Regulation of PTEN Phosphorylation and Stability by a Tumor Suppressor Candidate Protein* (cid:1)

The tumor suppressor PTEN plays an essential role in regulating signaling pathways involved in cell growth and apoptosis and is inactivated in a wide variety of tumors. In this study, we have identified a protein, referred to as PICT-1 (protein interacting with carboxyl terminus 1), that binds to the C terminus of PTEN and regulates its phosphorylation and turnover. Down-reg-ulation of PICT-1 in MCF7 cells by RNA interference enhances the degradation of PTEN with a concomitant decrease in its phosphorylation. PTEN C-terminal tu-mor-associated mutants, which are highly susceptible to protein degradation, have lost the ability to bind to PICT-1 along with their reduced phosphorylation, sug-gesting that their rapid turnover results from impaired binding to PICT-1. Our results identify PICT-1 as a PTEN-interacting protein that promotes the phosphorylation and stability of PTEN. These findings suggest a novel molecular mechanism underlying the turnover of PTEN, which also provides an explanation for the loss of PTEN function due to C-terminal mutations. The PTEN tumor suppressor antagonizes the actions of phos-phoinositide 3-kinase (PI3K) 1 by dephosphorylating the second messenger, phosphatidylinositol 3,4,5-trisphosphate

ϳ20% of all the known tumor-associated PTEN mutations, are localized to a 70-residue C-terminal segment (3,5). These types of mutations leads the rapid degradation of the mutant PTEN proteins in cells (6), indicating that this region is critical for controlling PTEN turnover. Although PTEN already has been shown to bind to several proteins through the C2 domain or a PDZ-binding motif at the extreme C terminus (7)(8)(9)(10)(11)(12), the involvement of these proteins in regulating the turnover of PTEN still remains elusive. Because recent studies have shown that the phosphorylation of specific Ser/Thr residues within the C-terminal region plays an important role in stabilizing PTEN (13)(14)(15)(16)(17)(18), it is crucial to unveil the molecular mechanism for the PTEN C-terminal phosphorylation. As a first step toward identifying proteins involved in regulating the phosphorylation and/or turnover of PTEN, we screened for PTEN C-terminalbinding proteins. In this screen we identified a novel PTENinteracting protein, referred to as PICT-1, encoded by a candidate tumor suppressor gene GLTSCR2 (19). PICT-1 appeared to regulate the phosphorylation and stability of PTEN through its interaction with the C-terminal region. Therefore disruption of the interaction between PICT-1 and PTEN resulted in the rapid degradation of PTEN. Our findings provide insight into the molecular mechanism(s) by which PTEN turnover is controlled.
Two-Hybrid Screen-To identify proteins that interact with the PTEN C-terminal region, we employed the MATCHMAKER yeast twohybrid screening system (Clontech). Yeast strain PJ69-2A harboring CT/pAS1 was mated with yeast strain Y187 pre-transformed with a human brain cDNA library according to the manufacturer's protocol. Mated yeast clones were initially subjected to histidine nutritional selection with 6 mM 3-amino-1,2,4-triazol, and positive clones from the first screening were then further subjected to adenine nutritional selection. Plasmid DNAs were prepared from positive clones and sequenced. A ␤-galactosidase assay was conducted according to the Clontech Yeast Handbook.
Pull-down Assay-Recombinant PTEN proteins were expressed in Escherichia coli strain JM109 as GST fusion proteins and purified as described previously (20,21) except that the protease cleavage step was omitted. PICT-1 was translated in vitro from Myc-PICT-1/pOPRSV plasmid DNA and simultaneously labeled with [ 35 S]methionine using the TNT T7 Quick-coupled Transcription/Translation System (Promega) according to the manufacturer's protocol. After the reaction, GST-fusioned PTEN proteins immobilized on glutathione-Sepharose beads (Amersham Biosciences) were added directly to the reaction mixture. After a 12-h incubation at 4°C, the beads were collected and washed with phosphate-buffered saline, and then bound proteins on beads were analyzed by SDS-PAGE and autoradiography.
Immunoprecipitation-Human breast cancer MCF7 cells (8 ϫ 10 6 ) were dispersed by sonication in 1 ml of immunoprecipitation buffer consisting of 20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM EGTA, 25 mM NaF, 10 mM ␤-glycerophosphate, 5 mM pyrophosphate, 120 mM NaCl, 1 mM orthovanadate, and 0.5 mM phenylmethylsulfonyl fluoride. After debris was removed, 7 g of rabbit IgG or affinity-purified anti-PICT-1 rabbit polyclonal antibody raised against the peptide corresponding to amino acids 37-51 of PICT-1 was added to cleared lysates along with 50 l of protein A/protein G plus agarose beads (Oncogene). After the samples were incubated for 10 h at 4°C, the beads were collected by centrifugation and washed with immunoprecipitation buffer, and then proteins remaining on the beads were resolved on SDS-PAGE. The resolved proteins were immunoblotted with anti-PTEN 26H9 antibody.
