761E198-LAP immunopurifications were used for in-solution digestion and analyzed by shotgun-LC-MS/MS PMID:20613862 Reciprocal immunoprecipitation experiments followed by mass spectrometry analyses of in-gel and in-solution digests confirmed the existence of a protein complex, which consists of at least five core proteins: KIAA0415, SPG11, SPG15, C20orf29, and DKFZp761E198 PMID:23825025 we show that the four AP-5 subunits can be coimmunoprecipitated with SPG11 and SPG15, both from cytosol and from detergent-extracted membranes, with a stoichiometry of βΌ1:1:1:1:1:1 |
| GO:0005515 protein binding | IPI PMID:20613862 A genome-scale DNA repair RNAi screen identifies SPG48 as a ... | MODIFY | Summary: AP5B1 co-purifies with AP5S1 (sigma-5), the small subunit of AP-5, in the same reciprocal BAC-tagged IP-MS experiment, with AP5B1 itself used as one of the baits. AP5S1 was identified in that study under its original name C20orf29. Reason: Same reasoning as the other intra-complex rows: the claim worth recording is that AP5B1 is a structural subunit of the AP-5 heterotetramer, not that it binds an unspecified protein. Knockdown of C20orf29/AP5S1 reproduces the AP5B1 knockdown phenotype, which is the functional counterpart of this physical association. GO:0005198 states the subunit-level contribution; GO:0044599 states the membership. Proposed replacements: structural molecule activity Supporting Evidence: PMID:20613862 Glycine eluates from KIAA0415-LAP, KIAA0415-NFLAP, SPG11-LAP, and DKFZp761E198-LAP immunopurifications were used for in-solution digestion and analyzed by shotgun-LC-MS/MS PMID:20613862 Reciprocal immunoprecipitation experiments followed by mass spectrometry analyses of in-gel and in-solution digests confirmed the existence of a protein complex, which consists of at least five core proteins: KIAA0415, SPG11, SPG15, C20orf29, and DKFZp761E198 PMID:22022230 Fourth, knocking down either KIAA0415/SPG48 or C20orf29 produces the same phenotype as knocking down ΞΌ5 or Ξ²5 |
| GO:0005515 protein binding | IPI PMID:22022230 The fifth adaptor protein complex. | MODIFY | Summary: The founding interaction of the complex, and the best-characterised binding event AP5B1 has. A yeast two-hybrid library screen with full-length mu-5 as bait repeatedly recovered residues 16-229 of AP5B1, the interaction was confirmed against bait and prey controls, and the mapped site corresponds to the N-terminal quarter of the beta subunit that makes most of the beta-mu contacts in the AP-1 and AP-2 structures. Depleting AP5B1 lowers mu-5 protein levels, the classic dependence of an AP medium subunit on its beta partner. Reason: This row should not stay as protein binding: it is the direct, mapped, orientation-correct beta-mu interaction on which the assignment of AP5B1 as the beta subunit of AP-5 rests, and it comes with a stability phenotype rather than only a binary interaction. GO:0005198 structural molecule activity is the informative MF that exists: AP5B1 contributes to the structural integrity of AP-5, with the mapped site and the destabilisation of mu-5 as the evidence. No more specific GO molecular function applies, because AP-5's own cargo-binding activity has never been demonstrated. Proposed replacements: structural molecule activity Supporting Evidence: PMID:22022230 Two clones encoding residues 16β229 of DKFZp761E198, isolated from a human placental cDNA library screen using C14orf108 as bait, were tested for specificity. PMID:22022230 The mapping of the binding site on DKFZp761E198 for C14orf108 to residues 16β229 is also consistent with a Ξ²-ΞΌ-like interaction PMID:22022230 knocking down DKFZp761E198 also decreased the intensity of the C14orf108 band |
| GO:0005515 protein binding | IPI PMID:23825025 Interaction between AP-5 and the hereditary spastic parapleg... | MODIFY | Summary: The study that established the native stoichiometry. All four AP-5 subunits, including AP5B1, coimmunoprecipitate with SPG11 and SPG15 at roughly 1:1:1:1:1:1 from both cytosol and detergent-extracted membranes, the six proteins colocalise on a late endosomal/lysosomal compartment, and knockdown of any of the six produces the same CIMPR trapping phenotype. SPG11's N-terminal beta-propeller-like domain binds AP-5 in vitro. Reason: Real, central and reproducible, but uninformative as protein binding. The target annotation would be CC membership of the AP-5/SPG11/SPG15 hexamer, which GO does not have (see proposed_new_terms); the available informative MF is GO:0005198, AP5B1 contributing to the structural integrity of the assembly it is a stoichiometric member of. Not marked over-annotated: this reference projects onto only four entities across all of GO, and the association is supported by a functional phenocopy, not just a pulldown. Proposed replacements: structural molecule activity Supporting Evidence: PMID:23825025 we show that the four AP-5 subunits can be coimmunoprecipitated with SPG11 and SPG15, both from cytosol and from detergent-extracted membranes, with a stoichiometry of βΌ1:1:1:1:1:1 PMID:23825025 Knockdowns of SPG11 or SPG15 phenocopy knockdowns of AP-5 subunits: all six knockdowns cause the cation-independent mannose 6-phosphate receptor to become trapped in clusters of early endosomes. PMID:29381698 However, AP-5 is predicted to be structurally very similar to APs 1-4, even though in its native form it exists as a heterohexamer rather than a heterotetramer. |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: CAMK2B, a calcium/calmodulin-dependent protein kinase subunit, recovered as a binary yeast