ADISSP is a small 174-residue protein that adipocytes release into the extracellular space, where it acts on fat tissue as a hormone. Its expression is essentially confined to adipose tissue and is highest in brown fat, and its release is triggered by beta-3-adrenergic stimulation, placing it downstream of the sympathetic signal that drives non-shivering heat production. The protein has no signal peptide and its mature form begins at residue 2 with an N-terminally acetylated alanine, so it reaches the outside of the cell by a non-classical route: blocking ER-to-Golgi transit increases rather than prevents its appearance in the medium, the behaviour seen for other leaderless secreted proteins. It therefore exists in both a cytosolic and an extracellular pool. Outside the cell it binds a saturable site on the adipocyte surface, and although the receptor has not been identified, engaging it raises cyclic AMP and activates protein kinase A independently of beta-adrenergic receptors, and separately activates a PI3K-PDK1/mTORC2-Akt cascade that drives GLUT4 to the plasma membrane without insulin. The physiological consequences, established in mice by both loss and gain of function, are browning of white adipose tissue, increased energy expenditure, protection from diet-induced obesity, and improved systemic glucose handling; adipose-specific deletion produces the mirror-image phenotype of obesity and hyperglycaemia, and the effects persist at thermoneutrality. Separately from this hormonal role, ADISSP binds the catalytic subunits of protein phosphatase 1 and the PP1 regulatory subunit PPP1R7. In colorectal cancer cells this interaction keeps PP1c away from its regulatory partner GADD34, leaving the TGF-beta receptor phosphorylated and increasing signalling through TAK1 to NF-kappaB, which promotes proliferation. Whether the intracellular phosphatase-binding activity and the extracellular hormonal activity are two faces of one mechanism or two independent functions of the same small protein is not known.
Definition: Any process that activates or increases the frequency, rate or extent of adaptive thermogenesis, the regulated production of heat in response to short term environmental changes such as stress, diet or reduced temperature.
Justification: GO currently offers regulation terms only for the cold-induced child of adaptive thermogenesis - GO:0120161, GO:0120162 and GO:0120163 - and none for the parent. That leaves no term for a gene product that increases adaptive thermogenesis by a route which is not cold-specific, which is exactly ADISSP's situation: its thermogenic effect persists in mice housed at thermoneutrality, so the cold-induced child would misdescribe the process regulated, while the parent process term GO:1990845 attributes heat production to a protein that does not produce heat. The gap is not peculiar to this gene - diet-induced thermogenesis (GO:0002024) and the futile creatine cycle (GO:0140651) are the parent's other children and are equally unserved by a cold-specific regulation branch. Adding the regulation trio at the GO:1990845 level, or at minimum the positive member, would let regulators be annotated to the level at which their evidence actually sits. Requested rather than worked around: the existing GO:1990845 rows are accepted here as the most specific expressible option.
Parent term: adaptive thermogenesis
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005576 extracellular region | IEA GO_REF:0000120 | ACCEPT | Summary: Accepted as a core location. ADISSP is a genuinely secreted protein, and unusually for a location row the claim is backed by measurement in both species rather than by a signal-peptide prediction. Reason: Secretion is measured, not predicted. In mouse the donor annotation is an IDA: endogenous Adissp was detected in brown-adipocyte conditioned medium, and a Flag knock-in at the native locus put the tag on the endogenous protein. The human protein itself was also shown to leave a cell - adenoviral expression in mouse liver put human ADISSP into the circulation at about 0.45 ug/mL - and the 2026 follow-up states that ADISSP is secreted by human adipocytes and present in human circulation. The route is non-classical: brefeldin A plus monensin, which blocks ER-to-Golgi transit, increased Adissp secretion while inhibiting adiponectin in the same experiment. That internal control matters here, because neither the human nor the mouse entry has a SIGNAL feature, the human mature chain begins at residue 2 with an N-acetylalanine, and the authors had to graft a heterologous signal peptide onto the construct to get secretion from HEK293 for purification. A leaderless protein with a cytosolic pool is exactly what those three facts describe. One caveat belongs on this row specifically, and it concerns the reference rather than the conclusion. GO_REF:0000120 is 'Combined Automated Annotation using Multiple IEA Methods', whose purpose is to record that several independent electronic pipelines produced the same annotation. Its WITH/FROM here names three tokens, and they look like three witnesses: the mouse ortholog, its Ensembl protein, and UniProtKB-SubCell:SL-0243 'Secreted'. They are not independent. The Ensembl token is the same mouse protein in another namespace, and the SubCell arm derives from the human entry's own SUBCELLULAR LOCATION line, which is itself ECO:0000250 by similarity to that identical mouse protein. All three arms reduce to Q9D1K7. The row is right, but the agreement it records is one source counted three ways, and a reader is entitled to be told that before treating a combined-methods reference as corroboration. