sdf-9 (eak-5, Y44A6D.4; UniProt G5EGA9) curation notes

Provenance / process

Deep research via falcon was not attempted: the falcon provider is known to fail for this
batch (HTTP 402). No -deep-research-*.md file exists; notes below are from the cached
primary literature plus a local sequence analysis.
PubMed search ("sdf-9 elegans OR eak-5 elegans") returned 5 PMIDs: 12783794, 16839187,
17660545, 15509773, 21209831. Cached 17660545 (Jensen et al. 2007, SDF-9/DAF-2) and
15509773 (NCR-1/NCR-2; abstract does not mention sdf-9, not cited). 21209831 is a broad
DAF-16 target RNAi screen, not used.
Full text available: PMID:16839187 (Hu et al. 2006, PLoS Genet) only. PMID:12783794
(Ohkura et al. 2003) and PMID:17660545 (Jensen et al. 2007) are abstract-only in cache.

345 aa; single PTP catalytic domain (PROSITE PRU00160, residues 33-284); PANTHER
PTHR19134:SF561. No transmembrane segment, no N-myristoylation motif
PMID:16839187.
Catalytic cysteine absent: PMID:16839187.
Own residue check (file:worm/sdf-9/sdf-9-bioinformatics/RESULTS.md): PTP1B HCSAGIGRSG
aligns gap-free to SDF-9 QSARGSSRAG (222-231). Catalytic Cys -> Ser223 (UniProt
CAUTION says "lysine at position 223"; the sequence has serine - the CAUTION text is wrong
about the substituting residue, which matches the PTHR19134 family review's note). Also
His214 -> Gln, P-loop Ser222 -> Ala, pTyr-loop Tyr46 -> Val. Arg221 (P-loop) retained.
Hu et al. count: PMID:16839187 of 19
invariant vertebrate PTP residues.
No direct phosphatase assay on SDF-9 is reported; PNPP assays were done on EAK-6 (inactive)
PMID:16839187.

Hypotheses only: substrate-trap-like pTyr binding / adaptor.
PMID:16839187
But experimental tests were negative (unpublished data within the paper):
PMID:16839187
and no phosphoinositide binding
PMID:16839187,
and no direct EAK-4/SDF-9/EAK-6 interaction in 293T coIP.
Jensen et al. 2007 propose a DAF-2 interaction / adaptor role, from genetics (abstract only)
PMID:17660545.
Conclusion: MF is unknown. No GO MF term is supportable; do not propose phosphotyrosine
binding (the only test was negative).

Expressed only in the two XXXL/R head cells PMID:12783794.
Plasma membrane (GFP fusion) PMID:16839187;
independent of DAF-2 signalling PMID:16839187.
Ohkura 2003 (UniProt) gives "Cytoplasm" plus dendrite-like structure; abstract does not state
this, full text not cached. The two are compatible (peripheral membrane protein at the cell
cortex of XXX cells).

Dauer: sdf-9 mutants form dauer-like larvae resembling daf-9/daf-12 Daf-c; enhanced by
cholesterol deprivation PMID:12783794.
Acts with DAF-9 (dafachronic acid synthesis) PMID:12783794.
Insulin pathway (eak screen): phenotypes suppressed by daf-16; act in parallel to akt-1;
eak-4, sdf-9, eak-6 same pathway/complex PMID:16839187.
Model: PMID:16839187.
Partial dauers, alae but no pharyngeal remodelling.

NOT PTP activity (IKR): ACCEPT - well supported by residue loss.
No positive PTP IBA/IEA rows exist in GOA for sdf-9 (PAINT already withholds GO:0004725 from
SF561), so no REMOVE needed for PTP activity.
IBA signal transduction: KEEP_AS_NON_CORE (true but uninformative).
CC rows: accept PM IDA; cytoplasm EXP accept with caveat; IEA PM accept; IEA cytoplasm keep non-core.
dauer larval development IMP/IGI/IEA: ACCEPT (standard WormBase usage; gene acts as negative
regulator of dauer entry).
No NEW annotations: insulin receptor signalling participation is not shown (genetic placement
only, "upstream of or in parallel to"). Raised as question.

Confirms SDF-9 as a valid exception to family-wide GO:0004725 (Ser at catalytic Cys
position 223, plus loss of His and P-loop Ser/Thr). GOA already carries NOT|enables
GO:0004725 (IKR) for sdf-9, consistent with the family review. Note EAK-6 (same SF561)
retains Cys but showed no PNPP activity and has other invariant-residue substitutions.