PTEN 2, a Golgi-associated Testis-specific Homologue of the PTEN Tumor Suppressor Lipid Phosphatase

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 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. phosphatidylinositol 3-kinase protein kinase B phosphatidylinositol 3-phosphate phospholipid 4, phosphatidylinositol 3,4-phosphate phospholipid 4,5, phosphatidylinositol 3,4,5-phosphate phospholipid membrane-associated guanylate kinase glutathioneS-transferase green fluorescence protein kilobase(s) base pair(s) dithiothreitol PSD95, discs large, Z01 domain phosphatase with tensin homology. The production of 3-phosphorylated phosphatidylinositol lipid products by the PI3K1 pathway is an important control point for the regulation of cell proliferation, growth, survival, and vesicular trafficking (1Toker, A., and Cantley, L. (1997) 387, 673–676.Google Scholar). 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 (2Chan T. Rittenhouse S. Tsichlis P. Annu. Rev. Biochem. 1999; 68: 965-1014Crossref PubMed Scopus (876) Google Scholar). These kinases enhance cell survival by phosphorylation of a number of substrates, including a subfamily of forkhead transcription factors (3Brunet A. Bonni A. Zigmond M.J. Lin M.Z. Juo P. Hu L.S. Anderson M.J. Arden K.C. Blenis J. Greenberg M.E. Cell. 1999; 96: 857-868Abstract Full Text Full Text PDF PubMed Scopus (5434) Google Scholar). 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) (1Toker, A., and Cantley, L. (1997) 387, 673–676.Google Scholar, 4Li J. Yen C. Liaw D. Podsypanina K. Bose S. Wang S.I. Puc J. Miliaresis C. Rodgers L. McCombie R. Bigner S.H. Giovanella B.C. Ittmann M. Tycko B. Hibshoosh H. Wigler M.H. Parsons R. Science. 1997; 275: 1943-1947Crossref PubMed Scopus (4285) Google Scholar, 5Steck P. Pershouse M.A. Jasser S.A. Yung W.K.A. Lin H. Ligon A.H. Langford L.A. Baumgard M.L. Hattier T. Davis T. Frye C. Hu R. Swedlund B. Teng D.H. Tavtigian S.V. Nat. Genet. 1997; 15: 356-362Crossref PubMed Scopus (2517) Google Scholar, 6Li D.M. Sun H. Cancer Res. 1997; 57: 2124-2129PubMed Google Scholar). 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 (7Cairns P. Okami K. Halachmi S. Halachmi N. Esteller M. Herman J.G. Isaacs W.B. Bova G.S. Sidransky D. Cancer Res. 1997; 57: 4997-5000PubMed Google Scholar, 8Rasheed B.K.A. Stenzel T.T. McLendon R.E. Parsons R. Friedman A.H. Friedman H.S. Bigner D.D. Bigner S.H. Cancer Res. 1997; 57: 4187-4190PubMed Google Scholar, 9Tashiro H. Blazes M.S. Wu R. Cho K.R. Bose S. Wang S.I. Li J. Parsons R. Ellenson L.H. Cancer Res. 1997; 57: 3935-3940PubMed Google Scholar). 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 (10Eng C. Peacocke M. Nat. Genet. 1998; 19: 223Crossref PubMed Scopus (73) Google Scholar). 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 (PIP3,4 and PIP3,4,5) (11Maehama T. Dixon J.E. J. Biol. Chem. 1998; 273: 13375-13378Abstract Full Text Full Text PDF PubMed Scopus (2601) Google Scholar). Importantly, many of the tumor-derived missense mutations observed in PTEN resulted in a complete loss of phospholipid phosphatase catalytic activity (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar). The loss of the PTEN lipid phosphates activity due to mutation was expected to result in increased levels of PIP3,4 and PIP3,4,5 and the up-regulation of the Akt/PKB cell proliferation/survival pathway, an event that might induce tumor resistance to chemotherapy and radiation (13Haas-Kogan D. Shalev N. Wong M. Mills G. Yount G. Stokoe D. Curr. Biol. 1998; 8: 1195-1198Abstract Full Text Full Text PDF PubMed Google Scholar, 14Whang Y.E. Wu X. Suzuki H. Reiter R.E. Tran C. Vessella R.L. Said J.W. Isaacs W.B. Sawyers C.L. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 5246-5250Crossref PubMed Scopus (563) Google Scholar, 15Wu X. Senechal K. Neshat M.S. Whang Y.E. Sawyers C.L. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 15587-15591Crossref PubMed Scopus (595) Google Scholar). Data supporting this conjecture demonstrated that glioblastoma cell lines mutated for the endogenous PTENlocus suffered deleterious effects on cell cycle progression (G1 arrest), proliferative capacity, and survival when transfected with a wild type but not a catalytically inactive, form of the phosphatase, suggesting that the enzymatic activity of the enzyme was involved with the regulation of this phospholipid signaling (6Li D.M. Sun H. Cancer Res. 1997; 57: 2124-2129PubMed Google Scholar, 16Furnari F. Lin H. Huang H. Cavanee W. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12479-12484Crossref PubMed Scopus (382) Google Scholar, 17Cheney I.W. Johnson D.E. Vaillancourt M.T. Avanzini J. Morimoto A. Demers G.W. Wills K.N. Shabram P.W. Bolen J.B. Tavtigian S.V. Bookstein R. Cancer Res. 1998; 58: 2331-2334PubMed Google Scholar). These studies suggested that the loss of this phosphatase in tumors induced the up-regulation of the Akt/PKB signaling pathway, which resulted in cell cycle progression and inhibition of apoptosis. A number of animal model studies supported an important role for PTEN in the control of proliferation, survival, and cell size. Importantly, mice with homozygous null mutations in the PTEN locus showed early embryonic lethality due to an apparent hyperproliferative effect, whereas heterozygous animals developed tumors postnatally with apparent loss of heterozygosity at the PTEN locus (18Dicristofano A. Pesce B. Cordoncardo C. Pandolfi P.P. Nat. Genet. 1998; 19: 348-355Crossref PubMed Scopus (1305) Google Scholar, 19Stambolic V. Suzuki A. Delapompa J.L. Brothers G.M. Mirtsos C. Sasaki T. Ruland J. Penninger J.M. Siderovski D.P. Mak T.W. Cell. 1998; 95: 29-39Abstract Full Text Full Text PDF PubMed Scopus (2107) Google Scholar). This latter result suggested that the loss of PTEN expression was an advantageous event, which allowed tumors to grow in a more unregulated manner after the accumulation of other oncogenic mutations. Cell lines derived from PTEN null embryonic mice demonstrated higher levels of PIP3 phospholipids and enhanced activation of the Akt/PKB kinase (19Stambolic V. Suzuki A. Delapompa J.L. Brothers G.M. Mirtsos C. Sasaki T. Ruland J. Penninger J.M. Siderovski D.P. Mak T.W. Cell. 1998; 95: 29-39Abstract Full Text Full Text PDF PubMed Scopus (2107) Google Scholar). These cells also showed significant resistance to a diversity of apoptotic stimuli, further endorsing a role for this phosphatase in the regulation of cell survival through the PI3K pathway. Additional evidence in both Caenorhabditis elegansand Drosophila supported a role for the PTEN phosphatase in the regulation of cell growth and survival. C. eleganscontains an insulin-like pathway, including an insulin-like receptor tyrosine kinase, a PI3K, PDK, and Akt/PKB kinases and a forkhead-like transcription factor, which is involved with dauer formation, a developmental stage where worms undergo a quiescent state. Genetic analysis of this pathway demonstrated that a worm homologue ofPTEN, termed DAF 18, could suppress upstream mutations in either the insulin-like receptor or PI3K, completely consistent with results found in the mammalian PI3K pathway (20Ogg S. Ruvkun G. Mol. Cell. 1998; 2: 887-893Abstract Full Text Full Text PDF PubMed Scopus (330) Google Scholar). The involvement of this pathway in the regulation cell size was further suggested by studies in Drosophila. These data demonstrated that the fly homologue of PTEN was involved with the determination of cell size, consistent with other studies, which established the importance of several components of the PI3K pathway, such as Akt/PKB (21Goberdhan D. Parico N. Goodman E. Mlodzik M. Wilson C. Genes Dev. 1999; 13: 3244-3258Crossref PubMed Scopus (293) Google Scholar, 22Gao X. Neufeld T. Pan D. Dev. Biol. 2000; 221: 404-418Crossref PubMed Scopus (220) Google Scholar). Together, these three separate animal models provided strong proof for the relevance of the PTEN lipid phosphatase in the regulation of various aspects of the PI3K pathway. X-ray crystallographic analysis of PTEN structure revealed that this phosphatase contains a novel substrate recognition pocket with positively charged residues potentially involved with the association of the phosphates on the inositol ring substrate (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar). Positioning of many tumor-derived mutations known to disrupt catalytic activity to the active site in part served to explain the mechanism of action of the phosphatase. The mechanism by which PTEN appears to be associated with its phospholipid substrates appears to be quite complex. Structural analysis revealed a