Identification of a Novel Function of PiT1 Critical for Cell Proliferation and Independent of Its Phosphate Transport Activity*♦

PiT1 is a Na+-phosphate (Pi) cotransporter located at the plasma membrane that enables Pi entry into the cell. Its broad tissue expression pattern has led to the idea that together with the closely related family member PiT2, PiT1 is the ubiquitous supplier of Pi to the cell. Moreover, the role of Pi in phosphorylation reactions, ATP production, DNA structure, and synthesis has led to the view that Pi availability could be an important determinant of cell growth. However, these issues have not been clearly addressed to date, and the role of either Pi or PiT proteins in cell proliferation is unknown. Using RNA interference in HeLa and HepG2 cells, we show that transient or stable PiT1 depletion markedly reduces cell proliferation, delays cell cycle, and impairs mitosis and cytokinesis. In vivo, PiT1 depletion greatly reduced tumor growth when engineered HeLa cells were injected into nude mice. We provide evidence that this effect on cell proliferation is specific to PiT1 and not shared by PiT2 and is not the consequence of impaired membrane Na+-Pi transport. Moreover, we show that modulation of cell proliferation by PiT1 is independent from its transport function because the proliferation of PiT1-depleted cells can be rescued by non-transporting PiT1 mutants. PiT1 depletion leads to the phosphorylation of p38 mitogen-activated protein (MAP) kinase, whereas other MAP kinases and downstream targets of mammalian target of rapamycin (mTOR) remain unaffected. This study is the first to describe the effects of a Pi transporter in cell proliferation, tumor growth, and cell signaling.

PiT1 belongs to the inorganic phosphate (P i ) transporter (PiT) 3 family (Transport Classification Database (TCDB) Number 2.A.20) (1) comprising conserved symporters throughout all kingdoms that use either sodium or proton gradients to transport P i . Well known examples of this family includes the Neurospora crassa Pho4 and Saccharomyces cerevisiae Pho89 or the Escherichia coli transporters PiTA and PiTB and Arabidopsis thaliana Pht2 (2). In mammals, the PiT family is comprised of only two members, PiT1 (SLC20A1) and PiT2 (SLC20A2), which were initially identified as receptors for retroviruses (3,4) and were subsequently found to possess electrogenic Na ϩ -P i symporter activity (5).
Phosphate has a structural role in phospholipids of cell membranes, nucleoproteins, and nucleic acids. It forms high energy ester bonds (e.g. in adenosine triphosphate and guanosine triphosphate) and plays a central role in cellular metabolic pathways and signal transduction through covalent phosphorylation of proteins and lipids. For these reasons, P i is essential for many vital functions, including storage and liberation of metabolic energy, delivery of oxygen to the peripheral tissues, muscle contractility, electrolyte transport, neurological functions, and integrity of bone (6). The broad tissue distribution of PiT1 and PiT2 has led to the proposal that these transporters could serve a housekeeping role for cellular P i homeostasis (5,7), although direct experimental evidence is lacking. Specifically, the consequences of variation of PiT expression at the cell surface on cell energy homeostasis and cell proliferation are not known. Moreover, although the consequences of variation of P i supply to the cell have been studied, P i starvation results in an up-regulation of PiT protein expression (5,8). For this reason, starving the cells of P i may not adequately uncover any potential role of the PiT proteins in cell proliferation.
In this study, we addressed the question of whether PiT1 and/or PiT2 could modulate cell proliferation. Through an RNA interference approach, we show that reduced expression of PiT1 results in decreased P i transport and cell proliferation in HeLa and HepG2 cells. PiT1 depletion delayed cell cycle and impaired mitosis and cytokinesis. In vivo, injection of PiT1depleted HeLa cells in nude mice results in reduced tumor growth. We generated PiT1 non-transporting mutants and showed that modulation of cell proliferation through PiT1 expression is independent from P i transport into the cell and cannot be compensated by PiT2 overexpression. PiT1 depletion did not affect the ERK1/2, JNK, and mammalian target of rapamycin (mTOR) signaling pathways, whereas p38 mitogenactivated protein (MAP) kinase was overphosphorylated, although it does not appear to be instrumental to the PiT1mediated effect on cell proliferation. This study is the first to demonstrate a direct role of PiT1 in cell proliferation, tumor growth, and cell signaling. Altogether, our data describe a novel function for PiT1, independent from its previously known transport activity. Discussion regarding PiT1 structure-function relationship in relation to this novel function is presented.

