PTPRN2 notes

2026-09-08 paired review

Selected accession A0A9L0T4W6; ordinary UniProt record is TrEMBL and cites genome sequencing/submission rather than a gene-specific functional experiment. Human evidence is summarized in the paired human investigation. Sequence comparison records the exact target gaps and transfer limits. The targeted Europe PMC search found no decisive gene-specific horse functional assay warranting a separate horse Edison investigation; general omics/association hits were not counted as mechanistic validation.

The 1996 cloning report states that recombinant IAR/PTPRN2 has phosphatase activity, but its abstract does not name the substrate. Later human cell work explicitly distinguishes lack of phosphoprotein-substrate activity from phosphoinositide phosphatase activity and tests the effect of a catalytic-domain mutant on migration and metastasis. Vesicle accumulation, insulin/neurotransmitter release, and membrane phosphoinositide regulation are not interchangeable with protein tyrosine dephosphorylation. Conversely, a biological-process annotation can describe regulation through another enzyme: losing intrinsic protein phosphatase activity alone does not refute all participation in protein dephosphorylation. The exact experimental annotation therefore remains unresolved while the inappropriate electronic intrinsic protein-phosphatase terms can be challenged.

Sources: PMID:8798755(https://pubmed.ncbi.nlm.nih.gov/8798755/), PMID:26620550(https://pubmed.ncbi.nlm.nih.gov/26620550/).

2026-10-20 — Full re-review, independent of the ProtNLM2 predictions

ProtNLM2 output was not opened or used. Deep research: just deep-research-falcon HORSE PTPRN2 --fallback perplexity-lite failed (Falcon "Failed to authenticate"; Perplexity HTTP 401 insufficient quota), so no horse deep-research file exists. The review uses primary literature cached in publications/.

Identity of the benchmark accession (alternative gene model)

compare_horse_models_rereview.py produced RESULTS-rereview.md and results-rereview.json. Findings:

• Horse PTPRN2 (Ensembl ENSECAG00000012762) has eight TrEMBL models, all in PANTHER PTHR46106:SF5. The canonical model in the UniProt gene-centric reference proteome is A0A9L0RAE3 (1054 aa, PTPRN2-205). The benchmark A0A9L0T4W6 (976 aa, PTPRN2-208, ENSECAT00000094767) is a non-canonical model.
• A0A9L0T4W6 keeps the signal peptide, the IA-2 luminal domain, the TM helix (568-592) and the catalytic loop HCSDGAGRSG (Cys897 = human Cys945). Its last coding exon (ENSECAE00000523854) is an alternative terminal exon. That exon replaces the 47 C-terminal residues conserved in human, mouse and the seven other horse models with 56 unrelated residues. The replaced stretch maps to human 969-1015, i.e. the final 37 residues of the PTP-like domain (human domain 745-1005) plus the cytoplasmic tail. The model also lacks 87 residues upstream of the catalytic motif compared with the canonical model.
• The earlier RESULTS.md description ("deletion of human residues 989-1003") is a pairwise-alignment artifact. The true change is the C-terminal replacement. RESULTS.md is left unchanged and marked DISPUTED in reference_review.

Evolutionary basis for transfer

• Rat phogrin (Ptprn2, Q63475) and human Q92932 are both in PTHR46106:SF5 with the horse protein, so subfamily-level transfer is justified. Rat phogrin biochemistry PMID:20097759 shows Cys-dependent PIPase activity and no activity on phosphoproteins or pNPP PMID:20097759. The effect on insulin secretion needs the catalytic Cys PMID:20097759.
• Mouse Ptprn2 knockouts: fewer DCVs and less insulin secretion PMID:21732083. IA-2beta also sits on synaptic vesicles PMID:19361477. Behavioural and lifespan phenotypes of the double knockout are downstream organismal effects and were not annotated.
• The horse accession carries no IBA rows. Human PTPRN2 gets regulation of secretion, secretory granule and synapse by IBA from PTN000439769 / PTN000439772, so the subfamily-level terms are PAINT-sanctioned.

Annotation decisions

• GO:0004725 and GO:0008138 (protein phosphatase activities, IEA): REMOVE. The subfamily lost phosphoprotein activity.
• GO:0016787 hydrolase activity: MODIFY to GO:0052866 phosphatidylinositol phosphate phosphatase activity (verified is_a hydrolase activity in OLS). This changes the previous UNDECIDED. The caveat that this exact model's C-terminus is altered is stated in the reason.
• GO:0030141 secretory granule: MODIFY to GO:0032127 dense core granule membrane.
• GO:0030658 transport vesicle membrane and GO:0045202 synapse: MODIFY to GO:0030672 synaptic vesicle membrane. OLS confirms GO:0030672 sits under both, and that secretory granule membrane is not under transport vesicle membrane.
• NEW GO:0051046 regulation of secretion (ISS, subfamily orthology; matches the PAINT IBA on human).

Relative to review next

A0A9L0T4W6 is a non-canonical alternative gene model. The canonical product of the same horse gene, A0A9L0RAE3, has no review. Suggested folder: genes/HORSE/PTPRN2__A0A9L0RAE3. The PANTHER family review (interpro/panther/PTHR46106) also places the horse protein in SF5, so there is no subfamily disagreement. But it uses the non-canonical A0A9L0T4W6 as the horse representative and repeats the "989-1003 deletion" description. No module in modules/ includes PTPRN2; modules/beta_cell_glucose_stimulated_insulin_secretion.yaml is the closest pathway context.

2026-10-21 — NEW rows withdrawn from the alternative gene model (PR #5303 review)

The re-review had proposed NEW rows (GO:0051046) on this benchmark accession because GOA propagates the
gene's annotations, including PAINT IBA rows, to one representative product per gene and not to
alternative transcript models. The same terms are already annotated in GOA on the canonical horse
product of this gene (PTPRN2__A0A9L0RAE3, reviewed in genes/HORSE/PTPRN2__A0A9L0RAE3/). Filling that propagation
convention on a second product of the same gene is not a curation gap, so the rows were withdrawn
for consistency across the cohort. The gene-level function is unchanged and is recorded on the
canonical product's review; this file's description and core functions still describe it.