Rat Pnkd (paroxysmal nonkinesigenic dyskinesia protein; myofibrillogenesis regulator 1, MR-1) encodes a glyoxalase-II-related metallo-beta-lactamase-fold protein of the PNKD-related subfamily (PANTHER PTHR11935:SF116), a lineage distinct from the canonical hydroxyacylglutathione hydrolase HAGH. D3ZXB0 is the 424-residue canonical rat product. It matches the mouse long isoform 3 almost exactly: the shared N-terminal exon, which carries the disease-mutated alanines and a cleavable mitochondrial targeting sequence, is followed by a 39-residue hydrophobic segment, also used by the short isoform MR-1S, and then by the complete metallo-beta-lactamase and HAGH_C domains with all seven zinc-ligand residues conserved. By orthology with the characterized human and mouse long isoform (PNKD-L), the protein is a probable zinc-dependent thioesterase. It hydrolyses S-D-lactoylglutathione only at a trace rate and cannot replace HAGH, so its physiological substrate, perhaps another alpha-hydroxythioester, is unknown. In neurons PNKD-L sits in presynaptic terminals, binds the active-zone scaffolds RIM1 and RIM2, and restrains synaptic-vesicle exocytosis. Reports disagree on whether the long isoform is mitochondrial (inner membrane, after MTS cleavage) or associated with the plasma membrane. None of these properties has been tested on the endogenous rat protein. Rat data consist of overexpression of exogenous PNKD in rat neurons and of human MR-1 in rat cardiomyocytes.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0004416 hydroxyacylglutathione hydrolase activity | IEA GO_REF:0000002 | MODIFY | Summary: InterPro2GO transfer of the HAGH-specific reaction to a PNKD-subfamily protein. The metallo-beta-lactamase fold and binuclear zinc site are conserved, which supports a thiolester hydrolase activity, but S-D-lactoylglutathione is not a physiological substrate of PNKD-L. Reason: In PAINT for PTHR11935 the curator placed GO:0016790 thiolester hydrolase activity at the root node PTN004322580, which reaches the PNKD-related subfamily SF116 (human Q8N490 and mouse Q69ZP3 both carry that IBA). GO:0004416 was placed at PTN000232795, a node that does not include the PNKD branch. Treat that omission as deliberate. It matches the biochemistry: human PNKD-L hydrolyses SLG only at a trace rate, and a PNKD-L transgene does not rescue HAGH-null flies. D3ZXB0 is the rat ortholog in SF116 and conserves all seven Zn(2+) ligands, so the parent activity is the defensible term. It is inferred by orthology and has not been measured on the rat protein. Proposed replacements: thiolester hydrolase activity Supporting Evidence: PMID:21487022 PNKD-L has extremely low but consistent SLG hydrolysis activity PMID:21487022 further supporting the idea that PNKD-L does not use SLG as an in vivo substrate file:rat/Pnkd__D3ZXB0/Pnkd__D3ZXB0-bioinformatics/RESULTS-d3zxb0.md | 291 | H | 330 | H | file:rat/Pnkd__D3ZXB0/Pnkd__D3ZXB0-bioinformatics/RESULTS-d3zxb0.md PAINT nodes seeded by human PNKD Q8N490 |
| GO:0005737 cytoplasm | IEA GO_REF:0000117 | MODIFY | Summary: True but uninformative. The best-supported intracellular compartment for a product that carries the PNKD N-terminal mitochondrial targeting sequence is the mitochondrion, which is part of the cytoplasm. Reason: The human "diffuse cytoplasmic" localization was reported for the short isoform MR-1S. Later work showed that MR-1L and MR-1S are imported into mitochondria through a cleavable 39-residue N-terminal MTS. D3ZXB0 has that N-terminal exon unchanged (residues 1-79 identical to the mouse long isoform), so mitochondrion is the specific supported term. The gene already carries it. Proposed replacements: mitochondrion Supporting Evidence: PMID:19124534 both MR-1L and MR-1S isoforms are mitochondrial proteins, imported into the organelle thanks to a 39 amino acid-long, N-terminal mitochondrial targeting sequence (MTS) |
| GO:0005737 cytoplasm | ISO GO_REF:0000121 | MODIFY | Summary: RGD ortholog transfer of the human cytoplasm annotation, which rests on the cytoplasmic and nuclear staining of the short isoform MR-1S. It is true but general for a product that carries the mitochondrial targeting sequence. Reason: Same claim as the IEA cytoplasm row, so the same action applies. Human PNKD is the 1:1 ortholog in PANTHER SF116, and the transfer is sound at gene level. For D3ZXB0, the specific supported compartment is the mitochondrion. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: UniProtKB:Q8N490 · PNKD SUPPORTS SOURCE BUT NOT TARGET Human cytoplasm EXP (PMID:15496428) describes MR-1S staining; for the MTS-bearing long product the specific compartment is mitochondrion. Proposed replacements: mitochondrion Supporting Evidence: PMID:15496428 MR-1S isoform is ubiquitously expressed and shows diffuse cytoplasmic and PMID:19124534 both MR-1L and MR-1S isoforms are mitochondrial proteins |
