Adipogenin (ADIG, also called SMAF1) is an 80-residue microprotein of the endoplasmic reticulum membrane, anchored by a single conserved N-terminal hydrophobic helix (residues 14-38) with its short C-terminal tail facing the cytosol. Its expression is essentially restricted to adipocytes and is a direct transcriptional output of PPAR-gamma, rising sharply during adipogenesis and falling when the adipocyte phenotype is lost; it is also induced in steatotic liver. ADIG's molecular activity is to bind seipin, the ER protein that organizes lipid droplet biogenesis at ER-droplet junctions. Seipin assembles into both undecameric and dodecameric rings, and ADIG selectively binds the dodecamer, bridging neighbouring subunits to stabilize the ring and promote its assembly; the resulting rigid seipin-ADIG particle favours droplet expansion over new droplet nucleation, redirecting triacylglycerol flux toward growing droplets and assisting the relocation of droplet-targeted proteins such as ACSL3 and GPAT4. Through this single structural role ADIG sets how much lipid an adipocyte stores: loss of adipogenin impairs adipocyte differentiation, yields leaner animals on a high-fat diet and in the leptin-deficient background, disrupts triglyceride accumulation in brown adipose tissue, while adipocyte-specific overexpression enlarges droplets and increases fat mass. Adipogenin deficiency additionally lowers fat-mass-adjusted circulating leptin and impairs leptin secretion from adipose explants, and human genetic variation at the ADIG locus associates with BMI-adjusted leptin levels.
Definition: A protein-containing complex of the endoplasmic reticulum membrane consisting of a dodecameric ring of seipin (BSCL2) together with adipogenin subunits bound between adjacent seipin protomers. The complex is rigid relative to adipogenin-free seipin oligomers and promotes the expansion of nascent lipid droplets.
Justification: GO has no term for any seipin-containing complex, so the defining assembly of this gene product cannot be recorded in the in_complex slot of its core function. The honest difficulty is that the stoichiometry is NOT fixed: the authors state plainly that [PMID:41196993 "the complex might be heterogeneous with respect to seipin:Adig stoichiometry."] and enumerate a set of coexisting species [PMID:41196993 "There may thus be several different types of seipin complexes: seipin-alone undecamers, seipin-alone dodecamers, Adig-poor seipin dodecamers, and Adig-rich seipin dodecamers"], with Adig exchanging on and off on a ~10 s timescale by FRAP, and the C12 cryo-EM map being a symmetrised average over many particles. A variable-occupancy, dynamically exchanging subunit is a harder case for minting a discrete complex term, not an easier one, and the proposal should be judged knowing that. The case for the term anyway: the Adig-bound dodecamer is compositionally and functionally distinct from the Adig-free undecamer and dodecamer - only it is selected by Adig, only it is rigidified, and it behaves differently toward nascent droplets - so it is a real biological entity that curation currently cannot name. A parent term for the seipin oligomer itself would also be useful, and would let the Adig-bound form be a child.