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0243 CIRCULAR OR REDUNDANT The UniProt controlled-vocabulary term 'Secreted'. On the human entry this is asserted as ECO:0000250|UniProtKB:Q9D1K7, i.e. by similarity to the very mouse protein that is the other arm of this WITH/FROM, so the SubCellular2GO route re-reads the Compara route's own conclusion. Recorded as redundant rather than wrong: the location is correct, the independence is not. UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:0005576 IDA from PMID:36496438. ensembl:ENSMUSP00000099477 CIRCULAR OR REDUNDANT The Ensembl protein of the same mouse gene. It is the identical entity as UniProtKB:Q9D1K7 in a second namespace, so it adds a token to the WITH/FROM list but no independent evidence. Supporting Evidence: PMID:36496438 we were able to detect endogenous Adissp present in conditioned medium of mouse brown adipocyte culture PMID:36496438 When probed with a Flag antibody, Adissp-Flag was detected in conditioned medium of differentiated knock-in adipocytes PMID:36496438 this injection resulted in acute expression of ADISSP in the liver and its secretion into circulation with a concentration of about 0.45 ΞΌg/mL PMID:36496438 While adiponectin secretion was inhibited by brefeldin A and monensin as expected, Adissp secretion was instead increased PMID:36496438 It has been reported that brefeldin A and monensin enhance the secretion of non-classically secreted proteins IL-1Ξ² and migration inhibitory factor PMID:42030391 Adissp is efficiently secreted by both mouse and human adipocytes (24, 25) and is present in human circulation (26, 27) (The human Protein Atlas). file:human/ADISSP/ADISSP-uniprot.txt FT CHAIN 2..174 file:human/ADISSP/ADISSP-uniprot.txt /note="N-acetylalanine" file:human/ADISSP/ADISSP-uniprot.txt CC -!- SUBCELLULAR LOCATION: Secreted {ECO:0000250|UniProtKB:Q9D1K7}. |
| GO:0005576 extracellular region | ISS GO_REF:0000024 | ACCEPT | Summary: Accepted as a core location. This single entry stands for two GOA rows that are identical in term, evidence code, reference, qualifier, assigner and WITH/FROM, and differ only in their annotation date (2022-12-15 and 2025-10-27); they are one annotation that was re-dated, not two. Reason: Secretion is measured, not predicted. In mouse the donor annotation is an IDA: endogenous Adissp was detected in brown-adipocyte conditioned medium, and a Flag knock-in at the native locus put the tag on the endogenous protein. The human protein itself was also shown to leave a cell - adenoviral expression in mouse liver put human ADISSP into the circulation at about 0.45 ug/mL - and the 2026 follow-up states that ADISSP is secreted by human adipocytes and present in human circulation. The route is non-classical: brefeldin A plus monensin, which blocks ER-to-Golgi transit, increased Adissp secretion while inhibiting adiponectin in the same experiment. That internal control matters here, because neither the human nor the mouse entry has a SIGNAL feature, the human mature chain begins at residue 2 with an N-acetylalanine, and the authors had to graft a heterologous signal peptide onto the construct to get secretion from HEK293 for purification. A leaderless protein with a cytosolic pool is exactly what those three facts describe. The similarity judgement underpinning the ISS is about as safe as an ortholog transfer gets: human and mouse ADISSP are both exactly 174 residues, 90.2% identical with no indels at all, and neither entry carries a SIGNAL feature, so the unusual leaderless topology is shared rather than assumed. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:0005576 IDA from PMID:36496438. Correct WITH/FROM semantics for ISS: the sequence-similar entity, and nothing else. Supporting Evidence: PMID:36496438 we were able to detect endogenous Adissp present in conditioned medium of mouse brown adipocyte culture PMID:36496438 When probed with a Flag antibody, Adissp-Flag was detected in conditioned medium of differentiated knock-in adipocytes PMID:36496438 this injection resulted in acute expression of ADISSP in the liver and its secretion into circulation with a concentration of about 0.45 ΞΌg/mL PMID:36496438 While adiponectin secretion was inhibited by brefeldin A and monensin as expected, Adissp secretion was instead increased PMID:36496438 It has been reported that brefeldin A and monensin enhance the secretion of non-classically secreted proteins IL-1Ξ² and migration inhibitory factor PMID:42030391 Adissp is efficiently secreted by both mouse and human adipocytes (24, 25) and is present in human circulation (26, 27) (The human Protein Atlas). file:human/ADISSP/ADISSP-uniprot.txt FT CHAIN 2..174 file:human/ADISSP/ADISSP-uniprot.txt /note="N-acetylalanine" |
| GO:0007189 adenylate cyclase-activating G protein-coupled receptor signaling pathway | IEA GO_REF:0000107 | MODIFY | Summary: Modified to positive regulation of cAMP/PKA signal transduction. The cAMP and PKA content of this term is well measured in the mouse donor; the G-protein-coupled receptor that the term names has never been identified. Reason: What PMID:36496438 measures and what GO:0007189 asserts are not the same thing. Measured, and measured well: cAMP rises in Adissp-transgenic inguinal WAT and falls in the adipose-specific knockout; PKA substrate and HSL phosphorylation move in both directions; purified Adissp protein activates PKA in a dose- and time-dependent manner; and labelled Adissp binds adipose sections and the brown-adipocyte surface in a competable way. Not measured: that the receptor is a G-protein-coupled receptor. The authors call it a putative receptor throughout and never identify it, and the entire link to a heterotrimeric G protein is one experiment with melittin, a membrane-active bee-venom peptide used at 1 uM for 24 h with n=3 - not a selective Gs tool. GO:0007189's definition is a G protein-coupled receptor signaling pathway in which the signal is transmitted via activation of adenylyl cyclase; its comment does permit annotating ligands, but only of a pathway known to be of that kind. The replacement keeps everything that was measured and drops only the receptor class. GO:0141163 positive regulation of cAMP/PKA signal transduction is a regulation term, so an extracellular ligand can be annotated to it without being a component of the cassette it regulates - which is what rules out GO:0141156 itself, defined as an intracellular signaling cassette. Four candidate homes were considered and three rejected. Generalising to an ancestor is impossible: every ancestor of GO:0007189 retains the G-protein-coupled receptor in its definition. GO:0045762 positive regulation of adenylate cyclase activity is closer but asserts more than the data: what PMID:36496438 measures is cAMP content in transgenic and knockout adipose tissue, not adenylate cyclase activity, and a rise in cAMP content does not discriminate increased synthesis from reduced phosphodiesterase activity. GO:0141163 is satisfied either way, so it asserts exactly the demonstrated content: the cAMP/PKA cassette runs harder when ADISSP is present and slower when it is absent. GO itself points here - the obsoletion comment on GO:0043950 directs users to GO:0141163 - and I verified that GO:0141163 is current and that neither it nor GO:0007189 is an ancestor of the other, so this is a move between branches rather than a change of granularity. REMOVE would have been wrong: the receptor's class is unmeasured, not refuted, since competable surface binding shows a receptor exists and a Gs-coupled one remains the most economical reading. The missing receptor identity is recorded in knowledge_gaps. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: SOURCE EVIDENCE WEAK Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:0007189 IDA from PMID:36496438. The status is SUPPORTS_TRANSFER because the donor's finding does cross the species boundary - recombinant Adissp activates Akt in human SGBS adipocytes, so a human cell carries the responding system. SOURCE_EVIDENCE_WEAK is recorded at row level and is scoped narrowly: the donor's cAMP and PKA evidence is strong, and it is only the G-protein step, resting on melittin alone, that is weak. ensembl:ENSMUSP00000099477 CIRCULAR OR REDUNDANT The Ensembl protein of the same mouse gene; the identical entity as UniProtKB:Q9D1K7 in a second namespace, so no independent support. Proposed replacements: positive regulation of cAMP/PKA signal transduction Supporting Evidence: PMID:36496438 Together, our data suggest that Adissp activates PKA signaling through binding to a putative receptor at adipocyte surface PMID:36496438 induction of Ucp1 expression by Adissp was reduced by a merely 2-h treatment with the PKA inhibitor H89 and was blocked by Melittin, an inhibitor for GΞ±s subunit of the heterotrimeric G protein PMID:36496438 These results together demonstrate that secreted Adissp activates a G protein-cAMP-PKA signaling pathway. PMID:36496438 a specific binding of Adissp to adipose tissue sections was detected, which can be competed away by incubation with Adissp-containing conditioned medium PMID:36496438 Adipose-specific Adissp deletion decreases cAMP content and PKA activity, suppresses inguinal WAT browning, and leads to HFD-induced obesity and hyperglycemia PMID:36496438 cAMP level was increased in primary inguinal adipocytes infected with Adissp adenovirus |
| GO:0007189 adenylate cyclase-activating G protein-coupled receptor signaling pathway | ISS GO_REF:0000024 | MODIFY | Summary: Modified to positive regulation of cAMP/PKA signal transduction. The cAMP and PKA content of this term is well measured in the mouse donor; the G-protein-coupled receptor that the term names has never been identified. Reason: What PMID:36496438 measures and what GO:0007189 asserts are not the same thing. Measured, and measured well: cAMP rises in Adissp-transgenic inguinal WAT and falls in the adipose-specific knockout; PKA substrate and HSL phosphorylation move in both directions; purified Adissp protein activates PKA in a dose- and time-dependent manner; and labelled Adissp binds adipose sections and the brown-adipocyte surface in a competable way. Not measured: that the receptor is a G-protein-coupled receptor. The authors call it a putative receptor throughout and never identify it, and the entire link to a heterotrimeric G protein is one experiment with melittin, a membrane-active bee-venom peptide used at 1 uM for 24 h with n=3 - not a selective Gs tool. GO:0007189's definition is a G protein-coupled receptor signaling pathway in which the signal is transmitted via activation of adenylyl cyclase; its comment does permit annotating ligands, but only of a pathway known to be of that kind. The replacement keeps everything that was measured and drops only the receptor class. GO:0141163 positive regulation of cAMP/PKA signal transduction is a regulation term, so an extracellular ligand can be annotated to it without being a component of the cassette it regulates - which is what rules out GO:0141156 itself, defined as an intracellular signaling cassette. Four candidate homes were considered and three rejected. Generalising to an ancestor is impossible: every ancestor of GO:0007189 retains the G-protein-coupled receptor in its definition. GO:0045762 positive regulation of adenylate cyclase activity is closer but asserts more than the data: what PMID:36496438 measures is cAMP content in transgenic and knockout adipose tissue, not adenylate cyclase activity, and a rise in cAMP content does not discriminate increased synthesis from reduced phosphodiesterase activity. GO:0141163 is satisfied either way, so it asserts exactly the demonstrated content: the cAMP/PKA cassette runs harder when ADISSP is present and slower when it is absent. GO itself points here - the obsoletion comment on GO:0043950 directs users to GO:0141163 - and I verified that GO:0141163 is current and that neither it nor GO:0007189 is an ancestor of the other, so this is a move between branches rather than a change of granularity. REMOVE would have been wrong: the receptor's class is unmeasured, not refuted, since competable surface binding shows a receptor exists and a Gs-coupled one remains the most economical reading. The missing receptor identity is recorded in knowledge_gaps. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: SOURCE EVIDENCE WEAK Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:0007189 IDA from PMID:36496438. The status is SUPPORTS_TRANSFER because the donor's finding does cross the species boundary - recombinant Adissp activates Akt in human SGBS adipocytes, so a human cell carries the responding system. SOURCE_EVIDENCE_WEAK is recorded at row level and is scoped narrowly: the donor's cAMP and PKA evidence is strong, and it is only the G-protein step, resting on melittin alone, that is weak. Proposed replacements: positive regulation of cAMP/PKA signal transduction Supporting Evidence: PMID:36496438 Together, our data suggest that Adissp activates PKA signaling through binding to a putative receptor at adipocyte surface PMID:36496438 induction of Ucp1 expression by Adissp was reduced by a merely 2-h treatment with the PKA inhibitor H89 and was blocked by Melittin, an inhibitor for GΞ±s subunit of the heterotrimeric G protein PMID:36496438 These results together demonstrate that secreted Adissp activates a G protein-cAMP-PKA signaling pathway. PMID:36496438 a specific binding of Adissp to adipose tissue sections was detected, which can be competed away by incubation with Adissp-containing conditioned medium PMID:36496438 Adipose-specific Adissp deletion decreases cAMP content and PKA activity, suppresses inguinal WAT browning, and leads to HFD-induced obesity and hyperglycemia PMID:36496438 cAMP level was increased in primary inguinal adipocytes infected with Adissp adenovirus |
