The Cellular Prion Protein PrPc Is Involved in the Proliferation of Epithelial Cells and in the Distribution of Junction-Associated Proteins

Background The physiological function of the ubiquitous cellular prion protein, PrPc, is still under debate. It was essentially studied in nervous system, but poorly investigated in epithelial cells. We previously reported that PrPc is targeted to cell–cell junctions of polarized epithelial cells, where it interacts with c-Src. Methodology/Findings We show here that, in cultured human enterocytes and in intestine in vivo, the mature PrPc is differentially targeted either to the nucleus in dividing cells or to cell–cell contacts in polarized/differentiated cells. By proteomic analysis, we demonstrate that the junctional PrPc interacts with cytoskeleton-associated proteins, such as gamma- and beta-actin, alpha-spectrin, annexin A2, and with the desmosome-associated proteins desmoglein, plakoglobin and desmoplakin. In addition, co-immunoprecipitation experiments revealed complexes associating PrPc, desmoglein and c-Src in raft domains. Through siRNA strategy, we show that PrPc is necessary to complete the process of epithelial cell proliferation and for the sub-cellular distribution of proteins involved in cell architecture and junctions. Moreover, analysis of the architecture of the intestinal epithelium of PrPc knock-out mice revealed a net decrease in the size of desmosomal junctions and, without change in the amount of BrdU incorporation, a shortening of the length of intestinal villi. Conclusions/Significance From these results, PrPc could be considered as a new partner involved in the balance between proliferation and polarization/differentiation in epithelial cells.


Introduction
The cellular prion protein (PrP c ) is a ubiquitous glycoprotein anchored to the outer leaflet of the plasma membrane, in raft domains, through a glycosylphosphatidylinositol (GPI) moiety [1]. Its central role in transmissible spongiform encephalopathies has been clearly demonstrated for many years [2][3][4] and efforts have been made to determine its biological role apart from pathological situations. Although many cells and tissues, such as blood lymphocytes, muscle, heart, kidney, digestive tract and skin, express PrP c [5][6][7][8][9], most of the studies concerning its physiological function have been performed on nerve cells. In these models, it has been established that PrP c binds copper [10] and can homodimerize [11] or interact with other proteins, among which are synapsin Ib, Grb2, Pint1, LRP/LR, and N-CAMs [12][13][14][15][16][17][18][19]. It has also been reported that PrP c , via interaction with phosphorylated Fyn [20], participates in cell redox homeostasis through ROS production [21]. In addition, it has been shown that multiple biochemical changes occur in prion protein knockout mice. They include increased levels of NF-kB and COX-IV and decreased levels of p53 and Cu, Zn superoxide dismutase activity, along with an increased neuronal sensitivity to oxidative stress in cultured cells isolated from these mice [22].
Much less is known about the role of PrP c in extra-neuronal tissues. In epithelial cells, PrP c was reported to be directed to basolateral membranes of MDCK and FRT epithelial cells [23]. We have shown that PrP c is expressed in enterocytes [24], which are highly polarized epithelial cells of the intestinal epithelium. Interestingly, we showed that, in polarized/differentiated enterocytes, PrP c is targeted to the lateral junctional complexes of adjacent cells where it interacts with Src kinase [24]. This tyrosine kinase is known targeted to cell-cell junctions where it phosphorylates substrates that induce adhesion turnover and actin remodeling [25]. Such a localization of PrP c , was also observed in human keratinocytes [24] and in endothelial cells [26], opening questions about the role of PrP c in cell-cell junctions of physiological barriers.
To address this question, we focused our study on intestinal epithelium and on enterocytes, the major cell population of this epithelium. Intestinal epithelium undergoes a rapid renewal throughout life (for review see [27]). Such a process requires a continuous coordination between proliferation, differentiation and death programs, along with a remodeling of cell-matrix and cellcell contacts responsible for cell polarization. In crypts are localized stem cells and dividing cells, which migrate up to the villus where differentiation takes place. In the present work, we have analyzed whether the sub-cellular localization of PrP c varied in relation with cell proliferation, cell polarization and the state of cell-cell junctions in human intestinal epithelium in vivo and in the human Caco-2/TC7 enterocytes [28], which reproduce in culture the sequence of division and polarization/differentiation. In this enterocyte model, we characterized the partners of PrP c in cellcell junctions. Finally, the impact of the invalidation of PrP c on the distribution of cell-cell junctions-associated proteins, the process of cell proliferation and the morphology of the intestinal epithelium was analyzed.

