The Prion Protein and Its Paralogue Doppel Affect Calcium Signaling in Chinese Hamster Ovary Cells

Abstract

The function of the prion protein (PrP^c), implicated in transmissible spongiform encephalopathies (TSEs), is largely unknown. We examined the possible influence of PrP^c on Ca²⁺ homeostasis, by analyzing local Ca²⁺ fluctuations in cells transfected with PrP^c and Ca²⁺-sensitive aequorin chimeras targeted to defined subcellular compartments. In agonist-stimulated cells, the presence of PrP^c sharply increases the Ca²⁺ concentration of subplasma membrane Ca²⁺ domains, a feature that may explain the impairment of Ca²⁺-dependent neuronal excitability observed in TSEs. PrP^c also limits Ca²⁺ release from the endoplasmic reticulum and Ca²⁺ uptake by mitochondria, thus rendering unlikely the triggering of cell death pathways. Instead, cells expressing Doppel, a PrP^c paralogue, display opposite effects, which, however, are abolished by the coexpression of PrP^c. These findings are consistent with the functional interplay and antagonistic role attributed to the proteins, whereby PrP^c protects, and Doppel sensitizes, cells toward stress conditions.

INTRODUCTION

The cellular prion protein (PrP^c) is a highly conserved cell surface glycoprotein, particularly expressed in the CNS with a still unrecognized function. A conformationally modified isoform (PrP^Sc) of PrP^c is the major component of prions, the etiological agent at the basis of fatal neurodegenerative disorders, called transmissible spongiform encephalopathies (TSEs). TSEs present as sporadic, genetic, and infectious illnesses, and include Creutzfeldt-Jakob disease (CJD) in humans, and bovine spongiform encephalopathy in cattle (Prusiner, 1998).

The mechanism of PrP^c conversion into PrP^Sc is not yet elucidated, nor it is clear if the disease progression relates to PrP^Sc toxic effect, or to the deprivation of PrP^c functionality. The most prominent evidence in support of the former hypothesis is that most PrP-knockout mice remain viable, and do not develop spontaneous neurodegeneration (Bueler et al., 1992; Manson et al., 1994), even if the PrP^c gene is postnatal deleted (Mallucci et al., 2002). An essential role of PrP^c in cell survival comes, however, from the finding that wild-type PrP transgenes abrogate the cerebellar degeneration and late-onset ataxia developed by some PrP-knockout lines over expressing a protein, named Doppel (Dpl; Sakaguchi et al., 1996; Li et al., 2000; Moore et al., 1999, 2001; Rossi et al., 2001). Dpl is normally absent in the CNS of adult animals and resembles truncated PrP^c; it lacks the copper-binding N-terminus (Brown et al., 1997) and is structurally and biochemically similar to PrP^c carboxyl end (Silverman et al., 2000; Mo et al., 2001). Also the role of Dpl in the CNS is obscure, but the capacity of PrP^c to restore the normal phenotype of Dpl-over expressing mice has suggested that PrP^c and Dpl have antagonistic functions and that PrP^c suppresses a death signal triggered by Dpl.

Comparative studies using cells from wild-type or PrP-knockout animals or infected by prions also argue in favor of “the-loss-of-function hypothesis,” suggesting that PrP^c protects cells by controlling the copper metabolism governing the cell resistance to oxidative stress (Brown et al., 2002) or by playing an antiapoptotic role (reviewed in Hetz et al., 2003a). Importantly, other data from electrophysiologic studies, and/or cell Ca²⁺ measurements, converge in supporting that PrP^c absence, or its recruitment into prions, leads to a compromised Ca²⁺ homeostasis and that such a defect ultimately impinges on a few Ca²⁺-dependent neurophysiologic functions, such as plasma membrane K⁺ currents, after-hyperpolarization potential (AHP) and depolarizing after-potential (DAP) events (Jefferys et al., 1994; Colling et al., 1996; Johnston et al., 1998; Barrow et al., 1999; Herms et al., 2000, 2001; Mallucci et al., 2002).