| GO:0042593 glucose homeostasis | IEA GO_REF:0000107 | ACCEPT | Summary: Accepted. Loss and gain of the mouse ortholog move whole-animal glucose handling in opposite directions, and the human protein is 90.2% identical over the full length. Reason: Glucose homeostasis is established in the mouse donor by perturbation in both directions and at the level of the whole animal. Adipose-specific knockout raises steady-state and fasting glucose and degrades glucose tolerance and insulin sensitivity; transgenic and adenoviral gain of function do the reverse; and the 2026 follow-up states flatly that endogenous Adissp is essential for glucose homeostasis, adding that basal Akt activity is reduced in knockout adipose tissue. GO:0042593 is deliberately broad - any process involved in maintaining an internal steady state of glucose - so an adipokine that sets basal Akt tone and glucose disposal in fat sits inside it comfortably. This is a physiological role rather than a molecular activity, and it is recorded as such: it appears in core_functions under directly_involved_in and never as a molecular function. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:0042593 IMP from PMID:36496438. ensembl:ENSMUSP00000099477 CIRCULAR OR REDUNDANT The Ensembl protein of the same mouse gene; the identical entity as UniProtKB:Q9D1K7 in a second namespace, so no independent support. Supporting Evidence: PMID:36496438 Remarkably, both steady-state and fasting glucose levels were higher, and glucose tolerance and insulin sensitivity were deteriorated in the ADKO mice PMID:36496438 Adipose-specific Adissp knockout mice are defective in WAT browning, and are susceptible to high fat diet-induced obesity and hyperglycemia PMID:42030391 Endogenous Adissp is essential for glucose homeostasis. |
| GO:0042593 glucose homeostasis | ISS GO_REF:0000024 | ACCEPT | Summary: Accepted. Loss and gain of the mouse ortholog move whole-animal glucose handling in opposite directions, and the human protein is 90.2% identical over the full length. Reason: Glucose homeostasis is established in the mouse donor by perturbation in both directions and at the level of the whole animal. Adipose-specific knockout raises steady-state and fasting glucose and degrades glucose tolerance and insulin sensitivity; transgenic and adenoviral gain of function do the reverse; and the 2026 follow-up states flatly that endogenous Adissp is essential for glucose homeostasis, adding that basal Akt activity is reduced in knockout adipose tissue. GO:0042593 is deliberately broad - any process involved in maintaining an internal steady state of glucose - so an adipokine that sets basal Akt tone and glucose disposal in fat sits inside it comfortably. This is a physiological role rather than a molecular activity, and it is recorded as such: it appears in core_functions under directly_involved_in and never as a molecular function. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:0042593 IMP from PMID:36496438. Supporting Evidence: PMID:36496438 Remarkably, both steady-state and fasting glucose levels were higher, and glucose tolerance and insulin sensitivity were deteriorated in the ADKO mice PMID:36496438 Adipose-specific Adissp knockout mice are defective in WAT browning, and are susceptible to high fat diet-induced obesity and hyperglycemia PMID:42030391 Endogenous Adissp is essential for glucose homeostasis. |
| GO:1990845 adaptive thermogenesis | IEA GO_REF:0000107 | ACCEPT | Summary: Accepted as the most specific currently expressible term, with the caveat that ADISSP regulates adaptive thermogenesis rather than performing it, and GO has no regulation term for this parent. Reason: Adissp drives the thermogenic program: adipose-specific knockout is defective in WAT browning, transgenic and adenoviral gain of function induce Ucp1 and beige morphology, and the protein's own secretion is stimulated by beta-3-adrenergic activation, which places it downstream of the sympathetic input that initiates adaptive thermogenesis. The term is accepted because it is the most specific thing GO can currently express, but it is not quite the right shape and that is worth stating. ADISSP does not produce heat; UCP1 does. What ADISSP does is positively regulate the process. GO has regulation terms only for the cold-induced child (GO:0120161/GO:0120162/GO:0120163), and those do not fit either, because the phenotype persists at thermoneutrality - transgenic and knockout mice housed at 30 C still differ in adipocyte morphology and Ucp1. So the regulatory term this gene needs does not exist, and a positive regulation of adaptive thermogenesis term is filed under proposed_new_terms rather than the existing cold-induced child being forced. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:1990845 IDA from PMID:36496438. ensembl:ENSMUSP00000099477 CIRCULAR OR REDUNDANT The Ensembl protein of the same mouse gene; the identical entity as UniProtKB:Q9D1K7 in a second namespace, so no independent support. Supporting Evidence: PMID:36496438 Adipose-specific Adissp knockout mice are defective in WAT browning, and are susceptible to high fat diet-induced obesity and hyperglycemia PMID:36496438 Adissp expression is adipose-specific and highly BAT-enriched, and its secretion is stimulated by Ξ²3-adrenergic activation PMID:36496438 Thus, Adissp-regulated thermogenesis is still maintained at thermoneutral condition. |
