IRAK4 (human, Q9NWZ3) review notes

Sources used

• UniProt Q9NWZ3 (IRAK4-uniprot.txt): 460 aa; N-terminal death domain (20-104), C-terminal
  protein kinase domain (186-454), active-site proton acceptor D311; Pelle subfamily of the TKL
  Ser/Thr kinases; EC 2.7.11.1; Mg2+ cofactor. Loss-of-function causes IMD67 (IRAK-4 deficiency).
• Cached GOA publications (all 26 GOA PMIDs present in publications/), and cached Reactome
  entries under reactome/.
• Deep research: IRAK4-deep-research-falcon.md (Falcon) arrived after the review was written;
  see the cross-check section below.

Deep-research cross-check (2026-09-30)

Compared the Falcon report against every annotation decision, the description and both core
functions.
- Agreement: active Ser/Thr kinase of the Myddosome; DD-mediated MyD88 recruitment then
IRAK1/IRAK2 recruitment; separable kinase and scaffold roles; scaffold integrates MYD88 and TRIF
at TRAF6 in TLR4 signalling (PMID:35977521, already cited); TLR3 is IRAK4-independent; IRAK4 is
cytosolic/receptor-proximal and "not a transmembrane, extracellular, or constitutively nuclear
protein" - consistent with REMOVE (extracellular region) and MARK_AS_OVER_ANNOTATED (nucleus).
The falcon file is now cited as context on those two rows. No annotation action changed.
- Additions (not verified in a cached primary paper, so no annotation change):
trans-autophosphorylation at T345/S346 as a Myddosome-maturation switch (Srikanth et al. 2024,
bioRxiv preprint); IRAK1 T209 as the initiating IRAK4 site, with T387 possibly IRAK1
autophosphorylation (secondary review); a MyD88-independent IRAK4-IRAK1 pathway after DNA
double-strand breaks (Li et al. 2023, not fetched). Raised as a suggested question.
- Nuance: the report states kinase activity is essential for MyD88 signalling largely from
mouse Irak4 kinase-dead data; the review already records that human macrophages/fibroblasts
are partly kinase-independent PMID:36865541. No conflict requiring change.
- Changes: falcon file added to references and cited on GO:0005576 and GO:0005634 rows; one new
suggested question.

Core biology (with provenance)

• Discovery and kinase activity: IRAK4 is the closest human homolog of Drosophila Pelle
  PMID:11960013. It phosphorylates IRAK1 in vitro
  PMID:11960013, on activation-loop sites
  PMID:11960013.
  Unlike other IRAKs it needs its kinase activity to activate NF-kB in overexpression
  PMID:11960013.
• Additional substrates: Pellino E3 ligases [PMID:12860405 "Pellino2 is one of the first substrates
  identified for IRAK1 and IRAK4."; PMID:19264966 "The E3 ubiquitin ligase Pellino can be activated by
  phosphorylation in vitro, catalyzed by IL-1 receptor-associated kinase 1 (IRAK1) or IRAK4."];
  IRAK2 PMID:36865541.
• Myddosome: IRAK4 death domain binds MyD88 death domains and in turn recruits IRAK2/IRAK1
  PMID:20485341.
  Disease/SNP variants in the IRAK4 DD abolish MyD88 binding PMID:24316379. Endogenous
  myddosomes in macrophages have IRAK4 and IRAK2 as core kinases PMID:38961291.
• Kinase vs scaffold: In mouse, kinase activity required for MyD88 signalling; the scaffold (not the
  kinase) is also required for TRAF6 activation downstream of both MYD88 and TRIF in TLR4 signalling
  PMID:35977521. In human cells the kinase is partly dispensable for cytokine output
  PMID:36865541. Both roles are part of the core function.
• Human genetics: IRAK4-deficient cells fail to respond to TLR7/8/9 (IFN induction)
  PMID:16286015, but respond to TLR3 (TRIF-only) agonists.
• Other: IRAK4 also mediates the MyD88-independent Unc5CL pathway PMID:22158417; contributes to
  non-transcriptional NLRP3 priming PMID:36865541.