Cells treatments
Cycloheximide treatment. When indicated, the cells were treated with cycloheximide (10 mM final concentration).
siRNA transfection. siRNA corresponding to the human Prnp gene from codon 399 to 417 was synthesized by MWG Biotech (Ebersberg, Germany). The specific human siRNA sequence used was: 59-GCC-GAG-UAA-GCC-AAA-AAC-CTT-39 (sense). Cells were seeded at 5000 cell/cm 2 on plastic or glass lamellae. siRNAs were mixed with Oligofectamine reagent (Invitrogen Life Technologies) for 15 min and Opti-MEM medium without serum was added according to the manufacturer's instructions. The final concentration of siRNA was 400 nM. After incubation for 6 hours at 37uC, Opti-MEM supplemented with 60% fetal calf serum was added to reach a final 20% serum concentration. A mouse PrP c siRNA sequence (59-GCC-CAG-CAA-ACC-AAA-AAC-CTT-39), inefficient on human PrP c RNA, and a scramble siRNA (59-CCG-AGA-AGU-AAA-GCC-AAC-CTT-39) were used as controls along with cells incubated with Oligofectamine reagent only. After 24 h, the medium was changed for the standard medium and the cells from all conditions were maintained in culture for the indicated times, the medium being changed every day.

GFP-PrP c plasmid construct and cell transfection
pEGFP-mouse PrP c plasmid [30] was obtained from MA Prado. GFP-PrP c was originally under the control of CMV promoter in this plasmid. To allow the expression of GFP-PrP c in differentiated cells, the corresponding sequence was subcloned into pGL2basic vector (Promega France) under the control of SV40 promoter. Caco-2/TC7 cells were transfected on day 2 with oligofectamine (Invitrogen, France) according to the manufacturer's instructions. After selection with G418 (Gibco BRL, France), transformed cells were allowed to expand prior to sorting, on the basis of GFP fluorescence, in an ALTRA cell sorter (Beckman Coulter, France).

Immunofluorescence analysis
Cells were washed twice with PBS containing 1 mM CaCl 2 and 0.5 mM MgCl 2 , and fixed for 30 min with 4% paraformaldehyde (wt/vol) in PBS at 4uC. After an extensive washing in 150 mM glycine in PBS (PBS-glycine), the cells were permeabilized by incubation for 30 min in 0.1% Triton X-100 in PBS and washed in PBS-glycine followed by PBS plus 1% BSA. Cells were incubated for 60 min at room temperature with primary antibodies in PBS supplemented with 1% BSA, washed with PBS and stained with secondary antibodies in PBS with 1% BSA for 60 min at room temperature in the dark. After extensive washing in PBS, cells were mounted in Fluoprep (BioMérieux, Marcy l'Etoile, France), and examined by epifluorescence microscopy (Axiophot microscope connected to Axiocam camera using Axiovision 4.5 software; Carl Zeiss). Confocal microscopy (LSM 510 microscope; Carl Zeiss, Jena, Germany) was used for the observation of cells cultured on microporous filters, which are strongly autofluorescent and generate excessive background in epifluorescence. X-Z planes resulted from 0.1 mm stacks.

