MAP3K2 (MEKK2) curation notes

UniProt Q9Y2U5, 619 aa. N-terminal PB1 domain (PB1_MEKK2/3, IPR034879), C-terminal STE11-family
Ser/Thr kinase domain. Close paralog of MAP3K3/MEKK3 (kinase domains ~94% conserved)
PMID:8621389.

Molecular function

• MAP3K that phosphorylates and activates MAP2Ks. Original cloning: immunoprecipitated MEKK2
  phosphorylated MEK1 and JNKK; overexpression activates ERK1/2 and JNK but not p38
  PMID:8621389.
• Best-established physiological substrate: MEK5 (MAP2K5), recruited through PB1-PB1 heterodimerization
  PMID:12912994,
  PMID:11073940.
• Direct phosphorylation of MEK5 (Ser311/Thr315) and MKK6 (Ser207), with a kinase-dead D483N control,
  and a 2.4 A kinase-domain crystal structure showing an alphaG-centred dimerization surface needed for
  autophosphorylation, conserved with MEKK3 PMID:41318559.
• MKK7/JNK branch in mast cells: MEKK2 disruption inhibits FcepsilonRI-induced MKK7 activation
  PMID:11274363.
  But in T cells Mekk2 loss does not block JNK (moderately enhances it) PMID:12138187.
  So the JNK branch is context-dependent; not proposed as a NEW term.

Biological process

• ERK5 cascade: MEKK2-MEK5-ERK5-MEF2C module in mast cells PMID:14515274.
  Free PB1 domain of MEKK2/3 selectively blocks ERK5 but not p38/JNK PMID:12912994.
• Hyperosmotic stress: MEKK2 mediates transient ERK activation, terminated by CHIP (STUB1)-dependent
  degradation PMID:20588253.
• Hippo: MEKK2/3 phosphorylate LATS1/2 and YAP/TAZ downstream of TNF, serum, actin dynamics PMID:33571521 (mostly
  combined MEKK2/3 perturbation; not proposed as NEW).
• Regulators: STK38 binds the catalytic domain and inhibits MEKK2 autophosphorylation PMID:17906693;
  SMYD3 methylates MEKK2 (Lys260) PMID:27066749; 14-3-3 binds phospho-Thr283 (deep research; secondary
  source) ; NEDD4L ubiquitination (deep research, not cached).

Localization

• Cytosolic at rest; translocates to nucleus on EGF PMID:15075238.
• HPA IDA: cytosol + nucleoplasm (GO_REF:0000052).

Annotation issues

• GO:0071260 cellular response to mechanical stimulus (IEP, PMID:19593445, UniProt 2010). The cached
  full text of PMID:19593445 (BAD in prostate cancer) contains no mention of MEKK2/MAP3K2 or of mechanical
  stimuli; the citation appears not to support the annotation. Left UNDECIDED rather than REMOVE because we
  cannot rule out supplementary data / identifier mix-up without the curator's source.
• GO:0004672 IDA from PMID:15001576 (TRAF7/MEKK3 paper): abstract only mentions MEKK3. Following the
  "do not overrule curators" rule: function is clearly correct for MEKK2, so ACCEPT.
• Protein binding IPIs: 14-3-3 partners (SFN, YWHAE) -> GO:0071889; MAP2K5 -> GO:0019901; STUB1 (CHIP E3)
  -> GO:0031625; SMYD3 (methyltransferase acting on MEKK2) -> REMOVE (no informative MF for being a
  substrate). Mirrors the MAP3K5 review.
• IDA intracellular signal transduction (PMID:14515274) -> MODIFY to GO:0070375 ERK5 cascade.
  ERK5 cascade is sparsely used in GO (7 annotations across human/mouse/rat in QuickGO, 2026-09), so its
  absence on MEKK2/3 is not a curatorial convention; MEKK2 catalyses the MEK5-phosphorylation step, so it
  participates (passes the participation test).

Audit (second pass)

• Replaced deep-research (secondary) support for the 14-3-3 protein-binding MODIFYs with primary papers:
  PMID:23963453,
  PMID:9452471. MEKK3 comparison now cites
  PMID:18308725 (pThr294) and PMID:16407301 (pSer526).
• Added reference_review to all references; tightened description (Ser519 autophosphorylation, Lad/RIBP, viable knockout
  PMID:12138187).
• Suggested question on MEKK2 vs MEKK3 dominance checked against PMID:11073940 abstract (kinase-inactive MEKK3, not
  MEKK2, blocked EGF/H2O2 ERK5 activation in Cos7/HEK293; reverse in D10 T cells).