APH1A curation notes

2026-06-19

Deep-research attempt with just deep-research-falcon human APH1A --fallback perplexity-lite timed out after 180 seconds with no generated research artifact, so this manual review uses cached UniProt, GOA, and publication evidence.
APH1A encodes a multipass non-catalytic gamma-secretase subunit. UniProt summarizes APH1A as a component of the endoprotease complex that cleaves integral membrane substrates including Notch receptors and APP, and says APH1A is required for normal complex assembly.
Direct APH1A evidence supports complex membership and substrate processing: the Lee et al. abstract states that mammalian APH-1 "physically associates with nicastrin and the heterodimers of the presenilin" and that APH-1 knockdown reduces amyloid-beta peptides plus APP and Notch intracellular domains [PMID:12297508 "physically associates with nicastrin and the heterodimers of the presenilin"; PMID:12297508 "reduction of gamma-secretase products (amyloid-beta peptides and the intracellular domains of APP and Notch)"].
APH1A should be curated as a non-catalytic adaptor/assembly/activator subunit, not as an independent peptidase. The complex contains presenilin, Pen-2, Aph-1, and nicastrin PMID:25043039, while presenilin is the catalytic aspartyl protease subunit in the complex.
APP and Notch processing are core outputs. The APH1A-containing complex cleaves APP-C100 in vitro PMID:25043039 and older RNAi/overexpression work describes APH1/PEN2/nicastrin as essential or limiting for gamma-secretase activity PMID:12763021.
The steady-state localization emphasis for APH1A is ER and cis-Golgi, with membrane and endomembrane annotations biologically consistent. Plasma-membrane Reactome annotations probably reflect substrate event locations rather than the strongest APH1A-specific localization evidence, so I treated them as retained but non-core.
Generic protein binding annotations capture real interactions but are not informative GO molecular-function statements for APH1A. I marked those as over-annotated rather than removing them, because several are experimental interaction datasets and complex membership is real.

2026-06-20 second-pass audit

The second-pass audit added manual reference_review metadata for the main APH1A/gamma-secretase papers documenting APH-1 association with presenilin/nicastrin, APP and Notch proteolysis, purified human gamma-secretase composition, and APH1 isoform effects on substrate cleavage. No annotation action changes were needed: APH1A remains curated as a non-catalytic scaffold/adaptor subunit required for gamma-secretase assembly and complex-level intramembrane protease activity, not as an independent protease.

Falcon deep research integration (2026-06-21)

A Falcon (Edison) deep-research report is now present in APH1A-deep-research-falcon.md; it corroborates the existing review well, reaffirming APH1A as a non-catalytic, multipass-membrane scaffold/stabilizer subunit of the four-component gamma-secretase complex (presenilin/PSEN, nicastrin/NCSTN, PEN-2/PSENEN) acting in APP processing/amyloid-beta generation and Notch processing, with no annotation actions overturned. The Falcon-sourced citations below are not yet independently verified against full text.

New or refined points beyond the existing notes/review:
- Membrane topology refined to seven transmembrane domains with N-terminus extracellular/luminal and C-terminus cytosolic, consistent with a structural core role [Malvankar & Wolfe, RSC Chem. Biol. 2026, doi:10.1039/d6cb00022c; Odorčić et al., Nat. Commun. 2024, doi:10.1038/s41467-024-48776-2].
- Cryo-EM architecture: APH-1 forms a major (~2,000 Å²) membrane-embedded interface with presenilin (APH-1 TM2-TM4 contacting PSEN1 TM1 and TM8-TM9; PSEN1 C-terminus inserts into the APH-1 helical bundle), giving a structural basis for the "presenilin-stabilization factor" alias [Odorčić et al., Nat. Commun. 2024, doi:10.1038/s41467-024-48776-2].
- Ordered stepwise assembly: NCSTN+APH1A form an early ER subcomplex, which then recruits full-length presenilin, after which PEN-2/PSENEN binds (via PSEN TMD4) to trigger presenilin autoproteolysis (NTF/CTF); complex matures in the Golgi with nicastrin glycosylation [Malvankar & Wolfe 2026, doi:10.1039/d6cb00022c; Ning et al., Front. Aging Neurosci. 2025, doi:10.3389/fnagi.2025.1637671].
- APH1A vs APH1B isoform distinction: APH1A(+PSEN1) complexes are associated with greater substrate processivity and relatively shorter Abeta products, whereas PSEN2/APH1B complexes favor longer aggregation-prone Abeta (Abeta42/43); APH-1 isoform divergence clusters in ~3 non-conserved regions including the TM3-TM4 loop implicated in substrate-induced allosteric coupling [Odorčić et al. 2024, doi:10.1038/s41467-024-48776-2; de Strooper & Karran, EMBO J. 2024, doi:10.1038/s44318-024-00057-w]. (NB: the Odorčić structure is of the APH-1B complex; APH1A is the comparison reference, so APH1A-specific claims are inferential.)
- Genetic essentiality: Aph1a knockout causes lethal Notch phenotypes in mouse embryogenesis vs. only mild deficits for Aph1b, marking APH1A as the predominant paralog for developmental Notch signaling [de Strooper & Karran 2024, doi:10.1038/s44318-024-00057-w].
- Broader substrate proteome: gamma-secretase processes >145-150 type I membrane substrates; the 2023 G-SECSI screen found 85 microglial substrates (59 novel), positioning APH1A within a wide regulated-intramembrane-proteolysis network [Hou et al., Mol. Cell 2023, doi:10.1016/j.molcel.2023.10.029].
- Pharmacology context: nirogacestat (first gamma-secretase inhibitor, FDA-approved 2023 for desmoid tumors) and complex-selective modulator concepts; some compounds discriminate APH1A- vs APH1B-containing complexes [de Strooper & Karran 2024, doi:10.1038/s44318-024-00057-w].

Discrepancies / annotations to revisit:
- No contradictions with the existing review or its annotation actions; the report reinforces every retained core function.
- The new seven-TM topology and cryo-EM interface data strengthen (do not change) the existing GO:0030674 protein-macromolecule adaptor activity and GO:0061133 endopeptidase activator activity calls, and the membership call GO:0070765 gamma-secretase complex.
- The report's substrate-proteome breadth (p75NTR, ERBB4, EphB2, etc.) is consistent with the existing Reactome-derived plasma-membrane annotations kept as non-core; nothing here argues for promoting plasma membrane (GO:0005886) to core, since APH1A steady-state localization remains ER/cis-Golgi-dominant.
- Optional consideration only (not required): the report's emphasis on APH1A's role in stepwise gamma-secretase assembly/maturation could motivate adding a protein-complex-assembly process term in proposed_new_terms (exact GO ID to be looked up via OLS, not guessed here); current core_functions already capture the adaptor/assembly intent, so this is left as a possible future refinement pending full-text verification.