Tissue analysis
PrP c knock out mice [2] backcrossed on C57Bl6 and their wild type C57Bl6 counterparts were purchased from CDTA (Orléans, France). After removal of intestine in wild type or PrP c knock out mice, 2 cm segments of duodenum and jejunum were cut, gently flushed with PBS and opened longitudinally. Rolled segments were frozen in cryo-embedding media (OCT) and stored at 280uC until cryostat sectioning (10-20 mm). Sections were applied onto gelatine-coated glass slides, fixed in paraformaldehyde solution (4%), permeabilized with Triton before DAPI staining and then mounted in fluoprep solution.
Mitotic crypt cells were labeled with an anti-phosphoS10-Histone H3, using the same protocol as for cell immuno-labeling. To label proliferating intestinal crypt cells in S-phase, PrP c knock out and wild-type C57Bl6 mice were given an intraperitoneal injection of 5-bromo-29-deoxyuridine (BrdU; Sigma; 120 mg/kg body weight) 90 min before sacrifice. Paraffin sections of alcoholformalin-acetic acid (AFA)-fixed jejunum were incubated for 30 min in 2.5 N HCl before processing for immunofluorescent labeling with anti-BrdU antibody.
Sections from paraffin-embedded human jejunum were sequentially treated with xylene (265 min), 100% EtOH (265 min), 95% EtOH (165 min) and then rinsed with water. Antigens retrieval was performed by boiling slides in 10 mM citrate buffer (pH 6) for 10 min. After washes in PBS, sections were fixed and then the same protocol as described above for immunofluorescence was used.

Immunoelectron microscopy
Caco-2/TC7 cells, grown on Thermanox coverslips (Agar scientific), were fixed with acetone. After incubation with 12F10 monoclonal antibody, gold (1 nm particles)-labelled goat anti mouse IgG (Amersham Biosciences) were used as secondary antibodies and the resulting signal was enhanced by the Intense TM M silver enhancement kit (Amersham Biosciences). After alcohol-graded dehydratation, sections were embedded in Epon and ultrathin sections were analyzed (Jeol 100 CX II). For desmosomal structure analyses, intestine was flushed with cold 0.1 M Phosphate buffer (pH 7.4), ligatured and filled with cold 2.5% glutaraldhehyde/0.5% tannic acid in 0.1M cacodylate buffer at pH 7.4 for 2H. Fine samples of intestine were cut and fixed by 0.6% glutaraldhehyde/0.5% tannic, which stains and preserve the ultrastructure of phospholipids [31] in the same buffer overnight at 4uC. Postfixation was carried out in 2% osmic acid in 0.1 M Phosphate buffer for 1h at 4uC. Samples were then dehydrated in graded alcohol and embedded in Epon resin (Poly/Bed 812, Polysciences Inc.Warrington, PA). Ultrathin sections of around 65 nm were counterstained with uranyl acetate (30 min, 40uC) and lead citrate (10 min, 25uC) using an LKB 2168 ultrostainer. Observations were made using a JEOL CX100 equipped with a Gatan Digital camera (3.11.0) and the migrographs were processed with Gatan software.

Sub-cellular fractionation
Preparation of detergent-insoluble membranes on sucrose gradient. Caco-2/TC7 cells were homogenized on ice for 30 min in 2 ml of 10 mM Tris-HCl pH 8, 150 mM NaCl buffer (TN) containing 1% Triton X100 or in 2 ml of 20 mM Tris-HCl pH 7.8, 250 mM sucrose, 1 mM CaCl 2 and 1 mM MgCl 2 without detergent [32]. Anti-protease cocktail and antiphosphatases (orthovanadate and beta-glycerophosphate) were added in both conditions. The cell homogenate was then adjusted to 40% sucrose by addition of 2 ml sucrose (80% in TN). The resulting 4 ml were covered with 4 ml of 30% sucrose and 4 ml of 5% sucrose and centrifuged for 16 h (39,000 rpm, 4uC) in a SW-41 rotor (Beckman Instruments, Gagny, France). Sequential 1 ml fractions were then collected from the top of the tube and the turbid fraction containing the floating detergent-insoluble membranes (fraction 4) was adjusted to 11 ml in TN buffer and centrifuged in a SW-41 rotor (35000 rpm, 1h). The pellet was dissolved in TN buffer containing 1% NP40, anti-proteases and anti-phosphatases and stored at 280uC until immunoprecipitations.
Nuclei and crude membrane preparations. Proliferating or polarized/differentiated Caco-2/TC7 cells were washed twice in 10 mM Tris-HCL pH 7.5 containing 20 mM KCl, 2 mM CaCl2, 2 mM MgCl2 and 0.2 mM spermidine (TKCM buffer) and scrapped in TKCM containing 1% Triton X-100, 1 mM phenylmethylsulfonylfluoride (PMSF), anti-proteases and antiphosphatases. Nuclei were pelleted by centrifugation at 1000g for 10 min at 4uC and supernatants corresponding to cytosolic and membrane proteins were stored at -80uC until analysis. The pellets were washed in TKCM buffer and nuclear proteins were extracted with 2M NaCl in TKCM buffer for 1h at 4uC. Excess NaCl was removed by overnight dialysis against PBS at 4uC.