functional C2 lipid binding domain in the carboxyl-terminal region of the protein, which was proposed to serve as a lipid association motif (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar). Many tumor-derived mutations have been mapped to the carboxyl terminus, and a fraction of these are involved with the formation of an interface between the phosphatase domain and the C2 domain. These latter results helped to explain why mutations in the carboxyl-terminal region appeared to affect catalytic activity. Human, mouse, and Drosophila PTEN all contain a PDZ binding motif ((S/T)XV) at their carboxyl termini, and yeast two-hybrid analyses established that the PTEN phosphatase binds to PDZ domains of a family of membrane-associated guanylate kinases (MAGUKs), peripheral membrane-associated proteins with multiple protein interaction domains, which function to juxtapose signaling molecules and position them near the plasma membrane (23Wu Y. Dowbenko D. Spencer S. Laura R. Lee J. Gu Q. Lasky L.A. J. Biol. Chem. 2000; 275: 21477-21485Abstract Full Text Full Text PDF PubMed Scopus (232) Google Scholar). Interestingly, some tumor-derived mutations are found at the extreme carboxyl terminus of PTEN, and these mutations would be expected to disrupt PDZ domain binding interactions, consistent with an important functional role for PDZ domain binding in PTEN tumor suppression. Because these MAGUKs are localized specifically to intercellular tight junction regions, these studies also suggested mechanisms for the positioning of the PTEN phosphatase to the lipid domains of subcellular regions such as the epithelial tight junction, a site known to be involved with the regulation of cell survival (24Watton S. Downward J. Curr. Biol. 1999; 9: 433-436Abstract Full Text Full Text PDF PubMed Scopus (264) Google Scholar). In this paper we describe a novel homologue of PTEN, termed PTEN 2, which has been identified using data base searches. Interestingly, this novel phosphatase is expressed uniquely in testis, and specifically in the secondary spermatocytes. In contrast to PTEN, PTEN 2 contains several potential transmembrane domains, which appear to target the phosphatase to the Golgi apparatus. Molecular modeling suggests that PTEN 2 is a lipid phosphatase with many active site residues conserved with PTEN, and enzymatic analysis demonstrates that the novel phosphatase actively dephosphorylates PIP3,5 and PIP3,4,5 in vitro. Together, these data suggest that PTEN 2 is a phospholipid phosphatase, which may play a role in the terminal stages of spermatocyte maturation by regulating intracellular levels of phosphatidylinositol 3-phosphate phospholipids. A human expressed sequence tag related to PTEN was initially identified from searches of public and private data bases. By using various protocols, including screening a testis cDNA library and PCR methods, we obtained a full-length sequence of human PTEN 2. A fragment of human PTEN 2 cDNA was used as a probe to isolate a full-length murine cDNA from a murine testis cDNA library. GST-PTEN 2 (amino acids 378–683) were subcloned into an expression vector containing a cytomegalovirus promoter. A catalytically inactive form of murine GST-PTEN 2 was constructed by changing Cys458 to Ser. To determine the localization of PTEN 2 in mammalian cells, a myc tag was placed at the carboxyl terminus of the gene. GFP-PTEN 2 was constructed by subcloning either the full-length cDNA or the amino or carboxyl terminus intoCLONTECH's pEGFPN3 vector. The PTEN 2 amino terminus includes amino acids 1–377, whereas the PTEN 2 carboxyl terminus includes amino acids 378–683. Fluorescence in situ hybridization mapping of mouse murine PTEN 2 was performed by SeeDNA Biotech Inc. The probe was an 8-kb BamHI genomic DNA containing the first exon of PTEN 2. For the molecular modeling, a sequence alignment was obtained by using ClustalW and a threading approach. The HOMOLOGY/MODELLER module from the Insight II package (version 98.0, MSI, San Diego, CA) was used for molecular construction and display. Docking of inositol (1,3,4,5)-tetrakisphosphate in the active site of PTEN 2 was manually performed as previously described for PTEN by Lee and co-workers (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar). In addition to Northern analysis, we