| GO:1990845 adaptive thermogenesis | ISS GO_REF:0000024 | ACCEPT | Summary: Accepted as the most specific currently expressible term, with the caveat that ADISSP regulates adaptive thermogenesis rather than performing it, and GO has no regulation term for this parent. Reason: Adissp drives the thermogenic program: adipose-specific knockout is defective in WAT browning, transgenic and adenoviral gain of function induce Ucp1 and beige morphology, and the protein's own secretion is stimulated by beta-3-adrenergic activation, which places it downstream of the sympathetic input that initiates adaptive thermogenesis. The term is accepted because it is the most specific thing GO can currently express, but it is not quite the right shape and that is worth stating. ADISSP does not produce heat; UCP1 does. What ADISSP does is positively regulate the process. GO has regulation terms only for the cold-induced child (GO:0120161/GO:0120162/GO:0120163), and those do not fit either, because the phenotype persists at thermoneutrality - transgenic and knockout mice housed at 30 C still differ in adipocyte morphology and Ucp1. So the regulatory term this gene needs does not exist, and a positive regulation of adaptive thermogenesis term is filed under proposed_new_terms rather than the existing cold-induced child being forced. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9D1K7 SUPPORTS TRANSFER Mouse Adissp (ADSSP_MOUSE), reviewed (Swiss-Prot), 174 aa - the same length as human ADISSP and 90.2% identical over an ungapped alignment with no indels. It is the sole experimental source behind every IEA and ISS row on this gene, and it carries its own IDA/IMP annotations from PMID:36496438 for all four transferred terms, so this is an ortholog transfer from measured data, not a family-level guess. For this term it holds GO:1990845 IDA from PMID:36496438. Supporting Evidence: PMID:36496438 Adipose-specific Adissp knockout mice are defective in WAT browning, and are susceptible to high fat diet-induced obesity and hyperglycemia PMID:36496438 Adissp expression is adipose-specific and highly BAT-enriched, and its secretion is stimulated by Ξ²3-adrenergic activation PMID:36496438 Thus, Adissp-regulated thermogenesis is still maintained at thermoneutral condition. |
| GO:0008157 protein phosphatase 1 binding | IDA PMID:32024300 C20orf27 Promotes Cell Growth and Proliferation of Colorecta... | ACCEPT | Summary: Accepted, and it is the only molecular function this gene has in either species. The interaction is real and independently replicated; the evidence code is the wrong one. Reason: The term is right and it is informative - GO:0008157 is isoform-agnostic, which matches the data, since the paper's mass spectrometry identified only 'PP1c' without resolving which catalytic subunit. Two things qualify it. First, the cited experiment is a co-immunoprecipitation between two transfected, tagged plasmids, which is not a direct assay at native levels; for a physical interaction the code should be IPI with the partner in WITH/FROM, and this row's WITH/FROM is empty, so the partner is recorded nowhere machine-readably. Second, and pulling the other way, the interaction does not rest on that one experiment: IntAct logs ADISSP against PPP1CA, PPP1CB, PPP1CC and the regulatory subunit PPP1R7 across seven publications that the annotation does not cite (PMIDs 24366813, 27173435, 27880917, 28330616, 28514442, 33961781, 40205054) by anti-tag co-IP, TAP and pull-down. That recurrence needs its null, and the null is unflattering: PP1 catalytic subunits are among the most frequently recovered proteins in affinity proteomics, with 1012 IntAct records for PPP1CA against 27 for ADISSP, so simply reappearing in tag pulldowns proves little. The informative direction is the subject-centric one - four of ADISSP's thirteen distinct IntAct protein partners are PP1-module components, so the PP1 module dominates this protein's own sparse interactome rather than ADISSP being one more name on PP1's long list. Finally, a topological objection that would ordinarily apply here does not. ADISSP is annotated to the extracellular region, and PP1c is cytosolic and nuclear, which on a signal-peptide protein in a secretory lumen would be a strong argument against the interaction. ADISSP is leaderless and secreted non-classically, so it necessarily has a cytosolic pool, and the two annotations are compatible. Supporting Evidence: PMID:32024300 CoIP experiments demonstrated that C20orf27 and PP1c can be co-precipitated, indicating an interaction between the two proteins PMID:32024300 the C20orf27 and PP1c plasmids with the marker were transfected into HCT15 and DLD-1 cells PMID:32024300 the interaction between PP1c and the regulatory subunit GADD34 was reduced with C20orf27 overexpression |
| GO:0030511 positive regulation of transforming growth factor beta receptor signaling pathway | IMP PMID:32024300 C20orf27 Promotes Cell Growth and Proliferation of Colorecta... | KEEP AS NON CORE | Summary: Kept, but as a non-core, context-restricted role. The measurement is sound and bidirectional; it is a downstream consequence of the PP1c interaction in colorectal cancer lines, not this gene's evolved function. Reason: The evidence code is right - overexpression in HCT15 and DLD-1 raised phospho-TGFbetaR1 and silencing in HT29 and SW480 lowered it, which is a mutant-phenotype experiment run in both directions. The term also survives a definition check that it might not have: GO:0007179, the process being regulated, is defined simply as signals initiated by a ligand binding a TGF-beta receptor and ending in regulation of a downstream process, with no requirement that the arm be Smad-mediated, so the TAK1 branch this paper follows is inside the term. Two limitations keep it out of core_functions. No exogenous TGF-beta was added, so what was measured is basal receptor phosphorylation in cancer lines with presumed autocrine ligand; and the effect is mechanistically downstream of ADISSP sequestering PP1c away from the TGFbetaR1-Smad7-GADD34 complex, which makes it a consequence of the molecular function rather than a second function. Nothing connects it to the adipose biology that the rest of this record describes. Supporting Evidence: PMID:32024300 with C20orf27 overexpression in HCT15 and DLD-1 cells, the expression of p-TGFΞ²R1, p-TAK1, p-IKK, p-IΔΈB, and p-p65 was increased PMID:32024300 in HT29 and SW480 cells with C20orf27 silencing, p-TGFΞ²R1, p-TAK1, p-IKK, p-IΔΈB, and p-p65 expressions were reduced PMID:32024300 C20orf27 inhibits the formation of PP1 holoenzyme by interacting with PP1c, thus reducing the inhibitory effect of PP1 on the TGFΞ²R-TAK1-NFΔΈB pathway. |