Immunoprecipitation and immunoblotting analyses
The raft fractions were immunoprecipitated with rabbit anti-PrP c (Ab 703) polyclonal antibodies, or non-specific rabbit immunoglobulins or anti-pan desmoglein antibodies coupled to protein A-sepharose 4B (Amersham Biosciences, Orsay, France). For SDS-PAGE, samples were boiled for 10 min in Laemmli buffer (2.5% SDS final concentration) and fractionated under reducing conditions in polyacrylamide gel. Proteins were transferred onto nitrocellulose membranes (Bio-Rad), blocked 2h with 5% non-fat dried milk in TBS/0.1% Tween 20 (TBST). After two washes in TBST, membranes were incubated (1h at room temperature) with specific antibodies. After three washes in TBST, the membranes were incubated (1h at room temperature) with peroxidase-labelled (HRP) secondary antibodies (Amersham Biosciences) in TBST. After three washes in TBST, bound antibodies were detected by chemiluminescent method (ECL, Amersham Biosciences). The quantitative analyses were performed with a high performance calibrated imaging densitometer (Bio-Rad GS-800) using PD Quest and Image Quant 5.2 softwares.
GPI anchor detection. After SDS-PAGE of immunoprecipitated materials and transfer onto nitrocellulose, membranes were incubated two hours in a binding buffer (50 mM NaH2PO4/0.3% Tween 20) before addition of biotinylated pro-aerolysin bacterial toxin. Biotinylated proteins were detected by blotting with HRPconjugated streptavidin.
Endo F treatment. Nuclear proteins and rafts extracts were treated with Endoglycosidase F (10 units/50 mg proteins) in 40 mM sodium phosphate buffer, pH 7.5 containing 0.4% SDS, 20 mM DTT and 0.8% NP40 for 16 hours at 37uC before immunoblot analysis.

MS analysis
SDS/PAGE separation and protein digestion. Raft fractions were immunoprecipitated with anti PrP c antibodies and separated on 4-12% SDS/polyacrylamide gels. After staining with colloidal Commassie blue (G250, Bio-Rad), the visualized bands were cut into slices of 1 mm. Gel slices were then reduced, alkylated and subjected to digestion with trypsin (Roche Diagnostics) as already described [33]. Extracted peptides were dried and solubilized in solvent A (95/5 water/acetonitrile in 0.1% (w/v) formic acid). The total digestion product of a gel slice was used per liquid chromatography-tandem MS (MS/MS) analysis.
Liquid chromatography-MS/MS analysis. The extracted peptides were concentrated and separated on a LC-Packing system (Dionex S.A.) coupled to the nano-electrospray II ionisation interface of a QSTAR Pulsar i (Applied Biosystems) using a PicoTip (10 mm i.d., New Objectives, Woburn, MA). The MS/ MS data was searched twice by using MASCOT (Matrix Science, London) and PHENYX (Geneva Bioinformatics S.A.) softwares on internal servers, first without taxonomic restriction to reveal the presence of proteins of interest and mammalian contaminants, then again the National Center for Biotechnology Information Human database (National Library of Medicine, Bethesda). All data are manually verified in order to minimise the errors in protein identification and/or characterization.