also determined the tissue distribution pattern of PTEN 2 using the PCR method. Mouse Multiple Tissue cDNA panels were purchased fromCLONTECH. After 35 cycles, the mouse PTEN 2 gene was only detected in testis. PCR primers (upper 5′-GAACTGGAACCATGGTG and lower 5′-TAGGAAGATTCGGAGAGAG) were designed to amplify a 423-bp fragment of PTEN 2. Primers included extensions encoding T7 and T3 RNA polymerase initiation sites to allow in vitro transcription of sense and antisense probes, respectively, from the amplified products. The hybridization was performed on 5-μm paraffin sections of formalin-fixed tissues. Prior to hybridization the sections were deparaffinized and treated with proteinase K for 15 min at 37 °C. [33P]UTP-labeled sense and antisense probes were hybridized to the sections at 55 °C overnight. Unbound probe was removed by treatment with RNase A for 30 min at 37 °C, followed by a high stringency wash (0.1× SSC for 2 h at 55 °C) and dehydrated in 70, 95, and 100% ethanol, respectively. The slides were dipped in NBT2 nuclear track emulsion (Eastman Kodak), exposed for 4 weeks at 4 °C, developed, and counterstained with hematoxylin and eosin. The intracellular localization of PTEN 2 gene was done in COS7 cells. 36 h after transfection, the cells were fixed using formaldehyde and stained using an anti-myc monoclonal antibody. The YFP-Golgi marker was purchased from CLONTECH. Brefeldin was purchased from Sigma Chemical Co., and the transfected cells were treated for 40 min before fixation. Approximately 5 × 108 293 transfected cells were collected and resuspended in 100 ml of 0.5% Triton X-100, 50 mm Tris, pH 7.5, 150 mm NaCl, 10% glycerol, 2 mm DTT, and protease inhibitors (Roche Molecular Biochemicals, 1836145). After sitting on ice for 15 min, the lysates were centrifuged at 10,000 × g and the supernatant was collected. The lysate was applied to a 2-ml reduced glutathione-Sepharose column and recirculated several times. The column was washed in 10 column volumes of 0.03% Brij35, 50 mmTris, pH 7.5, 0.5 m NaCl, 10% glycerol, and 2 mm DTT. The GST fusion protein was eluted with 50 mm Tris, pH 7.5, 150 mm NaCl, 2 mmDTT, 30% glycerol, and 10 mm reduced glutathione. The protein is stored in aliquots of 30 μl at −20 °C. Full-length human PTEN cDNA was cloned into the baculovirus expression vector, PH.hif, as a carboxyl-terminal HIS-tag fusion. The PCR primers used for this were: 5′-CATCGCGATCGCATGACAGCCATCATCAAAGAG-3′ and 5′-CTACGCGGCCGCTCAGACTTTTGTAATTTGTGTATGC-3′. Insect "Hi-Five" cells (Expression Systems) at 7.5 × 105/ml were infected with a multiplicity of 1.0 for 48 h and then harvested. Pelleted cells were resuspended in 100 ml of 50 mm Tris, pH 7.5, 300 mm NaCl, 250 mm sucrose, 1 mm DTT, and 5 mm imidazole. The suspension was sonicated for 1 min on ice and centrifuged at 10,000 ×g for 15 min. The supernatant was collected and recirculated over a 2-ml column of nickel-nitrilotriacetic acid Superflow (Qiagen, 1004493). The column was washed with 10 column volumes of lysis buffer and eluted with 5 × 1-ml steps of 50 mm Tris, pH 7.5, 250 mm sucrose, 150 mm NaCl, 2 mmDTT, and 250 mm imidazole. Enzyme was stored at −70 °C in 20-μl aliquots. The lipid-based phosphatase reactions were performed essentially as described previously (25Maehama T. Taylor G.S. Slama J.T. Dixon J.E. Anal. Biochem. 2000; 27: 248-250Crossref Scopus (99) Google Scholar). The reactions (50 μl) contained 100 mmTris (pH8.0), 10 mm dithiothreitol, 100 μmphosphatidylinositol phosphate substrates (Echelon), 1.0 mmphosphatidylserine (Avanti, 830052), and 50 μg/ml PTEN 2 or 10 μg/ml PTEN. Reactions were run at 37 °C for 3 h and centrifuged at 20,000 × g for 15 min. The supernatants were treated with malachite green (Biomol Green, AK-111), and absorbance was measured at A 650. Perusal of human and murine DNA sequence data bases demonstrated a closely related homologue of PTEN in both species. Cloning of cDNAs encoding both the human and murine forms of the homologue revealed that, whereas a diversity of sequences were expressed in the human, a single species was found in the mouse. While this work was being completed, a human homologue of the murine sequence was reported, and chromosomal mapping suggested that there were a number of differentPTEN-homologous genes encoded in the human genome, some of which appeared to be pseudogenes (26Chen H. Rossier C. Morris M.A. Scott H.S. Gos A. Bairoch A. Antonarakis S.E. Hum. Genet. 