Parent term: endoplasmic reticulum protein-containing complex
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005634 nucleus | IBA GO_REF:0000033 | REMOVE | Summary: Nuclear localization inherited through the PANTHER adipogenin node from mouse Adig, whose own nuclear call rests on a single overexpressed EGFP fusion; directly superseded by PMID:41196993. Reason: The nuclear placement descends entirely from one 2005 observation on an overexpressed SMAF1-EGFP fusion [PMID:15567149 "Transfection and localization studies of a SMAF1-EGFP fusion construct indicate nuclear localization, suggestive of a possible regulatory role."], and even that paper framed it as merely suggestive. ADIG is an 80-residue, ~10 kDa protein, far below the passive nuclear-pore diffusion limit, so an EGFP fusion reaching the nucleoplasm is uninformative about where the endogenous protein acts. PMID:41196993 went back to precisely this question, naming the nuclear report as one of the conflicting claims it set out to settle [PMID:41196993 "We next assessed the subcellular location of Adig because previous reports suggest different locations, including the plasma membrane (2), nuclei (3), and the surface of LDs (4)."], and resolved it with three orthogonal assays - EGFP fused at either terminus against an ER marker, APEX2 electron microscopy at near-endogenous levels in Adig-knockout adipocytes, and a digitonin protease-protection topology assay - concluding a single-pass ER topology [PMID:41196993 "Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail"]. UniProt has already retired the nuclear call and now lists only the ER membrane. This is an ISS/IBA propagation of a superseded result, not an experimental annotation whose full text a curator read, so removing it does not overrule a curator's reading. Propagation Review Root cause: SOURCE BAD Failure modes: SOURCE EVIDENCE WEAK Sources checked: PANTHER:PTN002213524 SOURCE BAD Ancestral node for the adipogenin family. The node carries a nuclear location that traces to the mouse Adig EGFP-fusion result in PMID:15567149; PMID:41196993 has since established a single-pass ER topology, so the node placement should be revised at source. MGI:MGI:2675492 · Adig SOURCE BAD Mouse Adig, the donor whose nuclear annotation seeds this node. Its nuclear evidence is an overexpressed SMAF1-EGFP fusion of a ~10 kDa protein, below the passive nuclear-pore diffusion limit. Supporting Evidence: PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail PMID:41196993 We next assessed the subcellular location of Adig because previous reports suggest different locations, including the plasma membrane (2), nuclei (3), and the surface of LDs (4). PMID:41196993 We found a strong APEX2 signal on the ER PMID:15567149 Transfection and localization studies of a SMAF1-EGFP fusion construct indicate nuclear localization, suggestive of a possible regulatory role. |
| GO:0005634 nucleus | ISS PMID:15567149 Cloning, expression, and differentiation-dependent regulatio... | REMOVE | Summary: Direct sequence-similarity transfer of the 2005 mouse SMAF1-EGFP nuclear call, which PMID:41196993 explicitly revisited and replaced with a single-pass ER topology. Reason: The nuclear placement descends entirely from one 2005 observation on an overexpressed SMAF1-EGFP fusion [PMID:15567149 "Transfection and localization studies of a SMAF1-EGFP fusion construct indicate nuclear localization, suggestive of a possible regulatory role."], and even that paper framed it as merely suggestive. ADIG is an 80-residue, ~10 kDa protein, far below the passive nuclear-pore diffusion limit, so an EGFP fusion reaching the nucleoplasm is uninformative about where the endogenous protein acts. PMID:41196993 went back to precisely this question, naming the nuclear report as one of the conflicting claims it set out to settle [PMID:41196993 "We next assessed the subcellular location of Adig because previous reports suggest different locations, including the plasma membrane (2), nuclei (3), and the surface of LDs (4)."], and resolved it with three orthogonal assays - EGFP fused at either terminus against an ER marker, APEX2 electron microscopy at near-endogenous levels in Adig-knockout adipocytes, and a digitonin protease-protection topology assay - concluding a single-pass ER topology [PMID:41196993 "Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail"]. UniProt has already retired the nuclear call and now lists only the ER membrane. This is an ISS/IBA propagation of a superseded result, not an experimental annotation whose full text a