EXPERIMENTAL PROCEDURES
Culture Conditions, Transfections, and Growth Curves-HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 5% FBS. The HepG2 cell line was cultured in a medium consisting of Dulbecco's modified Eagle's medium/Ham's F12, supplemented with 5 mg/liter insulin, 3.5 ϫ 10 Ϫ7 M hydrocortisone hemisuccinate, 2 mM L-glutamine, and 10% FBS. HepG2 cells were seeded at 5 ϫ 10 3 cells/cm 2 , and medium was renewed every day. For siRNA and shRNA transfections, cells were seeded 24 h before the experiment in antibiotic-free medium at 3 ϫ 10 5 or 5 ϫ 10 5 cells/well in a 6-well plate, respectively. Cells were transfected with 10 nM siRNAs or 4 g of plasmid using Lipofectamine 2000 in a serum-free medium. Four hours after transfection, 5% FBS was added to the medium. For growth curves, 25,000 cells were seeded in triplicate in 24-well plates. Cells were trypsinized and counted each day.
Production and Purification of an Anti-PiT1 Antibody-A 59-amino acid peptide sequence from the central intracellular loop of PiT1 was fused to GST or to thioredoxine-V5-His 6 , expressed in BL21AI E. coli (Invitrogen), and purified from bacterial lysates on GSTrapFF columns (for GST-PiT1) or by using the HisTrap kit (for thioredoxine-PiT1-V5-His 6 fusion protein) as per the manufacturer's instructions (GE Healthcare). Rabbits were immunized against human GST-PiT1 peptide (CovalAb), and the anti-PiT1 antibody was purified from rabbit serum by incubating the serum with pieces of polyvinylidene difluoride membrane blotted with 250 g of purified thioredoxine-PiT1-V5-His 6 protein. Elution was performed using 0.2 M glycine, pH 2, and the eluate was rapidly neutralized with 1 M Tris, pH 10. The purified anti-PiT1 antibody was diluted in Tris-glycine, pH 7.4, 50% glycerol, 0.1% bovine serum albumin.
RNA Interference-Transient inactivation of PiT1 was assayed using siRNA SMARTpool from Dharmacon (Chicago, IL) (catalog number L-007432-01) as per the manufacturer's instructions. Individual siRNA duplexes corresponding to the siRNA from the pool were then tested separately. Each individual siRNA gave comparable inactivation of PiT1 expression. Subsequent experiments were conducted with 10 nM PiT1 siRNA-A (5ЈP-UAUCAGUUCAGACCACU-UGUU-3Ј) and siRNA-B (5ЈP-UAUCUAUGCUGGUUUC-CUC UU-3Ј). Transient reduction of PiT2 expression was achieved using siRNA SMARTpool from Dharmacon (catalog number L-007433-01) used at 10 nM, as per the manufacturer's instructions. siCONTROL TM non-targeting siRNA 1 and siGLO TM RISC-free siRNA from Dharmacon were used as negative controls in transient transfection experiments. Stable knockdown of PiT1 expression was performed by cloning an shRNA corresponding to the sequence of siRNA-B into the pSUPER vector (9). The scramble sequence of shRNA-B was used as a negative control. HeLa cells were transfected with the pSUPER-shRNAs, plated at limiting density, and puromycinresistant clones were picked, expanded, and tested for PiT1 expression. The data presented herein are from individual clones displaying at least an 80% knockdown of PiT1 expression. Experiments were performed with 3-4 independent stable transfectants, and the data presented illustrate representative clones.
Cloning of Human PIT1 and Site-directed Mutagenesis-Human PIT1 was PCR-amplified from human kidney cDNA using the primers listed in Table 1. The PCR product was subcloned into pCR2.1 TA cloning vector (Invitrogen) and subsequently subcloned into the pcDNA6A expression plasmid (Invitrogen), in-frame with the V5 and His 6 C-terminal tags. The integrity of the construct was verified by sequencing. Site-directed mutagenesis was used to introduce three silent mutations in the PiT1 sequence at the siRNA-B binding site to render the cDNA resistant to siPiT1-B cleavage (PiT1-RNAiR). The transportdeficient mutants of PiT1, S128A and S621A, were constructed by site-directed mutagenesis (QuikChange; Stratagene) from the PiT1-RNAiR construct. The sequences of the respective primers are listed in Table 1. The human PIT2-expressing plasmid used in this study was previously described (10).
Gene Expression and Quantification-Total RNA was isolated from cells and tissue using NucleoSpin RNA columns (MACHEREY-NAGEL). Northern analysis of total RNA (25 g) from HeLa cells was performed as described previously (11). For PCR detection of Na ϩ -P i transporters, RNA (2 g) was reverse-transcribed with 200 units of M-MLV-RT (Invitrogen) and PCR-amplified using the primers listed in Table 1. PCR reactions contained 1ϫ reaction buffer, 0.2 mM dNTPs, 1.5 mM MgCl 2 , 0.25 M each primer, and 0.05 units/l Taq (Invitrogen). Cycle conditions were 94°C for 1 min of initial denaturation followed by 36 cycles of denaturation (94°C, 15 s), annealing (Table 1), and extension (72°C, 30 s). For gene quantification, RNA (2 g) was reverse-transcribed with 200 units of M-MLV-RT (Invitrogen). Real-time PCR was performed using SYBR Green chemistry (Thermo Scientific) on an ABI Prism 7700 detection system. The glucuronidase gene was used as the reference gene, and expression differences were calculated as described previously (12). Primer sequences are listed in Table 1.
