Protein kinase C signaling — chaperone-assisted maturation, PDPK1 and mTORC2 priming, isoform-class activation by DAG/Ca2+ or the PAR scaffold, RACK1/PICK1 anchoring and PHLPP termination; PRKCA/B/G/D/E/H/Q/Z/I, PDPK1, HSP90AA1/CDC37, PARD3/PARD6, RACK1, PICK1, PHLPP1/2

The protein kinase C (PKC) family comprises nine human AGC-group serine/threonine kinases, divided into three classes by the second messengers that activate them. All have an N-terminal autoinhibitory pseudosubstrate and a C-terminal kinase domain. Conventional PKCs (PRKCA, PRKCB, PRKCG) have a tandem diacylglycerol (DAG)-binding C1 domain and a calcium-binding C2 domain and need both DAG and Ca2+; novel PKCs (PRKCD, PRKCE, PRKCH, PRKCQ) have a C1 domain with higher DAG affinity and a C2 domain that does not bind calcium, so they need DAG alone; atypical PKCs (PRKCZ, PRKCI) have a single C1 domain that does not bind DAG and are controlled instead by protein-protein interactions, chiefly through the PB1 domain with PAR6 and with PAR3 in the PAR polarity complex. Newly synthesised conventional and novel PKCs are held by the HSP90-CDC37 kinase chaperone while they are processed by three ordered phosphorylations. PDPK1 (PDK1) phosphorylates the activation loop, and then the turn motif and hydrophobic motif in the C-terminal tail are phosphorylated in a step that requires mTORC2. The resulting mature, stable enzyme sits in the cytosol with its pseudosubstrate in the active site. When phospholipase C generates DAG and IP3-released calcium, conventional PKCs are recruited to the plasma membrane by the C2 domain and engage DAG through C1; novel PKCs go to DAG-containing membranes directly. Membrane binding releases the pseudosubstrate, and the active kinase is positioned at particular sites by anchoring proteins, including RACK1, which binds activated PKC, and the PDZ-BAR protein PICK1. The open, active conformation is also sensitive to phosphatases. PHLPP1 and PHLPP2 dephosphorylate the hydrophobic motif, which moves PKC to the detergent-insoluble fraction for degradation and so sets the cellular level of signalling-competent PKC. Atypical PKCs are phosphorylated by PDPK1 in an agonist-dependent manner and, within the PAR3-PAR6-aPKC complex, regulate cell polarity.

Derived from the RGD Pathway Ontology diagram PW:0000840 through the RGD_PATHWAY_ONTOLOGY project's diagram->module route. RGD curates 32 human genes to PW:0000840 (BTK CDC37 COPB1 FSCN1 FYN HSP90AA1 LCK LDB3 PARD3 PARD6A PARD6B PARD6G PCNT PDPK1 PHLPP1 PHLPP2 PICK1 PLCB1-4 PRKCA PRKCB PRKCD PRKCE PRKCG PRKCH PRKCI PRKCQ PRKCZ RACK1 RASA1). Twenty-one are modelled as participants here. The four PLC beta isoforms are covered by reference to another module. Seven are excluded with the reasons below. MTOR, RICTOR and MAPKAP1 (mTORC2) are added although RGD does not list them, because the RGD record text names mTORC2 as required for turn-motif phosphorylation. Boundary decisions. (1) No duplication of the Gq-PLC module. The receptor -> Galphaq -> PLC beta -> IP3/DAG/Ca2+ chain is already modelled in modules/gpcr_plc_calcium_pkc_signaling.yaml. RGD itself draws these steps as links to other pathway diagrams (PW:0000229 Galphaq signalling, PW:0001139/PW:0001140 calcium transport and signalling), not as part of PW:0000840. Second-messenger generation is therefore an optional part included by module_ref to that module's PLC beta node (gpcr_plc_calcium_pkc_signaling#plc_beta_activation). PLCB1-4 are not re-modelled. That part is optional because atypical PKCs need neither DAG nor Ca2+, and because PLC gamma downstream of receptor tyrosine kinases also produces DAG (mentioned in the RGD text but not modelled). The PW ids of the subpathways are mentioned here only and are not evidence source_ids. (2) The core of this module is the PKC life cycle: chaperoned maturation and priming (HSP90-CDC37, PDPK1, mTORC2), isoform-class activation, anchoring and PHLPP termination. The PRKCA node of the Gq-PLC module and the conventional variant here describe the same kinase activity. This module adds the isoform classes and the regulatory machinery around them. (3) The activation step is a variant set on the axis of second-messenger dependence, with selection ONE_OR_MORE because most cells express several classes. Molecular functions follow the experimental GO annotations. Conventional isoforms use calcium,diacylglycerol-dependent serine/threonine kinase activity GO:0004698 (IDA for PRKCA, PRKCB and PRKCG). Novel isoforms use diacylglycerol-dependent, calcium-independent serine/threonine kinase activity GO:0004699 (EXP/IDA for PRKCD, PRKCE and PRKCQ, and the PRKCQ gene-review core function). PRKCH carries GO:0004699 only by IEA, which is noted on its annoton. Atypical isoforms use protein serine/threonine kinase activity GO:0004674 (IDA for PRKCI). GO:0004699 is also IEA/ISS on PRKCI, which is not used because aPKCs do not respond to DAG. PRKCZ carries GO:0004698 by EXP (Reactome-derived), which is inconsistent with its DAG/calcium insensitivity and is flagged as a possible upstream annotation error below. (4) Comparator check for the pathway term. In human GOA, GO:0070528 protein kinase C signaling