curator read, so removing it does not overrule a curator's reading. Propagation Review Root cause: SOURCE BAD Failure modes: SOURCE EVIDENCE WEAK Sources checked: UniProtKB:Q8R400 · Adig SOURCE BAD Mouse adipogenin. The donor nuclear annotation derives from an overexpressed EGFP fusion in PMID:15567149 and has been superseded; UniProt now records only Endoplasmic reticulum membrane, Single-pass membrane protein for this entry's human ortholog. Supporting Evidence: PMID:15567149 Transfection and localization studies of a SMAF1-EGFP fusion construct indicate nuclear localization, suggestive of a possible regulatory role. PMID:41196993 We next assessed the subcellular location of Adig because previous reports suggest different locations, including the plasma membrane (2), nuclei (3), and the surface of LDs (4). PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail PMID:41196993 The localization of Adig-EGFP overlapped with that of the ER-mCherry-KDEL protein |
| GO:0005737 cytoplasm | IBA GO_REF:0000033 | MODIFY | Summary: Cytoplasm is true but uninformative for a single-pass ER membrane protein whose only cytoplasmic exposure is a short C-terminal tail; the ER membrane term already annotated carries the actual information. Reason: ADIG is an integral membrane protein of the ER with a transmembrane segment at residues 14-38 (UniProt Q0VDE8) and a cytosol-facing C-terminus [PMID:41196993 "Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail"]. GO:0005737 is not false - the ER and the cytosolic tail both lie within the cytoplasm - but as a location statement it is a very general parent that conveys nothing about where ADIG acts, and it was propagated alongside the nuclear call from the same 2005 fusion-protein study. The informative term is GO:0005789, which this gene already carries from two independent sources. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: PANTHER:PTN002213524 SUPPORTS TRANSFER The adipogenin ancestral node asserts cytoplasm, a parent so broad it is compatible with the superseded nuclear/LD claims and with the correct ER one; the family is uniformly ER-membrane anchored by a conserved N-terminal hydrophobic helix. MGI:MGI:2675492 · Adig SUPPORTS TRANSFER Mouse Adig donor; its cytoplasmic annotation predates the ER topology determination in PMID:41196993. Proposed replacements: endoplasmic reticulum membrane Supporting Evidence: PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail PMID:41196993 We found a strong APEX2 signal on the ER PMID:33691105 Adipogenin (Adig) is an adipocyte-enriched transmembrane protein. |
| GO:0005737 cytoplasm | ISS PMID:15567149 Cloning, expression, and differentiation-dependent regulatio... | MODIFY | Summary: Sequence-similarity transfer of a general cytoplasm term from the 2005 mouse study; superseded in specificity by the ER membrane localization. Reason: Same reasoning as the IBA cytoplasm row: correct only as a broad parent. The 2005 source characterized SMAF1 as a novel small protein [PMID:15567149 "SMAF1 encodes a wholly novel 10kDa protein."] without topology data; the transmembrane nature of the protein was established later [PMID:33691105 "Adipogenin (Adig) is an adipocyte-enriched transmembrane protein."] and its ER membrane topology definitively in PMID:41196993. Replace with the ER membrane term. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: UniProtKB:Q8R400 · Adig SUPPORTS TRANSFER Mouse adipogenin donor annotation; general cytoplasm, later refined to ER membrane in both UniProt and the primary literature. Proposed replacements: endoplasmic reticulum membrane Supporting Evidence: PMID:15567149 SMAF1 encodes a wholly novel 10kDa protein. PMID:33691105 Adipogenin (Adig) is an adipocyte-enriched transmembrane protein. PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail |
| GO:0005789 endoplasmic reticulum membrane | IEA GO_REF:0000044 | ACCEPT | Summary: ER membrane, mapped from the UniProt subcellular-location vocabulary. This is now the best-supported location for ADIG and should be treated as core. Reason: The UniProt SubCell mapping is correct and, unusually, is better evidence than the manual rows it sits beside, because UniProt has already incorporated PMID:41196993 while GOA has not. ADIG is a single-pass ER membrane protein: an N-terminal conserved hydrophobic helix [PMID:41196993 "Sequence analysis across species revealed a long conserved hydrophobic"] anchors it in the ER, EGFP fusions colocalize with an ER marker [PMID:41196993 "The localization of Adig-EGFP overlapped with that of the ER-mCherry-KDEL protein"], APEX2 electron microscopy at near-endogenous levels puts it on the ER [PMID:41196993 "We found a strong APEX2 signal on the ER"], and protease protection fixes the topology [PMID:41196993 "Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail"]. UniProt Q0VDE8 annotates TRANSMEM 14..38. Supporting Evidence: PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail PMID:41196993 The localization of Adig-EGFP overlapped with that of the ER-mCherry-KDEL protein PMID:41196993 We found a strong APEX2 signal on the ER PMID:41196993 Sequence analysis across species revealed a long conserved hydrophobic |
