The class I phosphoinositide 3-kinases (PI3K) are a family of lipid kinases that play a central role in an evolutionarily conserved signal transduction pathway/network of which PTEN is part (Hawkins et al. 2006). Their tightly controlled activity phosphorylates the relatively abundant phosphoinositide PtdIns(4,5)P2 on the D3 position of the inositol ring, to form the lipid second messenger PtdIns(3,4,5)P3 (Fig. 1). Class II and Class III PI3K enzymes also exist in mammalian cells, but although these phosphoinositide kinases also phosphorylate the D3 position of the inositol ring, they act upon different substrates and will not be further considered here.
The class I PI3Ks exist in cells as heterodimers, comprising a catalytic subunit of 110 kD in size and a regulatory subunit, which ranges in size from approximately 55 kD to 100 kD. The human and mouse genomes each encode four different catalytic p110 subunits, which display some differences in their expression patterns, regulatory partners, and signaling inputs (Hawkins et al. 2006). The p110 α, β, and δ enzymes bind with regulatory subunits, most commonly of 85 kD in size, that contain paired SH2 domains, and result in activation of PI3K activity through tyrosine kinase–based signaling mechanisms, in particular via receptor tyrosine kinase (RTK) signaling. In contrast, the p110 γ PI3K catalytic subunit binds to one of two unrelated regulatory subunits, p84 or p101, that allow regulation of kinase activity through G-protein coupled receptor (GPCR)-based signaling. It has been shown more recently that the p110 β isoform is also activated by many GPCRs, although the mechanism of this activation is currently unclear. Although these differences exist in the inputs into the PI3Ks via their regulatory subunits, each of the four catalytic subunits has a binding site for the activated form of the RAS small GTPases. Thus, PI3K activity appears to be tightly controlled, with low activity in the absence of stimulation that can be provided through mechanisms that include RTK, GPCR, and RAS signaling (Vanhaesebroeck et al. 2010).
The direct PI3K lipid product PtdIns(3,4,5)P3 propagates the effects of PI3K activation through a large and diverse group of effector proteins that are able to bind directly to this lipid with high selectivity (Fig. 1). The best studied downstream targets for PI3K/PIP3 activation are the AKT family of serine/threonine kinases that have an N-terminal PIP3 binding Pleckstrin Homology (PH) domain and a C-terminal kinase domain. Experiments in flies and worms have shown that the AKT kinases are evolutionarily conserved and key functional mediators of PI3K signaling. The three human AKT kinases have many substrates and are important regulators of cell growth, proliferation, survival, and metabolism (Manning and Cantley 2007). However, estimates are that there are between 25 and 50 selective PIP3 binding proteins encoded in the human genome and in many cases the relative contributions of these proteins to the PI3K signaling network is unclear (Hawkins et al. 2006).
Once synthesized by PI3K, the PIP3 signal can be removed either by one of two routes of metabolism. A family of phosphoinositide 5-phosphatases, the best characterized being the SHIP and SHIP-2 enzymes, dephosphorylate the 5-position of PIP3 to form the alternate phosphoinositide signal PtdIns(3,4)P2, which shares some selective binding partners with PIP3, including the AKT kinases, and is believed to contribute directly to the outcomes of PI3K activation (Fig. 1). PtdIns(3,4)P2 in turn appears to be removed at least in part, by the action of the INPP4 phosphoinositide 4-phosphatase enzymes.
The other main route of metabolism of PIP3 is by removal of its 3-phosphate, acting to terminate PI3K signaling by reproducing PtdIns(4,5)P2. It is this reaction that is performed by PTEN and which forms the basis for its tumor suppressor activity.
The PTEN Protein
In almost all cell types investigated, PTEN is found located throughout the cytosol, and in many cells also in the nucleus. The nuclear localization of PTEN, which appears to be regulated in part by monoubiquitination, will be discussed with more detail below. In a very few tissues, strong subcellular enrichment of PTEN has been demonstrated, most notably localization to apical junctions in the polarized epithelia of the chick epiblast and murine retinal pigment epithelium. However, the reported clear subcellular localization of PTEN with strong validating controls is very rare.
To access its membrane substrate, PTEN associates transiently with the plasma membrane, with an estimated average residence time of a few hundred milliseconds, but any membrane enrichment is not normally evident using normal epifluorescence microscopy. Basic surfaces on the phosphatase domain and the calcium-independent C2 domain appear to be involved in mediating interactions with acidic non-substrate membrane lipids and correctly positioning the active site relative to membrane surfaces. There also appears to be a specific PtdIns(4,5)P2 interacting motif at the N-terminus of PTEN that mediates a conformational change within the phosphatase and leads to activation. With emerging evidence for rapid regulated changes in the electrostatic surface charge of cellular membranes and PtdIns(4,5)P2 levels, it will be interesting to see whether such changes play an important role in PTEN regulation.
Several mechanisms of PTEN regulation have been identified, but it seems worth noting in advance that most of the mechanisms of PTEN regulation are through inhibition of the enzyme. Accordingly, unphosphorylated bacterially expressed PTEN has robust lipid phosphatase activity in vitro that can be inhibited by such mechanisms as phosphorylation, oxidation, and ubiquitination.