is carried only by PRKCA and PRKCB (IDA) and PRKCD (IEA). The regulators carry child regulation terms (GO:0090037/GO:0090038 on CD40, DGKQ, EGFR and HSPB1), and PDPK1, HSP90AA1, CDC37, PHLPP1/2, RACK1 and PICK1 carry neither. GO:0070528 is therefore placed only on the module concept and on the PKC kinase annotons. The maturation, anchoring and termination annotons carry their own molecular functions (and protein dephosphorylation for PHLPP) and are not asserted to be involved in GO:0070528. They act on PKC, the enzyme, not on the signal it relays. (5) Maturation. HSP90-CDC37 binding of newly made conventional and novel (not atypical) PKC is modelled as the HSP90-CDC37 chaperone complex GO:1990565. The complex's activity is ATP-dependent protein folding chaperone GO:0140662, the HSP90AA1 gene-review core function. CDC37 is an active unit without its own GO MF, because its only experimental MF rows are protein binding. PDPK1 activation-loop phosphorylation is protein serine/threonine kinase activity GO:0004674 (IDA). It is constitutive for cPKC/nPKC (PMID:9889098) and enhanced by PI 3-kinase products for PRKCZ (PMID:9768361). The mTORC2 step is modelled as the TORC2 complex GO:0031932, with MTOR GO:0004674 and RICTOR/MAPKAP1 enzyme-substrate adaptor activity GO:0140767, as in the human mTORC2 GO-CAM gomodel:66e382fb00001412. The evidence (PMID:18566586, PMID:18566587) shows that mTORC2 is required for turn-motif phosphorylation of conventional PKC and PKCepsilon. RGD states that whether mTORC2 phosphorylates the PKC tail directly is unclear, and this is recorded as a knowledge gap rather than written as a direct kinase-substrate edge on PKC. (6) Atypical PKC and the PAR complex. The PAR3-PAR6 scaffold is modelled inside the atypical variant as a PAR polarity complex (GO:0120157) annoton with signaling adaptor activity GO:0035591, the PARD6B/PARD6G gene-review core function. PARD6A's review uses the parent-adjacent protein-macromolecule adaptor activity GO:0030674. PARD3 has no review and no experimental MF beyond protein binding. Downstream polarity biology (Crumbs and Scribble complexes, apical-basal antagonism) is not modelled. No human or vertebrate polarity module exists in modules/. The closest is modules/dmel_epithelial_polarity_complexes.yaml (Drosophila Par/Crumbs/Scribble), which is not module_ref'd because it is a concrete fly module. A human PAR/Crumbs/Scribble polarity module is a candidate to which the aPKC variant could later point. (7) Anchoring is an optional part with RACK1 and PICK1. These are the two "PKC interactors" with gene reviews, and binding of activated PKC by each is documented (PMID:8302854 for RACK1; PMID:7844141 for PICK1). RACK1 uses its core GO:0060090 molecular adaptor activity. PICK1 uses protein kinase C binding GO:0005080, which its review accepts. PICK1 is also a PKC substrate, so it is kept for its anchoring role only. (8) Exclusions and the substrate-vs-participant rule. FSCN1 (fascin; PKC phosphorylates its actin-binding site) is a PKC substrate and is excluded: it does none of the work of PKC signalling. LDB3 (cypher/ZASP), PCNT (pericentrin) and COPB1 are listed by RGD as "PKC interactors". They are thought to act as compartment-specific anchoring/scaffold proteins (Z-disc, centrosome and Golgi coatomer respectively) but are not modelled, because no cached evidence for the PKC-binding role was assembled here. For COPB1 the coatomer RACK literature may refer to the beta-prime subunit (COPB2) rather than beta-COP. This was not verified and is recorded only as a question for RGD. BTK, LCK and FYN are tyrosine kinases with reciprocal cross-talk with PKC (tyrosine phosphorylation of novel PKCs, PKC phosphorylation of BTK). RASA1 (p120 RasGAP) is a Ras regulator. Both groups belong to antigen-receptor and Ras modules, not to the PKC life cycle. Under the CLAUDE.md rule, kinases that modify PKC or are modified by it do not get PKC-signalling participation unless they perform a step of the PKC module itself. (9) Signal termination by DGK conversion of DAG to phosphatidic acid and by receptor internalisation is mentioned by RGD as "not shown" and is not modelled. HSP70-mediated rephosphorylation and recycling of dephosphorylated PKC (RGD text) is described in prose only. Downstream outcomes drawn by RGD (proliferation, differentiation, cell death) are out of scope. Possible upstream errors noticed: the RGD PW:0000840 text says "Pdpk1 phosphorylates the activation loop of PLC", where PKC is plainly meant. PRKCZ (Q05513) carries GO:0004698 calcium,diacylglycerol-dependent serine/threonine kinase activity by EXP, which contradicts aPKC insensitivity to DAG and calcium. GO-CAM coverage is linked under gocam_associations for PRKCA, PRKCG (the project's best-overlap model gomodel:66b5638000002273), PRKCI, PDPK1/mTORC2 and PHLPP1. No GO-CAM in gocams/index.tsv covers PRKCB, PRKCD, PRKCE, PRKCH, PRKCQ, PRKCZ, HSP90AA1, CDC37, RACK1, PICK1, PARD3/PARD6 or PHLPP2. Gene reviews exist for HSP90AA1, PARD6A, PARD6B, PARD6G, PICK1, PRKCQ and RACK1, and their core functions were used. The other members have no review yet. PANTHER family ids and PTNs are deliberately not asserted.