Phosphate Uptake Measurements-Transport of phosphate was measured as described previously (13). Apparent affinity constant (K m ) and maximal transport rate (V max ) were calculated by non-linear curve fitting, assuming Michaelis-Menten kinetics.
DNA and Actin Staining-Cells grown on coverslips were washed with PBS, fixed with 3% paraformaldehyde at 20°C for 15 min, washed twice with PBS, and incubated for 10 min with 20 mM glycine. Cells were permeabilized in PBS, 0.1% saponin for 30 min at room temperature and washed three times in a drop of PBS/saponin. Actin was stained by incubating cells in a drop of Texas Red-X phalloidin (1:250) (Molecular Probes) for 30 min at 20°C in the dark. Cells were washed three times in PBS and incubated in a drop of Hoechst 33342 (Sigma) at 2 g/ml in PBS for 10 min at 20°C in the dark to stain DNA. After the incubation, coverslips were washed twice in PBS to remove non-adherent cells, mounted in Glycergel mounting medium (Dako), and photographed under a microscope (Nikon Eclipse 800).
Time-lapse Video Microscopy-HeLa cells were cultured in Petri dishes equipped with glass coverslips and placed in a humidified and thermo-regulated chamber maintained at 37°C on the stage of an inverted epifluorescence microscope (Axiovert 200, Carl Zeiss) equipped with a monochromatic light, a cooled CCD camera (ORCA, Hamamatsu Photonics), a motorized stage, a piezo z objective, and Plan-Apochromat 20ϫ/0.75 NA and Plan-Neofluar 40ϫ/0.75 NA objectives (Carl Zeiss). Images of dividing cells were acquired during 4 -16 h at a frequency of one image each 2 min. MetaMorph software (Universal Imaging, West Chester, PA) was used for computer-based image acquisition and analysis of live cell data. The single images shown were prepared using Adobe PhotoShop CS3.
Flow Cytometry Analysis and Synchronization of HeLa Cells-Cells were trypsinized and centrifuged at 1,000 ϫ g for 5 min in a conical tube. Cell pellets were thoroughly resuspended in 0.2 ml of PBS, fixed by slowly adding them onto 1.8 ml of ice-cold 70% ethanol, and incubated for 3 h at 4°C. For DNA content analysis, fixed cells were stained with 0.02 mg/ml propidium iodide in PBS containing 0.1% Triton X-100, 0.2 mg/ml DNasefree RNase A, for 30 min at 37°C. Synchronization of HeLa cells at G 1 /S transition was performed by double thymidine block, as described previously (14). Every 2 h after drug release, cells were fixed in ice-cold 70% ethanol and stained with 0.02 mg/ml propidium iodide, and their cycle profile was analyzed by flow cytometry analysis on a BD Biosciences FACSCalibur flow cytometer with CELLQuest software.
Confocal Microscopy-Confluent cells were fixed, and the actin and/or expressed proteins were stained. Immunodetection of hPIT1 was performed using our custom antibody described in this study and a secondary anti-rabbit antibody coupled to fluorescein isothiocyanate. Coverslips were mounted on glass slides using glycerol (Dako) containing 2.5% 1.4-diazabicyclo-(2.2.2) octane (Sigma) as a fading retardant. Confocal images were taken by using a Leica TCSSP (Wetzlar, Germany) laser scanning microscope equipped with a ϫ63 oil immersion objective.
Immunofluorescence on Living Cells-Cells were seeded on slides and cultured overnight. Slides were washed with cold PBS and placed in 35 l of PBS containing anti-V5 antibody (1:250) and 1% bovine serum albumin for 1 h at 4°C in the dark. After four washes with cold PBS, cells were incubated as above with Alexa Fluor 488 goat anti-mouse IgG (1:300 dilution) secondary antibody. Cells were washed with cold PBS and fixed in PBS containing 1% paraformaldehyde and 1% bovine serum albumin for 20 min at room temperature. Images were acquired under fluorescent illumination (Nikon Eclipse 800).
