Working journal for the human TTK gene review. Append new entries at the end.
UniProt P33981 TTK_HUMAN, 857 aa, "Dual specificity protein kinase TTK",
EC 2.7.12.1, synonyms MPS1 / MPS1L1. Kinase domain 525–791, active site Asp647,
ATP-binding 531–539 and Lys553. This is unambiguously the human Mps1 checkpoint
kinase, not a receptor tyrosine kinase despite the "TTK" (threonine/tyrosine
kinase) historical name. The deep research report independently confirmed the
identity [file:human/TTK/TTK-deep-research-falcon.md "The requested target is
human TTK protein kinase, UniProt P33981, also called MPS1, hMPS1,
or MPS1L1."].
Apical SAC kinase. TTK is recruited to unattached kinetochores and
phosphorylates the KNL1 MELT repeats to create the docking mark for Bub1–Bub3
PMID:22660415. The
MELT-phospho → BUB recruitment step is confirmed in human cells in the Ska
paper's introduction PMID:28441529. Downstream, BUB1 Thr461 (after CDK1 priming of Ser459) and
MAD1 Thr716 are TTK substrates that drive MAD1–MAD2 recruitment and CDC20
capture into the MCC [file:human/TTK/TTK-deep-research-falcon.md
"It phosphorylates KNL1, BUB1, and MAD1 in sequence, thereby catalyzing
production of the mitotic checkpoint complex (MCC)."]. UniProt records the
MAD1L1 phosphorylation and the MAD1L1/MAD2L1 interactions
(PubMed:29162720, not cached locally).
Attachment sensing via NDC80C. The mechanism that makes TTK an
attachment sensor rather than a constitutive kinase is competitive binding:
PMID:26068854 and
PMID:26068854. Consistent with this, PMID:28441529, and activity is highest where it matters
PMID:28441529.
This is what justifies ACCEPTing the GO:0043515 kinetochore binding
Ensembl-transferred row rather than treating it as a low-information
binding term: the binding event is the sensing mechanism.
Error correction. Two independent routes, at different confidence levels.
Borealin/CDCA8: PMID:18243099 and
PMID:18243099. The Ska route is the stronger and later account, and
it explicitly challenges the generality of the Borealin model: PMID:28441529, while showing a direct, Aurora-B-independent
effect PMID:28441529 mediated by SKA3 Ser34 PMID:28441529, opposed by PP2A-B56.
→ Both IDA rows for GO:0140273 were ACCEPTed; the tension between the two
papers was recorded in the Borealin row's reason rather than used as grounds
to downgrade either annotation.
Secondary roles. MCRS1/KIF2A spindle morphology PMID:30785839, with the site mapped and
inhibitor-validated PMID:30785839. Interphase nuclear pore pool PMID:19273613.
MPS1-dependent BLM Ser144 phosphorylation in mitosis PMID:16864798 — noted for completeness; there is no GOA row for it,
and I did not propose one (see "NEW terms considered" below).
Kinase activity family (8 rows). Ser/Thr kinase activity (GO:0004674, ×4:
IBA, 2× IDA, TAS) and serine kinase activity (GO:0106310, ×2: EXP, Rhea IEA) all
ACCEPT — every documented physiological substrate is a Ser/Thr site (Borealin,
KNL1 MELT, BUB1 T461, MAD1 T716, SKA3 S34, MCRS1 S64/S65, autophosphorylation
pT33/pS37 and pT360/pS363 PMID:28441529). GO:0004672 protein kinase activity
(InterPro2GO) ACCEPT as a correct broad parent. GO:0004712
protein serine/threonine/tyrosine kinase activity (IBA + IEA) ACCEPT — this is
the right level at which to record the family's dual specificity.
Tyrosine kinase activity (GO:0004713, ×2) → MARK_AS_OVER_ANNOTATED. The only
evidence is in vitro and marginal: the human kinase domain in E. coli gives
"markedly elevated" phosphoSer/phosphoThr but only "slightly increased"
phosphotyrosine PMID:1639825, and the TAS row's support is the yeast enzyme
PMID:7737118. No physiological tyrosine substrate is known. The deep research makes
the same point [file:human/TTK/TTK-deep-research-falcon.md "The nomenclature
therefore should not be taken to mean that tyrosine phosphorylation is equally
prominent in its established SAC function."]. This is exactly
MARK_AS_OVER_ANNOTATED territory — not wrong, but a stand-alone tyrosine-kinase
MF invites grouping TTK with real tyrosine kinases, and GO:0004712 already
carries the capability at the appropriate level. Deliberately not REMOVE, since
the in vitro activity is real.
GO:0008284 positive regulation of cell population proliferation (TAS) →
REMOVE. This is the one substantive REMOVE. The cited paper reports an
expression correlation PMID:1639825 — the expected pattern for any mitotic kinase, since
transcript abundance tracks mitotic index. The annotation converts a correlation
into a regulatory function. Applying the participation test from CLAUDE.md: which
entity performs a step of proliferation-promoting signalling? TTK performs no such
step; it restrains anaphase and corrects attachments. The deep research is
independently sceptical of the oncogenic reading: high TTK "often tracks
proliferation and mitotic fraction". Removal does not dispute the expression data.
GO:0016020 membrane (HDA) → MARK_AS_OVER_ANNOTATED. One HTP membrane-fraction
detection in YTS cells, where the authors themselves caution PMID:19946888.
