Phosphorylation keeps PTEN phosphatase closed for business.

P hosphatase and tensin homologue deleted on chromosome 10 (PTEN) is a phosphatase that suppresses many tumor types (1, 2). It plays a major role in the development of the nervous system and has been implicated in diseases such as autism (3). Despite its importance, we are only beginning to understand how this important protein is regulated. In 2000 and 2001, Vazquez et al. (4, 5) with great insight proposed that phosphorylation of the C terminus induces PTEN to assume a closed conformation with an inactive phosphatase domain. Upon dephosphorylation, PTEN assumes an open conformation, activating the phosphatase domain and enhancing degradation of PTEN protein. In this view, the phosphorylated closed form acts like a proenzyme, which is stable in the cytoplasm but ready for rapid activation, use, and degradation. Although not fully proven, this model (Fig. 1) has greatly influenced the PTEN field. The study in a recent issue of PNAS by Rahdar et al. (6) provides convincing proof for this model and new exciting insights into the interactions of PTEN with biological membranes. The PTEN protein is a phosphatidylinositol phosphate (PIP) phosphatase specific for the 3-position of the inositol ring (7). Although PTEN can dephosphorylate PI(3)P, PI(3,4)P2, or PI(3,4,5)P3, it is likely that PI(3,4,5)P3 is the most important substrate in vivo (8). The balance between PTEN and phosphoinositide 3-kinase (PI3K) controls PI(3,4,5)P3 basal levels in the plasma membrane, which in turn, regulates cell survival and proliferation. PTEN s action on PI(3,4,5)P3 requires membrane binding. PTEN can bind to a limited extent to zwitterionic phosphatidylcholine bilayers (9, 10). However, it has significantly stronger interactions with negatively charged lipids. The C2 domain binds phosphatidylserine, which is abundant in the inner leaflet of mammalian plasma membranes (9). A short N-terminal module binds selectively PI(4,5)P2 [PBM, PI(4,5)P2-binding module], which initiates a second conformational change that activates the PTEN phosphatase domain (11, 12). Finally, there is a C-terminus sequence that binds to PDZ domains. It should be noted that for several other interfacial enzymes, binding to membranes via multiple domains leads to enzymatic activation (13). Given these multiple PTEN-membrane contact points, it was surprising that most of PTEN is cytoplasmic. Detecting membrane-bound PTEN requires sophisticated microscopy (14). Using biochemical techniques, membranebound PTEN is detected only if very mild detergents are used to permeabilize the cells (15). The open/closed model provides a plausible explanation for the paucity of PTEN at the plasma membrane. In the closed conformation, the phosphorylated tail binds to the phosphatase domain as a pseudosubstrate and to the C2 domain (15), preventing membrane binding. According to this model, dephosphorylation of PTEN s tail leads to a release of the tail, allowing for an initial membrane association most likely through nonspecific electrostatic interactions with the negatively charged membrane, followed by selective interactions between the N-terminal PBM and PI(4,5)P2, and the C2 domain and phosphatidylserine. The PBM/PI(4,5)P2 interaction activates the phosphatase domain (11, 12, 16, 17). Consistent with this model it was found that PTEN s tail is sensitive to nonspecific proteases (9) and dephosphorylation enhances binding of the PTEN tail to PDZ domains (4). Despite the plausibility of this model, the key feature of the model (interactions between the phosphorylated tail and the remainder of the protein) is just now being assessed. Rahdar et al. (6) tested many predictions based on the open/closed model. Consistent with earlier studies, they found that PTEN 1-351 lacking the Cterminal tail bound more strongly to the plasma membrane than full-length PTEN. PTEN-A4, a mutant PTEN with