two-hybrid hit in a proteome-scale interactome map. There is no follow-up anywhere in the AP-5 literature: CAMK2B is not a subunit of AP-5, is not part of the AP-5/SPG11/SPG15 assembly, and appears in none of the targeted studies of this complex. Reason: Reproducible but uncharacterised high-throughput hit. Querying QuickGO by this reference and paginating fully returns 24,599 annotations across 937 distinct entities, so the row is one cell of a proteome-scale projection rather than a gene-specific finding. Nothing in it constrains AP5B1's function, and promoting it to an informative molecular function would assert a kinase-scaffolding role that no experiment supports. Not removed: the interaction may well be real, and CAMK2B is also recovered by an orthogonal method. Supporting Evidence: PMID:25416956 While currently available information is highly biased and only covers a relatively small portion of the proteome, our systematic map appears strikingly more homogeneous |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: NIF3L1, a NIF3-like protein of unknown function, recovered as a binary yeast two-hybrid hit in the same proteome-scale map. It is supported by one study only; the IntAct NbExp=3 for this pair is three experiments within that study, not three independent reports. Reason: A single high-throughput screen contributing one row of a 937-entity projection, with no orthogonal confirmation and no functional follow-up in either protein's literature. NIF3L1 is absent from every targeted study of AP-5 and of the AP-5/SPG11/SPG15 assembly. Marked over-annotated rather than removed, since a two-hybrid hit is evidence of something even when it is uninterpretable. Supporting Evidence: PMID:25416956 While currently available information is highly biased and only covers a relatively small portion of the proteome, our systematic map appears strikingly more homogeneous file:human/AP5B1/AP5B1-uniprot.txt Q2VPB7; Q9GZT8: NIF3L1; NbExp=3; IntAct=EBI-5917279, EBI-740897; |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | MARK AS OVER ANNOTATED | Summary: The same CAMK2B interaction recovered by a second high-throughput method, affinity-purification mass spectrometry in BioPlex. Two orthogonal screens agreeing makes the physical interaction more credible, but neither supplies any functional context, and the UniProt IntAct NbExp=4 for this pair counts experiments, not independent studies. Reason: Same judgement as the two-hybrid row. This reference projects onto 2,196 distinct entities across 9,514 annotations, so the row carries no gene-specific information. Recording it as an over-annotation rather than removing it preserves the observation while keeping it out of the functional picture: no study of AP-5 has ever reported a CaMKII connection, and AP5B1 has no kinase-related domain or motif. Supporting Evidence: PMID:33961781 The first, BioPlex 3.0, results from affinity purification of 10,128 human proteins-half the proteome-in 293T cells and includes 118,162 interactions among 14,586 proteins. file:human/AP5B1/AP5B1-uniprot.txt Q2VPB7; Q13554: CAMK2B; NbExp=4; IntAct=EBI-5917279, EBI-1058722; |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | MODIFY | Summary: AP5B1-SPG11 recovered independently in BioPlex, a proteome-scale affinity-purification mass spectrometry survey. Unlike the other two BioPlex-derived rows on this gene, this one names a partner that targeted work has established four separate times, and the IntAct record for the pair carries NbExp=8 across those studies. Reason: The reference is a 2,196-entity projection, so on its own it would be a high-throughput hit. It is not treated as one because the partner is independently established: AP5B1 and SPG11 are stoichiometric members of the same complex by reciprocal IP-MS, by quantitative coimmunoprecipitation, by phenocopy of knockdown, and by cryo-EM. The informative replacement is therefore the same structural-subunit statement as the other intra-assembly rows rather than a downgrade. Proposed replacements: structural molecule activity Supporting Evidence: PMID:33961781 Through affinity-purification mass spectrometry, we have created two proteome-scale, cell-line-specific interaction networks. file:human/AP5B1/AP5B1-uniprot.txt Q2VPB7; Q96JI7: SPG11; NbExp=8; IntAct=EBI-5917279, EBI-2822128; PMID:23825025 we show that the four AP-5 subunits can be coimmunoprecipitated with SPG11 and SPG15, both from cytosol and from detergent-extracted membranes, with a stoichiometry of βΌ1:1:1:1:1:1 |
| GO:0005515 protein binding | IPI PMID:40175557 Structural basis for membrane remodeling by the AP5-SPG11-SP... | MODIFY | Summary: The partner here is the mouse orthologue of mu-5 (Ap5m1, 490 aa, the same length as human AP5M1), recorded by IntAct as a cross-species interaction from the cryo-EM study of the AP5-SPG11-SPG15 assembly. AP5B1 residues 1-878 are chain B of the deposited structures 8YAB and 8YAH, so this is the same beta-mu contact seen in the 2011 two-hybrid work, now within a reconstituted and imaged complex. Reason: A cross-species reconstitution of the beta-mu contact is still evidence that AP5B1 is the structural subunit that holds mu-5 in the complex, and GO:0005198 is the informative molecular function for that. Kept as a MODIFY rather than downgraded: the structure is a targeted, low-projection study (10 distinct entities annotated from this reference across all of GO), not a high-throughput screen. The cached entry for this paper is abstract-only, so the review does not assert which AP-5 subunit contacts SPG11 or contributes the curvature-sensing surface. Proposed replacements: structural molecule activity Supporting Evidence: PMID:40175557 The AP5 complex is in a super-open conformation. file:human/AP5B1/AP5B1-uniprot.txt DR PDB; 8YAB; EM; 3.26 A; B=1-878. PMID:22022230 The mapping of the binding site on DKFZp761E198 for C14orf108 to residues 16β229 is also consistent with a Ξ²-ΞΌ-like interaction |