| GO:0005739 mitochondrion | IEA GO_REF:0000120 | ACCEPT | Summary: Supported by orthology. In human, the long and short PNKD isoforms are imported into mitochondria by an N-terminal MTS that is cleaved before insertion in the inner membrane. D3ZXB0 conserves that N-terminal exon, and PAINT placed mitochondrion on the PNKD clade node. Reason: Ghezzi et al. (2009) showed mitochondrial import and MTS processing for MR-1L and MR-1S. Shen et al. (2011) independently showed N-terminal cleavage of PNKD-L. PAINT placed GO:0005739 on PTN000940186 (Eumetazoa), seeded by human Q8N490, and the rat gene falls in that clade (SF116). Other studies place PNKD-L at the plasma membrane, so the localization is disputed. The mitochondrion is still the best-supported organelle for a product with the full MTS. Not tested in rat. Supporting Evidence: PMID:19124534 The MTS, which contains all three PNKD mutations, is then cleaved off the mature proteins before their insertion in the inner mitochondrial membrane. file:rat/Pnkd__D3ZXB0/Pnkd__D3ZXB0-bioinformatics/RESULTS-d3zxb0.md | 33 | A | 33 | A | |
| GO:0005739 mitochondrion | ISO GO_REF:0000121 | ACCEPT | Summary: RGD ortholog transfer from human PNKD mitochondrial localization. This is the same claim as the IEA row. Reason: Consistent with the IEA mitochondrion row. The transfer is sound because human Q8N490 is the SF116 ortholog and D3ZXB0 conserves the cleavable N-terminal MTS. Supporting Evidence: PMID:19124534 both MR-1L and MR-1S isoforms are mitochondrial proteins |
| GO:0005739 mitochondrion | ISO GO_REF:0000121 | ACCEPT | Summary: RGD ortholog transfer from human PNKD isoform 2 (MR-1S) mitochondrial localization. This is the same claim as the other mitochondrion rows. Reason: The MR-1S source is isoform-specific, but the MTS-containing exon behind mitochondrial import is shared with D3ZXB0. Ghezzi et al. reported import for both the long and the short isoform. Supporting Evidence: PMID:19124534 both MR-1L and MR-1S isoforms are mitochondrial proteins |
| GO:0005783 endoplasmic reticulum | ISO GO_REF:0000121 | MARK AS OVER ANNOTATED | Summary: Transferred from the human ER localization of the medium isoform MR-1M. MR-1M lacks the mitochondrial targeting exon and has a different N-terminus. D3ZXB0 is not that product. Reason: Ghezzi et al. attribute the ER, Golgi and plasma membrane localization to MR-1M and the mitochondrial localization to MR-1L and MR-1S. D3ZXB0 has the MR-1L/MR-1S N-terminal exon, identical over residues 1-79 to the mouse long isoform, and only about 75% identity to the medium-type products (mouse Q69ZP3-4, human Q8N490-3). The rat gene probably also makes an MR-1M-type product (the 360-residue F1M1E4), so ER localization may be true of the gene. It is not supported for this MTS-bearing product. Propagation Review Root cause: PROPAGATION BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB:Q8N490 · PNKD SUPPORTS SOURCE BUT NOT TARGET Human ER EXP (PMID:19124534) is for the medium isoform MR-1M, which has a different N-terminus from D3ZXB0. Supporting Evidence: PMID:19124534 the mutation-free MR-1M is localized in the Golgi apparatus, ER and plasma membrane file:rat/Pnkd__D3ZXB0/Pnkd__D3ZXB0-bioinformatics/RESULTS-d3zxb0.md | Q69ZP3-4 | 360 | 320 | 0.755 | |
| GO:0005794 Golgi apparatus | ISO GO_REF:0000121 | MARK AS OVER ANNOTATED | Summary: Transferred from the human Golgi localization of the medium isoform MR-1M, which has a different N-terminus from D3ZXB0. Reason: Same reasoning as the ER row. The Golgi signal belongs to MR-1M, while the MTS-bearing long and short products are mitochondrial. Propagation Review Root cause: PROPAGATION BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB:Q8N490 · PNKD SUPPORTS SOURCE BUT NOT TARGET Human Golgi EXP (PMID:19124534) is for the medium isoform MR-1M, which has a different N-terminus from D3ZXB0. Supporting Evidence: PMID:19124534 the mutation-free MR-1M is localized in the Golgi apparatus, ER and plasma membrane |