1999; 105: 399-409Crossref PubMed Scopus (67) Google Scholar). The sequence of the 664-amino acid (molecular mass 76,719 Da) mouse protein is illustrated in Fig. 1 and compared with the reported human PTEN 2 homologue and with PTEN. This figure illustrates that both the human and murine PTEN homologues are extended at their amino termini, and hydropathy analysis (Fig. 1) reveals that, in contrast to PTEN, the murine and human homologues appear to contain four high probability transmembrane domains followed by the catalytic domain. Interestingly, all four potential transmembrane domains contain charged residues embedded within the hydrophobic potential membrane-spanning sequence (Fig. 1). Using a structural algorithm, which predicts membrane topology, 2T. Wu, personal communication. we find that the murine homologue of PTEN may have a membrane-spanning structure, which is reminiscent of ion channels (Fig. 1). This sequence analysis suggests that the murine and human PTEN homologues appear to have extended amino termini, which may be involved with intracellular membrane association. Fluorescence in situ hybridization mapping revealed that the murine homologue mapped to a single locus on chromosome 8 between A3 and A4 (data not shown). The murine homologue appears to have a number of residues throughout the catalytic domain, which are conserved with PTEN (Fig. 1). Importantly, many of these residues are likely to be involved with substrate recognition and catalytic activity (Fig.2) (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar). For example, residues Asp426, His427, Cys458, Lys459, Gly461, Arg462, and Gln510, which are identical between the two proteins, were all proposed to be involved with substrate recognition in the structural analysis of PTEN. In addition, Fig. 2 also illustrates that the vast majority of the tumor-associated PTEN mutations, many of which are known to disrupt catalytic activity, occur at residues that are also either identical or conserved between the PTEN and the PTEN homologue. These comparative data suggest that the murine and human homologues are likely to have similar substrate specificity as PTEN, and we have therefore termed the new protein PTEN 2. The structure of PTEN has been recently solved by x-ray crystallography at 2.1-Å resolution (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar). This analysis revealed a molecule consisting of an amino-terminal phosphatase domain and a carboxyl-terminal C2 lipid binding motif, which were tightly packing against each other through a large interface. The phosphatase active site is similar to that observed in protein tyrosine phosphatases, including the essential catalytic residues, but is enlarged to allow for the binding of the larger phosphoinositide substrate. C2 domains have been shown to mediate membrane lipid association. Based on the 39% sequence identity between PTEN 2 and PTEN, a three-dimensional model of the phosphatase and C2 domains of PTEN 2 has been built by homology modeling and threading techniques using the PTEN structure as template (Fig. 2). The existence of a C2 domain in PTEN 2 (PTEN 2-C2) was established by using threading analysis and the high conservation of residues forming the phosphatase domain/C2 interface (Fig. 2). This domain presents a 23% sequence identity with the PTEN C2 domain (PTEN-C2). In PTEN 2 there is a five-residue insertion in the "T1 loop" that could allow this loop to establish more extensive contacts with the C2 domain as compared with PTEN (Fig. 2). The catalytic residues essential for the activity in all protein tyrosine phosphatases are conserved in PTEN 2 (Asp426, Cys458, and Arg464) and occupy the same position in both, PTEN and PTEN 2 (Fig. 2). The high conservation of residues forming the substrate binding site in PTEN and PTEN 2 has structural implications for substrate recognition. In particular, the positively charged residues that have been proposed to interact with the negatively charged phosphate groups of the phospholipid substrate in PTEN are conserved in PTEN 2 (His427, Lys459, and Lys462). Manual docking of inositol (1,3,4,5)-tetrakisphosphate in the