Statistical analysis
Statistical analyses were performed using student's t test.

Results
The cellular prion protein is localized in the nucleus in dividing cells and in cell-cell junctions in polarized epithelial cells We analyzed, by immunofluorescence and immunoelectron microscopy, the distribution of PrP c or of GFP-PrP c in exponentially growing or polarized Caco-2/TC7 enterocytes. Representative images of PrP c , E-cadherin and DAPI labeling of the nuclei in exponentially growing Caco-2/TC7 cells (day 3) are shown in Figure 1A. When cells have not yet established welldefined adherens junctions, as shown by the poor expression of Ecadherin at cell-cell contacts (left panel), PrP c was mainly detected intracellularly. Interestingly, this staining co-localized with DAPI labeling, in the nucleus (middle panel). Immunodetected PrP c appeared as dots that were distributed around the nucleolus (right panel). This localization was confirmed by immunoelectron microscopy where the PrP c signal appeared accumulated in the nucleus (Fig. 1B, N) and systematically excluded from the nucleolus (Fig. 1B, * ). The nuclear localization of the transfected mouse GFP-PrP c at this stage of the culture further strengthened the results obtained for the endogenous protein (Fig. 1C).
In confluent and polarized Caco-2/TC7 cells (day 10), when Ecadherin-dependent junctions were established, PrP c was detected at the lateral membrane, and did not co-localize with DAPI in 90-95% of the cells (Fig. 1D). GFP-PrP c was also found targeted to the lateral membranes of polarized cells (Fig. 1E). This PrP c localization was previously revealed by immunoelectron microscopy [24]. As cells were not synchronized, PrP c was found in the nucleus in the few dividing cells (5-10%) that were still present within the confluent cell layer. In all conditions, a significant proportion of trafficking PrP c , which corresponded to approximately 30-40% of the total protein, was also observed in the cytoplasm.
The cellular prion protein is differently localized in proliferative and differentiated compartments of human intestinal epithelium The sub-cellular localization of PrP c was analyzed in human intestinal epithelium and compared in crypts and villi, which correspond to the proliferative and differentiated compartments, respectively. In crypts, PrP c was found to co-localize with Ki-67, a nuclear marker of dividing cells (Fig. 2A). By contrast, in the cryptvillus transition compartment (Fig. 2B) and in villi (Fig. 2C), i.e. as soon as the process of cell division is stopped and the differentiation takes place, the nuclei, visualized by DAPI, were found devoid of Ki67 and of PrP c , which was visualized in the cytoplasm and in the lateral membranes of adjacent cells.
Nuclear and junctional PrP c isoforms exhibit similar posttranslational modifications but differ in their stability The differential sub-cellular localization of PrP c could result from different molecular properties of the protein. Thus, we first compared the stability of PrP c when localized in the nucleus or in the lateral membranes. The amount of PrP c was determined by western blot, in nuclear extracts from proliferative cells (day 3) and membrane PrP c -containing raft domains from differentiated Caco-2/TC7 cells (day 10), at different times after inhibiting translation by cycloheximide. Results reported in Figure 3A show that PrP c is much more stable when localized in the nucleus, where degradation could not be observed over a 3 hour period, than at the membrane, where 50% of the protein were degraded at 30 min.
PrP c is submitted to post-translational modifications such as Nlinked glycosylation and addition of a GPI anchor. Endo F treatment of nuclear or raft extracts resulted in an equivalent shift of PrP c bands, demonstrating that nuclear and raft PrP c are similarly N-glycosylated (Fig. 3B). Analysis of the presence of a GPI anchor was performed on PrP c immunoprecipitated by the specific polyclonal rabbit antibodies Ab 703, using the biotinylated pro-aerolysin bacterial toxin, which recognizes GPI anchors [29]. A specific band corresponding to the presence of a GPI anchor on PrP c was detected in both nuclear and raft extracts (Fig. 3C, middle panel), at the same molecular weight as the PrP c signal ( Fig. 3C upper panel), i.e. at 30 kDa. The absence of contamination of the nuclear and raft extracts with proteins derived from the other compartment was verified. Calnexin protein, which is known to be regularly buoyed with rafts [34] and the poly-(ADP-ribose) polymerase (PARP) protein that is exclusively expressed in the nucleus were used as markers of the respective compartments (Fig. 3C, lower panels).
Junctional PrP c is part of a complex involving desmosome-associated proteins and c-Src, and interacts with the structural proteins actin and annexin A2 A proteomic analysis was undertaken to determine the partners of PrP c in the junctional domains of enterocytes. Results presented in table 1 showed that the membrane PrP c interacts with five desmosome-associated proteins, among which desmoglein, plakoglobin and desmoplakin, and with gamma-and beta-actin and annexin A2, a structural protein that is known to participate in the regulation of actin cytoskeleton dynamics in junctions of epithelial cells [35]. The interaction with the desmosomal proteins and with annexin A2 was corroborated by western blots after purification of rafts in the presence of detergent and immunoprecipitation of PrP c by the specific rabbit polyclonal antibodies Ab 703 (Fig. 4A). Same results were obtained with rafts purified, after sucrose gradients, from cell extracts prepared without detergent (not shown). The absence of interaction between E-cadherin and PrP c [24] was confirmed, since E-cadherin was recovered exclusively in the immunoprecipitation supernatant (Fig. 4B). Interestingly, besides the already reported interaction of c-Src with the junctional PrP c [24], the immunoprecipitation of raft fraction with the antitransmembrane protein desmoglein antibody revealed a complex including c-Src, desmoglein, and PrP c (Fig. 4C), while, as expected, desmoglein and E-cadherin did not interact.