Immunoblots-Proteins resolved on SDS-PAGE were transferred to a FluoroTrans membrane (Pall), and then immunoblot analyses were performed according to the manufacturer's protocol. Antibodies used were anti-PTEN (Cascade Bioscience), anti-PTEN 26H9 (Cell Signaling Technology), anti-phospho-PTEN S380 (Cell Signaling Technology), and anti-actin (Sigma). The relative intensity of immunoreactive bands was measured using NIH Image, version 1.62. Typical images from repeated experiments are represented in each figure. All data represent the means Ϯ S.E. from duplicate determinations. Preparation of siRNA-21-Nucleotide RNA duplexes with UU overhangs at both 3Ј-ends were prepared by T7 RNA polymerase-based in vitro transcription as described by Yu et al. (22). Target sequences of GLT247 and GLT318 siRNAs correspond to nucleotides 247-267 and 318 -338 of GLTSCR2 (GenBank TM accession No. AF182076), respectively. GFP5 control siRNA was described by Yu et al. (22).
RT-PCR-Total RNA isolated from MCF7 cell was used as a template to synthesize cDNA utilizing Moloney murine leukemia virus reverse transcriptase (Toyobo). PCR was performed in a 10-l reaction mixture using 0.2 g of cDNA as a template. Primers used were GGA-CACTGGCTCCAAGGAAAAAG and TCTGGTGGTCTTCAAAGGATGG for PICT-1, CCTTTTGAAGACCATAACCCACC and ATCACCACACA-CAGGTAACGGC for PTEN, and GGAGAAAATCTGGCACCACACCT and AGGAAGGAAGGCTGGAAGAGTG for actin.

RESULTS AND DISCUSSION
Identification of PTEN C-terminal Region-binding Proteins-To identify proteins that interact with the C-terminal region of human PTEN, we performed a two-hybrid screen of a human brain cDNA library using the C-terminal region of PTEN (residues 338 -399) as bait. The extreme C terminus of PTEN (residues 400 -403), which contains a PDZ-binding motif, was deleted to selectively exclude interactions with PDZ domains. In this screen, we isolated 12 clones of which eight were determined to be false positives and two encoded unknown proteins. Of the two remaining clones, one corresponded to amino acids 10 -478 of a protein encoded by a human tumor suppressor candidate gene, GLTSCR2, and the other corresponded to amino acids 673-839 of human Vps16. We have referred to the gene product of GLTSCR2 as PICT-1 because of its ability to bind to the C-terminal region of PTEN. Although both proteins may exert a relevant biological function by binding to PTEN, we have focused herein on the characterization of PICT-1.
Interaction of PICT-1 with PTEN-To confirm our two-hybrid screen results and evaluate the relative contribution of the C-terminal region of PTEN to binding of PICT-1, we first conducted GST pull-down assays using wild-type and deletion mutants of PTEN. The full-length wild-type GST-PTEN fusion protein exhibited significant binding to radiolabeled PICT-1 produced by in vitro translation (Fig. 1A, WT), whereas GST alone did not bind (Fig. 1A, GST). Deletion of the C-terminal segment (338 -403) caused a remarkable decrease (by 93%) in the binding of PICT-1 (Fig. 1A, dCT). A construct in which only the C-terminal segment (338 -399) of PTEN was fused to GST also exhibited significant binding to PICT-1 (Fig. 1B, CT). However, this fusion protein did not bind PICT-1 as well as the GST-PTEN wild-type bait (32% of wild type). Although this result suggests that the C-terminal segment may not be the sole determinant of PTEN interaction with PICT-1, these data collectively indicate that the C-terminal region of PTEN plays an indispensable role for its interaction with PICT-1. Yeast two-hybrid assay demonstrated that deletion of the 338 -348 segment resulted in a complete loss of the binding ability of PTEN to PICT-1 (Fig. 1B), indicating that a critical amino acid residue(s) was located within this 11-amino acid segment. Further, the PDZ-binding motif at the C terminus of PTEN showed no effects on the binding (Fig. 1B).
