UniProt O43353. Human. Gene symbol RIPK2 (HGNC).
just fetch-gene-pmids human RIPK2 cached 47 publications (all GOA PMIDs).-deep-research-PROVIDER.md was hand-written (per CLAUDE.md).N-terminal Ser/Thr/Tyr kinase domain (~1-294) + intermediate (IM) region +
C-terminal CARD (~435-526). CARD mediates CARD-CARD recruitment by NOD1/NOD2 and
homotypic filament (RIPosome) assembly. Kinase catalytic residues: ATP-binding
Lys47; catalytic Asp146 (HRD->HHD motif) [PMID:30026309 "the catalytic aspartate
146 in the “HRD” motif (HHD in human RIPK2)"; "K47 and D146 are catalytic
residues for ATP hydrolysis."].
This is the crux and the literature converges (despite older claims):
- Kinase-dead K47M/K47A still activates NF-kB (Inohara 2000, cited in
PMID:18079694).
- RICK K63-polyUb (K209) does not require kinase activity; it is the step that
recruits TAK1 [PMID:18079694 "RICK polyubiquitination did not require the kinase
activity of RICK"; "the residue critical for kinase activity is not essential for
NF-kappaB activation, polyubiquitination of RICK and interaction of RICK with
NEMO"].
- Decisive reconstitution: two independent kinase-dead mutants (K47R, D146N)
restore NOD2 signaling to WT level [PMID:30026309 "introduction of both
kinase‐dead RIPK2 mutants restored NOD2 signaling and CXCL8 production to a
similar level as with WT RIPK2 in two independent RIPK2 KO clones, showing that
the catalytic function is not needed for RIPK2's role in NOD2‐dependent
inflammatory signaling"; abstract "RIPK2 kinase activity is dispensable for NOD2
inflammatory signaling"].
- Why inhibitors "work": kinase inhibitors block signaling by antagonizing XIAP
binding, not by blocking catalysis PMID:30026309. Confirmed independently PMID:29452636.
- Caveat / dissent: Tigno-Aranjuez reported Y474 autophosphorylation "necessary
for effective NOD2 signaling" PMID:21123652. This is the tyrosine-kinase
requirement claim. The weight of kinase-dead reconstitution (K47R + D146N, which
abolish both Ser/Thr and the ATP chemistry needed for any autophosphorylation)
showing FULL rescue (PMID:30026309) indicates catalysis per se is not required;
inhibitor phenotypes are explained by XIAP-binding disruption. Y474 may matter
as a conformational/structural feature of the kinase domain rather than as a
catalytic output. Older inhibitor-based "kinase required" claims
(Windheim 2007; Nembrini 2009; Tigno-Aranjuez 2010) are reinterpreted this way
PMID:30026309.
Position: RIPK2 is a genuine autokinase, but its catalytic activity is NOT the
essential step for NOD1/NOD2 -> NF-kB/MAPK output. The essential role is the
CARD-scaffold + ubiquitin-scaffold function.
MAP3K7 review (not edited): treats TAK1 as the ubiquitin-activated MAP3K
downstream of RIPK2 ubiquitination; "TAK1 is required for NOD2-mediated NF-kappaB
activation downstream of RIPK2 ubiquitination". Consistent with this review:
RIPK2 is the K63/M1-ubiquitinated scaffold that recruits TAK1-TAB; TAK1 does the
kinase work on IKK/MAP2Ks. No disagreement. MAP3K7 also handles generic
protein binding by MODIFY->ubiquitin protein ligase binding / REMOVE with the real
interaction captured elsewhere; I follow the same practice here.
modules/nlr_signaling.yaml RIPK2 adaptor annoton (ripk2_ripk2_kinase_adaptor):
- States "RIPK2 catalytic activity is dispensable for NOD2 signaling... so no
kinase function is asserted for RIPK2 in this module." AGREE on dispensability
for the NF-kB output, and with asserting the scaffold (GO:0035591) role as the
module function. The module's position is consistent with (b)/(c) here.
- Two minor points to report (NOT edit):
1. The annoton label has been relabelled "RIPK2 adaptor/scaffold" in the module (the annoton id
is unchanged to keep edges stable), since the module asserts no kinase function.
2. PANTHER family: the module's PTHR44329 grounding is correct. RIPK2's UniProt record gives
PTHR44329:SF9 (RECEPTOR-INTERACTING SERINE_THREONINE-PROTEIN KINASE 2), and its IBA anchors
PTN001908418, PTN002892787 and PTN002892800 are all PTHR44329 nodes
(interpro/panther/PTHR44329/PTHR44329-paint.tsv). An earlier draft of this note wrongly
placed those nodes in another family. The deep research "contributes" wording for kinase
activity is superseded by PMID:30026309/29452636 (dispensable); the module's current
edit is the better-supported position.
Rows: GO:0005783 EXP (PMID:28656966) and IEA (GO_REF:0000044, SL-0095), both located_in, both
previously ACCEPT.
Checks:
- UniProt O43353: "Endoplasmic reticulum {ECO:0000269|PubMed:28656966}", the only basis of the IEA.
- PMID:28656966 (Bist et al. 2017, full text): endogenous RIPK2 in HCT116 cells fractionated, ER vs
PM fractions controlled with markers PMID:28656966
PMID:28656966;
confocal with mCherry-Sec61beta used ectopic HA-RIP2. The ER is proposed as the site where the
ER-anchored E3 ZNRF4 ubiquitinates and degrades RIPK2.
- Soluble kinase (no signal peptide, TM or retention motif); core_functions place its activity
in the cytosol and at the plasma membrane.
Outcome: both rows ACCEPT -> KEEP_AS_NON_CORE. The ER pool is real (marker-controlled
fraction of endogenous protein) but recruited and regulatory, not where the kinase/scaffold
function is shown (BRSK2/A1CF precedent; contrast AMBRA1). Stays at the root.