Focused Curation Report — Dictyostelium discoideum tipA (UniProt Q94489)

Hypothesis evaluated: TipA is a functional PP2C-family (PPM), metal-dependent protein serine/threonine phosphatase. Focus type: function_assignment · Slug: tipa-pp2c-phosphatase Date: 2026-07-12 · Iterations: 1–3

Executive Judgment

Verdict: PARTIALLY SUPPORTED (family/fold supported; catalytic "functional phosphatase" claim UNRESOLVED — over-annotation risk if asserted with experimental confidence).

• The domain/fold assignment is well supported. Q94489 is annotated by multiple orthogonal signature databases as a PPM/PP2C-type phosphatase domain (InterPro IPR001932/IPR053287, Gene3D 3.60.40.10, SUPFAM SSF81606, PANTHER PTHR21586 "TIPA"; UniProt domain 388–706). AlphaFold models this region as a confident PP2C fold (domain mean pLDDT ≈ 81).
• The "functional / catalytically active phosphatase" part is not demonstrated. There is no biochemical evidence — no phosphatase assay, no metal-dependence test, and no substrate has ever been reported for TipA. The only functional study (Stege et al. 1999, 10402673) is a developmental-genetic phenotype (tip formation, cell sorting), not an enzymatic characterization.
• A conserved-motif red flag: the invariant PP2C metal-1/general-acid signature [VY]-D-G-H-x-G ("DGH") — present in human PPM1A and in the Dictyostelium functional PP2C Spalten/SpnA — is absent from TipA (the equivalent position reads ADGC, His→Cys).
• Counter-nuance from structure: despite the linear-motif loss, the AlphaFold model assembles a spatial cluster of aspartate carboxylates (D413, D430, D576, D642, D646) plus His645 into a pocket resembling a PP2C bimetal active site. So a divergent but real metal-dependent active site remains plausible; this is not a clearly collapsed pseudophosphatase.

• Family-wide catalytic-motif loss (over-annotation signal, added Iteration 2): across the entire TIPA-specific PP2C-like family (InterPro IPR053287, n=100; Dictyostelium, Drosophila CG9801, Caenorhabditis, Trichinella, cnidarians, molluscs, Entamoeba), the canonical DGHxG motif is present in 0% and a bare DGH in only 7%, versus 79% in the canonical PP2C family (IPR001932). The TIPA family is a divergent lineage that systematically lacks the invariant catalytic His, so family/domain membership cannot by itself justify a "functional phosphatase" call — this is a textbook paralog/frequency-bias over-annotation scenario. TipA's His645 is a Dictyostelium-specific substitution (family consensus has Ser here), i.e., a speculative species-specific candidate general acid, not a conserved catalytic signature.

• Metal binding vs catalysis (Iteration 3): the metal-coordinating aspartate scaffold is conserved family-wide (N-terminal motif 80%, C-terminal 86%, both 69% of 100 members) while the catalytic general-acid His is not (7%). This cleanly separates a plausibly retained metal-binding capacity from an unsupported catalytic capacity — the molecular signature of a metal-binding but catalytically divergent/pseudo-PP2C.

Bottom line for the curator: the PP2C domain membership can be annotated (computational/ISS-level), and metal ion binding (GO:0046872) is the best-supported computed inference. A protein Ser/Thr phosphatase activity term should NOT be asserted with experimental evidence; catalytic activity is unproven and the canonical catalytic His motif is degenerate across the whole TIPA lineage. Treat "functional metal-dependent Ser/Thr phosphatase" as a hypothesis to be tested, not an established fact.

Evidence Matrix

GO Curation Implications (leads — require curator verification)

• Molecular Function. Do not assign an experimentally-supported protein Ser/Thr phosphatase MF term. Options, in order of conservatism:
• Acceptable as computational/ISS or IBA (from PANTHER PTHR21586 "TIPA"): GO:0004722 protein serine/threonine phosphatase activity or the more generic GO:0016791 phosphatase activity — with an ISS/IEA evidence code and a caveat note that the canonical catalytic His motif is degenerate and activity is unassayed.
• GO:0046872 metal ion binding is the best-supported computed inference (the metal-coordinating aspartate scaffold is conserved family-wide: N-term 80%, C-term 86%, both 69%), but is still only sequence-level (ISS/IEA) — no metal-binding assay exists.
• If the review’s prior action asserted an experimental MF phosphatase term, that is too strong → recommend downgrading to computational or removing pending assay.
• Biological Process (retain). GO:0031154 culmination involved in sorocarp development (IMP) and GO:0030587 sorocarp development (HMP) are supported by Stege et al. 1999 (10402673). Keep.
• Cellular Component (retain). GO:0005737 cytoplasm (IDA, dictyBase). Keep.
• Net: the seed hypothesis, if used to add an experimentally-weighted MF phosphatase annotation, is too strong; a computational MF annotation with an explicit "atypical/unverified active site" caveat is defensible.

Mechanistic Scope

• Immediate molecular function under test: metal-dependent hydrolysis of phosphoserine/phosphothreonine on protein substrates by a PPM/PP2C catalytic domain. This requires a binuclear Mn²⁺/Mg²⁺ centre coordinated by conserved aspartates plus a general acid (9003755).
• What the evidence actually shows: TipA has the PP2C fold and can geometrically assemble a candidate metal-binding pocket, but no dephosphorylation, metal dependence, or substrate has been measured.
• Downstream/indirect layers (not the MF): the tipA-null phenotype (cell sorting, tip formation, delayed prespore/prestalk differentiation, aggregate morphogenesis) is a developmental outcome, not direct evidence of catalytic activity. Attributing the phenotype to phosphatase catalysis is inference, not demonstration.