We also performed a co-immunoprecipitation assay to determine whether an interaction between endogenous PICT-1 and PTEN proteins could be detected in intact cells. For this purpose, we utilized MCF7 cells that contain endogenous PTEN and PICT-1 proteins. As shown in Fig. 1C, the PICT-1 immunoprecipitated from MCF7 cell lysate by the anti-PICT-1 antibody was positive for interaction with endogenous PTEN protein, whereas no PTEN could be detected within the control rabbit IgG immunocomplex. Association of PTEN within the anti-PICT-1 immunocomplex was completely blocked by the corresponding PICT-1 peptide (Fig. 1C, PEP), confirming that endogenous PICT-1 and PTEN proteins can interact in intact cells.
Tumor-associated PTEN C-terminal Mutants Lose the Ability to Bind to PICT-1-The region that appeared to be responsible for the interaction with PICT-1 (Fig. 1, A and B) is nearly identical to a known "hot spot" (341-348) for C-terminal tumorassociated missense mutations (3,5). Intriguingly, one of these hot spot mutants, L345Q, was reported to exhibit extreme susceptibility to rapid protein turnover in cells (6). This implies the relevance of PICT-1 binding in the regulation of PTEN stability. To explore this possibility, we first tested the binding ability to PICT-1 of tumor-derived C-terminal mutants F341V, V343E, and L345Q. When tested using yeast two-hybrid assay ( Fig. 2A), we found that all of these mutations resulted in a complete loss of binding to PICT-1. A GST pull-down assay also showed significant decrease in the binding ability of these mutants to PICT-1 (Fig. 2B), a reduction comparable with that observed when the PTEN C-terminal segment (338 -403) was removed completely. We next carried out a cycloheximide pulse-chase experiment and determined the stability of these mutant PTEN proteins in MCF7 cells. MCF7 cells ectopically expressing PTEN mutant proteins were exposed to cycloheximide to abrogate protein synthesis, and then the level of PTEN protein was assessed by immunoblotting. As shown in Fig. 2C, reductions in mutant PTEN levels by 80 -90% were observed by 6-h cycloheximide treatment, whereas ϳ90% of wild-type PTEN remained by the same treatment. Collectively, these findings strongly suggest an essential role for the binding of PTEN to PICT-1 in the stabilization of PTEN protein and provide an explanation for the "loss-of-function" mutations that frequently occur in this region. The C-terminal PTEN mutants, including both deletion and missense mutants that no longer bind efficiently to PICT-1, cannot be protected from degradation and undergo rapid turnover.
PICT-1 Functions as a Regulator of PTEN Turnover-To confirm the role of PICT-1 in regulating PTEN turnover in cells, we next asked whether the down-regulation of PICT-1 directly affects the degradation of wild-type PTEN. For this purpose, we employed an RNA interference method utilizing PICT-1-targeted siRNAs GLT247 and GLT318, which substantially inhibited the ectopic expression of PICT-1 (data not shown). MCF7 cells transfected with FLAG-PTEN/pCMV and either GLT247 or GLT318 were exposed to cycloheximide, and then the level of FLAG-PTEN was assessed by immunoblotting to determine the rate of PTEN degradation (Fig. 3A). In control cells, a 5-h cycloheximide treatment did not significantly affect the level of PTEN, whereas reductions in PTEN levels by 21 and 48% were observed in GLT247-and GLT318-treated cells, respectively (Fig. 3A). Because these findings collectively indicate that PICT-1 plays an essential role in regulating PTEN turnover, we next tested the effect of PICT-1 down-regulation on the steady-state expression level of endogenous PTEN in MCF7 cells. Treatment of MCF7 cells with these siRNAs reduced the level of PTEN by 40 -50%, whereas the control siRNA had no effect (Fig. 3B). It is of note that in this experiment MCF7 cells were cultured longer (48 h) after the siRNA transfection than in the cycloheximide-chase experiment (24 h; see Fig. 3A) to observe a significant effect of PICT-1 downregulation on endogenous PTEN levels. The residual PTEN protein observed in GLT247-and GLT318-treated samples is likely to be derived from cells into which siRNAs were not effectively delivered and/or because of the relatively long halflife of PTEN protein in MCF7 cells (data not shown, but see Figs. 2C and 3A). The overall reduction in the PTEN expression was coincident with a reduction in the PICT-1 transcript, whereas the level of PTEN and actin mRNAs remained unaffected (Fig. 3C). From these data, we conclude that the reduction in the level of PTEN protein induced by the PICT-1 RNA interference is likely because of changes in the rate of protein degradation. The involvement of PICT-1 in maintaining PTEN stability also raises the intriguing possibility that PICT-1 loss-of-function mutations may result in the reduction of cellular