The atypical protein kinase C (PKC) member PKC-has been implicated in several signal transduction pathways regulating differentiation, proliferation or apoptosis of mammalian cells. We report here the identification of a cytoplasmic and membrane-associated protein that we name zetainteracting protein (ZIP) and that interacts with the regulatory domain of PKC-but not classic PKCs. The structural motifs in ZIP include a recently defined ZZ zinc finger as a potential protein binding module, two PEST sequences and a novel putative protein binding motif with the consensus sequence YXDEDX 5 SDEE͞D. ZIP binds to the pseudosubstrate region in the regulatory domain of PKC-and is phosphorylated by PKCin vitro. ZIP dimerizes via the same region that promotes binding to PKC-suggesting a competitive situation between ZIP:ZIP and ZIP:PKC-complexes. In the absence of PKC-proper subcellular localization of ZIP is impaired and we show that intracellular targeting of ZIP is dependent on a balanced interaction with PKC-. Taking into account the recent isolation of ZIP by others in different contexts we propose that ZIP may function as a scaffold protein linking PKC-to protein tyrosine kinases and cytokine receptors.The intracellular propagation of pleiotropic signals is the major field of activity of the protein kinase C (PKC) family. Besides the classic (␣, ␤I, ␤II, ␥) and the novel (␦, , , ) members the PKC family comprises two atypical members (, ͞) that are distinguished structurally by the presence of only a single PKC zinc finger module in their regulatory domain and biochemically by their inability to bind and to respond to phorbol esters and diacylglycerol (1-3). Most cells and tissues express several PKC enzymes suggesting that the members of this family do not have overlapping functions (4). While the classic and novel members are expected to participate in signal transduction from cell surface receptors that trigger the generation of diacylgycerol by activating phospholipases C the mode of activation and the function of the atypical members is much less clear (5, 6). The search for a function has connected PKC-with two distinct signaling pathways as the most attractive sites of action of this enzyme. First, the finding that phosphatidylinositol-3,4,5-trisphosphate can activate PKC-in vitro pointed to the possibility that PKC-participates in phosphorylation events downstream of phosphatidylinositol 3-kinase activation by receptor tyrosine kinases (7). Secondly, PKC-has been proposed as a mediator of the growth inhibitory and apoptotic actions of ceramide, an intracellular messenger generated by hydrolysis of sphingolipids (8). Evidence for effects of the sphingomyelin cycle on PKC-activity, specifically in tumor necrosis factor ␣ signaling, has been presented recently (9-11).In addition, overexpression of PKC-has been shown to be necessary and sufficient to deregulate growth control in mouse fibroblasts supporting a crucial role for PKC-in ras-induced mitogenic signaling (12), albeit conflicting data have been re...

No abstract

Cdc42, a Rho-family GTPase, has been implicated in several signal transduction pathways, including organization of the actin cytoskeleton, activation of the c-Jun N-terminal MAP kinase (JNK) and stimulation of the nuclear transcription factor kappa B (NF(kappa)B). We report here that exposure of fibroblasts to the inflammatory cytokines tumor necrosis factor (alpha) (TNF(alpha)) and interleukin-1 (IL-1) triggers the activation of Cdc42 leading first to filopodia formation and subsequently to Rac and Rho activation. Inhibition of Cdc42 completely suppresses cytokine-induced actin polymerization, but not activation of JNK or NF(kappa)B. The latent membrane protein 1 of Epstein-Barr virus, LMP1, is thought to mimic constitutively activated TNF family receptors. When expressed in fibroblasts, LMP1 stimulates Cdc42-dependent filopodia formation as well as JNK and NF(kappa)B activation. Using LMP1 mutants, we show that activation of Cdc42 and JNK/NF(kappa)B occur through distinct pathways and that Cdc42 activation is independent of LMP1's interaction with TRADD and TRAF proteins.