AAMDC is a small cytoplasmic Mth938-domain protein whose biochemical activity is still unknown; it is the founding member of the AAMDC family and, despite having a solved structure, has no assigned catalytic or binding activity. Its name comes from work on the mouse ortholog, where the protein is highly expressed in white adipose tissue and is both sufficient and required for preadipocyte differentiation: overexpression drives 3T3-L1 differentiation without insulin, by raising CREB phosphorylation and transcriptional activity, while knockdown causes caspase-3-dependent preadipocyte apoptosis. In human cells AAMDC is best known as an oncogene within the 11q13.5-14.1 amplicon that defines the IntClust2 subgroup of estrogen receptor-positive breast cancer. There it drives PI3K-AKT-mTOR signalling - ectopic expression alone is enough to activate AKT and to support estrogen-independent tumour growth - and through that axis controls the translation of ATF4 and MYC and the expression of enzymes of the one-carbon folate, methionine and lipid metabolic pathways. AAMDC has also been reported to interact with the Rab GTPase-activating protein RABGAP1L and to colocalise with RABGAP1L and RAB7A at endolysosomes. All of the described effects are downstream cellular consequences; what AAMDC does biochemically to produce them remains the central open question about this protein.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005737 cytoplasm | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic inference from MGI:MGI:1913523, which is mouse Aamdc - the true ortholog, not a paralog. The mouse gene carries its own IDA cytoplasm annotation (PMID:21622130), so this is a well-founded transfer rather than an over-propagation, and it agrees with the UniProt subcellular location for the human protein. Correct core compartment. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt SUBCELLULAR LOCATION: Cytoplasm {ECO:0000250}. Note=Diffuse |
| GO:0045600 positive regulation of fat cell differentiation | IBA GO_REF:0000033 | ACCEPT | Summary: A well-founded IBA, and the only annotation on this gene with substantial evidence behind it. The WITH/FROM field cites MGI:MGI:1913523, which is mouse Aamdc itself, and that gene carries both IDA and IMP annotations to this same term from PMID:22279136. Reading the source shows the evidence is bidirectional: overexpression of the mouse protein was sufficient to drive 3T3-L1 preadipocyte differentiation in the absence of insulin, acting through CREB phosphorylation and transcriptional activity, while siRNA knockdown caused preadipocyte apoptosis during adipogenesis. This is precisely the situation IBA exists for - a solid ortholog phenotype transferred to a human gene that has no experimental data of its own - and it is why UniProt's human FUNCTION line is an ECO:0000250 by-similarity statement about adipogenesis. Accept as the core biological process. Supporting Evidence: PMID:22279136 Our results indicated that LI2 was sufficient to drive preadipocyte differentiation via modulating the phosphorylation level and transcriptional activity of CREB, coincident with expression of several adipogenic regulators and mature adipocyte markers, without insulin treatment. PMID:22279136 knockdown of the LI2 protein resulted in preadipocyte apoptosis via caspase-3 activation during adipogenesis file:human/AAMDC/AAMDC-uniprot.txt May play a role in preadipocyte differentiation and |
| GO:0005737 cytoplasm | IEA GO_REF:0000120 | ACCEPT | Summary: Automatic annotation combining UniProt SubCell mapping with orthology to mouse Aamdc (UniProtKB:Q8R0P4). Consistent with the IBA and ISS cytoplasm annotations and with the UniProt subcellular location. Correct core compartment. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt SUBCELLULAR LOCATION: Cytoplasm {ECO:0000250}. Note=Diffuse |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: Interaction with ACY3 (UniProtKB:Q96HD9) from a proteome-scale yeast two-hybrid map. Reported by a single publication with no functional follow-up - the NbExp=3 in the UniProt line below counts assay replicates within that one study, not independent studies - and bare "protein binding" conveys nothing about AAMDC's activity in any case. Notably, none of the eight protein-binding annotations on this gene recovered RABGAP1L or RAB7A, the only AAMDC interaction that has been followed up functionally, so the GOA binding record is orthogonal to what is known about the protein. Reason: Isolated high-throughput two-hybrid hit; uninformative term with no supporting biology. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; Q96HD9: ACY3; NbExp=3; IntAct=EBI-10308705, EBI-3916242; |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: Interaction with GORASP2/GRASP55 (UniProtKB:Q9H8Y8) from the same proteome-scale two-hybrid map. Reported by that one publication only (the NbExp=3 below counts assay replicates within it), no functional follow-up, and no described connection between AAMDC and Golgi stacking or unconventional secretion. Reason: Isolated high-throughput two-hybrid hit; uninformative term with no supporting biology. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; Q9H8Y8: GORASP2; NbExp=3; IntAct=EBI-10308705, EBI-739467; |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Interaction with VPS9D1 (UniProtKB:Q9Y2B5) from the human binary interactome reference map. Reported by that one publication only (the NbExp=3 below counts assay replicates within it), with no functional follow-up. A VPS9-domain protein is at least topically compatible with the reported endolysosomal colocalisation of AAMDC, but nothing connects the two observations and no study has pursued this pair. Reason: Isolated high-throughput hit; uninformative term. Recorded rather than removed because the partner is not implausible, but it carries no annotation value as it stands. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; Q9Y2B5: VPS9D1; NbExp=3; IntAct=EBI-10308705, EBI-9031083; |
| GO:0005515 protein binding | IPI PMID:32814053 Interactome Mapping Provides a Network of Neurodegenerative ... | MARK AS OVER ANNOTATED | Summary: Interaction with APP isoform 2 (UniProtKB:P05067-2), one of five partners this single screen contributed for AAMDC. The screen was one systematic yeast two-hybrid pass over ~500 neurodegeneration-related baits, and a small uncharacterised protein scoring against APP, huntingtin, ataxin-3, dynamin-2 and GDAP1 all within that one pass is more consistent with a promiscuous hit than with five distinct biological interactions. None of the five has been independently replicated or followed up, and AAMDC has no described role in neurodegeneration. (The paper's title phrase "Widespread Protein Aggregation in Affected Brains" refers to aggregation observed in postmortem patient tissue, not to the interaction dataset; that phrase is not evidence about these Y2H hits and is no longer relied on here.) Reason: One of a five-partner block contributed by a single systematic Y2H pass over ~500 neurodegeneration baits; no independent replication, no follow-up, and an uninformative term. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; P05067-2: APP; NbExp=3; IntAct=EBI-10308705, EBI-17264467; |
| GO:0005515 protein binding | IPI PMID:32814053 Interactome Mapping Provides a Network of Neurodegenerative ... | MARK AS OVER ANNOTATED | Summary: Interaction with huntingtin (UniProtKB:P42858) from the same neurodegeneration interactome screen. Same assessment as the APP entry: part of a five-partner block from one systematic Y2H pass, not independently replicated and never followed up. Reason: One of a five-partner block contributed by a single systematic Y2H pass over ~500 neurodegeneration baits; no independent replication, no follow-up, and an uninformative term. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; P42858: HTT; NbExp=3; IntAct=EBI-10308705, EBI-466029; |
| GO:0005515 protein binding | IPI PMID:32814053 Interactome Mapping Provides a Network of Neurodegenerative ... | MARK AS OVER ANNOTATED | Summary: Interaction with dynamin-2 isoform 2 (UniProtKB:P50570-2) from the same neurodegeneration interactome screen. Same assessment as the other members of that five-partner block from a single Y2H pass. Reason: One of a five-partner block contributed by a single systematic Y2H pass over ~500 neurodegeneration baits; no independent replication, no follow-up, and an uninformative term. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; P50570-2: DNM2; NbExp=3; IntAct=EBI-10308705, EBI-10968534; |
| GO:0005515 protein binding | IPI PMID:32814053 Interactome Mapping Provides a Network of Neurodegenerative ... | MARK AS OVER ANNOTATED | Summary: Interaction with ataxin-3 (UniProtKB:P54252) from the same neurodegeneration interactome screen. Same assessment as the other members of that five-partner block from a single Y2H pass. Reason: One of a five-partner block contributed by a single systematic Y2H pass over ~500 neurodegeneration baits; no independent replication, no follow-up, and an uninformative term. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; P54252: ATXN3; NbExp=3; IntAct=EBI-10308705, EBI-946046; |
| GO:0005515 protein binding | IPI PMID:32814053 Interactome Mapping Provides a Network of Neurodegenerative ... | MARK AS OVER ANNOTATED | Summary: Interaction with GDAP1 (UniProtKB:Q8TB36) from the same neurodegeneration interactome screen, completing the five-partner block. Same assessment as the others: one Y2H pass, no independent replication, no follow-up. Reason: One of a five-partner block contributed by a single systematic Y2H pass over ~500 neurodegeneration baits; no independent replication, no follow-up, and an uninformative term. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt Q9H7C9; Q8TB36: GDAP1; NbExp=3; IntAct=EBI-10308705, EBI-11110431; |