| GO:0005789 endoplasmic reticulum membrane | ISS GO_REF:0000024 | ACCEPT | Summary: ER membrane transferred by curator judgment from mouse Adig; independently confirmed for the orthologous protein and consistent with the conserved transmembrane helix. Reason: Correct and core. The mouse donor localization is itself now anchored in direct structural and imaging evidence [PMID:41196993 "Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail"], and the anchoring helix is conserved across species, so the similarity transfer to human ADIG is sound. Human Q0VDE8 carries TRANSMEM 14..38 with TOPO_DOM 1..13 and 39..80, matching the mouse topology. Supporting Evidence: PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail PMID:41196993 Sequence analysis across species revealed a long conserved hydrophobic PMID:33691105 Adipogenin (Adig) is an adipocyte-enriched transmembrane protein. |
| GO:0005811 lipid droplet | IBA GO_REF:0000033 | MODIFY | Summary: Lipid droplet localization rests on overexpressed HA-tagged Smaf1; the definitive study places ADIG on the ER and shows that Adig-bound seipin complexes specifically do not form stable LD contacts. Reason: ADIG unquestionably acts in lipid droplet biology, but not by residing on the droplet. The LD placement traces to one overexpression immunolocalization [PMID:26427354 "immunolocalization studies using HA-tagged Smaf1 reveal enrichment at adipocyte lipid droplets"]. PMID:41196993 listed the LD surface among the conflicting prior claims it set out to adjudicate [PMID:41196993 "We next assessed the subcellular location of Adig because previous reports suggest different locations, including the plasma membrane (2), nuclei (3), and the surface of LDs (4)."] and placed the protein on the ER instead. More pointedly, live-cell tracking showed the opposite of LD residence for the Adig-containing species [PMID:41196993 "they formed stable contacts with seipin complexes without Adig, while seipin-Adig complexes moved rapidly through their vicinity"] - the Adig-bound complex is the mobile form that does not engage nascent droplets stably. The correct location is the ER membrane, from which ADIG acts on droplet formation in trans through seipin. The affinage record is cited in supported_by for this row not as evidence FOR lipid droplet localization but as the provenance of the claim being declined: its mechanism_profile still lists GO:0005811 as the location, and a reader should be able to see what was considered and rejected. Propagation Review Root cause: SOURCE BAD Failure modes: SOURCE EVIDENCE WEAK Sources checked: PANTHER:PTN002213524 SOURCE BAD The adipogenin ancestral node asserts lipid droplet as a location. The underlying evidence is overexpressed HA-tagged mouse Smaf1; higher-resolution work localizes the protein to the ER membrane and shows Adig-bound seipin complexes do not form stable LD contacts. MGI:MGI:2675492 · Adig SOURCE BAD Mouse Adig donor for the LD location, from PMID:26427354 overexpression immunolocalization. Proposed replacements: endoplasmic reticulum membrane Supporting Evidence: PMID:26427354 immunolocalization studies using HA-tagged Smaf1 reveal enrichment at adipocyte lipid droplets PMID:41196993 We next assessed the subcellular location of Adig because previous reports suggest different locations, including the plasma membrane (2), nuclei (3), and the surface of LDs (4). PMID:41196993 they formed stable contacts with seipin complexes without Adig, while seipin-Adig complexes moved rapidly through their vicinity PMID:41196993 Thus, Adig is a single-pass ER protein with a cytosolic C-terminal tail file:genes/human/ADIG/ADIG-deep-research-affinage.md **localization:** GO:0005811 lipid droplet |
