PIGH (Q14442) review notes

Human PIGH — Phosphatidylinositol N-acetylglucosaminyltransferase subunit H (PIG-H).
188 aa, ER membrane protein (UniProt: multi-pass; two predicted TM helices 40-60 and
63-83). HGNC:8964. Chromosome 14. Gene MIM 600154; disease MIM 618010 (GPIBD17).

Deep research: falcon provider was OUT OF CREDITS (HTTP 402) at review time, so no
-deep-research-falcon.md was generated. This review is grounded in the UniProt record
(PIGH-uniprot.txt), the seeded GOA (PIGH-goa.tsv), the cached primary publications,
and the Reactome entry.

Core biology (verified)

PIGH is a non-catalytic (accessory) subunit of the GPI-GlcNAc transferase (GPI-GnT)
complex, the multi-subunit ER enzyme that catalyzes the first committed step of
GPI-anchor biosynthesis: transfer of GlcNAc from UDP-GlcNAc onto phosphatidylinositol
(PI) to yield GlcNAc-PI, on the cytoplasmic face of the ER.

• The catalytic subunit is PIGA (homologous to a bacterial GlcNAc transferase); PIGH
  is one of the required accessory subunits.
• The complex comprises at least PIGA, PIGC, PIGH, PIGP, PIGQ (GPI1), PIGY and DPM2.
  PMID:16162815
  [UniProt SUBUNIT: "composed at least by PIGA, PIGC, PIGH, PIGP, PIGQ, PIGY and DPM2"]

Provenance of key claims

• First step / GlcNAc transfer / complex:
  PMID:8900170
  PMID:8900170
  PMID:9463366
• PIG-H is ER-associated, cytoplasmically oriented; forms complex with PIG-A:
  PMID:8900170
  (UniProt models the mature protein as a multi-pass ER membrane protein; GOA carries
  ER membrane EXP/IDA. I defer to GOA/UniProt for the ER-membrane CC.)
• Interactions: PIGH interacts with PIGA (P37287) and PIGQ (Q9BRB3) within the complex.
  [UniProt INTERACTION: "Q14442; P37287: PIGA" and "Q14442; Q9BRB3: PIGQ"]
• PIG-P / DPM2 components: PMID:10944123

Molecular function

The GOA carries no catalytic MF term for PIGH — only GO:0005515 protein binding
(IPI, from complex/interactome studies). This is consistent with PIGH being a
non-catalytic subunit; the GlcNAc-transferase catalytic activity belongs to the complex
(and specifically to PIGA). I therefore do NOT assign a catalytic MF in core_functions;
I record PIGH's contribution via directly_involved_in the BP and in_complex the
GPI-GnT complex, plus ER-membrane location.

Disease

Biallelic PIGH variants cause GPI biosynthesis defect 17 (GPIBD17) [MIM:618010], an
autosomal-recessive inherited GPI-deficiency disorder (developmental/epileptic
encephalopathy): developmental delay, epilepsy/seizures, microcephaly, autistic features,
skeletal manifestations. Variants: start-codon disruption (PMID:29573052),
p.Ser103Pro (PMID:29603516, PMID:33156547 — reduced GPI-anchored protein surface
expression), p.Arg163Trp (PMID:33156547, PMID:35445667). Disease refs are not in the
GOA and are used only for the top-level description/context (no supporting_text quotes
needed since they are not cached).

GOA annotation inventory (21 lines -> 20 review entries)

MF: GO:0005515 protein binding (IPI) x6 references (7 GOA lines; two PMID:33961781 lines
with different partners P37287/Q9BRB3 collapse to one review entry).
CC: GO:0000506 GPI-GnT complex (IBA, IEA, IPI, IDA x2) ; GO:0005789 ER membrane
(IEA, EXP, IDA, TAS) ; GO:0005783 ER (TAS).
BP: GO:0006506 GPI anchor biosynthetic process (IBA, IEA, IDA, TAS).

Curation decisions summary

• CC GPI-GnT complex (all evidence): ACCEPT — core, well-supported.
• CC ER membrane / ER (all evidence): ACCEPT — core localization.
• BP GPI anchor biosynthetic process (all evidence): ACCEPT — core.
• MF protein binding (all IPI): MARK_AS_OVER_ANNOTATED — uninformative bare
  protein binding; the real informative content (subunit of GPI-GnT) is captured by the
  complex CC term. Per policy, do NOT REMOVE bare protein-binding IPIs.