Immunoblot Analysis and Antibodies-Cells were detached and the cell pellet was incubated for 30 min in ice-cold lysis buffer (150 mM NaCl, 10 mM Tris-HCl, 5 mM EDTA, 1% Non- idet P-40, 0.1% SDS, 0.5% deoxycholate, 1 mM Na 3 VO 4 , 1 mM NaF, 5 mM sodium pyrophosphate, 0.2 M phenylmethylsulfonyl fluoride, and a protease inhibitor mixture). After centrifugation at 14,000 rpm for 15 min, the protein extracts (supernatants) were boiled in 1ϫ SDS loading buffer prior to SDS-PAGE. Proteins were transferred to polyvinylidene difluoride membrane and blocked with 5% milk/TBST (10 mM Tris, 154 mM NaCl, 0.15% Tween 20) for 1 h. Blots were probed with primary antibodies in 5% bovine serum albumin or milk/TBST overnight followed by secondary antibodies for 1 h, detected using ECL Western blotting detection reagent, and exposed on ECL hyperfilm (GE Healthcare). Monoclonal anti-14-3-3 clone CS112-2AB was from Upstate Biotechnology, mouse monoclonal anti-V5 antibody was from Invitrogen, and anti-bromodeoxyuridine antibody was from Amersham Biosciences. Xenographic Growth of HeLa Tumor in Athymic Mice and PCNA Immunodetection-Outbred male athymic nude mice were obtained at 7 weeks of age from Elevage Janvier (Le Genest Saint Isle, France) and used for experiments at 8 weeks of age. Tumor formation was assayed by subcutaneously injecting 5 ϫ 10 6 cells suspended in 100 l of sterile PBS. Groups of five mice were injected at two sites per mouse. To measure the rate of tumor growth, the size of the tumor was monitored weekly using the formula: volume ϭ (length ϫ width 2 )/2 (15). Mice and excised tumors were weighed 70 days after implantation (at time of sacrifice). Tumors were fixed in paraformaldehyde, and proliferating cell nuclear antigen (PCNA) immunodetection was performed using the M.O.M. TM immunodetection kit (Vector Laboratories) and anti-PCNA antibody (Dako), according to the manufacturer's instructions.
Statistics-All graphs are plotted as mean Ϯ S.E. Statistics for dual comparisons were generated using Student's t tests, whereas statistics for multiple comparisons were generated using one-way analysis of variance followed by a suitable post hoc t test; *, p Ͻ 0.05, **, p Ͻ 0.01, for all statistics in the legends for Figs. 1 and 5-8.

RESULTS
The Na ϩ -P i Transporter PiT1 Is Required for Proliferation of HeLa and HepG2 Cells-Quantification of Na ϩ -P i transporter expression in HeLa cells using real-time PCR revealed that PiT1 and PiT2 were the only P i transporters expressed in these cells (Fig. 1A). Other described mammalian P i transporters, namely NPT1, NPT2a, NPT2b, and NPT2c (16), were not detected. PiT1 expression was 2.1-fold higher than PiT2 in HeLa cells, in agreement with previous data obtained from cDNA microarrays hybridizations (17). The lack of expression of other P i transporters in HeLa cells, together with the fact that inoculation of nude mice with HeLa cells is a classical model of tumorigenesis, make them an attractive system for the study of PiT protein function. As shown in Fig. 1B, knockdown of PiT1 expression in HeLa cells using two different siRNA constructs (siRNA-A and -B) was effective, as evidenced by Western blot analysis using a custom anti-PiT1 antibody (Fig. 2). More detailed analysis of RNAi-mediated knockdown of PiT proteins is presented in Fig. 3. Northern analysis showed that PiT1 and PiT2 siRNAs induce a specific knockdown of the respective transporter expression, with no compensatory up-regulation of the remaining PiT (Fig. 3, A and B). To study the prolonged effects of PiT1 inactivation, we generated an shPiT1 corresponding to the sequence of the siRNA-B and selected HeLa cell clones stably transfected with shPiT1-B plasmids. Realtime PCR data showed that in HeLa cells stably transfected by PiT1 shRNA, PiT1 mRNA expression was significantly decreased, whereas PiT2 expression was unchanged (Fig. 1C). A similar reduction in PiT1 protein levels was seen with shRNA as with transient siRNA experiments (Fig. 3, C and D), and immunochemistry confirmed that plasma membrane labeling with anti-PiT1 antibody almost completely disappeared in HeLa cells stably transfected with PiT1 shRNA (Fig. 3E). As a result, both in transient and in stable RNAi experiments, Na ϩ -P i transport function was reduced by 65-70% in HeLa cells (Fig. 1D). More detailed analysis of transport function reveals that the reduction in P i transport was due to a reduced transport capacity (V max shScramble ϭ 5.7 Ϯ 0.3 nmol⅐mg of protein Ϫ1 ; V max shPiT1 ϭ 2.4 Ϯ 0.2 nmol⅐mg of protein Ϫ1 ) rather than a change in transport affinity (K m shScramble ϭ 150 Ϯ 16 M; K m shPiT1 ϭ 132 Ϯ 11 M). Because PiT1 and PiT2 have the same affinity for P i (18,19), the reduction in Na ϩ -P i uptake in HeLa cells is consistent with a decrease in PiT1 protein abundance at the cell membrane.