| GO:0005886 plasma membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Plasma-membrane association was reported for the human brain-specific long isoform PNKD-L, which D3ZXB0 matches outside the 39-residue exon. A later study found it to be mitochondrial instead. Keep as a disputed, non-core location. Reason: Lee et al. (2004) and Shen et al. (2011) place PNKD-L at the cell membrane. Lee et al. (2012) found one isoform membrane-associated in mouse brain, and PNKD-L is found in presynaptic terminals. Ghezzi et al. (2009) found MR-1L in mitochondria, and the conflict has not been resolved. The term is as specific as the evidence allows, but it does not describe where the core activity is established, so it is kept as non-core. Supporting Evidence: PMID:15496428 The MR-1L isoform is specifically expressed in brain and is localized to the cell membrane PMID:25730884 The PNKD long isoform (PNKD-L) is CNS-specific and localizes to the cell membrane in vitro. |
| GO:0005886 plasma membrane | ISO GO_REF:0000121 | KEEP AS NON CORE | Summary: RGD ortholog transfer of the human PNKD-L plasma-membrane localization. This is the same claim as the IEA row. Reason: Consistent with the IEA plasma membrane row. The localization is disputed between plasma membrane and mitochondria and has not been tested in rat. Supporting Evidence: PMID:15496428 The MR-1L isoform is specifically expressed in brain and is localized to the cell membrane |
| GO:0016020 membrane | IEA GO_REF:0000107 | MODIFY | Summary: True but general. The specific membranes supported for the long PNKD product are the plasma membrane and the mitochondrial inner membrane. Reason: Membrane association is well supported. Lee et al. (2012) found it in mouse brain, and D3ZXB0 has a predicted hydrophobic segment at residues 74-92. Mitochondrial import is followed by insertion in the inner membrane (Ghezzi et al. 2009). The plasma membrane association reported for PNKD-L is already carried by the gene. Both supported specific terms are listed, because the literature disputes which applies. Proposed replacements: plasma membrane mitochondrial inner membrane Supporting Evidence: PMID:22214848 One PNKD isoform is a membrane-associated PMID:19124534 before their insertion in the inner mitochondrial membrane file:rat/Pnkd__D3ZXB0/Pnkd__D3ZXB0-uniprot.txt TRANSMEM 74..92 |
| GO:0016020 membrane | ISO GO_REF:0000121 | MODIFY | Summary: RGD ortholog transfer of the human membrane annotation. This is the same claim as the IEA membrane row. Reason: Consistent with the IEA membrane row. The term is correct, and the specific supported membranes are the plasma membrane and the mitochondrial inner membrane. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: UniProtKB:Q8N490 · PNKD SUPPORTS TRANSFER Human membrane IDA (PMID:22214848) is sound; plasma membrane or mitochondrial inner membrane are the informative terms. Proposed replacements: plasma membrane mitochondrial inner membrane Supporting Evidence: PMID:22214848 One PNKD isoform is a membrane-associated |
| GO:0032225 regulation of synaptic transmission, dopaminergic | ISO GO_REF:0000121 | MARK AS OVER ANNOTATED | Summary: Transferred from a mouse IMP. The source is a transgenic model that carries the dominant human disease mutations and shows reduced striatal extracellular dopamine. That is a phenotype of a pathogenic allele, not evidence that PNKD performs part of dopaminergic transmission. Reason: The mouse curator used a weak qualifier, acts_upstream_of_or_within. The underlying model expresses A7V/A9V mutant Pnkd, and the dopamine changes are a circuit-level outcome downstream of the presynaptic role. The mechanistic process supported for PNKD is negative regulation of neurotransmitter secretion through RIM, which applies to synapses generally and is not specific to dopamine. Transferring the allele-phenotype term to rat adds no information about what the protein does. I am not overruling the mouse annotation; I am marking the rat transfer as over-annotation. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: CONTEXT OR TISSUE MISMATCH ROLE CONFLATION Sources checked: MGI:MGI:1930773 · Pnkd SUPPORTS SOURCE BUT NOT TARGET Mouse IMP (PMID:22214848) is a phenotype of a dominant disease-allele transgene (reduced striatal dopamine), not participation in dopaminergic transmission. Supporting Evidence: PMID:22214848 Transgenic mice carrying mutations in the mouse Pnkd locus equivalent to those found in patients with PNKD recapitulated the human PNKD phenotype. |