PrP c invalidation impairs the sub-cellular localization of junction-associated proteins and desmosome structure
Based on its interaction with desmosomal proteins, studies were undertaken to determine whether PrP c could be involved in the organization of cell-cell junctions. Enterocytes were thus treated with human PrP c -siRNA, before the onset of cell polarity and the formation of cell-cell junctions. A kinetic analysis of PrP c levels in transfected cells, by immunofluorescence, showed that in 50 to 60% of the cells the levels of the endogenous PrP c were dramatically decreased 2 days (not shown) or 3 days after transfection (Fig. 5A) as compared to control cells, i. e. cells transfected with an inefficient mouse PrP c -siRNA or with a scramble siRNA or cells incubated with Oligofectamine only, and returned to the control levels from 4 days after transfection (not shown). In a first attempt, the junctional state was assessed 3 and 7 days after transfection by the analysis of the expression and localization of E-cadherin, which does not interact with PrP c but is the most studied marker of cell-cell junctions. Three days after transfection (i.e 6 days after seeding), junctional complexes were already formed in the center of expanding cell clusters. In cells where PrP c expression was specifically impaired, the E-cadherin still localized to the lateral membranes of adjacent cells but its labeling intensity was decreased (Fig. 5A). Moreover, in the same fields, the cells appeared enlarged as compared with the three control conditions, as clearly noticeable in phase contrast picture (Fig. 5A). When the effect of human siRNAs on the expression of PrP c was no longer observed (day 7, Fig. 5A), the size of the cells and the amount of the junctional E-cadherin were rescued. We then analyzed the expression and the sub-cellular localization of the junctional partners of PrP c by immunofluorescence (Fig. 5B). To better compare cells that still expressed PrP c (40-50%) or not (50-60%), fields that combined the two cell populations of human PrPc-siRNA transfected cells are shown (Fig 5B, right panels) along with pictures representative of the results obtained in the three control conditions (Fig. 5B, left panels). Cells exhibiting a net decrease of PrP c levels were systematically enlarged. In these cells, the amount and/or the sub-cellular localization of the junctional partners of PrP c were modified: c-Src was found essentially intracellular, desmoglein, plakoglobin and desmoplakin labelings were lowered, and numerous actin stress fibers could be visualized in large cells that no longer expressed PrP c .
The role of PrP c on structural organization of cell-cell junctions was further analyzed in intestinal epithelium of PrP c knock out mice. Ultrastructural analysis revealed drastic changes in the length of the desmosomal plaque (Fig. 5C upper panel), which was not compensated by their number (not shown). Quantification of the length of desmosomes indicated a distribution of their size concentrated between 83 to 134 nm in knock out mice instead of 100 to 183 nm in wild type mice (Fig. 5C lower panel).