| GO:1901224 positive regulation of non-canonical NF-kappaB signal transduction | IMP PMID:32024300 C20orf27 Promotes Cell Growth and Proliferation of Colorecta... | MODIFY | Summary: Modified to positive regulation of canonical NF-kappaB signal transduction. Every NF-kappaB readout in the cited paper is a canonical marker; nothing diagnostic of the non-canonical cascade was measured. Reason: The two branches are distinguished by GO on mechanism, and the definitions decide this cleanly. GO:0038061 non-canonical NF-kappaB signal transduction requires NIK-dependent processing: NIK activates IKKalpha, IKKalpha phosphorylates p100, and p100 is processed to release p52. GO:0007249 canonical NF-kappaB signal transduction is IKK-complex-dependent activation running through IkappaB to the RelA-containing dimer. What PMID:32024300 measures, in both the overexpression and the silencing direction, is phospho-IKK, phospho-IkappaB and phospho-p65 - RelA. NIK, p100/p52 processing and RelB appear nowhere in the paper. Those are canonical markers exclusively. This is a wrong term rather than an imprecise one, and I checked rather than assumed: I fetched the is_a/part_of ancestor closures of GO:1901224 and GO:0043123 and neither contains the other, so they are siblings naming different cascades, not a parent quietly covering the data. Note for anyone re-running that check that neither regulation term has its target cascade in its is_a closure either, because GO links them by positively_regulates, which does not subsume. The replacement asserts exactly what was measured. Like the TGF-beta row this remains non-core: it is the cancer-cell-line consequence of the PP1c interaction. Proposed replacements: positive regulation of canonical NF-kappaB signal transduction Supporting Evidence: PMID:32024300 with C20orf27 overexpression in HCT15 and DLD-1 cells, the expression of p-TGFΞ²R1, p-TAK1, p-IKK, p-IΔΈB, and p-p65 was increased PMID:32024300 in HT29 and SW480 cells with C20orf27 silencing, p-TGFΞ²R1, p-TAK1, p-IKK, p-IΔΈB, and p-p65 expressions were reduced |
| GO:0005575 cellular_component | ND GO_REF:0000015 | REMOVE | Summary: Removed. This is a 2012 placeholder meaning no cellular-component information was available, and the same GOA record now carries three extracellular region annotations, so it contradicts its own file. Reason: The ND evidence code against a root term is not a statement about the protein but about the state of knowledge on the day it was made. GO:0005575's own comment says the term is recommended for gene products whose cellular component is unknown and that using it indicates no information was available. This row was assigned by BHF-UCL on 2012-07-16, a decade before PMID:36496438. ADISSP now has three GO:0005576 extracellular region rows, one of them tracing to a mouse IDA, so the assertion the ND row encodes is no longer true and GO's own convention is that it should be retired once any annotation in that aspect exists. Nothing biological is lost by removing it. It is worth distinguishing this from the other actions in this review: it is not a propagation failure of any kind - there is no WITH/FROM and no source to inspect - just a placeholder that outlived its purpose. Supporting Evidence: file:human/ADISSP/ADISSP-uniprot.txt CC -!- SUBCELLULAR LOCATION: Secreted {ECO:0000250|UniProtKB:Q9D1K7}. |
| GO:0048018 receptor ligand activity | ISS GO_REF:0000024 | NEW | Summary: New. The gene's headline activity - a secreted protein that engages an adipocyte cell-surface receptor and changes its output - has no molecular function term in either species, and this is the gap. Reason: The coverage gap is real and I checked it rather than inferring it: querying QuickGO for molecular function annotations returns one for human ADISSP and two for mouse Adissp, and all of them are GO:0008157 protein phosphatase 1 binding, transferred between the two species. Neither species has any term for the adipokine activity that both papers are about. What is measured supports GO:0048018 squarely. Labelled and SEAP-fused Adissp binds adipose tissue sections and the brown-adipocyte surface, and the binding is competed away by unlabelled protein, so a saturable surface site exists. Purified protein then changes that cell's behaviour: PKA substrate phosphorylation rises dose- and time-dependently, and in the 2026 study recombinant protein activates Akt and AS160 with appreciable activity at 25 nM. That study also answers the obvious objection about what molecule was assayed - it is full-length recombinant protein, and untagged protein reproduces the effect, so the result is not an artefact of the His tag or of the purification. The same protein activates Akt in human SGBS adipocytes, so the responding system is present in human cells. GO:0048018 is the right term because its definition covers a gene product that interacts with a receptor to change the receptor's activity, explicitly including ligands that are secreted and diffuse to a receiving cell, and it does not require the receptor to have been identified. I verified its placement rather than assuming it: GO:0048018 is not a descendant of GO:0005102 signaling receptor binding, and not under GO:0005488 binding at all - it sits under GO:0140677 molecular function activator activity and GO:0098772 molecular function regulator activity. Evidence is ISS from mouse Q9D1K7 because the functional assays used the mouse protein; WITH/FROM carries the sequence-similar entity only, as ISS requires. The unidentified receptor is recorded in knowledge_gaps and is the reason no downstream pathway term is proposed alongside this one. Supporting Evidence: PMID:36496438 a specific binding of Adissp to adipose tissue sections was detected, which can be competed away by incubation with Adissp-containing conditioned medium PMID:36496438 Together, our data suggest that Adissp activates PKA signaling through binding to a putative receptor at adipocyte surface PMID:42030391 rAdissp robustly activated Akt and induced phosphorylation of its substrate AS160, key regulator of GLUT4 trafficking (30), in a dose-dependent manner, with appreciable activity at 25 nM and substantial effect around 300 nM PMID:42030391 This activation was not attributable to the presence of the His tag or any potential contaminants inherent in the purification process, as rAdissp lacking the His tag similarly activated Akt PMID:42030391 Mature human SGBS adipocytes treated with 300 nM rAdissp for 30 min (p-Akt) or 12 hours (Ucp1). |
Loading supporting contentβ¦
Download this section (compressed HTML)Q: For GOA and FlyBase: all three of this gene's experimental annotations (GO:0008157, GO:0030511, GO:1901224) are assigned by FlyBase from PMID:32024300, a human colorectal cancer study with no Drosophila content. A reference-projection query confirms the reference annotates only this one entity, so it is a targeted curation rather than a bulk import, but is the assigning database correct?