| GO:0005737 cytoplasm | ISS GO_REF:0000024 | ACCEPT | Summary: Sequence-similarity transfer of the cytoplasmic location from mouse Aamdc (UniProtKB:Q8R0P4), which has this location by IDA. Redundant with the IBA and IEA entries but correct, and it rests on the same well-supported ortholog relationship. Supporting Evidence: file:human/AAMDC/AAMDC-uniprot.txt SUBCELLULAR LOCATION: Cytoplasm {ECO:0000250}. Note=Diffuse |
| GO:0051897 positive regulation of phosphatidylinositol 3-kinase/protein kinase B signal transduction | IMP PMID:33772001 The oncogene AAMDC links PI3K-AKT-mTOR signaling with metabo... | NEW | Summary: Proposed new annotation, and the only biological process for which there is direct human experimental evidence. AAMDC is an oncogene within the 11q13.5-14.1 amplicon of IntClust2 estrogen receptor-positive breast cancer, and ectopic expression of AAMDC alone is sufficient to activate AKT signalling and to support estrogen-independent tumour growth. The same study shows AAMDC controls PI3K-AKT-mTOR signalling more broadly, regulating the translation of ATF4 and MYC. The evidence is bidirectional: the converse shRNA knockdowns inhibit AKT activation (reduced p-S473 and p-T308) and downregulate the PI3K-AKT-mTOR axis, so the direction of effect is established by both gain and loss of function rather than by a correlative association. GOA currently records no human-derived biological process for this gene at all. Reason: Adds the only human experimentally supported process; complements the mouse-derived adipogenesis IBA, which describes a different cellular context. Evidence code is IMP: the experiments are ectopic overexpression and shRNA knockdown, both of which are mutant/altered-gene-product phenotypes under GO evidence-code guidance, not direct assay of the gene product (IDA). Supporting Evidence: PMID:33772001 Ectopic AAMDC expression is sufficient to activate AKT signaling, resulting in estrogen-independent tumor growth. PMID:33772001 The AAMDC KDs inhibited activation of AKT relative to controls, as indicated by a reduction in p-S473 and p-T308 levels PMID:33772001 AAMDC knockdowns downregulate the PI3K-AKT-mTOR axis through translational suppression of MYC and ATF4 PMID:33772001 We show that AAMDC controls PI3K-AKT-mTOR signaling, regulating the translation of ATF4 and MYC and modulating the transcriptional activity of AAMDC-dependent promoters. file:human/AAMDC/AAMDC-deep-research-affinage.md AAMDC is an oncogenic signaling regulator amplified in estrogen receptor-positive breast cancer that drives estrogen-independent tumor growth by activating the PI3K-AKT-mTOR axis |
| GO:0036019 endolysosome | IDA PMID:33772001 The oncogene AAMDC links PI3K-AKT-mTOR signaling with metabo... | NEW | Summary: Proposed new annotation, offered with an explicit caveat. AAMDC was reported to interact with the Rab GTPase-activating protein RABGAP1L, and AAMDC, RABGAP1L and RAB7A were shown to colocalise in endolysosomes. RAB7A is a canonical late endosome/lysosome marker, so the compartment assignment is coherent. However the evidence is imaging colocalisation from a single study, and the authors themselves hedge ("we provide evidence that AAMDC can interact"). Colocalisation is weaker than co-residence in a complex, so this is proposed as a secondary, non-core location that does not displace the well-supported cytoplasmic annotation. Reason: Records a specific, mechanistically interesting compartment that GOA lacks entirely, while flagging that it rests on single-study colocalisation rather than fractionation or a defined complex. "Non-core" here means secondary to the cytoplasmic annotation, not excluded: the term is retained in core_functions.locations for the PI3K-AKT-mTOR function alongside GO:0005737, because it is the only compartment reported for the RABGAP1L/RAB7A association that function describes. Note also that GO:0036019 endolysosome is a specific hybrid intermediate compartment; light-microscopy colocalisation with RAB7A cannot by itself distinguish it from late endosome (GO:0005770) or lysosome (GO:0005764), so the exact compartment term should be treated as provisional pending fractionation or marker-resolved imaging. Supporting Evidence: PMID:33772001 we provide evidence that AAMDC can interact with the RabGTPase-activating protein RabGAP1L, and that AAMDC, RabGAP1L, and Rab7a colocalize in endolysosomes. |
Loading supporting content…
Download this section (compressed HTML)Q: What does AAMDC actually do biochemically? This is the central gap, and a family-wide analysis run for this review (see file:human/AAMDC/AAMDC-bioinformatics/RESULTS.md) shows it is not specific to this gene: of the 13 reviewed Swiss-Prot members of Pfam PF04430 (aptly named DUF498, "domain of unknown function"), zero carry any experimentally supported molecular function term and zero have a UniProt CATALYTIC ACTIVITY block, while four carry an explicit GO:0003674 ND. The protein has a solved structure and a defined Mth938 domain, yet every described effect - adipogenesis, AKT activation, ATF4/MYC translation - is a downstream cellular consequence rather than an activity.