PrP c invalidation impairs completion of cell division and results in a shortening of intestinal villi
When compared to the three control conditions, we noticed an enlargement of human PrP c siRNA-treated dividing enterocytes (Fig. 5A) that could reflect the impact of PrP c -expression on cell proliferation. Analysis of cell growth (Fig. 6A) showed an arrest of the increase in cell numbers between 2 and 3 days after transfection, i.e during the period when PrP c expression was significantly decreased by human siRNAs. Surprisingly, DAPI staining of the nuclei showed that growth arrest was paralleled with a net increase of the number of mitosis, without apoptosis, in cells that did not express PrP c anymore (Fig. 6B). In parallel, the morphological examination of the intestinal epithelium of PrP c knock out mice revealed a net decrease in the length of the villi in both duodenum and jejunum segments in comparison with their wild type counterparts (Fig. 6C). To analyze the impact of PrP c expression on cell proliferation within the intestinal epithelium, PrP c knock out and wild type mice were pulse-labeled with BrdU for 1.5 hour prior to sacrifice. No significant difference in the number of BrdU-labeled cells in sections of jejunal crypts was observed between wild type and knock out mice (Fig. 6D), suggesting that S phase was not affected. By contrast, labeling of the mitosis-associated phospho-H3 revealed a net increase of mitotic cells in the crypt compartments of intestine from PrP c knock out mice as compared with wild type mice, while no obvious change in the size of the crypts was observed. Bands obtained in western blots (SAF 32 antibody) were quantified by scanning densitometry. E-cadherin and PARP were used to normalize the values obtained in membrane and nuclear fractions respectively, since both proteins were found stable for the duration of CHX treatment. Histograms correspond to the ratio (%) between PrP c and E-cadherin or PARP from the corresponding scanned bands at each time (mean6SD from 3 independent experiments), the value obtained at time 0 being set at 100. (B): To determine the glycosylation state, rafts and nuclear extracts were treated (+) or not (2) with endo F and PrP c was analyzed by western blot (SAF 32 antibody). Molecular weight in KDa are indicated (C): The presence of a GPI anchor was analyzed after immunoprecipitation of PrP c from rafts or nuclear extracts, SDS-PAGE, transfer and overlay with biotinylated proaerolysin bacterial toxin. To check the purity of the extracts, the expression of calnexin (membrane marker) and PARP (nuclear marker) was analyzed by western blot. Molecular weight in KDa are indicated. doi:10.1371/journal.pone.0003000.g003