| GO:0005515 protein binding | IPI PMID:40175557 Structural basis for membrane remodeling by the AP5-SPG11-SP... | MODIFY | Summary: AP5B1-SPG11 from the cryo-EM study of the assembled AP5-SPG11-SPG15 complex, in which AP5B1 is chain B of both deposited structures. The paper reports that the N-terminal region of SPG11 is required for AP-5 interaction and assembly, and that AP-5 adopts a super-open conformation within the assembly. Reason: Structural evidence for membership in the hexamer, recorded as protein binding for want of a GO term for that complex. Replaced with the structural-subunit molecular function on the same grounds as the other intra-assembly rows. The cached entry for this paper is abstract-only, so this review does not claim that AP5B1 is the subunit SPG11 contacts; the abstract attributes the requirement to SPG11's N-terminal region and describes the interaction at the level of the AP-5 complex. Proposed replacements: structural molecule activity Supporting Evidence: PMID:40175557 the N-terminal region of SPG11 is required for AP5 complex interaction and assembly PMID:40175557 The AP5 complex is in a super-open conformation. file:human/AP5B1/AP5B1-uniprot.txt DR PDB; 8YAB; EM; 3.26 A; B=1-878. |
| GO:0005764 lysosome | NAS PMID:40175557 Structural basis for membrane remodeling by the AP5-SPG11-SP... | ACCEPT | Summary: Lysosomal localisation of the AP-5 complex, asserted by ComplexPortal for CPX-5181 on the basis of the cryo-EM study. It is independently supported for AP5B1 itself: an anti-AP5B1 antibody gives punctate staining in human retinal pigment epithelium that overlaps Rab7-positive structures, and the complex colocalises on a late endosomal/lysosomal compartment. Reason: Correct and core. AP-5 is a peripherally associated coat component that cycles between cytosol and the endolysosomal surface, so located_in lysosome is the standard treatment. The NAS code understates the evidence available in 2025: subunit-specific immunofluorescence now exists, which is why this review treats the location as established rather than asserted. Supporting Evidence: PMID:40175557 These studies provide insights into the structure and function of the spastic paraplegia AP5-SPG11-SPG15 complex, which is essential for the initiation of autolysosome tubulation. PMID:40081374 Immunofluorescence imaging of those sections confirmed the punctate staining patterns of AP5Z1, AP5M1, and AP5B1, revealing their strongest co-localization with Rab7, a marker of late endosomes, and partial overlap with TGN46-positive Golgi-derived vesicles PMID:23825025 In addition, AP-5, SPG11, and SPG15 colocalize on a late endosomal/lysosomal compartment. |
| GO:0005765 lysosomal membrane | HDA PMID:17897319 Integral and associated lysosomal membrane proteins. | ACCEPT | Summary: AP5B1 was detected in purified human placental lysosomal membranes in 2007, four years before anyone knew what the protein was; it was one of the twelve novel proteins of unknown function reported in that study. The localisation has since been confirmed three times by targeted work on the assembled complex and, most recently, on AP5B1 itself. Reason: This is a broad proteomics projection - querying QuickGO by this reference and paginating fully returns 246 annotations across 242 distinct entities - so on its own it would be a candidate for MARK_AS_OVER_ANNOTATED. It is accepted instead because it is corroborated for this specific protein by independent, targeted evidence spanning 2013 to 2025, and because a peripherally associated coat subunit recovered in a lysosomal membrane fraction is exactly what AP-5 biology predicts. Retained as core: lysosomal membrane is where AP5B1 acts, and it appears in core_functions locations. Supporting Evidence: PMID:17897319 In membranes purified from placental lysosomes, we identified 58 proteins, known to reside at least partially in the lysosomal membrane. PMID:17897319 Among these, 12 novel proteins of unknown functions were found. PMID:23825025 In addition, AP-5, SPG11, and SPG15 colocalize on a late endosomal/lysosomal compartment. |
| GO:0005770 late endosome | NAS PMID:22022230 The fifth adaptor protein complex. | ACCEPT | Summary: Late endosomal localisation of AP-5, asserted by ComplexPortal for CPX-5181 from the paper that discovered the complex. In that paper the compartment was defined through tagged mu-5, which colocalises substantially with LAMP1; tagged AP5B1 could not be localised at the time. Reason: Correct and core. The compartment assignment has held up across every subsequent study of the complex and has now been confirmed for AP5B1 protein directly, by anti-AP5B1 immunofluorescence in human RPE showing strongest overlap with Rab7. The caveat worth recording rather than hiding is that the original assignment was made on a partner subunit; it is no longer the only evidence. Supporting Evidence: PMID:22022230 Tagged AP-5 localises to a late endosomal compartment in HeLa cells. PMID:22022230 substantial colocalisation can be seen between tagged C14orf108 and LAMP1, a protein associated with late endosomes and lysosomes PMID:40081374 Immunofluorescence imaging of those sections confirmed the punctate staining patterns of AP5Z1, AP5M1, and AP5B1, revealing their strongest co-localization with Rab7, a marker of late endosomes, and partial overlap with TGN46-positive Golgi-derived vesicles |
