ABHD14A is an uncharacterised member of the alpha/beta-hydrolase superfamily, assigned to the ABHD14 family and carrying only a generic hydrolase EC number (3.-.-.-). No molecular function has been established for it: a 2023 study of the family states plainly that ABHD14A still lacks any functional annotation. It is closely related in sequence to ABHD14B, which is a lysine deacetylase. UniProt places ABHD14A in the cytoplasm and in membranes, annotating it as a single-pass type II membrane protein with a signal-anchor helix at residues 35-55 and N-glycosylation at 67 and 201 - a topology that would put its hydrolase domain on the lumenal side and that distinguishes it from the soluble ABHD14B. It offers only a by-similarity suggestion of a role in granule neuron development, which derives from work on the rodent ortholog Dorz1, a transcript expressed in differentiating cerebellar granule neurons and positively regulated by the transcription factor Zic1. Whether the protein has any function in that setting has not been tested.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005737 cytoplasm | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic inference that ABHD14A is active in the cytoplasm. Accepted, but with the contrary features on record rather than passed over. Topology, which the paralog framing must not obscure: UniProt annotates ABHD14A as a single-pass type II membrane protein with a signal-anchor transmembrane helix at 35-55 and N-glycosylation sites at 67 and 201. Type II topology with glycosylated residues C-terminal to the anchor places the hydrolase domain on the lumenal or extracellular side, not free in the cytosol. ABHD14B, by contrast, is soluble. So the two proteins are not interchangeable even at the level of where they sit, and 'the family behaves this way' is not a reason to accept a cytoplasmic call for this one. The IBA is nonetheless retained: it is evidence independent of the by-similarity UniProt call, and a type II membrane protein does present a short cytosolic N-terminal tail, so a cytoplasmic assignment is not incompatible with the topology. What the IBA does not establish is where the catalytic domain faces - which is the question that matters for this gene, and which the excluded preprint's Golgi result would bear on. Supporting Evidence: file:human/ABHD14A/ABHD14A-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm {ECO:0000250}. Membrane {ECO:0000305}; file:human/ABHD14A/ABHD14A-uniprot.txt Single-pass type II membrane protein {ECO:0000305}. |
| GO:0005737 cytoplasm | IEA GO_REF:0000044 | ACCEPT | Summary: Electronic mapping from the UniProt cytoplasm location. UniProt's own location is ECO:0000250, by similarity, so this is an inference on an inference, and the similarity in question runs to a family whose other characterised member, ABHD14B, is soluble - exactly the transfer this review flags as the gene's chief hazard. Accepted on the strength of the independent IBA rather than on family behaviour, and with the topology recorded above as a standing caveat. Supporting Evidence: file:human/ABHD14A/ABHD14A-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm {ECO:0000250}. Membrane {ECO:0000305}; file:human/ABHD14A/ABHD14A-uniprot.txt Single-pass type II membrane protein {ECO:0000305}. |
| GO:0016020 membrane | IEA GO_REF:0000044 | MARK AS OVER ANNOTATED | Summary: Membrane, from the UniProt SubCell mapping. The underlying assertion is ECO:0000305 - a curator inference rather than an observation - and no particular membrane is named, so the annotation places the protein in an unspecified bilayer. That is the grounds for marking it over-annotated: not that nothing is known about its membrane association, but that this term records none of what is. What is annotated, and is more informative than this term, is the topology: a signal-anchor transmembrane helix at residues 35-55 and single-pass type II orientation, with N-glycosylation at 67 and 201. A specific membrane and a topology-aware term would both be improvements; neither can be justified from an ECO:0000305 inference alone, and no experimental localisation exists for this protein. Reason: Unspecified membrane from a curator-inferred UniProt statement. Marked over-annotated for lack of specificity rather than for lack of membrane association - the protein does carry an annotated signal-anchor and type II topology, which this term fails to convey. Supporting Evidence: file:human/ABHD14A/ABHD14A-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm {ECO:0000250}. Membrane {ECO:0000305}; file:human/ABHD14A/ABHD14A-uniprot.txt Single-pass type II membrane protein {ECO:0000305}. file:human/ABHD14A/ABHD14A-uniprot.txt FT /note="Helical; Signal-anchor for type II membrane protein" |
Loading supporting content…
Download this section (compressed HTML)Q: What does ABHD14A do? The 2023 family study states directly that it still lacks any functional annotation, and that remains true: UniProt assigns only a generic EC 3.-.-.- and a by-similarity guess at a role in granule neuron development. No substrate, no reaction and no phenotype has been established for the human protein.
Q: Which side of the membrane does the hydrolase domain face? UniProt annotates a signal-anchor at 35-55 with single-pass type II topology and N-glycosylation at 67 and 201, which together place the catalytic domain lumenally or extracellularly - unlike the soluble ABHD14B. That prediction is ECO:0000255 and untested, but it would determine what substrates the protein could reach and would reconcile with the Golgi localisation reported in the excluded preprint.
Q: Is ABHD14A being confused with ABHD14B? This is the specific hazard for this gene rather than a general caution. ABHD14B is a lysine deacetylase, sequence similarity between the two is high, and PMID:37974539 was written because automated databases frequently assign them the same function. Any future functional annotation on ABHD14A should be checked against the sequence determinants that paper defines, and any transfer from ABHD14B should be treated as a paralog transfer, not an ortholog one.
Q: Is the granule-neuron connection about the protein or only the transcript? The link comes from Dorz1, a rodent transcript expressed in differentiating cerebellar granule neurons and regulated by Zic1. That is expression biology - being downstream of a transcription factor is not participation in a process - and no ABHD14A protein-level experiment in neurons has been reported.
Q: Is the protein expressed at all? A recent preprint reports that endogenous ABHD14A is undetectable across immortalized cell lines and adult mouse tissues despite transcriptomic predictions. If that holds under peer review it would reframe the whole gene, and would explain why no function has been found.
Experiment: Profile recombinant ABHD14A against ABHD14B with activity-based protein profiling and a broad ester and thioester substrate panel, using the sequence determinants from PMID:37974539 to confirm construct identity, so that any activity found is attributable to ABHD14A rather than to contaminating or misassigned paralog.
Hypothesis: ABHD14A is a serine hydrolase with a substrate preference distinguishable from ABHD14B.
Type: enzymology
Experiment: Raise or validate an antibody against a region divergent from ABHD14B and survey developing cerebellum, where the Dorz1 transcript is enriched, alongside the adult tissues where the protein has been reported undetectable. A negative result across development would be as informative as a positive one.
Hypothesis: Endogenous ABHD14A protein is expressed in a restricted setting rather than not at all.
Type: expression analysis
Experiment: Knock out Abhd14a in mouse and examine cerebellar granule neuron precursor proliferation, differentiation and migration, comparing with the Zic1 phenotype to test whether the gene does anything downstream of the transcription factor that regulates it.
Hypothesis: ABHD14A has a function in cerebellar granule neuron differentiation.
Type: mouse genetics
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)