The Tumor Suppressor, PTEN/MMAC1, Dephosphorylates the Lipid Second Messenger, Phosphatidylinositol 3,4,5-Trisphosphate*

Phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3) is a key molecule involved in cell growth signaling. We demonstrated that overexpression of PTEN, a putative tumor suppressor, reduced insulin-induced PtdIns(3,4,5)P3 production in human 293 cells without effecting insulin-induced phosphoinositide 3-kinase activation. Further, transfection of the catalytically inactive mutant of PTEN (C124S) caused PtdIns(3,4,5)P3 accumulation in the absence of insulin stimulation. Purified recombinant PTEN catalyzed dephosphorylation of PtdIns(3,4,5)P3, specifically at position 3 on the inositol ring. PTEN also exhibited 3-phosphatase activity toward inositol 1,3,4,5-tetrakisphosphate. Our results raise the possibility that PTEN acts in vivo as a phosphoinositide 3-phosphatase by regulating PtdIns(3,4,5)P3 levels. As expected, the C124S mutant of PTEN was incapable of catalyzing dephosphorylation of PtdIns(3,4,5)P3 consistent with the mechanism observed in protein-tyrosine phosphatase-catalyzed reactions.

A recently identified candidate tumor suppressor gene, PTEN/MMAC1, shares sequence identity with the family of protein-tyrosine phosphatases (PTPases) 1 (1). Deletions and mutations within the PTEN gene have been observed in several cancer cell types and tumor cell lines (2,3). Additional evidence that PTEN functions as a tumor suppressor was obtained by Furnari et al. (4), who showed that PTEN had a growth suppressor activity in glioma cells. PTEN encodes the active site consensus motif HCXXGXXR(S/T) found in all PTPases. In contrast, the recombinant protein is a poor catalyst toward both phosphoproteins and peptide substrates with the highest activity of PTEN observed toward the highly negatively charged, multiply phosphorylated polymer of (Glu-Tyr) n (5,6). Based on these observations we thought it possible that PTEN could catalyze the dephosphorylation of acidic nonproteinaceous substrate. Identification of possible in vivo substrates would not only suggest a possible physiological function of PTEN, but they might also provide insight into how PTEN functions as a tumor suppressor.

EXPERIMENTAL PROCEDURES
Determination of PtdIns (3,4,5)P 3 in Vivo-The coding sequence of human PTEN and the C124S mutant of PTEN (gift from Yi Zhao) were amplified by polymerase chain reaction using 5Ј primer (5Ј-CCGGTA-CCGCCACCATGGACTACAAGGACGACGATGACAAGACAGCCATC-ATCAAAGAG-3Ј) and 3Ј primer (5Ј-CCGTCGACTCAGACTTTTGTAA-TTTGTG-3Ј). The product was cleaved with KpnI and SalI and ligated into the KpnI/SalI sites of pCMV5 (gift from David W. Russell) to produce FLAG-tagged PTEN/pCMV5 and PTEN(CS)/pCMV5. Human 293 cells were cultured on a 6-well plate, and transfection of the cells were performed as described (12) using 1 g of the constructs. The efficiency of transfection was about 80% in this condition. 48 h after the transfection, the cells were serum-starved and labeled with [ 32 P]P i (100 Ci/ml) for 4 h. The cells were stimulated by incubation with insulin (0.1 g/ml) for 2.5 min at 37°C, and the stimulation was quenched by the addition of 0.93 ml of CH 3 OH/CHCl 3 /1% HClO 4 (50/25/18, v/v/v). The lipids were extracted and separated on a TLC plate as described (13) to determine the amount of [ 32 P]PtdIns(3,4,5)P 3 . To analyze the expression of FLAG-tagged PTEN protein, the transfected cells were lysed in Laemmli sample buffer and subjected to SDS-polyacrylamide gel electrophoresis. The samples were transferred to Immobilon filter (Millipore) and immunoblotted with anti-FLAG M2 antibody (Kodak), and the signal was visualized by Enhanced Chemiluminescence (Amersham Pharmacia Biotech) using the manufacturer's recommended protocols.
(Amersham Pharmacia Biotech) was then added to the supernatant. After the incubation for 2 h, the resin was packed in a column and washed with 100 ml of the lysis buffer, and the glutathione S-transferase-fusioned PTEN was eluted with 10 ml of the lysis buffer containing 10 mM glutathione. After overnight incubation with thrombin, the eluate was dialyzed against the lysis buffer and passed through glutathione-Sepharose 4B column, followed by a p-aminobezamidine-agarose column for adsorption of glutathione S-transferase and thrombin, respectively. Then the eluate was diluted with equal volume of TED buffer (20 mM Tris-HCl (pH 8), 2 mM EDTA, 2 mM DTT) and applied to MonoQ HR5/5 (Amersham Pharmacia Biotech) column equilibrated with TED buffer containing 150 mM NaCl. PTEN was eluted with a linear gradient of NaCl (150 -500 mM, 20 ml), followed by concentration using Centricon-30 (Amicon) and stored at Ϫ80°C until use.