TTK has no TM segment, signal peptide or lipid anchor in the UniProt feature
table, and no membrane function is described anywhere. Chose
MARK_AS_OVER_ANNOTATED over REMOVE because the peptide detection itself is not
in doubt — the interpretation as a localization is.
GO:0005515 protein binding (IPI, TPR) → REMOVE. Repository policy for bare
GO:0005515: resolve to an informative MF if the evidence supports one, otherwise
REMOVE as uninformative without asserting the interaction is false. Here the
paper's finding is that TPR/Megator acts as a spatial regulator delivering Mps1
to kinetochores PMID:19273613 —
that is a property of TPR, not a molecular activity of TTK, so no informative
replacement MF can be justified. The interaction is recorded in UniProt and is
not disputed; the interphase NPC localization it reports is instead used to
support the nucleus row.
GO:0005819 spindle (TAS) → MODIFY to GO:0000776 kinetochore. The cited paper
is a locus-mapping and mutation-screening study whose only TTK content is
PMID:10366450 — no localization data at all. The structure TTK acts on is the
kinetochore, so the accurate CC is the term GOA already has by IBA/IEA.
GO:0007051 spindle organization (TAS) → MODIFY to GO:0007052 mitotic spindle
organization. Same uninformative reference, but unlike the spindle-localization
row the process is genuinely supported by later human work (MCRS1/KIF2A), so
MODIFY to the mitosis-specific descendant rather than REMOVE. GO:0007052 itself
(TAS from the yeast SPB paper) → KEEP_AS_NON_CORE: supported in human cells by
PMID:30785839 but secondary to the checkpoint role.
GO:0000280 nuclear division (ARBA IEA) → MODIFY. Correct but at a uselessly
high altitude; proposed the two mitosis-specific descendants GOA already carries.
Non-core keeps. GO:0007059 chromosome segregation (IBA) and GO:0098813
nuclear chromosome segregation (ARBA) are accurate outcome-level parents —
KEEP_AS_NON_CORE, with the mechanism captured by GO:0007094/GO:0140273.
GO:0005634 nucleus (IBA) and GO:0005737 cytoplasm (IDA) are real pools but not
sites of characterised function — KEEP_AS_NON_CORE. GO:0042802 identical protein
binding: self-association is genuine (trans-autophosphorylation PMID:28441529) but low-information — KEEP_AS_NON_CORE.
GO:0033316 meiotic SAC (IBA from fly + mouse) — KEEP_AS_NON_CORE: sound
phylogenetic inference from characterised metazoan orthologues, no human data.
GO:0060391 positive regulation of SMAD protein signal transduction (TAS) →
UNDECIDED. The cited reference is an abstract-only perspective piece on
noncanonical Smad roles that does not mention TTK in the cached text
PMID:19018011. The underlying
primary claim (Mps1 phosphorylation of Smad2/3) is not in the cache and is not
corroborated by the deep research. Per the enum, inability to access the relevant
publication → UNDECIDED, not REMOVE.
All five IBA rows (GO:0000776, GO:0004674, GO:0004712, GO:0007059, GO:0007094,
GO:0033316) were treated as considered PAINT node placements, not similarity
transfers. Note that TTK's own accession UniProtKB:P33981 appears in the
WITH/FROM of the GO:0004674, GO:0007059 and GO:0007094 IBA rows. Per CLAUDE.md
this is correct and expected — TTK's own experimental annotations are among
the descendant evidences the PAINT curator used to place the IBD — so this was
recorded as a marker of experimental grounding on the target, never flagged as
circular. No propagation_review blocks were added: none of the IBA rows has a
propagation defect to classify.
Two nodes are in play: PTN001122082 (the tighter Mps1/SAC node, cited for
kinetochore, Ser/Thr/Tyr kinase activity, mitotic and meiotic SAC) and
PTN000540737 (a deeper node including plant AT5G20930, worm, trypanosome and
other kinases, cited for Ser/Thr kinase activity, nucleus and chromosome
segregation). The nucleus IBA comes from the deeper node, which is part of why it
was graded non-core rather than accepted as a site of action.
GO:0007100 mitotic centrosome separation / centrosome-duplication terms.GO:0034501 protein localization to kinetochore. Tempting, since TTKproposed_new_terms: [].
Three, all at the kinetochore:
1. Ser/Thr kinase activity → mitotic SAC signalling (the MELT→BUB→MAD1→MCC cascade).
2. Kinetochore binding (GO:0043515) → mitotic SAC signalling — the NDC80C-competitive
binding that makes the kinase an attachment sensor. Kept as a separate core
function because it is a distinct, mechanistically essential activity, not a
restatement of the catalysis.
3. Ser/Thr kinase activity → repair of mitotic kinetochore microtubule attachment
defect (SKA3 Ser34, Borealin).
No in_complex: TTK is a kinase that acts on kinetochore substrates rather than a
stoichiometric subunit of a named GO complex. MCC (GO:0033597) would be wrong —
TTK catalyses its assembly but is not a component.
just validate human TTK → Valid, 0 errors, 0 warnings (the initial
"no annotations reference available deep research files" warning was cleared by
adding file:human/TTK/TTK-deep-research-falcon.md support to the GO:0007094 IBA
and GO:0004713 IEA rows). Two references not in the GOA seed were added to
references because they are cited in supported_by: PMID:22660415 (KNL1 MELT)
and PMID:26068854 (NDC80C competition). Every reference carries a
reference_review; PMID:19018011 is the only UNVERIFIED one.
Status set to COMPLETE.