We next evaluated the effect of PiT1 knockdown on the proliferation of HeLa cells. Transient transfection of two different PiT1 siRNAs led to a significant reduction in cell number, whereas unrelated siRNA (siControl) and PiT2 siRNA had no effect on cell proliferation (Fig. 1E). Stable expression of PiT1 shRNA showed that 4 or 5 days after equal seeding (i.e. at the time of exponential growth), the number of shPiT1 HeLa cells was half that of wild-type or shScramble HeLa cells (Fig. 1F). This result indicates that a wild-type level of PiT1 is necessary for normal proliferation of HeLa cells.
To exclude the possibility that the effect of PiT1-knockdown is specific to HeLa cells, PiT1 knockdown was performed in the non-tumorigenic HepG2 cell line, from hepatic origin. Expression analysis of the Na ϩ -P i transporters showed that although PiT1 was the main transporter expressed in these cells, there was a high expression of NPT1 and a weaker expression of PiT2 (Fig. 4A). Stable inactivation of PiT1 using shRNA (Fig. 4B) resulted in reduced Na ϩ -P i transport due to a decrease in the transport capacity (V max shScramble ϭ 2.08 Ϯ 0.16 nmol⅐mg of protein Ϫ1 ; V max shPiT1 ϭ 1.31 Ϯ 0.12 nmol⅐mg of protein Ϫ1 ) rather than a change in the transport affinity (K m shScramble ϭ 75.1 Ϯ 10.9 M; K m shPiT1 ϭ 75.3 Ϯ 15.3 M) (Fig. 4C). Although the decrease in Na ϩ -P i transport following shPiT1 knockdown was less prominent than in HeLa cells, the proliferation of HepG2 cells was similarly affected, and the cell number was half that of shScramble HepG2 clones 4 days after seeding (Fig. 4D).
In HeLa Cells, PiT1 Depletion Delays the Cell Cycle and Impairs Mitosis and Cytokinesis-Hoechst staining of stably transfected shRNA HeLa clones revealed the presence of giant polyploid cells in the PiT1-depleted population and the frequent observation of lagging chromosomes in mitotic figures (Fig. 5A). A representative series of time-lapse images shows FIGURE 2. Anti-PiT1 antibody production and characterization. A polyclonal antibody against human PIT1 was generated as described under "Experimental Procedures." Briefly, a 59-amino acid peptide located in the intracellular loop of PiT1 was expressed as a GST fusion protein and used for immunization. Rabbit sera were then purified from anti-GST antibodies. A, total, membranous, and cytosolic protein extracts from untransfected (Cont; 1 g of protein extract) or hPIT1-transfected (T; 1 g) COS7 cells, as well as untransfected HEK cells (HEK; 20 g of protein extract), were separated by SDS-PAGE, and the blotted membrane was probed with the purified anti-PiT1 antibody (1/500). B, glycosylation of endogenous versus transfected PiT1 protein. Proteins extracts from COS7 or HEK cells were treated (ϩ) with 500 units of (PNGaseF) for 1 h at 37°C or not treated (Ϫ) and then separated by SDS-PAGE, and the blotted filter was probed with the purified anti-PiT1 antibody (1/500). C, extracts from COS7 cells transfected with hPIT1-V5 were separated by SDS-PAGE, and the membranes were immunoblotted (IB) with either the anti-PiT1 (1/500) or the anti-V5 antibody (1/5000), as indicated. D, GST-PiT1 peptide used for immunization and an unrelated fusion peptide (GST-klotho) were separated by SDS-PAGE, and the blots were hybridized with immune serum or purified antibody, as indicated. E, competition with the immunogen peptide. Extracts from HEK cells (50 and 20 g) were separated by SDS-PAGE, and the blotted filters were probed with anti-PiT1 (1/1000) antibody, which was preincubated for 1 h at room temperature with or without 0.5 g of GST-PiT1 peptide, as indicated.
abnormal cytokinesis without cell separation, generating binucleated cells (Fig. 5B and supplemental Movie 1). Other figures of division show delayed progression through mitosis due to an extended metaphase stage (supplemental Movie 2) and generation of several nuclear poles (supplemental Movies 3 and 4), leading to either increased cell division duration or abnormal polyploid cells. In accordance with the reduced proliferation of PiT1-depleted HeLa cells, the mitotic index of shPiT1 clones was reduced by half (Fig. 5C). Mitotic figures were characterized by an increased proportion of cells in metaphase and a decreased proportion in anaphase and telophase, together with the appearance of numerous abnormal images of anaphase and telophase (Fig. 5D), consistent with the presence of polyploid cells in the PiT1-depleted population. FACS analysis of transiently or stably PiT1-depleted cells was characterized by a marked reduction in the proportion of PiT1-depleted cells in G 1 phase, together with an increase in S and G 2 /M phases ( Fig. 5E and Table 2), which, according to the reduction in the mitotic index, may indicate a delay in the G 2 phase. By synchronizing cells at G 1 /S by double thymidine block (Fig. 5F), we found that entry in G 2 /M was delayed in shPiT1 HeLa cells, consistent with our observations. A significantly higher proportion of cells in the PiT1-depleted cell population (16.5 versus 4.2%, when compared with HeLa cells) was still found in the G 2 /M cell population (having 4n chromosomes) after synchronization (Fig. 5F), but they most likely represent tetraploid cells arrested at the G 1 /S transition (having two sets of 2n chromosomes). Pulse bromodeoxyuridine staining of asynchronous shScramble and shPiT1 HeLa cells showed that large polyploid PiT1-depleted cells were still cycling normally, whereas apoptosis end-stage terminal deoxynucleotidyltransferase-mediated dUTPbiotin nick end-labeling staining showed that these cells did not enter apoptosis more frequently than their normal counterpart (data not shown). Therefore, an increase in the apoptosis rate of these cells cannot account for the reduced number of cells seen after PiT1 depletion.