| GO:0019915 lipid storage | ISS GO_REF:0000024 | ACCEPT | Summary: Lipid storage, transferred from mouse Adig. Well supported and core: ADIG sets the amount of triacylglycerol stored by controlling seipin-dependent droplet expansion. Reason: This is the process the definitive study concludes with [PMID:41196993 "Thus, Adig can modulate lipid storage through its structural and functional interactions with seipin."], and it is supported bidirectionally in vivo [PMID:41196993 "In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues."] and by the independent knockout [PMID:33691105 "Adig-/- mice are leaner than wild-type mice when fed a high-fat diet and when crossed with Ob/Ob hyperphagic mice"]. One caveat worth recording: acute siRNA knockdown in 3T3-L1 adipocytes did not reproduce a storage defect [PMID:26427354 "Smaf1 does not have a major role in adipocyte triglyceride accumulation, lipolysis or insulin-stimulated pAkt induction"], which at first sight conflicts with the knockout. That tension is already resolved in the knockout paper itself [PMID:33691105 "these observations can be reconciled by the timing of the initiation of Adig KD, as we clearly show that early KD significantly impairs lipid accumulation in 3T3-L1 adipocytes"], [PMID:33691105 "whereas late (day 5 onward) KD in 3T3-L1 cells does not clearly impair differentiation"] - the requirement is developmental, and the 2015 study knocked down in cells that had largely already differentiated. Recording the answer rather than posing it as an open question. Supporting Evidence: PMID:41196993 Thus, Adig can modulate lipid storage through its structural and functional interactions with seipin. PMID:41196993 In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues. PMID:33691105 Adig-/- mice are leaner than wild-type mice when fed a high-fat diet and when crossed with Ob/Ob hyperphagic mice |
| GO:0045444 fat cell differentiation | IEA GO_REF:0000002 | ACCEPT | Summary: Fat cell differentiation from the InterPro adipogenin-domain mapping. Correct, though it is the general parent of the white/brown terms already annotated. Reason: The InterPro IPR027938 to GO mapping is sound: the adipogenin family is adipocyte-restricted and genetically required for adipogenesis [PMID:33691105 "Data from Adig-deficient cells suggest that Adig is required for adipogenesis."]. The term is broader than the white and brown fat cell differentiation terms this gene already carries, but a domain-based electronic mapping is appropriately placed at the parent, and a general IEA sitting above more specific manual annotations is not a defect. Supporting Evidence: PMID:33691105 Data from Adig-deficient cells suggest that Adig is required for adipogenesis. PMID:26427354 siRNA studies reveal that expression of Smaf1 in adipocytes is wholly dependent on PPARγ |
| GO:0045600 positive regulation of fat cell differentiation | ISS PMID:15567149 Cloning, expression, and differentiation-dependent regulatio... | ACCEPT | Summary: Positive regulation of fat cell differentiation. The conclusion is right, but the cited 2005 reference only shows correlated expression; the genetic evidence comes later. Reason: The term is correct. The cited source, however, is purely correlative - it reports that SMAF1 expression tracks adipogenic conversion [PMID:15567149 "Northern blot analysis of brown and white adipogenic conversion reveals that expression of SMAF1 closely parallels emergence of an adipocyte phenotype"], which establishes co-regulation, not a positive regulatory role. The claim is nevertheless supported by loss-of-function evidence published afterwards [PMID:33691105 "Data from Adig-deficient cells suggest that Adig is required for adipogenesis."]. Keeping the annotation but re-anchoring it to PMID:33691105 would strengthen it. Supporting Evidence: PMID:33691105 Data from Adig-deficient cells suggest that Adig is required for adipogenesis. PMID:15567149 Northern blot analysis of brown and white adipogenic conversion reveals that expression of SMAF1 closely parallels emergence of an adipocyte phenotype |
| GO:0050872 white fat cell differentiation | IBA GO_REF:0000033 | ACCEPT | Summary: White fat cell differentiation, inherited through the adipogenin ancestral node. Supported by mouse loss-of-function data. Reason: Sound node placement: adipogenin is an adipocyte-restricted PPAR-gamma target across mammals [PMID:26427354 "siRNA studies reveal that expression of Smaf1 in adipocytes is wholly dependent on PPARγ"], and loss of Adig impairs adipogenesis and reduces white fat mass [PMID:33691105 "Adig-/- mice are leaner than wild-type mice when fed a high-fat diet and when crossed with Ob/Ob hyperphagic mice"]. Human ADIG lies inside the clade that inherited the function and shows no evidence of divergence or loss. Supporting Evidence: PMID:33691105 Data from Adig-deficient cells suggest that Adig is required for adipogenesis. PMID:33691105 Adig-/- mice are leaner than wild-type mice when fed a high-fat diet and when crossed with Ob/Ob hyperphagic mice PMID:26427354 siRNA studies reveal that expression of Smaf1 in adipocytes is wholly dependent on PPARγ |