| GO:0005770 late endosome | NAS PMID:40175557 Structural basis for membrane remodeling by the AP5-SPG11-SP... | ACCEPT | Summary: The same late endosomal assignment, asserted by ComplexPortal from the 2025 cryo-EM study rather than from the 2011 discovery paper. A duplicate GO term from a second reference is not a problem; it records that the localisation was reasserted on independent grounds fourteen years later. Reason: Correct and core, for the same reasons as the row from PMID:22022230. The recruitment mechanism is now known - coincidence detection of PI3P and Rag GTPases, enhanced by starvation - which explains why the complex is found on late endosomes and lysosomes rather than earlier compartments. Supporting Evidence: PMID:40175557 SPG11-SPG15 can cooperate with the fifth adaptor protein complex (AP5) involved in membrane sorting of late endosomes. PMID:33464297 GDP-locked RagC promotes recruitment of AP-5/SPG11/SPG15, while GTP-locked RagA prevents its recruitment. |
| GO:0007040 lysosome organization | NAS PMID:40175557 Structural basis for membrane remodeling by the AP5-SPG11-SP... | ACCEPT | Summary: A role in lysosome organisation, asserted by ComplexPortal from the cryo-EM study, which reports that the AP5-SPG11-SPG15 complex binds PI3P, senses membrane curvature and drives membrane remodelling in vitro, and is essential for initiating autolysosome tubulation. The loss-of-function counterpart is the accumulation of aberrant multilamellar endolysosomes in AP-5-deficient cells and patient fibroblasts. Reason: Well supported at the level of the complex, from both directions: a gain-side in vitro membrane-remodelling activity and a loss-side storage phenotype severe enough to be classified as a lysosomal storage disease. GO:0007040 covers assembly, arrangement of parts and disassembly of a lysosome, which is the right scope for autolysosome tubulation and lysosome reformation. Accepted rather than made more specific: the abstract-only cache for the structural paper does not let me establish which subunit provides the curvature-sensing surface, and no more specific GO term matches autophagic lysosome reformation. Supporting Evidence: PMID:40175557 Our findings reveal that the AP5-SPG11-SPG15 complex can bind PI3P molecules, sense membrane curvature and drive membrane remodeling in vitro. PMID:40175557 These studies provide insights into the structure and function of the spastic paraplegia AP5-SPG11-SPG15 complex, which is essential for the initiation of autolysosome tubulation. PMID:26085577 Loss of AP-5 results in accumulation of aberrant endolysosomes: defining a new type of lysosomal storage disease. |
| GO:0016192 vesicle-mediated transport | NAS PMID:22022230 The fifth adaptor protein complex. | MODIFY | Summary: Vesicle-mediated transport, asserted by ComplexPortal for CPX-5181. True but the least informative statement that could be made about this gene: GO:0016192 is the grandparent of GO:0016197 endosomal transport, which this gene already carries by IMP, IBA and IEA. Reason: Too general to convey anything the gene does not already carry more specifically. The pathway is now known: AP-5 acts in retrieval out of late endosomes back to the Golgi, established by CRISPR knockout of AP5Z1 with subcellular fractionation profiling and quantitative mass spectrometry, and by the CIMPR trapping phenotype shared by all six subunit knockdowns including AP5B1's. GO:0042147 retrograde transport, endosome to Golgi is a verified descendant of GO:0016197 and is the level this row should sit at. Proposed replacements: retrograde transport, endosome to Golgi Supporting Evidence: PMID:29381698 Immunolocalisation showed that loss of AP-5 led to impaired retrieval of the cation-independent mannose 6-phosphate receptor (CIMPR), GOLIM4, and GOLM1 from endosomes back to the Golgi region. PMID:29381698 Together, our findings suggest that AP-5 functions in a novel sorting step out of late endosomes, acting as a backup pathway for retromer. |
| GO:0016197 endosomal transport | IBA GO_REF:0000033 | ACCEPT | Summary: A self-referential IBA, and correctly so. The PAINT slice for PTHR34033 contains exactly one node, PTN002145545, placed at taxon:2759 (Eukaryota), carrying two IBD assertions, and the only seed of both is UniProtKB:Q2VPB7 - human AP5B1 itself. The two terms are precisely the two experimental annotations AP5B1 held in 2017, the IMP for this term and the IDA for GO:0030119, both from PMID:22022230. The target appearing in its own WITH/FROM is the expected signature of a node seeded by the target's own experimental evidence. Reason: Kept at the general term, deliberately and divergently from the IMP row for the same term, which is refined to GO:0042147. The direction of the transport step is established in human HeLa cells only; propagating retrograde endosome-to-Golgi transport to every eukaryotic AP5B1 orthologue would assert more than the evidence supports, whereas 'endosomal transport' is exactly the level at which the ancestral claim is safe. That is a deliberate divergence of action on one GO term, made on biological grounds rather than by evidence code. Correct and core. What the IBA adds over the IMP it derives from is a phylogenetic judgement: the curator placed the assertion at the eukaryotic root of the family rather than on the human branch, i.e. endosomal transport is the ancestral function of the AP5B1 clade rather than a primate innovation. The comparative genomics in the same paper supports that placement, with AP-5 subunits found in all five eukaryotic supergroups. Human AP5B1 is trivially inside the inheriting clade, since it is the seed; there is no IRD or IKR anywhere in the family; and the four other reviewed members of PTHR34033:SF1 are full-length orthologues of 876-883 residues with no sign of degeneration. A one-donor WITH/FROM here is not weak evidence, it is the whole family's experimental grounding. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002145545 Β· PTHR34033 AP-5 COMPLEX SUBUNIT BETA-1 IBD node at taxon:2759 SUPPORTS TRANSFER The only node in the PTHR34033 PAINT slice, placed at taxon:2759 (Eukaryota) and carrying exactly two IBD assertions, of which this is one. Its seed list is a single entry, UniProtKB:Q2VPB7, so the node encodes the judgement that AP5B1's experimentally demonstrated endosomal transport role is ancestral to the whole eukaryotic AP5B1 clade. Human AP5B1 is inside that clade by construction. UniProtKB:Q2VPB7 Β· AP5B1 (human, the target) SUPPORTS TRANSFER The target's own accession, and the sole gene-level seed of the IBD. Its IMP for this term (PMID:22022230, four independent siRNAs causing CIMPR and Vps26 to accumulate in enlarged puncta) is the descendant evidence the PAINT curator used. Expected and valid, not circular: it marks that the family's experimental grounding sits on the target itself. Supporting Evidence: file:human/AP5B1/AP5B1-uniprot.txt DR PANTHER; PTHR34033; AP-5 COMPLEX SUBUNIT BETA-1; 1. PMID:22022230 AP-5 subunits can be found in all five eukaryotic supergroups, but they have been co-ordinately lost in many organisms. PMID:22022230 Thus, AP-5 is an evolutionarily ancient complex, which is involved in endosomal sorting, and which has links with hereditary spastic paraplegia. |
| GO:0016197 endosomal transport | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from IPR038741 to endosomal transport. I checked the source: the InterPro API returns IPR038741 as a family-type entry named 'AP-5 complex subunit beta-1' whose only GO mapping is GO:0016197 and whose description reads 'AP5B1 is part of the AP-5 adaptor protein complex, which may be involved in endosomal transport'. UniProt carries the signature on this entry as IPR038741, AP5B1. Reason: A correct, appropriately hedged, family-specific mapping. IPR038741 is the AP5B1 family signature rather than a pan-adaptin one, so there is no risk of the promiscuous-mapping failure mode: nothing that is not an AP5B1 orthologue matches it. The term is general, but that is the right level for a signature-based inference, and it agrees with the gene's own IMP. Accepted as-is; the more specific retrograde term is proposed only on the rows that carry direct experimental evidence for the direction, so this row and the IBA for the same term are intentionally ACCEPT while the IMP is MODIFY. Refining a family signature mapping to a direction demonstrated in one human cell line would be a change to InterPro's model, not to this gene's annotation. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR038741 Β· InterPro AP5B1 family signature SUPPORTS TRANSFER Verified at the InterPro API: a family-type entry named 'AP-5 complex subunit beta-1' whose sole GO mapping is GO:0016197 and whose description is 'AP5B1 is part of the AP-5 adaptor protein complex, which may be involved in endosomal transport'. Family-specific rather than pan-adaptin, so the mapping cannot reach non-orthologues. Supporting Evidence: file:human/AP5B1/AP5B1-uniprot.txt DR InterPro; IPR038741; AP5B1. PMID:22022230 Together with the C14orf108 localisation data, these observations point to a role for the C14orf108-DKFZp761E198 complex in endosomal trafficking. |
| GO:0016197 endosomal transport | IMP PMID:22022230 The fifth adaptor protein complex. | MODIFY | Summary: The AP5B1-specific loss-of-function evidence. Four independent siRNAs against AP5B1, and the SMARTpool, cause the cation-independent mannose 6-phosphate receptor and the retromer subunit Vps26 to accumulate in enlarged perinuclear puncta, phenocopying mu-5 knockdown, with swollen multivesicular bodies bearing emanating tubules by electron microscopy. In 2011 the direction of the transport defect could not be determined, so the annotation was made at the general term. Reason: The essence is right and the evidence is this gene's own, but the term is too general given what is now known. The direction was resolved by CRISPR knockout of AP-5 with spatial proteomics: AP-5 acts in retrieval from late endosomes back to the Golgi, with CIMPR, GOLIM4 and GOLM1 as the affected cargo and retromer as the parallel pathway. GO:0042147 retrograde transport, endosome to Golgi is a descendant of GO:0016197 - verified against the QuickGO ancestors endpoint - so this is a refinement of the same claim, made on the same CIMPR readout that the AP5B1 knockdown produced. Proposed replacements: retrograde transport, endosome to Golgi Supporting Evidence: PMID:22022230 Double labelling for the CIMPR and the retromer subunit Vps26 in control cells (a), cells depleted of C14orf108 (b), and cells depleted of DKFZp761E198 (c). PMID:22022230 The finding that the DKFZp761E198 knockdown phenocopies the C14orf108 knockdown provides further evidence not only that the knockdown phenotypes are specific, but also that the interaction between C14orf108 and DKFZp761E198 is physiologically relevant. PMID:29381698 Immunolocalisation showed that loss of AP-5 led to impaired retrieval of the cation-independent mannose 6-phosphate receptor (CIMPR), GOLIM4, and GOLM1 from endosomes back to the Golgi region. |