RESULTS AND DISCUSSION
Although PTEN has the consensus sequence of a PTPase, it dephosphorylates p-nitrophenylphosphate and other artificial protein substrates poorly, having the highest catalytic activity with the highly negatively charged, multiply phosphorylated polymer of (Glu-Tyr) n (5,6). This observation raised the distinct possibility that PTEN might utilize acidic substrates other than Tyr or Ser/Thr phosphoproteins. In order to explore this possibility, we transfected PTEN into 293 cells and analyzed the changes in cellular phospholipids.
PtdIns(3,4,5)P 3 3-Phosphatase Activity of PTEN-To investigate the possibility that PTEN has PtdIns(3,4,5)P 3 phosphatase activity, the recombinant enzyme was expressed in E. coli and purified to homogeneity (data not shown). Radiolabel was incorporated in position 3 of the substrate, [ 32 P]PtdIns(3,4,5)P 3 , using PtdIns(4,5)P 2 , PI 3-kinase and [␥-32 P]ATP. When [ 32 P]PtdIns(3,4,5)P 3 was incubated with the purified PTEN, the radiolabel of [ 32 P]PtdIns(3,4,5)P 3 rapidly disappeared from the lipid phase ( Fig. 2A, inset) while coincidentally appearing in the aqueous phase ( Fig. 2A), suggesting the release of inorganic phosphate. In order to conclusively prove that the only phosphate that had been cleaved was at position 3 of the phosphoinositide, as opposed to other possible cleavages that could also generate a water-soluble radiolabel, we used the product of the reaction of PTEN as a substrate for PI 3-kinase. When the product was treated with PI 3-kinase,