Yet, if PrP^c is involved in the correct handling of cell Ca²⁺, the above-reported findings are hardly surprising, given that Ca²⁺ is an ubiquitous intracellular messenger regulating a large amount of cell functions and that subtle alterations of endoplasmic reticulum (ER) and mitochondrial Ca²⁺ pools control initiation of cell death pathways. For example, both overload and depletion of the ER Ca²⁺ concentration ([Ca²⁺]) can impinge on correctly folded protein levels, thereby inducing (ER–) stress responses that may determine apoptosis (Orrenius et al., 2003). Accordingly, perturbed ER Ca²⁺ regulation contributes to neuronal degeneration (reviewed in Mattson et al., 2000; Paschen, 2001), including Alzheimer's disease and TSEs (Nakagawa et al., 2000; Hetz et al., 2003b). Also mitochondrial Ca²⁺ fluxes are integrated modulators of the cell Ca²⁺ signaling and intimately govern the cell fate. Indeed, mitochondria respond to high [Ca²⁺] microdomains that form in proximity of plasma membrane, and/or sarco/endoplasmic reticulum, Ca²⁺ channels upon cell stimulation (Brini, 2003). By activating the production of ATP, such a relayed signal satisfies immediate physiological energy demands (Jouaville et al., 1999). However, it can also turn into a death signal, given that excessive Ca²⁺ accumulation triggers abnormal mitochondrial membrane permeability (Bernardi, 1999) and release of proapoptotic factors to the cytoplasm (Li et al., 1997).

In view of the many implications related to the fine tuning of local Ca²⁺ changes, we have investigated on the involvement of PrP^c in Ca²⁺ homeostasis, using CHO cells transiently transfected with PrP^c and the recombinant Ca²⁺-sensitive photoprotein aequorin (AEQ) targeted to specific cell compartments. Furthermore, within the proposed functional antagonism of PrP^c and Dpl, we also monitored the Ca²⁺ metabolism in AEQ-transfected cells over expressing Dpl, either alone or together with PrP^c.

MATERIALS AND METHODS

Expression Plasmids

PrP^c and Dpl isoforms used in all experiments were the murine (m) and human (h) ones, respectively. The expression vector for eukaryotic cells containing the coding sequence for hDpl was constructed as in Massimino et al. (2004), whereas that coding for mPrP^c was constructed using a mouse genomic DNA as template, and by Prnp amplification by PCR with primers 5′ GCGCTAGCATGGCGAACCTTGGCTACTGGCTGC 3′; and 5′ CGGAATTCATCCCACGATCAGGAAGATGAG 3′. Amplified products were cut at NheI and EcoRI sites and cloned directly in the phEGFP plasmid (Clontech, Palo Alto, CA) between the same restriction sites to obtain plasmid pcDNAm-PrP. To express mPrP^c carrying the epitope specific for monoclonal antibody (mAb) 3F4, L108M, and V111M point mutations were introduced in pcDNAmPrP by inverse PCR, using primers 5′ CATATGGCAGGGGCTGCGGCAGCTGGGC 3′ and 5′ CTTCATGAAGGTTTTTGGTTTGCTGGGCT 3′. The amplified product was itself ligated after T4 polynucleotide kinase treatment, to yield plasmid pmPrP3F4. The expression plasmid for the fusion protein between the green fluorescent protein (GFP) and the glycosylphosphatidylinositol (GPI) attachment signal of bovine PrP^c (GFP-GPI_PrP) was as described in Cereghetti et al. (2004), whereas plasmids for the photoprotein AEQ targeted to the various cell compartments were obtained according to Brini et al. (1995) for cytAEQ; Montero et al. (1995, 2000) for erAEQ; Marsault et al. (1997) for pmAEQ; and Rizzuto et al. (1992) for mtAEQ. Recombinant (rec) mPrP (23–230; containing the mAb 3F4 epitope) was generated in, and purified from Escherichia coli, following the method described in Negro et al. (2000) for the bovine PrP isoform, whereas hDpl_rec (28–152) was obtained as in Cereghetti et al. (2004).