| GO:0050872 white fat cell differentiation | ISS GO_REF:0000024 | ACCEPT | Summary: White fat cell differentiation transferred from mouse Adig by curator judgment; duplicates the IBA claim from an independent route and is equally well supported. Reason: Same biology as the IBA row, reached by direct orthology to mouse Adig rather than by node placement. The mouse evidence is genetic [PMID:33691105 "Data from Adig-deficient cells suggest that Adig is required for adipogenesis."], and the human protein is a clear one-to-one ortholog, so the transfer holds. Supporting Evidence: PMID:33691105 Data from Adig-deficient cells suggest that Adig is required for adipogenesis. PMID:33691105 Adig-/- mice are leaner than wild-type mice when fed a high-fat diet and when crossed with Ob/Ob hyperphagic mice |
| GO:0050873 brown fat cell differentiation | IBA GO_REF:0000033 | ACCEPT | Summary: Brown fat cell differentiation via the ancestral adipogenin node. Now directly supported by a brown-adipose phenotype in Adig-deficient mice. Reason: The brown-fat branch of this annotation has become the better-evidenced one: deleting Adig disrupts triglyceride accumulation specifically in brown adipose tissue [PMID:41196993 "In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues."], and adipogenin expression accompanies brown as well as white adipogenic conversion [PMID:15567149 "Northern blot analysis of brown and white adipogenic conversion reveals that expression of SMAF1 closely parallels emergence of an adipocyte phenotype"]. Supporting Evidence: PMID:41196993 In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues. PMID:15567149 Northern blot analysis of brown and white adipogenic conversion reveals that expression of SMAF1 closely parallels emergence of an adipocyte phenotype |
| GO:0050873 brown fat cell differentiation | ISS GO_REF:0000024 | ACCEPT | Summary: Brown fat cell differentiation transferred from mouse Adig by curator judgment; supported by the brown-adipose-specific triglyceride phenotype. Reason: Correct. The mouse donor phenotype is brown-adipose specific [PMID:41196993 "In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues."], and there is no indication that human ADIG has diverged. Supporting Evidence: PMID:41196993 In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues. |
| GO:0050873 brown fat cell differentiation | ISS PMID:15567149 Cloning, expression, and differentiation-dependent regulatio... | ACCEPT | Summary: Third brown-fat row, anchored to the 2005 expression study. Term correct; this particular reference supports only co-expression during brown adipogenic conversion. Reason: The term is right and duplicates are acceptable, but this row's evidence is the weakest of the three: the 2005 paper reports that SMAF1 transcript accompanies brown adipogenic conversion [PMID:15567149 "Northern blot analysis of brown and white adipogenic conversion reveals that expression of SMAF1 closely parallels emergence of an adipocyte phenotype"], which is co-regulation rather than participation. The functional brown-fat evidence is in PMID:41196993 ["In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues."]; re-anchoring this row there would be an improvement. Supporting Evidence: PMID:15567149 Northern blot analysis of brown and white adipogenic conversion reveals that expression of SMAF1 closely parallels emergence of an adipocyte phenotype PMID:41196993 In transgenic mice, adipocyte-specific overexpression of Adig increased fat mass and enlarged lipid droplets, whereas Adig deletion disrupted triglyceride accumulation in brown adipose tissues. |
