PDIA4 (ERp72) is an endoplasmic reticulum luminal oxidoreductase of the protein disulfide isomerase (PDI) family (PANTHER subfamily PTHR18929:SF210, protein disulfide-isomerase A4). Mammalian PDIA4 has an N-terminal signal peptide, an acidic N-terminal stretch, three catalytic thioredoxin-like a-type domains carrying CGHC redox-active sites, two non-catalytic b-type domains, and a C-terminal KDEL-type ER retrieval signal. It catalyses thiol-disulfide oxidation, reduction and isomerization of client proteins during oxidative folding in the ER, and works within a large ER chaperone network with HSPA5 (BiP), HSP90B1 (GRP94), PPIB (cyclophilin B), PDIA6 and other folding factors that binds unfolded secretory proteins. It is an ER stress-inducible protein, and small amounts are released and found at the cell surface, for example on activated platelets. The donkey UniProt entry A0A8C4LTV9 (1197 aa) is a merged gene model: residues ~581-1197 are the PDIA4 product (90% identical to human PDIA4 over residues 31-645, all three CGHC sites conserved, C-terminal REEL), but they are preceded by a KRAB domain and ten C2H2 zinc fingers that are most similar to human ZNF786, the gene immediately upstream of PDIA4 on the same strand in human, and the PDIA4 signal peptide is missing. The zinc-finger part is not PDIA4 biology.
Reason: The PDIA4 portion of the donkey entry sits in PANTHER subfamily PTHR18929:SF210 (PDIA4) and retains all three CGHC redox-active sites of human ERp72. PAINT curators placed protein disulfide isomerase activity at the PDI family root node PTN000432607, from which PDIA4 orthologs inherit it, and human ERp72 has measured peptide thiol-disulfide oxidase activity comparable to PDI. This is the core catalytic function of PDIA4.
we have recombinantly expressed and purified six members of the family, PDI, PDIp, ERp57, ERp72, P5, and PDIr from a single organism, human. An examination of the pH-dependence and nature of the rate limiting step for the peptide thiol-disulfide oxidase activity of these enzymes reveals that, with the exception of PDIr, they are all remarkably similar.
file:EQUAS/PDIA4/PDIA4-bioinformatics/RESULTS.md
DISULFID Redox-active: human [91, 94] CC -> donkey [643, 646] CC
file:EQUAS/PDIA4/PDIA4-uniprot.txt
PANTHER; PTHR18929:SF210; PROTEIN DISULFIDE-ISOMERASE A4; 1.
file:EQUAS/PDIA4/PDIA4-bioinformatics/RESULTS.md
DISULFID Redox-active: human [555, 558] CC -> donkey [1107, 1110] CC
file:EQUAS/PDIA4/PDIA4-uniprot.txt
SIMILARITY: Belongs to the protein disulfide isomerase family.
Reason: PDIA4 is an ER luminal enzyme with no nuclear function. The nuclear location in this entry comes from the KRAB-C2H2 zinc-finger segment at its N-terminus, which matches human ZNF786 (64% identity) rather than PDIA4; in human, ZNF786 lies immediately upstream of PDIA4 on the same strand, and the donkey annotation has no separate ZNF786 gene, so the segment reflects a merged ZNF786-PDIA4 gene model. The term does not describe PDIA4 biology.
Supporting Evidence:
file:EQUAS/PDIA4/PDIA4-uniprot.txt
InterPro; IPR001909; KRAB.
file:EQUAS/PDIA4/PDIA4-bioinformatics/RESULTS.md
Donkey A0A8C4LTV9 (1197 aa) aligned globally to human P13667 (645 aa): 88.37% identity over 645 paired residues; paired coverage 100.0% of human and 53.88% of donkey.
Reason: Correct (it is the PAINT IBD placed at the PDI family root node PTN000432607), but the accepted, more specific location endoplasmic reticulum lumen entails it.
Reason: Human PDIA4 is an ER luminal resident retained by its C-terminal KDEL-type signal; the donkey PDIA4 portion ends with the equivalent REEL motif. The donkey gene model lacks the signal peptide needed for ER entry, which is attributed to the gene-model fusion, not to the biology of donkey PDIA4.
GO:0006355 regulation of DNA-templated transcription
IEA GO_REF:0000002
REMOVE
Summary: Regulation of DNA-templated transcription (InterPro KRAB domain mapping).
Reason: This annotation comes from the InterPro KRAB domain of the N-terminal zinc-finger segment, which matches human ZNF786 and is a gene-model merger, not part of PDIA4. PDIA4 is an ER luminal oxidoreductase with no role in transcription regulation.
Reason: PDIA4 catalyses disulfide formation and rearrangement during oxidative folding of secretory proteins in the ER and acts within the BiP/GRP94 chaperone network that binds unfolded substrates; its complex with cyclophilin B accelerates immunoglobulin folding. PAINT placed protein folding at the PDI family root node (PTN000432607), which the TreeGrafter row inherits.
Protein folding in the endoplasmic reticulum is often associated with the formation of native disulfide bonds, a process which in vivo is one of the rate limiting steps of protein folding and which is facilitated by the enzyme protein disulfide isomerase (PDI).
Reason: Human ERp72 is released by platelets and relocates to the platelet surface on activation. PAINT placed cell surface at node PTN001448731, whose only experimental donors are human and mouse PDIA4, so the TreeGrafter transfer to the donkey PDIA4 subfamily member is phylogenetically sound. It is a minor, non-core location of an ER-resident protein.
ERp72, ERp57, ERp44 and ERp29 are released by platelets and relocate to the cell surface following platelet activation.