| GO:0030119 AP-type membrane coat adaptor complex | IBA GO_REF:0000033 | ACCEPT | Summary: The second of the two assertions on node PTN002145545, seeded by the same single donor, human AP5B1 itself, and mirroring AP5B1's 2017 IDA for the same term from PMID:22022230. Complex membership propagated at the eukaryotic root of a family whose every member is an AP-5 beta subunit. Reason: Correct as a family-level claim. The one thing worth checking here is whether AP-1/AP-2/AP-3/AP-4 beta-subunit terms could leak onto AP5B1 through the tree, and they cannot: AP1B1, AP2B1, AP3B1 and AP4B1 are all in PTHR11134, a 17,626-protein beta-adaptin family, while AP5B1 is the sole occupant of PTHR34033 with 1,227 proteins and one subfamily. No clathrin-binding term from the classical beta adaptins can reach this gene by descent. Kept at the parent rather than modified to GO:0044599: an IBA is deliberately conservative, the node reflects the seed's own 2017 term, and propagating the human-defined four-subunit complex to every eukaryote in the family would assert more than the node was placed to assert. The corresponding human-specific IDA is modified to GO:0044599 instead. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002145545 Β· PTHR34033 AP-5 COMPLEX SUBUNIT BETA-1 IBD node at taxon:2759 SUPPORTS TRANSFER The same single eukaryote-root node, carrying its second assertion. PTHR34033 contains only AP5B1 orthologues - 1,227 proteins, one subfamily SF1, five reviewed members spanning human, mouse, rat, cow and Xenopus at 876-883 residues - so the node has no paralogous subfamily it could mis-propagate into, and no beta-adaptin from PTHR11134 is downstream of it. UniProtKB:Q2VPB7 Β· AP5B1 (human, the target) SUPPORTS TRANSFER The target, and the sole gene-level seed. Its IDA for this term from PMID:22022230 is the descendant evidence behind the IBD. Self-reference here is the correct signature of a node placed on the target's own experimental annotation. Supporting Evidence: file:human/AP5B1/AP5B1-uniprot.txt DR PANTHER; PTHR34033:SF1; AP-5 COMPLEX SUBUNIT BETA-1; 1. file:human/AP5B1/AP5B1-bioinformatics/RESULTS.md AP1B1, AP2B1, AP3B1 and AP4B1 are all `PTHR11134` (ADAPTOR COMPLEX SUBUNIT BETA FAMILY MEMBER, 17,626 proteins); AP5B1 is the sole occupant of `PTHR34033` file:human/AP5B1/AP5B1-bioinformatics/RESULTS.md PAINT therefore cannot leak AP-1/2/3/4 beta-subunit terms onto AP5B1 through the family tree. |
| GO:0030119 AP-type membrane coat adaptor complex | IDA PMID:22022230 The fifth adaptor protein complex. | MODIFY | Summary: Direct assay evidence that AP5B1 is part of an AP-type membrane coat adaptor complex, from the paper that established the complex: a mapped direct interaction with mu-5, destabilisation of mu-5 on AP5B1 knockdown, a matching knockdown phenotype, and co-purification with the other two subunits. Reason: The essence is right but the term is a level too general now that the specific term exists. GO:0044599 AP-5 adaptor complex is a child of GO:0030119 whose definition explicitly names beta5, and it is already on this gene from ComplexPortal with the weaker NAS code. The strongest evidence should carry the most specific term, so this IDA is the row that should be moved. Note also that GO:0030119's definition describes AP complexes as linking clathrin or another coat-forming molecule to a membrane - AP-5 does not use clathrin, which GO:0044599 acknowledges and GO:0030119 does not. Proposed replacements: AP-5 adaptor complex Supporting Evidence: PMID:22022230 Two clones encoding residues 16β229 of DKFZp761E198, isolated from a human placental cDNA library screen using C14orf108 as bait, were tested for specificity. PMID:22022230 knocking down DKFZp761E198 also decreased the intensity of the C14orf108 band PMID:22022230 The lack of any identifiable clathrin binding sites in Ξ²5 and the lack of colocalisation between tagged AP-5 and clathrin also indicate that AP-5 is clathrin-independent. |
| GO:0044599 AP-5 adaptor complex | NAS PMID:22022230 The fifth adaptor protein complex. | ACCEPT | Summary: Membership of the AP-5 adaptor complex - the defining fact about this gene product. The cited paper is the one that identified AP5B1 as a beta adaptin, showed it binds mu-5 directly, and proposed the name AP-5 for the complex they form. ComplexPortal curates it as CPX-5181 with 1:1:1:1 stoichiometry. Reason: Core, correct and at the right level of specificity. GO:0044599's definition names beta5 as one of the four subunits, so the term was written with this protein in it. The only thing wrong with the row is its NAS evidence code, which understates the direct experimental support in the cited paper; the IDA row for the parent term GO:0030119 from the same reference is modified to this term for that reason. Supporting Evidence: PMID:22022230 Therefore, we suggest that C14orf108 and DKFZp761E198 should be renamed ΞΌ5 and Ξ²5, respectively, and that the complex that they form should be called AP-5. PMID:22022230 In addition, C14orf108 binds to a previously uncharacterised protein, DKFZp761E198, which is homologous to the Ξ²-adaptins and has an almost identical predicted secondary structure. file:human/AP5B1/AP5B1-uniprot.txt DR ComplexPortal; CPX-5181; AP-5 Adaptor complex. |