PTEN has two well-documented groups of phosphorylation sites: a cluster of serines and threonines surrounded by acidic residues between amino acids 380 and 385 that appear to be phosphorylated by the protein kinase, CK2, and also threonine 366, which seems to be phosphorylated by Glycogen Synthase Kinase 3 (GSK3) after priming phosphorylation of serine 370 by CK2 (Fig. 2).
It has been shown by many studies that phosphorylation of the 380–385 cluster sites stabilizes a “closed” conformation of PTEN due to a strengthened interaction between the flexible C-terminal tail of PTEN and the phosphatase and C2 domains (Ross and Gericke 2009). This closed conformation interferes with membrane binding, so reduces lipid phosphatase activity, but also enhances PTEN stability, it seems as an indirect consequence of this reduced membrane localization (Maccario et al. 2010). It should be noted that although a small fraction of cellular PTEN has been identified in an unphosphorylated state (at these 380–385 sites) and incorporated in high molecular weight complexes, most cellular PTEN appears to be found in the phosphorylated “closed” state correlating with its cytosolic localization. A similar picture emerges in studies of the 366/370 phosphorylation sites, which have been shown capable of affecting PTEN stability in some, but not all, cells, and also controlling the phosphospecific binding of a nucleolar protein, MSP58 to PTEN. However, as with the 380–385 sites, although at least a little is known about the consequences of phosphorylation, there is no clear picture of how regulated changes in phosphorylation fit into any system of controlling PTEN function.
The catalytic mechanism of PTEN, shared with the rest of the PTP family, makes it potentially sensitive to regulation by oxidation of the reactive catalytic cysteine. The oxidation of a fraction of cellular PTEN occurs in several circumstances in which cells stimulated by growth factors and other receptor ligands are driven to produce endogenous reactive oxygen (ROS) species and is accompanied by a ROS-dependent activation of downstream signaling (Leslie et al. 2008). Although studies in both cardiac and skeletal muscle have found a correlation between PTEN oxidation and Akt activation in vivo, as with many areas of redox signaling, compelling evidence for the functional significance of any observed oxidation is currently lacking (Leslie et al. 2008).
PTEN, like many proteins, has its degradation controlled in large part through ubiquitination and subsequent proteolysis by the proteasome. Two lysine residues have been proposed to be sites for ubiquitination, Lys13 and Lys289, located within the N-terminal PtdIns(4,5)P2 binding site and within the C2 domain, respectively (Fig. 2), and two different E3 ubiquitin ligases have been proposed to act on PTEN, the HECT domain ligase NEDD4.1, and the RING domain protein XIAP. PTEN ubiquitination, and in particular, monoubiquitination of Lys289, appears to enhance nuclear localization (Salmena et al. 2008). Significantly, ubiquitination of PTEN in vitro leads to a substantial loss of phosphatase activity, indicating that ubiquitination may act to directly inhibit catalytic activity, in addition to other effects on PTEN function (Maccario et al. 2010). Given the current data, it seems appealing to speculate that there may be significant complexity in the area of PTEN ubiquitination and that individual ubiquitination events on multiple sites, mediated by different regulated ubiquitin ligases, may have distinct functional consequences, with ubiquitination differentially controlling activity, stability, and nuclear localization. The functional consequences of nuclear localization are currently rather unclear, with several PtdInsP3-independent nuclear functions having been proposed, along with a simple model that nuclear PTEN may be a stable protein pool that cannot access its membrane substrate.
PTEN Functions in Health and Disease: The Significance of PTEN as a Tumor Suppressor
As discussed above, PTEN is a core component and critical regulator of a signaling pathway that controls many diverse cellular processes. As such, PTEN plays a role in a vast array of physiological processes such as neuronal development and function, central control of adiposity, and cardiac hypertrophy and contractility (Chang et al. 2007; Oudit and Penninger 2009; Plum et al. 2006). However, by far the best recognized role for PTEN in human disease is as an important tumor suppressor in many forms of cancer. Analyses of the PTEN coding regions in tumor samples indicate that it is the second most frequently mutated tumor suppressor across all cancers after p53 (Salmena et al. 2008; Keniry and Parsons 2008). It is also very common that tumors carry deletions of one, or less frequently, both copies of the PTEN gene. The pathological significance of loss of the PTEN gene in different tumor types, including the loss of just one allele, has been supported by the generation of many different lines of transgenic mice lacking PTEN, throughout the body or in selected tissues, most of which have greatly increased risks of tumor formation (Salmena et al. 2008; Suzuki et al. 2008). It is outside the scope of this entry to describe the wealth of research over the last decade or more that has provided increasing evidence for the significance and mechanisms of PTEN functional loss in cancer, a deeper understanding of the downstream pathways through which PTEN loss affects tumor development and has described the development of therapies aimed at treating the many tumors lacking PTEN. However, these topics have been recently reviewed by others (Keniry and Parsons 2008; Chalhoub and Baker 2009; Garcia-Echeverria and Sellers 2008).
PTEN’s status as one of the most important tumor suppressors in human cancer has driven a great deal of research into its function. In particular, the discovery that PTEN appears to act in a dose-dependent, haploinsufficient manner as a tumor suppressor has broad implications for the potential importance in cancer of pathways that regulate PTEN activity and expression. Also, great efforts are going into understanding what can be learned from a tumor’s PTEN status in terms of predicting outcome and response to treatments. It is to be hoped that our developing understanding of PTEN function and the increasing number of drugs targeting the PI3K pathway will lead to an improvement in the outcome for this patient group.