Suggested experts: GOA curators, FlyBase curators
Q: For GOA: the GO:0008157 row is coded IDA with an empty WITH/FROM, but the underlying experiment is a co-immunoprecipitation between two transfected tagged plasmids. Should it be IPI with the PP1 catalytic subunit recorded in WITH/FROM? As it stands the partner - the only informative part of a binding annotation - is not machine-readable anywhere in the record.
Suggested experts: GOA curators
Q: For GOA and UniProt: the GO:0005576 IEA row cites GO_REF:0000120, the combined-multiple-IEA-methods reference whose purpose is to record that independent pipelines agreed. Its WITH/FROM names three tokens, but all three reduce to one source: mouse Q9D1K7, the same protein's Ensembl identifier, and UniProtKB-SubCell:SL-0243, which on the human entry is itself asserted as ECO:0000250 by similarity to that same mouse protein. Should the combined reference distinguish pipelines that reached a conclusion independently from pipelines that re-read each other's output? At present a reader has no way to tell which kind of agreement a GO_REF:0000120 row records.
Suggested experts: GOA curators, UniProt curators
Q: For UniProt: the Q9GZN8 entry still carries DR GO; GO:0005615; C:extracellular space, and GO:0005615 was obsoleted on 2026-03-06 as representing the same concept as GO:0005576 extracellular region. GOA has already moved to GO:0005576, so the cross-reference in the UniProt entry is stale.
Suggested experts: UniProt curators
Q: For UniProt: would a CAUTION or a note on the non-classical secretion route be warranted? The entry carries the Secreted keyword with no SIGNAL feature, which reads as an omission, whereas it is a measured property - brefeldin A and monensin increase rather than block secretion, the mature chain begins at an N-acetylated residue 2, and the authors of PMID:36496438 had to graft a heterologous signal peptide onto the construct to purify secreted protein. Recording the route would stop future readers treating the missing signal peptide as evidence against secretion.
Suggested experts: UniProt curators
Q: For GO: no term exists for positive regulation of adaptive thermogenesis; the regulation branch stops at the cold-induced child. Filed in full under proposed_new_terms.
Suggested experts: GO Central curators
Q: For GO curators comparing adipokines: ADIPOQ and LEP both carry GO:0005179 hormone activity, and ADIPOQ additionally GO:0005125 cytokine activity, whereas the molecular function proposed here for ADISSP is GO:0048018 receptor ligand activity. The two are not alternatives: GO:0005179 is a verified descendant of GO:0048018 (checked against QuickGO - GO:0048018, GO:0140677, GO:0098772 and GO:0030545 are all in GO:0005179's is_a/part_of ancestor closure, and GO:0005102 and GO:0005488 are not), so the parent was chosen deliberately. GO:0048018 asserts only what has been measured - engagement of a receptor with a consequence for that receptor's output - whereas hormone activity adds an endocrine claim that holds for the mouse protein but has not been tested for the human one. Once recombinant human ADISSP has been assayed, refining this row to GO:0005179 would bring ADISSP into line with ADIPOQ and LEP without retracting anything asserted here.
Suggested experts: GO Central curators
Q: For whoever maintains the PAINT worklist: projects/paint/human-no-IBA-simple.csv contains both 'human,Q9GZN8,ADISSP' and 'human,Q6UWT4,C5orf46'. These are different genes on different chromosomes - ADISSP is HGNC:15873 at 20p13, previously C20orf27, renamed 2022-12-12; C5orf46 is HGNC:33768 at 5q32, whose aliases are SSSP1 and AP-64. Anything that treats ADISSP as 'formerly C5orf46/SSSP1' has conflated two entries in the same file.
Suggested experts: ai-gene-review project maintainers
Q: For GO/GOA coverage: PMID:42030391 (Sci Adv 2026) carries zero GO annotations. It establishes a second signalling output for this protein - insulin-independent PI3K-PDK1/mTORC2-Akt activation and GLUT4 translocation - which is currently absent from the GO record for either species.