| GO:0001895 retina homeostasis | IMP PMID:40081374 Bi-allelic variants in three genes encoding distinct subunit... | NEW | Summary: AP5B1 is a macular dystrophy gene, established since the GOA snapshot and annotated nowhere. Bi-allelic AP5B1 variants cause recessive inherited retinal disease in three independent cohorts: two families in the 2025 study that first linked three AP-5 subunit genes to hereditary macular dystrophy, five probands in an independent 2026 replication, and 22 affected individuals from 20 families in a 2026 cohort defined by the recurrent p.Leu785Pro founder allele. The AP5B1 alleles in the first cohort are truncating and remove more than two-thirds of the protein. The same study localised AP5B1 protein in human retinal pigment epithelium with a subunit-specific antibody, where it is punctate and overlaps Rab7 and TGN46. Reason: A well-evidenced human loss-of-function phenotype with a matching expression and localisation pattern in the affected tissue, from three independent groups, and entirely absent from GOA. Coded IMP against human variant evidence, which is the appropriate code here because the evidence is human rather than a model organism. Stated plainly so the basis is not mistaken later: the IMP rests on human genetic association - bi-allelic truncating and founder missense alleles segregating with the phenotype across three cohorts - plus subunit-specific localisation in the affected tissue, not on an engineered mutant or a knockout cell line. That is standard GOA practice for inherited retinal disease genes, and a curator who prefers a weaker code should read it as such rather than as an unsupported claim. This is deliberately recorded as a tissue-level, non-core consequence rather than a molecular core function: what AP5B1 does is contribute to endolysosomal retrieval as a structural subunit of AP-5, and RPE degeneration is what happens downstream when that fails in a post-mitotic, phagocytosis-loaded cell type. Noted for completeness: unlike AP5Z1, which causes spastic paraplegia SPG48, AP5B1 has not been reported as a hereditary spastic paraplegia gene, and p.Leu785Pro falls inside the C-terminal appendage domain PF21590 (781-873). Supporting Evidence: PMID:40081374 New WES, new WGS screening, or a reanalysis of previous sequencing data by using a customized analytical pipeline led to the identification of assortments of bi-allelic variants in AP5Z1 (14 families), AP5M1 (three families), and AP5B1 (two families) PMID:40081374 they all remove more than two-thirds of the protein's full sequence and are therefore extremely likely to represent full LoF alleles PMID:40081374 Immunostaining of RPE cells revealed a punctate pattern of AP5Z1, AP5M1, and AP5B1 staining and co-localization with markers of late endosomes and the Golgi, suggesting a role of AP-5 in the normal physiology of this tissue. PMID:42568187 Here, we describe 22 affected individuals from 20 families with AP5B1-associated IRD, all carrying the recurrent missense variant c.2354T>C PMID:41830174 Our findings provide additional evidence that AP5Z1 and AP5B1 are critical for retinal function, and their deficiency leads to retinal degeneration, with or without extra-retinal phenotypes. |
Loading supporting contentβ¦
Download this section (compressed HTML)Q: Does AP-5 recognise a sorting signal in cargo directly, or only through spastizin and spatacsin? Fifteen years after the complex was described, no motif and no direct cargo contact has been reported, and the one proposed link runs through spastizin.
Suggested experts: Jennifer Hirst, Margaret S. Robinson
Q: Why has no accessory protein been found for the beta-5 appendage, which uniquely retains both the beta-sandwich and the platform subdomain, when the equivalent domains of the AP-1 and AP-2 beta subunits are among the busiest interaction platforms in membrane traffic?
Suggested experts: Margaret S. Robinson, Jennifer Hirst
Q: Why do bi-allelic null alleles of AP5B1 give macular dystrophy without the spastic paraplegia seen with AP5Z1 loss, when the two proteins are obligate partners in the same complex?
Suggested experts: Carlo Rivolta, Petra Liskova, Rui Chen
Q: Which subunit of the AP-5/SPG11/SPG15 assembly contributes the curvature-sensing surface that drives membrane remodelling in vitro, and is beta-5 part of it or purely structural?
Suggested experts: Goran Stjepanovic, Min-Yuan Su
Experiment: Express and purify the AP-5 core and the full AP-5/SPG11/SPG15 assembly, and test binding to the cytosolic tails of CIMPR, GOLIM4, GOLM1 and sortilin by ITC or biolayer interferometry. Use a mu-5 pocket mutant, and an AP5B1 subunit-deletion complex, as specificity controls. A positive result would license GO:0140312 cargo adaptor activity for the complex, which the present evidence does not.
Hypothesis: AP-5 binds cargo directly through a mu-5 pocket homologous to the tyrosine-motif pocket of mu-1 and mu-2, and the retrieval defect on AP5B1 loss reflects failure of that recognition rather than failure of coat assembly alone.
Type: in vitro reconstitution and binding
Experiment: Use the isolated beta-5 appendage (approximately residues 690-878) as a GST or biotinylated bait against brain and iPSC-RPE lysate with quantitative mass spectrometry, running the p.Leu785Pro appendage as a paired bait. In parallel, express full-length wild-type and p.Leu785Pro AP5B1 in AP5B1-null cells and measure protein stability, AP-5 assembly by blue-native PAGE, and rescue of CIMPR retrieval. This distinguishes a hypomorphic interface allele from a folding null.
Hypothesis: The beta-5 appendage is a functional accessory-protein docking platform, and p.Leu785Pro disrupts it rather than destabilising the whole protein.