PiT1 Depletion in HeLa Cells Reduces Tumor Growth in Nude
Mice-To evaluate whether PiT1 depletion affects tumor growth in vivo, nude mice were injected subcutaneously with shScramble, shPiT1 stably transfected HeLa cells, or parental HeLa cells (Fig.  6A). At 7 days after inoculation, all animals developed a palpable tumor at the injection site. However, the rate of tumor growth in the shPiT1 groups (two independent clones) was significantly slower than the control groups when comparing either tumor volume (Fig. 6A) or tumor size (Fig. 6B). This difference became significant (p Ͻ 0.05) by day 35. By day 70, the mean tumor size in shPiT1 mice was 0.41 Ϯ 0.09 and 0.26 Ϯ 0.07 cm 3 , whereas in the control groups, it had reached 1.10 Ϯ 0.13 and 1.4 Ϯ 0.27 cm 3 , corresponding to a 62-80% inhibition of tumor growth rate. Tumor weight was also decreased by 66 -82% (Fig. 6C). Terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end-labeling staining of tumor sections revealed no difference in apoptotic cell number in the different tumors (not shown), whereas the number of PCNApositive cells was much lower in shPiT1 HeLa cell-derived tumors (Fig. 6D). Taken together, these data indicate that reduced expression of PiT1 significantly inhibited tumor growth through decreased proliferation but did not affect tumor induction in nude mice.
Rescue of Cell Morphology, Proliferation, and Transport by RNAi-resistant PiT1 Mutants-Although we demonstrated, both in transient transfection with two different siRNAs and in stable transfection with separate shPiT1 clones, that depleting PiT1 from cells resulted in reduced proliferation, we conducted rescue experiments to rule out any possible off-target effects due to the RNA interference approach. To do this, we cloned the coding region of human PIT1 into the pcDNA6 expression vector and mutated the target sequence of siRNA-B to produce a human PIT1-RNAi-resistant plasmid. HeLa cells stably transfected with shPiT1 were transfected with PiT1-RNAiR or empty pcDNA6 plasmids and were grown in selective medium. As shown in Fig. 7A, expression of PiT1-RNAiR in shPiT1 HeLa cells rescued their cell morphology. The number of polyploid cells was low and not different from shScramble HeLa cells. These rescued cells appeared to have a restored proliferation rate, similar to that of shScramble HeLa cells (Fig. 7B). Moreover, the rescued cell morphology and proliferation was accompanied by a rescue of Na ϩ -P i transport (Fig. 7C). These experiments demonstrate that expressing sufficient PiT1 levels in the cells is enough to restore normal cell morphology, cell proliferation, and Na ϩ -P i transport. However, this does not show whether or not the transport function of PiT1 per se is instrumental in this rescue.
PiT1 Modulates Cell Proliferation Independently of Its Transport Function-Because the main reported role of PiT1 is to couple inward P i uptake to the Na ϩ gradient across the cell plasma membrane (27), we evaluated whether this function underlies the effect of PiT1 knockdown on cell proliferation. To test this hypothesis, we attempted to rescue the proliferation of PiT1-depleted HeLa cells by transfecting a plasmid encoding for human PiT2 in these cells. As shown on Fig. 8A, and in contrast to the results obtained with PiT1-RNAiR (Fig. 7), cell proliferation of shPiT1 HeLa cells could not be rescued by PiT2 (Fig. 8A), although PiT2 overexpression resulted in a significant increase in Na ϩ -P i transport (Fig. 8B). This result demonstrated that restoring a normal P i transport in shPiT1 HeLa cells is not sufficient to rescue a normal proliferation and suggests that PiT1 mediates its effects on cell proliferation independently from its transport activity. To test this hypothesis, we generated PiT1 mutants (using the PiT1-RNAiR backbone) in which the serines at position 128 or 621 were replaced by alanine residues (see Fig. 10A). This strategy was based on previous structure-function studies conducted on human PIT2 in which corresponding serine 113 and 593 were shown to be essential for Na ϩ -P i transport activity and did not affect protein expression and plasma membrane localization (20). As expected, despite adequate cell membrane localization and expression levels identical to that of wild-type PiT1 (see Fig. 10, B and C), the S128A mutant of PiT1 did not transport P i and hence was unable to rescue normal P i transport in PiT1-deficient cells (Fig. 8B). Despite the lack of Na ϩ -P i transport function, transient transfection of the S128A mutant in PiT1deficient HeLa cells restored cell proliferation to a level identical to that observed in their wild-type counterpart (Fig.  8C). These results clearly show that PiT1, but not PiT2, selectively modulates cell proliferation, independently of its transport function.