| GO:0140042 lipid droplet formation | ISS GO_REF:0000024 | ACCEPT | Summary: Lipid droplet formation, transferred from mouse Adig. Core, and now mechanistically explained: ADIG binds and stabilizes the dodecameric seipin complex that builds droplets. Reason: This is ADIG's central process and the best-supported annotation in the set. ADIG binds seipin [PMID:41196993 "The most enriched protein with substantial sequence coverage was seipin"], selects and stabilizes one of its two oligomeric states [PMID:41196993 "Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits."], and the resulting complex drives droplet development [PMID:41196993 "Functionally, this complex promoted lipid droplet development at both early and late stages."]. ADIG contributes the structural work rather than merely being required, which satisfies the participation test for this process. One qualification the bare term hides: the effect is not uniform promotion. ADIG-loaded complexes suppress nucleation [PMID:41196993 "the seipin complex with higher Adig content (12:4) slowed down the rate of TAG nucleation, and in the fully saturated seipin-Adig complex (12:12) (Movie S5), TAG was completely blocked from migrating to the center of the seipin complex"] and nascent droplets form on ADIG-free complexes [PMID:41196993 "nascent LDs were associated with seipin complexes lacking a visible Adig signal in A431 cells"], with the net result that droplets become fewer and larger [PMID:41196993 "As the expression of Adig increased, the number of LDs decreased (Fig. 5D), but their size increased"]. So ADIG redirects flux from nucleation to expansion rather than increasing droplet formation across the board. GO:0034389 lipid droplet organization is an arguable alternative framing for that. Supporting Evidence: PMID:41196993 Functionally, this complex promoted lipid droplet development at both early and late stages. PMID:41196993 Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits. PMID:41196993 The most enriched protein with substantial sequence coverage was seipin |
| GO:0044877 protein-containing complex binding | ISS PMID:41196993 Adipogenin promotes the development of lipid droplets by bin... | NEW | Summary: ADIG has no molecular function annotation at all. It binds the seipin complex and, critically, discriminates between its two oligomeric states, selecting the dodecamer. Reason: The 16 existing annotations contain not one molecular function term, which is the largest gap in this gene's curation. ADIG's activity is binding to an assembled multi-subunit complex, and it is oligomer-state selective, which is exactly what GO:0044877 (binding to a macromolecular complex) expresses and what the generic protein binding term could not. Seipin was the top hit in unbiased pulldowns [PMID:41196993 "The most enriched protein with substantial sequence coverage was seipin"], the interaction forms a defined particle [PMID:41196993 "Here, we found that Adig interacts with seipin to form a stable, rigid complex."], and the selectivity is explicit [PMID:41196993 "Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits."], resolved in a ~3 angstrom cryo-EM structure. UniProt Q0VDE8 already records "Forms a dodecameric complex with seipin (BSCL2)". Evidence code ISS because the structural work is on the mouse and human seipin-Adig complex and transfers to human ADIG by orthology. Supporting Evidence: PMID:41196993 The most enriched protein with substantial sequence coverage was seipin PMID:41196993 Here, we found that Adig interacts with seipin to form a stable, rigid complex. PMID:41196993 Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits. |
| GO:0065003 protein-containing complex assembly | ISS PMID:41196993 Adipogenin promotes the development of lipid droplets by bin... | NEW | Summary: ADIG does the work of seipin oligomer assembly - it is built into the growing complex and bridges adjacent subunits - so this is participation in assembly, not regulation of it. Reason: ADIG is not a passive partner of seipin: it increases the extent of seipin oligomer assembly and the mechanism is structural [PMID:41196993 "Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits."]. Bridging two adjacent subunits is a direct contribution to the assembly step, so the participation test is satisfied - ADIG supplies the structure the step depends on rather than merely being required for it. Grounded on GO:0065003 (participation) rather than GO:0031334 (positive regulation) deliberately: a regulator acts on an assembly it is not part of, whereas ADIG is a constituent of the product complex and the source describes it as actively participating in assembling newly translated seipin. Annotating it as a regulator of a complex it belongs to would misdescribe the mechanism. This is distinct from GO:0044877, which captures the binding activity, and from GO:0140042, which captures the downstream droplet outcome. Supporting Evidence: PMID:41196993 Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits. PMID:41196993 Here, we found that Adig interacts with seipin to form a stable, rigid complex. |