Curation decisions and rationale

• Kinase MF: accept GO:0004674 (EXP, TAS, IEA); generic protein kinase activity and kinase activity MODIFY to GO:0004674 (the IDA paper shows Ser/Thr phosphorylation of IRAK1 peptides).
• Scaffold: IPI protein-binding rows with MyD88 and IRAK2 from the structural/biophysical papers
  (PMID:20485341, PMID:24316379) are MODIFIED to GO:0070513 death domain binding (DD-DD
  interactions are what those papers measure). High-throughput Y2H/variant screens and the
  Pellino/IRAK3/VSIG8/TRIM7/ZNF597/IL36RN rows are REMOVED as uninformative protein binding
  (the substrate relationship to Pellino is already captured by the kinase MF).
• Project question (Toll vs TLR): the IBA to GO:0008063 Toll signaling pathway (node
  PTN000701353, donors FlyBase tub FBgn0003882 and pll FBgn0010441) transfers a term defined by
  the Drosophila Toll receptor/Spaetzle to a vertebrate protein. GO:0008063 is not an ancestor of
  GO:0002224 (its parent is cell surface receptor signaling pathway; GO:0002224 sits under pattern
  recognition receptor signaling). MODIFY to GO:0002755 MyD88-dependent TLR signaling pathway,
  which is the vertebrate equivalent. The IBA to plasma membrane from the same node is kept as
  non-core (IRAK4 is peripheral, recruited to activated receptor complexes).
• GO:0038172 IL-33 signalling (IDA, PMID:11960013): the full text studies IL-1 in 293RI cells
  and never mentions IL-33/ST2 (IL-33 was described after 2002). MODIFY to GO:0070498 IL-1
  mediated signaling; IL-33 signalling through IRAK4 is biologically plausible but is not what
  this paper shows.
• cell surface (IDA, PMID:22851693): IRAK4 is a cytosolic protein recruited to the cytoplasmic
  face of the IL-1R complex; the paper describes membrane-associated complex I. Results section not
  in cache. MODIFY to GO:0031234 extrinsic component of cytoplasmic side of plasma membrane.
• NLRP3 inflammasome complex assembly (NAS): IRAK4 acts in priming, it does not assemble the
  inflammasome; MODIFY to GO:1900227 positive regulation of NLRP3 inflammasome complex assembly.
• Ensembl Compara IEA from rat (extracellular region, positive regulation of smooth muscle cell
  proliferation): REMOVE. IRAK4 has no signal peptide and is intracellular; the SMC-proliferation
  claim is an indirect downstream phenotype.
• IL-1 receptor binding (IEA from mouse Irak4): IRAK4 is recruited through MyD88, not by direct
  receptor binding; MARK_AS_OVER_ANNOTATED.
• Nucleus IBA (PTN000784925; donors IRAK1, IRAK3, Pelle relatives, plant kinases): no IRAK4-specific
  nuclear evidence; UniProt says cytoplasm. MARK_AS_OVER_ANNOTATED.
• Neutrophil migration / neutrophil mediated immunity (IMP, PMID:19663824, abstract not available):
  kept as non-core; phenotype of patient cells downstream of TLR/IL-1R signalling.

Paralog consistency (IRAK3)

IRAK3 (genes/human/IRAK3) is a pseudokinase; its kinase annotations are an IRAK-family
propagation issue. IRAK4 is the catalytically active apical kinase (canonical catalytic motifs
conserved in IRAK1 and IRAK4 but not IRAK2/IRAK3 PMID:33238146), so kinase terms are accepted here. Note for IRAK3's review
(not edited): the IRAK3 review keeps several UNDECIDED rows on magnesium ion binding etc.; not
changed here.