Glycogen synthase kinase-3 (GSK3) is implicated in the regulation of several physiological processes, including the control of glycogen and protein synthesis by insulin, modulation of the transcription factors AP-1 and CREB, the specification of cell fate in Drosophila and dorsoventral patterning in Xenopus embryos. GSK3 is inhibited by serine phosphorylation in response to insulin or growth factors and in vitro by either MAP kinase-activated protein (MAPKAP) kinase-1 (also known as p90rsk) or p70 ribosomal S6 kinase (p70S6k). Here we show, however, that agents which prevent the activation of both MAPKAP kinase-1 and p70S6k by insulin in vivo do not block the phosphorylation and inhibition of GSK3. Another insulin-stimulated protein kinase inactivates GSK3 under these conditions, and we demonstrate that it is the product of the proto-oncogene protein kinase B (PKB, also known as Akt/RAC). Like the inhibition of GSK3 (refs 10, 14), the activation of PKB is prevented by inhibitors of phosphatidylinositol (PI) 3-kinase.

Insulin activated endogenous protein kinase B alpha (also known as RAC/Akt kinase) activity 12‐fold in L6 myotubes, while after transfection into 293 cells PKBalpha was activated 20‐ and 50‐fold in response to insulin and IGF‐1 respectively. In both cells, the activation of PKBalpha was accompanied by its phosphorylation at Thr308 and Ser473 and, like activation, phosphorylation of both of these residues was prevented by the phosphatidylinositol 3‐kinase inhibitor wortmannin. Thr308 and/or Ser473 were mutated to Ala or Asp and activities of mutant PKBalpha molecules were analysed after transfection into 293 cells. The activity of wild‐type and mutant PKBalpha was also measured in vitro after stoichiometric phosphorylation of Ser473 by MAPKAP kinase‐2. These experiments demonstrated that activation of PKBalpha by insulin or insulin‐like growth factor‐1 (IGF‐1) results from phosphorylation of both Thr308 and Ser473, that phosphorylation of both residues is critical to generate a high level of PKBalpha activity and that the phosphorylation of Thr308 in vivo is not dependent on phosphorylation of Ser473 or vice versa. We propose a model whereby PKBalpha becomes phosphorylated and activated in insulin/IGF‐1‐stimulated cells by an upstream kinase(s).

Stimulation of cells with growth factors and cytokines, or exposure to cellular stresses, activates several signal transduction pathways that have specific physiological roles. These include at least three in which a mitogen-activated protein kinase (MAPK) 1 homologue is involved. In one pathway, cell stimulation leads to the sequential activation of p21 ras and the protein kinases c-Raf, MAP kinase kinase-1 and -2 (MAPKK1, MAPKK2), and p42 and p44 MAP kinases (p42 MAPK , p44 MAPK ). These MAPKs phosphorylate a variety of proteins in vivo including MAP kinase-activated protein (MAPKAP) kinases 1␣ and 1␤ (also known as Rsk-1 and Rsk-2 (1)). The sustained activation of p42/p44MAPK is not only required, but it is sufficient to induce the proliferation or differentiation of several cells (2).In order to dissect MAPK pathways and to elucidate their physiological roles, one approach has been to generate dominant negative mutants and overexpress them in cells. For example, dominant negative forms of p21 ras , c-Raf, and MAPKK1 all inhibit the activation of p42/p44 MAPK and the growth factor-induced proliferation or differentiation of several cells (3). However, although dominant-negative mutants are useful, the generation of cell lines that stably express them is time consuming, and their expression may lead to erroneous conclusions. For example, overexpression of an inactive form of MAPKK1 that can be phosphorylated by Raf may not only prevent the activation of endogenous wild-type MAPKK1, but also the activation of other cellular substrates of Raf that might lie in distinct signaling pathways. The need to express a dominant negative mutant for many hours may also result in unwanted secondary effects. Similarly, the use of dominant negative mutants of Raf may affect Ras-dependent processes that are independent of Raf.An alternative strategy is to identify small cell-permeant molecules that are specific inhibitors of particular protein kinases. An advantage of this approach is that the effects of these inhibitors can be investigated in any cell in vivo. Moreover, these inhibitors may have therapeutic potential as anti-cancer, or anti-inflammatory agents, or as immunosuppressants. Several such inhibitors have recently been described, including an inhibitor of the epidermal growth factor (EGF) receptor tyrosine kinase (4), which may be useful for treating human tumors that overexpress this receptor, and a specific inhibitor of the MAP kinase homologue termed reactivating kinase (RK) or p38 (5). The latter inhibitor prevents the synthesis of interleukin-1