| GO:0050821 protein stabilization | ISS PMID:41196993 Adipogenin promotes the development of lipid droplets by bin... | NEW | Summary: ADIG holds seipin in its functional assembled state and prevents it from aggregating - the definition's aggregation clause is met literally. Reason: GO:0050821 covers maintaining a protein's structure and integrity and preventing degradation or aggregation. Both halves apply: ADIG stabilizes adjacent seipin subunits within the oligomer [PMID:41196993 "Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits."], and its presence determines whether seipin forms functional complexes or aggregates [PMID:41196993 "Thus, when expressed with Adig, seipin appeared to form functional seipin-Adig complexes rather than seipin aggregates."]. That anti-aggregation result is co-overexpression gel filtration, the same class of evidence this review discounts for the lipid droplet localization, so it is not leaned on. The stronger support is physical: molecular dynamics shows Adig damps the complex's motion [PMID:41196993 "We observed that the Root Mean Square Fluctuation (RMSF) for the seipin-Adig structure was lower compared to the Adig-free seipin dodecameric structure, suggesting that Adig suppresses dynamic and thermal fluctuations, thereby stabilizing the seipin complex"]. Worth recording that stabilization is mutual [PMID:41196993 "multiple domains of seipin, except for the luminal and C-terminal domains, stabilize Adig"]. Proposed as a companion to GO:0065003 rather than a substitute: assembly and stabilization of the assembled product are separable claims. Supporting Evidence: PMID:41196993 Adig selectively bound to the dodecameric form and enhanced seipin assembly by bridging and stabilizing adjacent subunits. PMID:41196993 We observed that the Root Mean Square Fluctuation (RMSF) for the seipin-Adig structure was lower compared to the Adig-free seipin dodecameric structure, suggesting that Adig suppresses dynamic and thermal fluctuations, thereby stabilizing the seipin complex PMID:41196993 multiple domains of seipin, except for the luminal and C-terminal domains, stabilize Adig PMID:41196993 Thus, when expressed with Adig, seipin appeared to form functional seipin-Adig complexes rather than seipin aggregates. |
Loading supporting content…
Download this section (compressed HTML)Q: Does human ADIG bind human seipin with the same dodecamer selectivity shown for the mouse protein? Every human ADIG annotation is ISS or IBA, and the structural work used mouse Adig, so the human orthology assumption is untested even though it is very likely.
Q: The preprint of PMID:41196993 reported that Adig overexpression elevates thermogenesis during cold exposure and that adipocyte-specific knockout impairs cold tolerance, but neither claim appears anywhere in the published Science paper. Was the thermogenic phenotype withdrawn, or simply cut for space? It is not annotated here because the published version does not support it.
Q: Is the reduced leptin secretion in Adig null mice a direct consequence of altered droplet architecture, or a separate role for the protein at the ER? The GWAS association of the ADIG locus with BMI-adjusted leptin suggests the link also holds in humans.
Q: ADIG is induced in steatotic liver by hepatic PPAR-gamma. Does it drive hepatocyte droplet enlargement there through the same seipin mechanism, and is that a tractable target in fatty liver disease?
Q: Given that GO has no seipin complex term, should a small family of terms be created for the seipin oligomer and its adipogenin- and LDAF1-bound forms, which are functionally distinct?
Experiment: Reconstitute human ADIG with human seipin and determine by cryo-EM whether the human complex is also dodecamer-selective, converting the entire human annotation set from ISS to direct evidence.
Experiment: Structure-guided separation-of-function: mutate the ADIG residues that bridge adjacent seipin protomers and test whether droplet expansion, but not ER targeting or seipin binding, is lost. This would isolate the assembly-promoting activity from mere complex binding.
Experiment: Degron-mediated acute depletion of ADIG in mature adipocytes, contrasted with germline loss, to resolve whether the storage requirement is developmental or ongoing - the discrepancy between the siRNA and knockout results.
Experiment: Measure seipin oligomer state distribution by native mass spectrometry in ADIG-null versus ADIG-overexpressing adipocytes, to test directly that ADIG shifts the undecamer/dodecamer equilibrium in cells rather than only in vitro.
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)