Increased Phosphorylation of p38 MAPK in PiT1-depleted HeLa Cells-Because PiT1 Na ϩ -P i activity per se is not involved in the phenotype seen in shPiT1 cells, we questioned whether the knockdown of PiT1 could change the phosphorylation status of several signaling pathways implicated in cell proliferation. mTOR is a kinase that integrates signals from nutrients and growth factors to regulate cell growth and cell cycle progression coordinately. It is known to regulate translation mainly through the S6K1 (ribosomal protein S6 kinase 1) and the 4EBP1 proteins (eukaryotic translation initiation factor 4E-binding protein 1) (21). Study of the phosphorylation of S6, S6K1, and 4EBP1 in basal culture conditions, after starving of serum or restimulation, did not detect any differences between cells transfected with siControl, siPiT1, or siPiT2 siRNAs, suggesting that the mTOR pathway was unaffected (Fig. 9A). Similarly, as shown in Fig. 9B, depletion of PiT1 in HeLa cells affected neither the phosphorylation of the ERK (extracellular signalregulated kinase) protein nor the phosphorylation of other components of the cascade such as c-Raf and 14-3-3. This was true in our basal culture conditions or after stimulation with either serum or 10 nM epidermal growth factor (EGF). The same observation was made for phosphorylation of the JNK MAPK (Fig. 9B). In contrast, significant changes were observed in the phosphorylation status of p38 MAPK under basal culture conditions as well as following serum stimulation (Fig. 9C). In PiT1-depleted cells, the phosphorylation status of p38 was strongly increased when compared with control cells, whereas the total amount of p38 protein was not changed. Importantly, phosphorylation of p38 was unaffected following transfection of the cells by a PiT2 siRNA (Fig. 9C), showing that increased phosphorylation of p38 was not due to a decrease in Na ϩ -P i FIGURE 4. Knockdown of PiT1 in HepG2 cells leads to reduced P i transport and impaired proliferation. A, relative levels of NPT mRNAs in HepG2 cells using real-time RT-PCR reveals that PiT1 and NPT1 transporters are most abundantly expressed in HepG2 cells followed by that of PiT2; other NPTs are not significantly detected. Error bars indicate S.E. B, stable knockdown of PiT1 in HepG2 cells using shRNA. Proteins from HepG2 cells were extracted from untransfected cells (HepG2) or clones stably transfected with shScramble or shPiT1 in pSUPER vector, as indicated. PiT1 expression was analyzed by Western blot using anti-PiT1 antibody. C, Na ϩ -dependent P i uptake in shScramble (square) and shPiT1 (circle) HepG2 cells were calculated by subtracting uptake measured under Na ϩ -free conditions from that observed in the presence of Na ϩ at each KH 2 PO 4 concentration. D, HepG2 cells were grown in complete medium, stably transfected with shScramble (square), or shPiT1 (circle), and proliferation was assessed by counting the cells on the indicated days.
transport. To gain insight into the possible link between PiT1 and p38 phosphorylation, we monitored the proliferation of PiT1-depleted HeLa cells in the presence of 2.5-20 M p38 inhibitors SB202190 and SB203580. The results show that under these conditions, the proliferation rate of PiT1-depleted HeLa cells was not restored to wild-type levels (data not shown). We confirmed these data by knocking down p38 RNA synthesis using 100 mM p38 siRNA, arguing against a direct role of p38 in the PiT1-mediated effect on cell proliferation.

DISCUSSION
In this work, we have identified a novel function of PiT1 that is related to cell proliferation and is independent from its previously known Na ϩ -P i transport activity. Our results further indicate that PiT2, the second member of the mammalian PiT family, does not share this function. We propose that this function may represent one of the main physiological roles of PiT1 and that sequence-specific regions of the PiT1 protein could underlie the differences observed between PiT1 and PiT2 protein function (Fig. 10).
P i supply to the mammalian cell relies on the activity of the ubiquitously expressed Na ϩ -dependent P i cotransporters PiT1 and PiT2, which utilize the Na ϩ electrochemical gradient as the driving force to mediate the uphill import of P i . Given the combined role of P i in cellular processes and the function of PiTs as ubiquitous suppliers of P i to the cell, we addressed the question of the role of these transporters in cell proliferation. In the course of our study, we showed that under normal physiological conditions and with an unchanged P i availability to the cell, PiT1 expression was instrumental to cell proliferation. To our knowledge, this is the first observation that a Na ϩ -P i transporter is directly involved in regulating cell proliferation. More importantly, we were able to dissociate the Na ϩ -P i transport activity of PiT1 from its effect on cell proliferation, revealing a new distinct function for this membrane protein. This result is of fundamental importance because it opens new avenues on the physiological relevance of such transporters.
