SDF-9 catalytic-residue check

Script: catalytic_residues.py (run with uv run python genes/worm/sdf-9/sdf-9-bioinformatics/catalytic_residues.py
from the repo root). It reads the SDF-9 (G5EGA9) sequence from sdf-9-uniprot.txt, fetches human
PTP1B/PTPN1 (P18031) from UniProt, does a BLOSUM62 local alignment (gap open -10, extend -0.5) of the
PTP1B catalytic region (residues 1-300) against SDF-9, and reports the SDF-9 residue aligned to each PTP1B
landmark. Raw output: catalytic_residues_output.tsv.

Result

PTP1B residue	Role	SDF-9 residue
Y46	pTyr-recognition loop (KNRY)	V37
D48	pTyr-recognition loop	K39
D181	WPD-loop general acid	E171 (WPD loop region poorly aligned; low confidence)
H214	P-loop, lowers Cys pKa	Q222
C215	catalytic nucleophile	S223
S216	P-loop	A224
G220	P-loop	S228
R221	P-loop phosphate binding	R229 (retained)
S222	P-loop	A230
Q262	Q-loop	P269 (gapped region; low confidence)

The signature motif aligns without gaps: PTP1B HCSAGIGRSG (214-223) vs SDF-9 QSARGSSRAG (222-231).

Interpretation

The catalytic cysteine position is Ser223 in the current SDF-9 sequence. UniProt's CAUTION text
says "lysine at position 223"; the sequence in sdf-9-uniprot.txt has S at 223, so the CAUTION text is
inaccurate about the identity of the substituting residue (the conclusion, loss of the nucleophile, holds).
Besides Cys->Ser, SDF-9 also lost the P-loop His (->Gln) and the P-loop Ser/Thr (->Ala) and the
pTyr-recognition Tyr46 (->Val). Loss of multiple catalytic and substrate-recognition residues supports a
catalytically dead pseudophosphatase.
The P-loop arginine (R229) is retained, so a residual phosphate/anion-binding pocket is possible, but the
pTyr-recognition loop is degenerate; there is no sequence-level basis to assert phosphotyrosine binding.
Limitation: pairwise alignment to a single PTP; WPD- and Q-loop assignments fall in gapped regions and are
low confidence. The P-loop assignment is unambiguous.