SNURF (Q9Y675) review notes

2026-10-08 — MICROPROTEINS Tier 2 review (claude-code)

HGNC SNURF (HGNC:11171), UniProt Q9Y675 (Swiss-Prot, PE1), 71 aa, highly basic
(~17% Arg), conserved nuclear localization motif (KRRR). Encoded by the upstream ORF of
the bicistronic SNURF-SNRPN transcript in the Prader-Willi imprinting centre (15q11-q13);
the downstream ORF encodes SmN (SNRPN). Separate UniProt entry and HGNC symbol, so it has
its own folder (CLAUDE.md alt-ORF convention).
Symbol collision: "SNURF" was also the original name of RNF4 (small nuclear RING
finger protein, 194 aa) PMID:9710597. Most PubMed hits for "SNURF" are RNF4 papers
(PMID:10617653, 10934038, 11319220, 15707587 ...). None of them concern Q9Y675.

Only primary paper: Gray et al. 1999 PMID:10318933; GFP fusion PMID:10318933; protein
detected in normal but not PWS cells. Function unknown PMID:10318933. Wobble-position-biased substitution pattern argues
for coding selection.
HPA immunofluorescence: nuclear speckles, reliability "Supported" (proteinatlas.org JSON for
ENSG00000273173, queried 2026-10-08).
PubMed search (2026-10-08) for "SNRPN upstream reading frame" found no later study of the
protein's function; later papers address the locus/imprinting centre or RNF4.

Mouse source is MGI IPI to Snurf, WITH Rad54l2/ARIP4 (Q99NG0), PMID:12058073 (Rouleau 2002,
ARIP4 paper; cache is abstract-only). QuickGO shows MGI also gave ARIP4 "protein binding"
IPI WITH Q9WU12 from the same paper.
Full text read through PMC (WebFetch, PMC117628): the paper defines "small nuclear RING
finger protein (SNURF)" and uses a "SNURF residues 20–177" VP16 fusion as an AR-binding
control in two-hybrid. Residues 20-177 cannot exist in a 71-aa protein, so the SNURF in the
paper is RNF4. The MGI rows are a symbol-collision mapping error (should be Rnf4, O88846,
if an ARIP4 interaction was shown at all). Removed here; upstream fix is at MGI.