PTEN in vivo. The GLTSCR2 gene, which encodes PICT-1, is located in a 150-kb minimal common deletion region for human gliomas, and thus it was originally identified as a candidate tumor suppressor gene (19). Down-regulation of PICT-1 expression FIG. 2. Degradation-susceptible C-terminal PTEN mutations reduce the ability to bind to PICT-1. A, yeast Y190 strain harboring pAS1 bait plasmid encoding various PTEN C-terminal mutants as indicated and either of pACT2 (open columns) or PICT-1/pACT2 (closed columns) was subjected to ␤-galactosidase assay. B, purified recombinant GST-PTEN wild type (WT), missense mutants F341V, V343E, and L345Q, and C-terminal deletion mutant (dCT) fusion proteins were tested for their ability to bind to PICT-1 using the GST pull-down assay as described for Fig. 1A. Proteins retained on the glutathione-Sepharose beads were detected by autoradiography (upper panel) or Coomassie Blue staining (lower panel). C, MCF7 cells expressing the indicated PTEN mutants were treated with 100 g/ml cycloheximide (CHX) for 6 h, and then PTEN proteins were detected by immunoblotting. may function to destabilize PTEN and subsequently deregulate PI3K/PIP 3 -mediated signals, even if no PTEN mutations are present. Our findings suggest a novel tumorigenic pathway that is dependent on the loss of PTEN function but independent of genetic lesions in the PTEN gene. The effect of impaired PICT-1 on PTEN signaling and function could be quite complex. This would be in part because the potential of tissuespecific expression of PICT-1 and PTEN, as well as the phosphorylation of the PTEN C-terminal region, may have multiple effects in addition to promoting PTEN stability (13,16,18,23,24). The significance of PICT-1 as a regulator of PI3K/PIP 3mediated signals is currently under investigation, and further study will be required to demonstrate a biological role for PICT-1.
PICT-1 Regulates PTEN C-terminal Phosphorylation-We next looked into the underlying mechanism for the stabilization of PTEN protein by PICT-1. A number of reports have shown that the phosphorylation of Ser/Thr residues within the C-terminal segment of PTEN plays a crucial role in stabilizing the PTEN protein (13)(14)(15)(16)(17)(18). The increased turnover rate of PTEN associated with down-regulation of PICT-1 raises the possibility that PICT-1 might affect the phosphorylation state of PTEN. To test this possibility, we analyzed the effect of PICT-1 on the phosphorylation of Ser-380 of PTEN, which is one of the crucial phosphorylation sites necessary for stabilizing PTEN in cells. As shown in Fig. 4A, the level of endogenous PTEN protein was comparable in each siRNA-transfected cell at 24 h post-transfection, and down-regulation of PICT-1 by RNA interference significantly reduced the level of PTEN phosphorylation at Ser-380. Moreover, remarkable reductions (52 Ϯ 4.3% for F341V; 38 Ϯ 1.2% for V343E; 76 Ϯ 5.8% for L345Q) in Ser-380 phosphorylation were also associated with these degradation-susceptible mutants (Figs. 2B and 4B), indicating that their rapid degradation resulted from reduced phosphorylation caused by an inability to interact with PICT-1. These results indicate that PICT-1 can regulate the phosphorylation of PTEN at Ser-380, and thus the binding of PTEN to PICT-1 governs its turnover via phosphorylation of the C-terminal region. Recent studies have shown that five Ser/Thr residues within the C-terminal segment, including Ser-380, are phos-phorylated and may contribute to the stabilization of PTEN protein. In particular, phosphorylation at Thr-382 and Thr-383 of PTEN has been suggested to be the most critical for stabilization (13)(14)(15)(16)(17)(18). However, in the MCF7 cells utilized in this study, phosphorylation of neither Thr-382 nor Thr-383 was likely to contribute significantly to PTEN stabilization, because a phosphorylation-resistant mutant, T382A/T383A, retained its stability in MCF7 cells and its degradation rate was indistinguishable from that of wild-type PTEN (data not shown). Although we cannot rule out a mutually exclusive effect between binding of PICT-1 and phosphorylation at other Ser/Thr residues within the C-terminal region, our findings suggest that Ser-380 is one of the phosphorylation sites that plays a crucial role in the regulation of PTEN turnover.
Our findings also raise the question as to which kinases and phosphatases are involved in the regulation of PTEN turnover. Casein kinase 2 has been implicated in the phosphorylation of the PTEN C-terminal region; however, Ser-380 is a poor match for the casein kinase 2 phosphorylation recognition sequence. It is therefore possible that unidentified kinase(s) may be responsible for this process, and PICT-1 may affect Ser-380 phosphorylation by activating a kinase and/or inhibiting a phosphatase. More detailed study will be required to address these possibilities.