It has long been known that the membrane potentials in nonexcitable cells fluctuate during their progression through the cell cycle (for review, see Ref. 22). Typically, hyperpolarized somatic cells do not undergo mitosis, whereas developing and cancerous cells tend to be depolarized and mitotically active  (23,24). A recent study providing a thorough investigation of PiT1 transport properties shows that the apparent K m for P i was independent of the test potential, suggesting that P i did not interact with the transmembrane electric field (18). Similarly, the maximum electrogenic activity (I P i max ) showed a rectifying behavior with no evidence of rate-limiting behavior at the hyperpolarizing limit (18). Although a specific study is required, this may indicate that the activity of PiT1 may only be slightly influenced by membrane potential, meaning that the PiT-mediated P i transport may not be related to the proliferation state of the cell. Consistent with this hypothesis, we show that decreasing Na ϩ -P i transport does not decrease cell proliferation per se because cells displaying different V max for P i could proliferate at the same rate, providing that they have a normal PiT1 expression.
Our work identifying a novel function of PiT1 raises questions about the physiological function of this protein. Indeed, instead of being a ubiquitous supplier of P i for cellular needs, PiT1 could be a regulator of cell proliferation, at least in certain cell types or under certain physiological conditions. Considering the essential role of P i in metabolic and structural processes, one might have expected that a proliferation-related role of PiT1 would have been in relation to its P i transport function. We show that this is not the case. The quantity of P i molecules transported by the high affinity low capacity Na ϩ -P i transporter PiT1 may actually not represent the bulk of P i necessary for cellular needs, and alternative large capacity low affinity P i transporters may exist, as described for renal cells (25,26). Hence, although PiT proteins do possess the ability to transport P i , their main physiological functions may not be to provide cells with a sufficient amount of inorganic phosphate. In line with this hypothesis, it has been shown previously that transport-deficient mutants of PiT2

TABLE 2 FACS analysis of HeLa cells transiently or stably depleted from PiT1
For transient transfection, cells were transiently transfected with 10 nM siControl, siPiT1, or siPiT2 siRNA. Forty-eight hours after transfection, asynchronous cells were stained with propidium iodide, and their DNA content was determined by FACS analysis. For stable transfection, the DNA content was determined by FACS analysis of asynchronous stable HeLa cell clones stably transfected with the indicated shRNAs, which was determined 48 h after seeding. Results are expressed as the percentage of cells in the respective phase. *, p Ͻ 0.05 versus siControl or shScramble.  33) elegantly showed that the intracellular loop of GLUT2 could convey a glucose signal from the plasma membrane to the nucleus. Our data show that when PiT1 is depleted from the cell, there is an increase in the phosphorylation status of p38 MAPK. These results are in line with the reported key role of p38 in delaying the G 2 /M transition (34,35) under various environmental stress conditions and the inhibitory roles in cell proliferation and tumor progression (36). Moreover, our results obtained under basal culture conditions are consistent with a recent report showing that p38 could regulate the timing of mitotic entry under non-stress conditions (37). Despite these facts, we were unable to rescue the decrease in proliferation following PiT1 depletion by blocking p38 activity or synthesis, arguing against a direct role of p38 in the PiT1-mediated effect on cell proliferation.
In conclusion, we have identified PiT1 as a critical protein for cell proliferation. This constitutes a novel function, as this effect does not stem from the Na ϩ -P i cotransport activity of PiT1. We show that this function is specific to PiT1 because PiT2 does not share this function. Although p38 MAPK shows an increase in phosphorylation status, the mechanisms by which PiT1 modulates cell proliferation require further inves- Other residues (Asp-43, Glu-70, Asp-534, and Glu-603) indicated in blue are conserved residues corresponding to those that have been shown to be essential for transport function of PiT2 (31, 38) but have not been investigated for PiT1 transport function. Conserved GANDVANA signature sequences critical for transport (31) are indicated in light blue. Light green shading indicates the region important for GALV binding (39). Yellow shading indicates the regions that are not present in members of the PiT family from plants and bacteria. Extensive discussion on PiT topology can be found in Refs. 31 and 27. Ext., extracellular; Int., intracellular. B, stable expression of a cleavage-resistant version of PiT1 (PiT1-RNAiR) in shPiT1 HeLa cells. Immunofluorescence on living cells shows that S128A and S621A mutants of PiT1 are expressed at the plasma membrane of HeLa cells. C, Western blot analysis of serine mutants of PiT1 shows that they are expressed at a level comparable with the parent version of PiT1. tigation, notably through the identification of PiT1 protein partners.