active site of PTEN 2 was performed as described by Lee and collaborators (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar) and indicated that this phospholipid might bind to PTEN 2 in the same binding mode as that proposed for PTEN (Fig.3). Like the PTEN-C2, PTEN 2-C2 lacks the residues present in other Ca2+-dependent C2-containing proteins that coordinate Ca2+ and regulate membrane binding. A patch of five lysines or "CBR3 loop" has been proposed by Lee and collaborators (12Lee J. Yang H. Georgescu M. Di Cristofano A. Maehama T. Shi Y. Dixon J. Pandolfi P. Pavletich N. Cell. 1999; 99: 323-334Abstract Full Text Full Text PDF PubMed Scopus (880) Google Scholar) to mimic the Ca2+charge in PTEN. In PTEN 2, none of the five lysine residues in the CBR3 loop of PTEN is conserved (Pro607, Pro610, Tyr613, Asp614, and Cys616). A second "positive patch" in helix c2 (Lys644, Lys646, and Lys649) is conserved in both PTEN and PTEN 2. This region is similar to a helix in phospholipase A2 that has been shown to contribute to membrane binding. Together, these analyses suggest that many of the functional characteristics of the PTEN structure are conserved in PTEN 2. Northern blot analysis (Fig. 3) of various murine tissues using a PTEN 2 probe revealed a discrete ∼2.7-kb transcript, which was specifically expressed in testis, in agreement with results for the human PTEN 2 sequence (26Chen H. Rossier C. Morris M.A. Scott H.S. Gos A. Bairoch A. Antonarakis S.E. Hum. Genet. 1999; 105: 399-409Crossref PubMed Scopus (67) Google Scholar). The specificity of testis expression is emphasized by the observation that PCR analysis of multiple murine tissues revealed a signal only in testis RNA, even after a large number of PCR cycles (Fig. 3). Because testis contains a diversity of cell types, some of which (spermatocytes) pass through a number of developmental stages (27Kerr J. Microsc. Res. Tech. 1995; 32: 364-384Crossref PubMed Scopus (34) Google Scholar), we decided to analyze this tissue using in situ hybridization. Isotopic in situhybridization was performed on adult testis, as well as on testes representing various stages of adolescence. In the adult testis a positive signal was observed in germ cells within seminiferous tubules, whereas cells in the interstitium of the testis were negative. The positive signal within the seminiferous tubules showed an uneven distribution with some tubules displaying a strong signal, whereas others were completely negative (Fig. 4). A positive in situ hybridization signal appeared to correlate with the presence of a specific cell population. The positively reacting cells reside in the adluminal portion of the tubule, are of small to medium size, have a round nucleus with evenly distributed, finely granular chromatin, and a distinct nucleolus. Based on morphological criteria, these cells are most consistent with secondary spermatocytes and/or very early spermatids. The abundance of the positive signal, together with the short half-life of secondary spermatocytes, makes it likely that both cell types express PTEN 2. Sertoli cells, spermatogonia, primary spermatocytes, or mature spermatids did not express PTEN 2 RNA. Expression in these cell types was ruled out on the basis of their location within the seminiferous tubules, size and shape of cell body and nucleus, and chromatin pattern. Expression of PTEN 2 during testicular maturation is not detected until day 19 of postnatal development (Fig. 4). In a time course experiment we were unable to demonstrate PTEN 2 RNA in testes removed on days 3, 7, 10, and 16. The expression of PTEN 2 RNA therefore slightly precedes sexual maturity in the male mouse, consistent with an association of PTEN 2 with late events of spermatogenesis. Examination of the PTEN 2 sequence suggested that four potential transmembrane domains are found in the protein, consistent with the possibility that PTEN 2 is a membrane-associated molecule that passes through the secretory pathway. To examine the subcellular localization of PTEN 2, a plasmid encoding a form of the protein with a carboxyl-terminal myc epitope tag was transfected together with a plasmid encoding a yellow fluorescence protein-trans/medial Golgi marker (encoding the first 81 amino acids of β1,4-galactosyltransferase) into COS cells, an