Type: affinity proteomics plus structure-function rescue
Experiment: Generate isogenic null lines for each of AP5Z1, AP5B1, AP5M1 and AP5S1 in both iPSC-RPE and iPSC-derived cortical neurons. Measure, for each, the abundance and assembly state of the remaining subunits and of SPG11 and SPG15, CIMPR retrieval, and endolysosomal morphology by electron microscopy. If all four nulls are equivalent in both cell types, the phenotype difference is not cell-autonomous complex biology and must be sought in expression or in modifier alleles.
Hypothesis: The subunit-specific phenotype spectrum reflects quantitative differences in residual complex assembly rather than moonlighting functions of individual subunits.
Type: isogenic knockout panel with proteomic and morphological readouts
Experiment: In AP5B1-null iPSC-RPE monolayers fed photoreceptor outer segments, follow autolysosome tubulation by live imaging after a starvation-refeeding protocol, with SPG15-GFP as the recruitment reporter and LAMP1 tubule counts as the readout. Compare with wild-type and with AP5Z1-null lines. This tests whether the disease-relevant defect is the reformation step identified in vitro, in the cell type that actually degenerates.
Hypothesis: AP-5 recruitment and autolysosome tubulation are the step that fails first in AP5B1-null retinal pigment epithelium, before any degeneration is visible.
Type: live-cell imaging in a disease-relevant cell model
What is not known β curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: GO has no cellular component term for the AP-5/SPG11/SPG15 complex, so the annotation that would correctly describe five of AP5B1's twelve protein-binding rows cannot be made.
OPEN ONTOLOGYCURATION CC_DARK
What is known: The complex is not hypothetical and not obscure. Its composition and 1:1:1:1:1:1 stoichiometry were published in 2013, it has a cryo-EM structure, its recruitment mechanism is characterised, and it is the form in which AP-5 is understood to act. It is curated in ComplexPortal as CPX-20045 "AP5-Spastizin-spatacsin complex", built from CPX-5181 (AP-5) plus CPX-26503 (spastizin-spatacsin), citing PMID:40175557, PMID:25365221 and PMID:23825025. GO represents only the tetrameric core, GO:0044599, which has no children; QuickGO and OLS searches for spatacsin, SPG11 and spastizin-spatacsin return no GO class, and CPX-20045 carries no GO cross-reference.
Significance: The gap has a visible cost in this gene's annotations: the AP5B1-SPG11 and AP5B1-SPG15 interactions are recorded four and one times respectively as bare GO:0005515 protein binding, with no way to express that they are co-membership of a defined complex. The same gap affects SPG11 and ZFYVE26, and it blocks any GO-CAM representation of the coat.
What would resolve it: Create a child of GO:0032991 for the hexamer, as proposed in proposed_new_terms, with an exact mapping to ComplexPortal CPX-20045, and reannotate the five co-membership rows as part_of that term.
Provenance (the field's own admissions):
Gap: No ligand or function is known for the AP5B1 appendage domain, which is the most complete appendage in the family and the location of the recurrent disease allele.
OPEN BIOLOGY MF_DARK
What is known: Established: AP5B1's appendage contains both the beta-sandwich subdomain and the alpha/beta platform subdomain, unlike the AP-4 beta subunit, which has only the platform, or the AP-1 gamma subunit, which has only the sandwich; the corresponding domains in the AP-1 and AP-2 beta subunits are docking platforms for accessory proteins. Also established: the recurrent founder allele p.Leu785Pro falls inside this region, within Pfam PF21590 (781-873). Not established: any binding partner for it, or any consequence of the substitution beyond the genetic association.
Significance: Retaining both appendage subdomains is a conspicuous feature for a complex that has no known accessory proteins. If the appendage has ligands, they would be the missing regulatory layer of AP-5; if it does not, the retention is a puzzle. Either answer bears directly on how a missense change in it causes macular dystrophy while truncating alleles cause the same phenotype.
What would resolve it: Appendage-domain pulldowns from brain or RPE lysate, as were used to find the AP-1 and AP-2 ear ligands, with the p.Leu785Pro appendage as a paired bait; and a stability and complex-assembly assay comparing wild-type and p.Leu785Pro AP5B1.
Provenance (the field's own admissions):
Gap: Why AP-5 subunit loss produces a retinal phenotype in some genes and a spastic paraplegia phenotype in others is unexplained, and for AP5B1 specifically no spastic paraplegia has been reported despite bi-allelic null alleles.
NARROWING BIOLOGY BP_DARK
What is known: Established: AP5Z1 loss of function causes hereditary spastic paraplegia SPG48, and bi-allelic loss of function in AP5Z1, AP5M1 and AP5B1 causes recessive macular dystrophy. Established for AP5B1 in particular: at least five distinct truncating alleles across three cohorts, plus the recurrent p.Leu785Pro, all presenting with retinal disease. The subunits are obligate members of one complex, so a purely complex-level model predicts one shared phenotype, and that is not what is observed.
Significance: A subunit-specific phenotype in an obligate complex implies either a function outside the complex for one or more subunits, or quantitative differences in residual complex assembly between subunit nulls. Distinguishing these would change how the gene's core functions should be scoped.
What would resolve it: Side-by-side null cell lines for each of the four subunits, in RPE and in neurons, with quantitative assembly of the residual complex measured and the CIMPR retrieval assay as a common readout.
Provenance (the field's own admissions):
Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)