Suggested experts: GOA curators, GO Central curators
Experiment: Identify the receptor. The binding assay already developed in PMID:36496438 - labelled protein binding adipose sections and adipocyte surface, competed by unlabelled protein - is the assay to scale up: photo-crosslinkable or proximity-labelling ADISSP on primary adipocytes followed by mass spectrometry, or a CRISPR knockout screen selecting for loss of the cAMP or Akt response. Naming the receptor is what would let the adenylate-cyclase-activating GPCR annotation be either justified or replaced with the correct receptor-class term.
Experiment: Test recombinant human ADISSP in the mouse assays. Every functional measurement to date used the mouse protein; the human protein has only been shown to be secreted. Running human protein through the PKA, cAMP, Akt/AS160 and GLUT4-translocation assays, ideally in human SGBS adipocytes where mouse protein is already known to work, would convert eight ortholog-transferred rows into direct human evidence.
Experiment: Replace melittin with a specific test of G-protein dependence. The entire link between ADISSP and a heterotrimeric G protein is one experiment with 1 uM melittin, a membrane-active peptide. Gnas-deleted or Gnas-silenced adipocytes, or a selective Gs antagonist, would establish or exclude Gs coupling; pertussis and cholera toxin would discriminate the coupling class. Until then the receptor-class element of GO:0007189 is unmeasured rather than established.
Experiment: Ask whether protein phosphatase 1 binding matters in adipocytes, and whether it is connected to secretion. The PP1c interaction was characterised only in colorectal cancer lines with overexpressed tagged protein. Endogenous co-immunoprecipitation in differentiated adipocytes, plus a separation-of-function mutant that loses PP1 binding while retaining secretion, would show whether the two activities are one mechanism or two. A candidate route to such a mutant is to map the PP1-docking surface, which is currently unknown - note that a first-pass motif scan run for this review was discarded because its own positive controls failed, so no motif claim is made here.
Experiment: Determine the unconventional secretion route. That brefeldin A and monensin increase rather than block release places ADISSP with IL-1beta and MIF, but the specific pathway is untested. Candidate-directed knockdowns (GRASP55, TMED10, autophagy and secretory-lysosome components) with the existing conditioned-medium assay, and the FCCP sensitivity already observed, would narrow it, and would tell whether any GO biological-process term about the secretion mechanism is applicable to this cargo.
What is not known β curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: The receptor ADISSP acts through is unknown. Its existence is inferred from specific, competable binding to the adipocyte surface and from two downstream cascades that the purified protein switches on, but no candidate has been named, and the authors describe it as putative. This is the single fact whose absence propagates furthest through the record: it is why the adenylate-cyclase-activating GPCR term was moved to a regulation term that makes no claim about receptor class, and why the molecular function proposed here is the generic receptor ligand activity rather than anything about receptor class or coupling.
OPEN BIOLOGY RESIDUAL_SUBGAP
What is known: Known: a saturable, competable binding site is present on adipocytes and on adipose tissue sections but not on liver, skeletal muscle or brain; engaging it raises cAMP and activates PKA independently of beta-adrenergic receptors, and separately activates PI3K-PDK1/mTORC2-Akt and GLUT4 translocation; human SGBS adipocytes respond, so the receptor is present in human cells. Unknown: the identity of the receptor, whether one receptor serves both cascades, and whether it is a G-protein-coupled receptor at all - the only evidence bearing on that is a single melittin experiment.
Significance: Without it, the gene's molecular function can only be stated at the level of 'is a ligand for something', and the mechanism terms that GO would otherwise offer cannot be justified.
Provenance (the field's own admissions):
Gap: No functional experiment has been performed on the human protein. Human ADISSP has been shown to leave a cell - adenoviral expression in mouse liver put it into the circulation - and human adipocytes secrete it and human plasma contains it. But every assay of what the protein does, from receptor binding through PKA and Akt activation to the metabolic phenotypes, used the mouse orthologue. The eight IEA and ISS rows on this gene therefore rest on an ortholog transfer, which is well licensed at 90.2% identity over identical length with no indels, but is a transfer nonetheless.
NARROWING BIOLOGYCURATION MF_DARK
What is known: Known for the human protein: it is expressed, detected at protein level by direct sequencing and mass spectrometry, secreted, present in circulation, lower in adipocytes of obese than lean subjects, and it binds PP1c. Known for the mouse protein only: receptor binding, cAMP/PKA activation, Akt activation, GLUT4 translocation, browning, and the glucose and obesity phenotypes. Narrowing rather than open because recombinant mouse protein does activate Akt in human SGBS adipocytes, so the human receiving system is demonstrably competent even though the human ligand has not been tested.
Significance: Testing recombinant human ADISSP in the same assays would convert eight transferred rows into direct human evidence and is a small experiment relative to that gain.
Provenance (the field's own admissions):
Gap: Nothing connects the two activities recorded for this protein. ADISSP is a leaderless secreted hormone acting on a surface receptor, and it is also a cytosolic binder of protein phosphatase 1; the two literatures do not cite each other and no experiment addresses whether the phosphatase interaction is part of the hormonal mechanism, a regulator of the protein's own secretion, or an unrelated second job. The leaderless route makes the coexistence physically possible, which removes the easy way of dismissing one of them, and leaves the question genuinely open.
OPEN BIOLOGY RESIDUAL_SUBGAP
What is known: Known: the protein has a cytosolic pool because it is secreted non-classically; it binds PP1 catalytic subunits and PPP1R7; and it acts from outside the cell on an adipocyte receptor. Unknown: whether PP1 binding influences secretion, whether it operates in adipocytes at all rather than only in the cancer lines where it was studied, and whether the adipose and cancer phenotypes share any mechanism.
Significance: The answer determines whether protein phosphatase 1 binding belongs in the same functional account as the hormonal activity or is a separate core function that happens to reside in the same 174 residues.
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)