GO:0034976 response to endoplasmic reticulum stress
IEA GO_REF:0000118
KEEP AS NON CORE
Summary: Response to endoplasmic reticulum stress (TreeGrafter).
Reason: ERp72 is one of the classical ER stress proteins, and PAINT placed this term at the PDI family root node PTN000432607 (the separate NOT rows in the family tree sit on other clades, since human PDIA4 retains the IBA). PDIA4 contributes to the response as a folding catalyst for accumulating client proteins rather than as a signalling component, so it is kept as non-core.
Reason: Based on mass-spectrometric detection of human PDIA4 in melanosome proteomes (about 1500 proteins). ER luminal proteins are common in such organelle preparations and there is no evidence of a melanosome-specific function.
Reason: The thioredoxin-fold CGHC sites of PDIA4 act on substrate-protein disulfides during oxidative folding in the ER. This is not regulation of cellular redox state. The generic InterPro2GO mapping over-states a role in redox homeostasis.
Protein folding in the endoplasmic reticulum is often associated with the formation of native disulfide bonds, a process which in vivo is one of the rate limiting steps of protein folding and which is facilitated by the enzyme protein disulfide isomerase (PDI).
GO:0140096 catalytic activity, acting on a protein
IEA GO_REF:0000117
TOO GENERAL
Summary: Catalytic activity, acting on a protein (InterPro mapping).
Reason: True but uninformative; entailed by the accepted protein disulfide isomerase activity.
Supporting Evidence:
file:EQUAS/PDIA4/PDIA4-bioinformatics/RESULTS.md
DISULFID Redox-active: human [91, 94] CC -> donkey [643, 646] CC
Core Functions
ER luminal protein disulfide isomerase that catalyses formation, reduction and rearrangement of disulfide bonds in client proteins during oxidative folding of secretory proteins, acting within the BiP/GRP94 ER chaperone network.
we have recombinantly expressed and purified six members of the family, PDI, PDIp, ERp57, ERp72, P5, and PDIr from a single organism, human. An examination of the pH-dependence and nature of the rate limiting step for the peptide thiol-disulfide oxidase activity of these enzymes reveals that, with the exception of PDIr, they are all remarkably similar.
Protein folding in the endoplasmic reticulum is often associated with the formation of native disulfide bonds, a process which in vivo is one of the rate limiting steps of protein folding and which is facilitated by the enzyme protein disulfide isomerase (PDI).
Q: Does the donkey PDIA4 locus encode a conventional ER-targeted ERp72 with a signal peptide (as expected), so that the KRAB-zinc-finger N-terminus of the Ensembl models A0A8C4LTV9/A0A8C4PKQ4 is an annotation fusion with a neighbouring zinc-finger gene?
Q: Should the Ensembl donkey gene model ENSEASG00005009014 be split into separate ZNF786 and PDIA4 genes, and are the equivalent horse models (e.g. the 1248-aa PDIA4 entry) affected by the same merger?
Suggested Experiments
Experiment: RT-PCR or long-read cDNA sequencing across the 5' end of donkey PDIA4 transcripts to establish whether any transcript joins zinc-finger exons to PDIA4 and to recover the signal-peptide exon.
External Prediction Reviews
These computational predictions are reviewed separately from the GOA annotation set used for this review. The assessments below are from this project and do not constitute official GO annotations or endorsement by GO/UniProt. They are not included in the existing annotation review above.
Assessment of the single ProtNLM2 GO prediction for donkey PDIA4 (A0A8C4LTV9). The UniProt entry is a 1197-aa merged gene model in which a KRAB domain and ten C2H2 zinc fingers most similar to human ZNF786 (the gene immediately upstream of PDIA4 in human) precede a complete PDIA4 (ERp72) protein disulfide isomerase lacking its signal peptide. The predicted double-stranded DNA binding derives from the zinc-finger segment and does not describe PDIA4, an ER luminal oxidoreductase (NPI, WRONG_INPUT_SEQUENCE); the gene review removes the other zinc-finger-derived annotations (nucleus, regulation of DNA-templated transcription) on the same grounds, and a PubMed search for PDIA4/ERp72 and DNA binding returned no report of DNA binding by PDIA4. ProtNLM2 also emitted function text (not scored) - "Participates in the folding of proteins containing disulfide bonds, may be involved in glycosylation, prolyl hydroxylation and triglyceride transfer" - whose first clause fits PDIA4; the remaining clauses resemble descriptions of canonical PDI (P4HB) and are not established for PDIA4 (the evidence key for this text is a phmmer match to alfalfa PDI, P29828). No location was emitted.
Prediction method: ProtNLM2 · Version: UniProt API snapshot 2026-10-14
WRONG INPUT SEQUENCE
Review rationale: The input sequence fuses a KRAB-C2H2 zinc-finger segment (residues 1 to about 580, 64% identical to human ZNF786) to PDIA4; the ProtNLM2 evidence for this GO term is a weak phmmer match to mouse zinc finger protein 148 (Q61624), i.e. the zinc-finger part. Human PDIA4 is a 645-aa ER luminal disulfide isomerase with no zinc-finger region, the PDIA4 part of the donkey entry aligns to human residues 17-645, and the signal peptide needed for ER entry is missing, which together indicate a gene-model fusion rather than a real PDIA4 protein. A second donkey entry from the same Ensembl gene shows the same KRAB-type N-terminus. DNA binding is not a PDIA4 function (ERp72 is an abundant ER luminal protein), so the prediction is incorrect for this gene; the cause is the chimeric input sequence, not paralog confusion.