Reply to Agarwal: Activity against nuclear substrates is not necessarily mediated by nuclear Akt

The phosphorylation of Akt substrates in subcellular compartments other than the plasma membrane has previously been proposed to be mediated by the diffusion of activated Akt (1, 2). While this neatly accounts for the observation that Akt is activated by growth factors primarily at the plasma membrane, it poses problems for the cell in controlling the spatial and temporal dynamics of Akt substrate phosphorylation, since Akt would essentially be uncoupled from its activating stimulus. We recently showed that Akt activity is in fact confined to membranes enriched in either PI(3,4,5)P₃ or PI(3,4)P₂ (3, 4). By restricting Akt activity to the engagement of its activating lipid second messengers, both spatial and temporal control of Akt signaling is not lost. However, Agarwal (5) raises the question of how so-called “nuclear substrates” of Akt can be phosphorylated under these conditions. While more than 200 substrates of Akt have been identified, few have been carefully validated as bona fide Akt substrates and none that have been is exclusively nuclear. The FOXO family of transcription factors are, however, well-established Akt substrates, and insulin signaling via Akt2, in particular, leads to the termination of transcriptional programs for gluconeogenesis. Inactivation of Akt2 leads to insulin resistance and a severe form of inherited diabetes in humans (6). Phosphorylation of FOXO by Akt regulates its subcellular localization: unphosphorylated FOXO resides in the nucleus, where it drives transcription, while phosphorylated FOXO is sequestered in the cytoplasm by 14-3-3 proteins (7). The phosphorylation of FOXO by Akt may thus occur in the nucleus, thereby driving its nuclear export, or in the cytoplasm, thereby promoting its sequestration. Although early studies proposed the phosphorylation of FOXO by Akt in the nucleus (8), evidence for nuclear pools of PI(3,4,5)P₃ or PI(3,4)P₂ is lacking. This contrasts with the now-clear evidence for endomembrane pools of PI(3,4,5)P₃ and PI(3,4)P₂ that drive Akt signaling (3, 9). Furthermore, while Akt readily forms a complex with a model pseudosubstrate on endomembranes in the cytoplasm, we could not observe complex formation in the nucleus (3). Phosphorylation of a small pool of unphosphorylated FOXO in the cytoplasm and its sequestration by 14-3-3 proteins could drive nuclear exclusion simply by shifting the nucleo-cytoplasmic equilibrium. In general, the phosphorylation of substrates that are freely diffusible is entirely compatible with membrane-restricted Akt activity, while the phosphorylation of membrane-bound substrates such as TSC2 can be readily accounted for by the membrane trafficking of both PI(3,4,5)P₃ and PI(3,4)P₂, which results from the internalization of membrane receptors and their associated PI3K activity (10). In such cases, reactivation of Akt by PDK1 and mTOCR2 would not be necessary, since membrane-bound Akt is protected from dephosphorylation (4). Alternatively, Akt may be activated by phosphorylation on subcellular membranes distinct from the plasma membrane (9, 11). What is clear, however, is that phosphorylation of Akt alone does not relieve the steric occlusion of the substrate binding cleft mediated by its PH domain in the absence of PI(3,4,5)P₃ or PI(3,4)P₂ (4).