Discussion
Our present results demonstrate that the cellular prion protein, PrP c , exhibits a dual distribution between the nucleus, in actively dividing cells, and cell-cell adhesion sites in polarized/differentiated cells. Interestingly, the membrane PrP c interacts with desmosomal proteins as well as with actin and actin-binding proteins at cell-cell junctions. Furthermore, we show that down regulation of PrP c expression in Caco-2/TC7 enterocytes lead to a complex pattern of alterations in both cell architecture and completion of the cell division process. These results are strengthened by the analysis of the intestinal epithelium of PrP c knock out mice, in which intestinal villi were found shortened and the size of enterocyte desmosomes decreased as compared to wild type mice.
Contrary to abnormal prion proteins [36][37][38], a targeting of the normal PrP c isoform to the nucleus has been rarely reported, [39,40]. We demonstrate here that in intestinal epithelial cells, such a nuclear targeting is observed only in actively dividing cells, both in cultured enterocytes and in the intestinal epithelium in vivo. The characterization of the nuclear pattern of PrP c isoforms revealed that it is similar to that of membrane-associated PrP c in terms of glycosylation and presence of a GPI anchor. Analysis of protein stability shows a much longer half-life of PrP c in the nucleus than in plasma membrane, where junctional proteins are rapidly recycled. Nuclear PrP c stability could rely on the association with particular sub-nuclear compartments, such as PML bodies [41], a localization compatible with the pattern of nuclear PrP c staining that we observed. Contrasting with the misfolded protein [38], we show an exclusion of normal PrP c from the nucleolus. Altogether, our results asked the question of its biological role in the nucleus of dividing cells. Upon PrP c down regulation in cultured enterocytes, we observed modifications of cell morphology and an arrest of cell growth. This observation was consistent with the decreased villus size of the intestinal epithelium that we observed in PrP c null mice. In several mouse models, it has been shown that reduced intestinal crypt cell proliferation is associated with shorter villi [42,43]. The growth arrest observed in PrP c siRNA-transfected cells was paralleled by an increase of mitotic cells. Further analysis of cell cycle perturbations was rendered difficult by the moderate transfection efficiency of Caco-2/TC7 cells by siRNA (50%). Nevertheless, the absence of S phase overt perturbation in crypts from PrP c null mice, as shown by BrdU incorporation experiments, along with the important increase of the number of mitotic cells in the crypt compartment suggest that PrP c invalidation would affect the mitosis process. PrP c has been shown to associate with tubulin [44] [45]. In addition, desmoplakin, that we identified as a PrP c partner, participates to the organization of microtubules in keratinocytes, through the recruitment at cell-cell junctions of a centrosomal protein, which is required for microtubule anchoring [46]. However, it cannot be concluded yet whether all the phases of mitosis or more particularly the last step of cytokinesis are slowed down through PrP c invalidation.    Another important finding of our study is the identification of desmosomal proteins as PrP c partners at cell-cell junctions. We had already identified c-Src kinase as a partner of PrP c in polarized enterocytes and demonstrated that PrP c does not interact with E-cadherin [24], asking the question of the partner that could link PrP c , which is anchored in the outer leaflet of the lateral membrane, to Src, which is localized in the inner one. Our data are consistent with the existence of a molecular complex in which the transmembrane desmosomal cadherin desmoglein is this link between PrP c and c-Src. Involvement of PrP c in the regulation of desmosome organization, and more generally in cell architecture, is further supported by our demonstration of alterations in the junctional targeting of Src and desmosomal proteins upon PrP c down-regulation by siRNA during the establishment of enterocyte cell-cell contacts. Importantly, absence of PrP c in vivo results in alterations in the size of desmosomes in intestinal epithelial cells. This structural perturbation could reflect modifications of assembly and/or stability of the desmosomal-protein complex in the absence of PrP c , as described for invalidation or mutation of other desmosome-associated proteins, such as desmoglein, plakoglobin or plakophilin families, in murine models or human pathologies (for review, see [47]). Desmosomal proteins have emerged as adhesion molecules that not only play critical structural roles but are also involved in signaling pathways [48]. Our data further support the hypothesis of the involvement of PrP c , in association with its desmosomal partners, in a signaling platform of the junctional state.
In conclusion, our results indicate that the normal cellular prion isoform PrP c , by its differential nuclear or junctional localization, must be considered as a potential actor of the balance between proliferation and polarization/differentiation of epithelial cells, through interaction with c-Src and with desmosome-and cytoskeleton-associated proteins. As such PrP c could be involved in the homeostasis of renewing epithelia.