The tumor suppressor PTEN is a phosphatidylinositol phospholipid phosphatase, which indirectly down-regulates the activity of the protein kinase B/Akt survival kinases. Examination of sequence data bases revealed the existence of a highly conserved homologue of PTEN. This homologue, termed PTEN 2, contained an extended amino-terminal domain having four potential transmembrane motifs, a lipid phosphatase domain, and a potential lipid-binding C2 domain. Transcript analysis demonstrated that PTEN 2 is expressed only in testis and specifically in secondary spermatocytes. In contrast to PTEN, PTEN 2 was localized to the Golgi apparatus via the amino-terminal membrane-spanning regions. Molecular modeling suggested that PTEN 2 is a phospholipid phosphatase with potential specificity for the phosphate at the 3 position of inositol phosphates. Enzymatic analysis of PTEN 2 revealed substrate specificity that is similar to PTEN, with a preference for the dephosphorylation of the phosphatidylinositol 3,5-phosphate phospholipid, a known mediator of vesicular trafficking. Together, these data suggest that PTEN 2 is a Golgi-localized, testis-specific phospholipid phosphatase, which may contribute to the terminal stages of spermatocyte differentiation.The production of 3-phosphorylated phosphatidylinositol lipid products by the PI3K 1 pathway is an important control point for the regulation of cell proliferation, growth, survival, and vesicular trafficking (1). The activation of this pathway by various growth factors, extracellular matrices, or oncogenic events results in a diversity of signals, including the up-regulation of the catalytic activity of the Akt/PKB kinases (2). These kinases enhance cell survival by phosphorylation of a number of substrates, including a subfamily of forkhead transcription factors (3). A novel mechanism for the control of the Akt/PKB pathway was identified when genetic evidence pointed to a tumor suppressor locus on chromosome 10 at q23-25. Analysis of a candidate tumor suppressor gene from this region demonstrated that the locus encoded a phosphatase, which was termed PTEN (also called MMAC and TEP) (1, 4 -6). Further studies demonstrated that PTEN was mutated in a large percentage of brain, endometrial, and prostate tumors as well as a smaller percentage of other tumors (7-9). In addition, Cowden disease and Bannayan-Zonana syndrome, which are both characterized by increased susceptibility to breast and thyroid tumors, showed a range of germline PTEN mutations, which were similar to those observed in tumors (10). Enzymatic studies demonstrated that PTEN is a lipid phosphatase, which down-regulates the PI3K pathway by removing the 3-phosphate from the phosphatidylinositol 3,4(and 3,4,5)-phosphate phospholipids (PIP 3,4 and PIP 3,4,5 ) (11). Importantly, many of the tumor-derived missense mutations observed in PTEN resulted in a complete loss of phospholipid phosphatase catalytic activity (12). The loss of the PTEN lipid phosphates activity due to mutation was expected to result in increased level...