Q: Does AAMDC help assemble a protein complex? The family analysis gives a concrete direction. The other branch of PF04430 is NDUFAF3, whose function is consistent across every organism studied: an assembly factor for mitochondrial complex I, a protein that binds subunits and helps build a complex without being part of the finished product and without any catalytic activity. If the MTH938 fold is an assembly-factor fold, the right question for AAMDC is not what it catalyses but what it helps assemble - which would also explain why no molecular function has ever been found for any member of this family.
Q: Do the Mth938 domain and its bacterial and archaeal homologues suggest a conserved activity that could be tested directly? A structure-guided comparison with characterised Mth938-fold proteins might propose a candidate biochemical function where cell-biological approaches have not.
Q: Is the AAMDC-RABGAP1L interaction direct and stoichiometric, and does AAMDC regulate RABGAP1L's GAP activity toward RAB7A? If so, AAMDC would acquire a genuine molecular function and a mechanistic link between its endolysosomal localisation and its effect on mTOR signalling, which is itself regulated at the lysosome.
Q: Are the adipogenesis role in mouse and the PI3K-AKT-mTOR role in human cancer cells the same underlying activity in two contexts, or genuinely different functions? AKT signalling is central to adipogenesis, so a single mechanism is plausible and would consolidate two currently disconnected literatures.
Q: Does human AAMDC promote adipocyte differentiation? The GO:0045600 annotation is entirely IBA from mouse. Direct testing in human preadipocytes would convert this gene's best-supported annotation into experimental evidence, or reveal a species difference.
Experiment: Reconstitute purified AAMDC, RABGAP1L and RAB7A in vitro and measure RAB7A GTP hydrolysis with and without AAMDC, then test in cells whether AAMDC depletion changes the RAB7A-GTP load and mTORC1 recruitment to lysosomes.
Hypothesis: AAMDC modulates RABGAP1L GAP activity toward RAB7A, coupling it to lysosomal mTORC1 signalling.
Type: in vitro reconstitution and cell biology
Experiment: Screen purified recombinant AAMDC against broad metabolite, nucleotide and metal panels by thermal shift and native mass spectrometry to identify any bound ligand or cofactor, combined with structure-based comparison to characterised Mth938-fold proteins. Aimed squarely at the missing molecular function.
Hypothesis: The Mth938 domain has a conserved biochemical activity detectable by unbiased screening.
Type: biophysics and structural bioinformatics
Experiment: Compare the transcriptional and phosphoproteomic response to AAMDC gain and loss of function in 3T3-L1 preadipocytes and in IntClust2 breast cancer lines, testing in particular whether the adipogenic effect requires AKT and whether the oncogenic effect requires CREB.
Hypothesis: AAMDC's adipogenic and oncogenic roles are the same activity in different contexts.
Type: comparative functional genomics
Experiment: Re-test the AAMDC interactions with APP, HTT, ATXN3, DNM2 and GDAP1 by co-immunoprecipitation at endogenous expression levels, and check whether AAMDC scores as a promiscuous partner across unrelated aggregation-prone baits. A negative result would justify removing these annotations outright rather than only flagging them.
Hypothesis: The five neurodegeneration-panel interactions are aggregation artefacts, not biology.
Type: interaction validation
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)