| GO:0033617 mitochondrial respiratory chain complex IV assembly | ISO GO_REF:0000121 | MARK AS OVER ANNOTATED | Summary: Transferred from the human short isoform MR-1S, which was identified as a COX assembly factor that works with PET100 and PET117. D3ZXB0 is a long product. It shares MR-1S residues 1-118 but lacks the MR-1S-specific C-terminus and carries the whole hydrolase domain. Whether this product takes part in COX assembly has not been tested. Reason: Vidoni et al. (2017) name the short isoform specifically, and the source row is isoform-tagged (Q8N490-2). At gene level, rat Pnkd probably also produces an MR-1S equivalent, so the function may hold for the gene. For the long D3ZXB0 product it is an untested extrapolation. It is not a core function of the hydrolase-domain product, and I have not read the full text. Propagation Review Root cause: PROPAGATION BAD Failure modes: CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:Q8N490-2 · PNKD isoform 2 (MR-1S) SUPPORTS SOURCE BUT NOT TARGET COX assembly role (PMID:28199844) is specific to the short isoform MR-1S; D3ZXB0 is a long, hydrolase-domain product lacking the MR-1S C-terminus. Supporting Evidence: PMID:28199844 we identified the short isoform of the myofibrillogenesis regulator 1 (MR-1S) as a new COX assembly factor file:rat/Pnkd__D3ZXB0/Pnkd__D3ZXB0-bioinformatics/RESULTS-d3zxb0.md Exact substring of D3ZXB0: True; position in D3ZXB0: 80 |
| GO:0042053 regulation of dopamine metabolic process | ISO GO_REF:0000121 | MARK AS OVER ANNOTATED | Summary: Transferred from a mouse IMP on the dominant-mutant transgenic model. Altered dopamine levels are a downstream phenotype, not a PNKD activity in dopamine metabolism. Reason: PNKD has no known role in dopamine synthesis or catabolism. The mutant-allele model shows reduced extracellular striatal dopamine, which follows from altered release. The transfer to rat over-annotates an indirect phenotype. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: CONTEXT OR TISSUE MISMATCH ROLE CONFLATION Sources checked: MGI:MGI:1930773 · Pnkd SUPPORTS SOURCE BUT NOT TARGET Mouse IMP (PMID:22214848) reports altered dopamine levels in mutant-allele mice; PNKD has no known dopamine-metabolic activity. Supporting Evidence: PMID:22214848 reduced extracellular dopamine levels in the striatum |
| GO:0046929 negative regulation of neurotransmitter secretion | IEA GO_REF:0000120 | ACCEPT | Summary: Core neuronal role of PNKD-L, established in mouse and with human constructs. PNKD-L binds the active-zone proteins RIM1 and RIM2. Overexpressed PNKD halves synaptic-vesicle exocytosis, and disease mutants are less effective. Pnkd knockout mice have lower RIM levels and impaired synaptic strength. Reason: Experimental support comes from the human IMP and mouse IGI annotations from Shen et al. (2015). The rat gene is the 1:1 ortholog in PANTHER SF116, and D3ZXB0 contains the PNKD-L sequence, including the C-terminal epitope used to immunoprecipitate the mouse protein. In that study the exocytosis assay was run in cultured rat hippocampal neurons, but with exogenous PNKD constructs, so endogenous rat Pnkd function has not been measured. The term describes regulation of release, which is the level the evidence supports. Supporting Evidence: PMID:25730884 Overexpressed PNKD protein suppresses release, and mutant PNKD protein is less effective than wild-type at inhibiting exocytosis. PMID:25730884 In PNKD KO mice, RIM1/2 protein levels are reduced and synaptic strength is impaired. |
| GO:0046929 negative regulation of neurotransmitter secretion | ISO GO_REF:0000121 | ACCEPT | Summary: RGD transfer of the mouse IGI annotation (Pnkd and RIM genetic interaction). This is the same claim as the other negative regulation of neurotransmitter secretion rows. Reason: Consistent with the IEA and human-derived ISO rows. Mouse Pnkd is the SF116 ortholog. Supporting Evidence: PMID:25730884 We propose a model wherein PNKD stabilizes RIM and inhibits the RIM-dependent pathway to suppress presynaptic neurotransmitter release. |
| GO:0046929 negative regulation of neurotransmitter secretion | ISO GO_REF:0000121 | ACCEPT | Summary: RGD transfer of the human IMP annotation. Wild-type PNKD suppresses synaptic-vesicle exocytosis, and disease mutants are less effective. Reason: Consistent with the other rows for this term. Supporting Evidence: PMID:25730884 Overexpressed wild-type PNKD significantly reduces neurotransmitter release to 50% that of controls transfected with empty vector |