Conflicts and Alternatives

1. Paralog confusion (important). The validated Dictyostelium developmental PP2C is Spalten/SpnA (9585512), not TipA. Spalten retains the intact DGH catalytic motif; TipA does not. A curator should ensure phosphatase evidence is not being transferred from Spalten (or the broader PP2C family) onto TipA by homology/family inference.
2. Pseudophosphatase possibility. Loss of the invariant general-acid His (DGH→DGC) is a classic hallmark of degenerate/pseudo-phosphatases. TipA could function as a non-catalytic scaffold/adaptor whose PP2C-like fold binds phospho-proteins without turning them over.
3. Divergent-but-active possibility. The AlphaFold carboxylate+His cluster leaves open that TipA is a sequence-atypical active phosphatase (His role relocated to His645). This cannot be resolved without an assay.
4. Prediction-quality caveats. Family-level annotation (PANTHER "TIPA") is prone to frequency/paralog bias; AlphaFold does not model catalysis; the linear-motif call ignores 3D. Each tool, alone, is weak — hence the "unresolved" MF verdict.
5. Family-wide degeneracy quantified (Iteration 2). The catalytic His loss is not idiosyncratic to TipA but is shared by the whole TIPA/IPR053287 lineage (0/100 with DGHxG vs 79/100 in canonical PP2C). This means any "functional phosphatase" annotation propagated from family/domain membership is unsupported by conserved catalytic residues; if the review's assignment derives from InterPro2GO/PANTHER family inference, that is the specific over-annotation to flag.

Knowledge Gaps

Discriminating Tests (most efficient first)

1. In vitro phosphatase assay of recombinant TipA (or its 388–706 domain) on a generic substrate (pNPP or phosphopeptide) with and without Mn²⁺/Mg²⁺ — directly settles "functional metal-dependent phosphatase."
2. Structure-guided catalytic mutagenesis (candidate metal ligands D413/D430/D576/D642/D646 and His645) → test loss of activity and whether mutants fail to rescue the tipA-null developmental phenotype (links catalysis to biology).
3. Rigorous PPM-family MSA/HMM with metal-site mapping (e.g., against solved PP2C structures) to formally score active-site integrity rather than relying on a single linear motif.
4. Phosphoproteomics of tipA⁻ vs WT during aggregation/tip formation to identify candidate substrates (hyperphosphorylated targets).
5. Comparative check of the PANTHER "TIPA" subfamily (PTHR21586, incl. Drosophila CG9801) for conservation/loss of catalytic residues to see if the whole subfamily is degenerate.

Curation Leads (require curator verification)

• Candidate references to cite/verify:
• PMID 10402673 (Stege 1999) — snippet: "null mutations in tipA showed a primary defect in cell sorting and the formation of tips on the developing mound" → supports BP/CC only, not MF.
• PMID 9585512 (Aubry & Firtel 1998) — snippet: "a carboxy-terminal domain that encodes a functional PP2C" → the validated developmental PP2C is Spalten, useful as paralog control (do not transfer to TipA).
• PMID 9003755 (Das 1996) — snippet: "Mn²⁺- or Mg²⁺-dependent protein Ser/Thr phosphatase … binds two manganese ions … provide a nucleophile and general acid" → defines the catalytic requirements TipA has not been shown to meet.
• Candidate GO actions:
• Keep BP GO:0031154, GO:0030587; keep CC GO:0005737.
• If an MF phosphatase term exists with experimental evidence → downgrade to ISS/IEA or remove (too strong).
• If adding MF, use ISS/IBA GO:0004722 (or GO:0016791) with a caveat that the catalytic His motif is atypical and activity is unverified. Avoid a bare "protein binding" fallback.
• Suggested curator questions:
• Is any experimental phosphatase evidence for TipA in the source, or is the assignment homology-only?
• Is the MF being transferred from Spalten/family membership (paralog carry-over)?
• Suggested experiments: see Discriminating Tests 1–2.

Provenance / Reproducibility

• UniProt Q94489 (759 aa; domain 388–706), P35813 (PPM1A), O15743 (Spalten/SpnA) fetched via UniProt REST.
• Active-site motif scan (DG[HN].G, [ILVMFC]DG[ILVMFAWI]): DGH present in PPM1A@60 and Spalten@749; absent in TipA (ADGC@430).
• AlphaFold AF-Q94489-F1 v6: domain mean pLDDT 80.8; carboxylate cluster D413/D430/D576/D642/D646 + His645 within ~4–6 Å.
• InterPro family motif census (Iteration 2): IPR053287 (TIPA) n=100 → DGHxG 0%, DGH 7%; IPR001932 (canonical PP2C, reviewed) n=100 → DGH 79%. Conserved TIPA-family motifs: "ADG[VC]NWG" and "(T/I)SDG[IV]xDN"; His645 is a Dictyostelium-specific substitution of the family-consensus Ser.
• Pairwise BLOSUM62 global alignment was explicitly judged unreliable (failed on the functional Spalten control) and was therefore not used for residue-level conclusions.

Limitations: all sequence/structure analyses are computational; no wet-lab activity data exist. Conclusions distinguish direct results (motif presence/absence, fold confidence) from inference (catalytic competence).