FIG. 1. Effects of PTEN expression on insulin-induced PtdIns(3,4,5)P 3 production and PI 3-kinase activation in 293 cells.
A, B, and D, human 293 cells were transfected with indicated amount of pCMV5 and PTEN/pCMV5 (A and B) or 1 g of PTEN(CS)/ pCMV5 (D), then serum-starved, and labeled with [ 32 P]P i as described under "Experimental Procedures." The cells were incubated for 2.5 min in the presence or absence of insulin, and then the lipids were extracted and separated on a TLC plate as described under "Experimental Procedures." The plate was subjected to autoradiography (A and D), and the radioactivity of [ 32 P]PtdIns(3,4,5)P 3 (PIP 3 ) was estimated using a PhosphorImager (B). Error bars represent the differences between duplicate determinations in a typical study. Immunoblot of expressed FLAG-tagged PTEN using anti-FLAG antibody is shown in the inset of panel B. C, human 293 cells were transfected with indicated vector, starved, and stimulated by insulin as described under "Experimental Procedures." After the stimulation, the cells were lysed, and the PI 3-kinase activity of the immunoprecipitants with anti-phosphotyrosine antibody was assayed as described under "Experimental Procedures." Similar results were obtained in a repeated experiment.
Inositol Phosphate 3-Phosphatase Activity of PTEN-To more carefully dissect the specific nature of the catalytic activity of PTEN toward PtdIns(3,4,5)P 3 , we asked if PTEN displayed activity toward inositol phosphates. PTEN can dephosphorylate Ins(1,3,4,5)P 4 , whereas tyrosine-specific (PTP1D) and the dualspecific phosphatases (Cdc25B, VHR) exhibited no activity toward this inositol phosphate (Fig. 3A). Again, the PTEN-catalyzed reaction was specific for the position 3 of Ins(1,3,4,5)P 4 . Other inositol phosphates that do not have a phosphate at the position 3 on the inositol ring were not dephosphorylated by PTEN (Fig. 3A). The dephosphorylated product was identified as Ins(1,4,5)P 3 using the Ins(1,4,5)P 3 -binding protein (Fig. 3B). These results demonstrate that PTEN also has 3-phosphatase activity toward inositol phosphate. Both Ins(1,4,5)P 3 and Ins(1,3,4,5)P 4 have been proposed to be a functional second messenger responsible for the intracellular calcium signaling (17). Interestingly, Ins(1,3,4,5)P 4 can associate and activate a GTPase-activating protein (18). In contrast to PtdIns(3,4,5)P 3, Ins(1,3,4,5)P 4 is water-soluble and therefore was used to assess the significance of our in vitro observations by determining the kinetic parameters for the PTEN-catalyzed dephosphorylation occurred at position 3 on the inositol ring. The K m and V max values for Ins(1,3,4,5)P 4 were 98.9 M and 8.49 nmol/min/mg (k cat , 0.49 min Ϫ1 ), respectively (Fig. 3B). The K m value of 98.9 M is 250-fold lower than the K m of p-nitrophenylphosphate, which is 25.6 mM. Similar comparisons with the phosphorylated polymer (Glu-Tyr) n were difficult to assess because a detailed kinetic analysis was not performed with this substrate (6), and it is likely to be phosphorylated at more than one site. The low V max for the PTENcatalyzed dephosphorylation of Ins(1,3,4,5)P 4 implies that this may not be the preferred substrate in vitro for this enzyme. These in vitro assays may not be reflective of the in vivo activity of PTEN because the phosphatase could be regulated by phosphorylation, subcellular localization, and/or interaction with other cellular proteins. It is noteworthy that the C terminus of the phosphatase has a consensus PDZ binding site, and we have recently shown that PTEN interacts with several PDZ-containing proteins. 2 The activity of PTPases and dual-specific phosphatases to-2 Y. Zhao and J. E. Dixon, unpublished data. . At the indicated time, the reaction was terminated, and [ 32 P]PtdIns(3,4,5)P 3 was extracted and separated on a TLC plate as described under "Experimental Procedures." Inset, autoradiography of the plate. Released [ 32 P]P i in the aqueous phase was pooled during the extraction, and the radioactivity was counted. B, PtdIns(3,4,5)P 3 pretreated (30 min) with PTEN or corresponding buffer was subjected to the phosphorylation by PI 3-kinase using immunopurified PI 3-kinase and [␥-32 P]ATP as described under "Experimental Procedures." After the incubation for 30 min, the reaction was terminated, and the phospholipids were extracted and separated on a TLC plate. The radioactivity of PtdIns(3,4,5)P 3 produced was determined using a PhosphorImager. Inset, autoradiography of the plate. Similar results were obtained in a repeated experiment. ward all protein substrates is dependent upon an essential cysteine residue that forms a phosphoenzyme intermediate during catalysis (1). Because PTEN has the HCXXGXXR(S/T) motif conserved in all tyrosine or dual-specific phosphatases, we were interested in determining whether the cysteine was essential for the Ins(1,3,4,5)P 4 phosphatase activity. Mutation at Cys 124 of PTEN (C124S) resulted in a complete loss of enzyme activity toward Ins(1,3,4,5)P 4 (Fig. 3A). This mutation also resulted in a loss of phosphoinositide phosphatase activity (data not shown). Additionally, PTEN is extremely labile in the absence of thiols in the assay buffer. Optimum concentration of DTT for the PTEN-catalyzed reactions was 10 mM (data not shown). Therefore, we propose that PTEN-catalyzed dephosphorylation of inositol phosphate and phosphoinositide proceeds via a mechanism that is consistent with that described for other PTPases (1).
Potential Role of PTEN in Intracellular PtdIns (3,4,5)P 3 Regulation-We have established that the recombinant PTEN has phosphoinositide 3-phosphatase and inositol phosphate 3-phosphatase activities. The data shown in Fig. 1 suggest that suppression of insulin-induced PtdIns(3,4,5)P 3 production by overexpression of PTEN is due to its phosphoinositide phosphatase activity. In addition, as shown in Fig. 1D, overexpression of the CS mutant caused PtdIns(3,4,5)P 3 accumulation without insulin stimulation. These results strongly suggest that PTEN can act as a regulator of PtdIns(3,4,5)P 3 in vivo. Insulin activates PI 3-kinase via tyrosine phosphorylation of insulin receptor substrate-1 catalyzed by the insulin receptor (Fig. 4). PtdIns(3,4,5)P 3 produced by the PI 3-kinase can then activate Akt-mediated signals (Fig. 4). PtdIns(3,4,5)P 3 levels reached a plateau within 3-5 min after the stimulation, and then PtdIns(3,4,5)P 3 is degraded by unknown mechanisms (data not shown). Although our results suggest that PTEN can alter PtdIns(3,4,5)P 3 levels in 293 cells, it is clear that there are other cellular mechanisms that can also alter phosphoinositide concentrations. For example, Guilherme et al. (11) reported that a phosphoinositide 5-phosphatase was activated by insulin stimulation, and SHIP (SH2 domain-containing inositol 5-phosphatase) is also well known as a regulator of PtdIns(3,4,5)P 3 (10,19). If PTEN functions in vivo as a PtdIns(3,4,5)P 3 phosphatase, it follows that a homozygous deletion/mutation of this tumor suppressor gene could lead to a tumorigenic state through activation of the proto-oncogene, Akt. Specifically, loss of PTEN function would increase cellular levels of PtdIns(3,4,5)P 3 , thereby resulting in enhanced activation of Akt. In conclusion, although the physiological function of PTEN needs further clarification, we propose that: (i) members of the PTPase family of enzymes having an active site motif HCXXGXXR(S/T) such as PTEN are candidates to regulate intracellular levels of nonproteinaceous substrates as has also been reported for the RNA 5Ј-triphosphatase activity of CEL-1 (20); (ii) the dephosphorylation of phosphoinositide and inositol phosphate by PTEN is specific for position 3 on the inositol ring; and (iii) there are likely to be additional activating and/or localization mechanisms for PTEN within the cell regulating the catalytic activity of this enzyme.