| GO:0050884 neuromuscular process controlling posture | ISO GO_REF:0000121 | MARK AS OVER ANNOTATED | Summary: Transferred from the dystonia-like phenotype of mice that carry the dominant PNKD disease mutations. This is an organismal motor phenotype of a pathogenic allele, not a process the protein performs. Reason: The paroxysmal dystonic attacks of mutant-transgenic mice follow from disordered basal-ganglia transmission and do not show that PNKD takes part in a postural neuromuscular process. Campaign rules exclude downstream organismal phenotypes as process participation, so the rat transfer is marked as over-annotation and the mouse IMP is left to its curators. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: CONTEXT OR TISSUE MISMATCH Sources checked: MGI:MGI:1930773 · Pnkd SUPPORTS SOURCE BUT NOT TARGET Mouse IMP (PMID:22214848) is the dystonic organismal phenotype of the dominant mutant transgene. Supporting Evidence: PMID:22214848 Transgenic mice carrying mutations in the mouse Pnkd locus equivalent to those found in patients with PNKD recapitulated the human PNKD phenotype. |
| GO:0051596 methylglyoxal catabolic process | IEA GO_REF:0000002 | REMOVE | Summary: InterPro2GO transfer of the glyoxalase-II pathway role. PNKD-L does not perform the glyoxalase-II step in vivo, and PAINT propagated no glyoxalase-pathway process term to the PNKD branch. Reason: The glyoxalase-II step of methylglyoxal detoxification is hydrolysis of S-D-lactoylglutathione. HAGH performs it. A human PNKD-L transgene does not restore SLG metabolism in HAGH-null flies, and the authors conclude that PNKD-L does not use SLG in vivo. In PTHR11935 the only related process IBD (GO:0006749 glutathione metabolic process) is on the HAGH node PTN002631193, not on the PNKD subfamily. That a PNKD hydrolase detoxifies some other alpha-hydroxythioester remains a hypothesis and does not support this term. Supporting Evidence: PMID:21487022 introduction of a human PNKD-L transgene into the same null background shows no rescue of SLG metabolism PMID:21487022 This indicates that hydrolysis of SLG in vivo by PNKD-L is unlikely to be physiologically relevant. |
| GO:0099523 presynaptic cytosol | IEA GO_REF:0000107 | ACCEPT | Summary: Ensembl Compara transfer from the mouse IDA (immuno-electron microscopy), which places Pnkd in presynaptic terminals with no signal in Pnkd-null tissue. This is the compartment where the release-regulating role is performed. Reason: Mouse Q69ZP3 is the SF116 ortholog. D3ZXB0 is about 99% identical to mouse isoform 3 and contains the PNKD-L C-terminus. The presynaptic localization fits the RIM interaction. It has not been tested in rat. Supporting Evidence: PMID:25730884 Here, we show that PNKD is a novel synaptic protein that interacts with RIMs and can be found at presynaptic terminals. PMID:25730884 Because virtually no labeling of presynaptic profiles was found in Pnkd −/− samples |
| GO:0099523 presynaptic cytosol | ISO GO_REF:0000121 | ACCEPT | Summary: RGD transfer of the same mouse IDA. This is the same claim as the IEA row. Reason: Consistent with the IEA presynaptic cytosol row. Supporting Evidence: PMID:25730884 PNKD localizes to the presynaptic terminal |
Loading supporting content…
Download this section (compressed HTML)Q: Is the long PNKD product mitochondrial (inner membrane, after MTS cleavage) or plasma-membrane/synaptic, and does the 39-residue hydrophobic exon present in D3ZXB0 (but absent from the 385-residue PNKD-L) change its targeting?
Q: What is the physiological alpha-hydroxythioester (or other thioester) substrate of PNKD, and is the hydrolase activity needed for its presynaptic role?
Q: Is the rat D3ZXB0 (mouse isoform 3-type) transcript the major brain product in rat, compared with the 385-residue PNKD-L-type and the MR-1S/MR-1M products?
Experiment: Purify rat D3ZXB0 and a 385-residue PNKD-L construct, measure zinc content and thioesterase activity on a panel of alpha-hydroxythioesters, and run untargeted metabolomics on Pnkd-null rodent brain to identify the substrate that accumulates.
Experiment: Do isoform-resolved localization of endogenous rat Pnkd in brain (subcellular fractionation with mitochondrial inner-membrane and synaptosomal markers, plus immuno-EM) using isoform-specific antibodies or epitope knock-ins.
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)