Cell Cultures and Transfection

Experiments were carried out with CHO cells, cultured in 75-cm² flasks (37°C, 5% CO₂ atmosphere), using Ham's F12 medium (EuroClone, Devon, United Kingdom) supplemented with 10% fetal calf serum (FCS), penicillin (100 U/ml), and streptomycin (100 μg/ml). The cells, seeded onto glass coverslips at 4.5 × 10⁴ and 3.5 × 10⁵ cells/cm² for fura-2 and aequorin measurements, respectively, were transiently transfected preferentially with the Lipofectamine Plus reagent (Invitrogen, Milano, Italy), following manufacturer's instructions (in double and triple transfection experiments the quantity of each added plasmid was proportionally reduced). To optimize PrP^c and Dpl expression, 6 h after transfection the medium was changed, and cells were kept at 30°C for 48 h before fura-2-, or aequorin-based, Ca²⁺ measurements, or immunolabeling assays. Importantly, identical results were obtained when cells were transfected using the Ca²⁺ phosphate-based procedure (Rizzuto et al., 1995) and a total amount of 3 μg of plasmid DNA (1.5:1.5 μg in the case of double transfection). To generate CHO cells stably expressing mPrP^c, the method was as described in Massimino et al. (2004).

Immunocytochemistry, Western Blotting, and Densitometric Analysis

Immunocytochemistry and image analysis of intact cells, immunoblotting, and densitometric analyses were carried out as in Negro et al. (2001) and Massimino et al. (2004), except that 20 μg of cell lysates were loaded in each gel lane.

Antibodies

For immunocytochemistry and immunoblotting of PrP^c and Dpl, the following antibodies were used (dilutions are given in parenthesis): anti-PrP mAb 8H4, raised against the human PrP 173–185 sequence (a kind gift of Dr. M.S. Sy, Case Western University, Cleveland, OH; 1/300 for immunolocalization; 1/8000 for immunoblotting), or mAb 3F4 (DAKO, Glostrup, Denmark; 1/100 for immunolocalization; 1/1000 for immunoblotting), which gave the same results as with mAb 8H4; and anti-Dpl mAb Dpl-79 (kindly provided by Dr. J. Grassi, Commissariat a l'Energie Atomique, Saclay, France) raised against hDpl_rec (28–152; 1/100 for immunolocalization; 1/1000 for immunoblotting). The immunoblotting of the sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) and β-actin was carried out using commercially available mAbs against the ubiquitous isoform (SERCA2; clone IID8, Affinity Bioreagents, Golden, CO; 1/1000), and β-actin (Sigma, St. Louis, MO; 1/4000), respectively, and anti-calreticulin polyclonal antibody (Affinity Bioreagents; 1/1000) was used for calreticulin immunolabelings.

Ca²⁺ Measurements: Aequorin

Although the cytAEQ plasmid coded for the wild-type AEQ, to reduce the affinity for Ca²⁺, erAEQ, mtAEQ, and pmAEQ plasmids expressed aequorins carrying a mutation in one Ca²⁺ binding site (Montero et al., 1995). Functional erAEQ was reconstituted with a modified prosthetic group (coelenterazine n, Molecular Probes, Eugene, OR), after depleting ER Ca²⁺ (Montero et al., 1995; Barrero et al., 1997). To this end, cells were incubated (1 h, 4°C) in KRB (Krebs-Ringer modified buffer: 125 mM NaCl, 5 mM KCl, 1 mM Na₃PO₄, 1 mM MgSO₄, 5.5 mM glucose, 20 mM HEPES, pH 7.4) supplemented with ionomycin (5 μM), EGTA (600 μM), and coelenterazine n (5 μM). After extensive washings with KRB containing 2% bovine serum albumin and 1 mM EGTA, the coverslip with the cells was transferred to the thermostatted (37°C) chamber of a purpose-built luminometer. Conversely, transfected cytAEQ, mtAEQ, and pmAEQ were reconstituted by incubating cells (1–3 h, 37°C, 5% CO₂ atmosphere) with wild-type coelenterazine (5 μM, Molecular Probes) in DMEM containing 1% FCS. Measurements started by perfusing cells with KRB supplemented with 1 mM CaCl₂. Alternatively, pmAEQ was reconstituted in cells incubated with coelenterazine in KRB containing EGTA (100 μM). In this case, experiments started with cells in the EGTA-supplemented KRB, which was then replaced by CaCl₂ (1 mM)-containing KRB. Other additions were as specified in the figures. Experiments ended by lysing cells with digitonin (100 μM) in a hypotonic Ca²⁺-rich solution (10 mM CaCl₂ in H₂O), to discharge the remaining aequorin pool. The light signal was collected and stored in an IBM-compatible computer for further analyses. Aequorin luminescence data were calibrated off-line into [Ca²⁺] values, using a computer algorithm based on the Ca²⁺ response curve of wild-type and mutant aequorins, as previously described (Brini et al., 1995).

Fura-2

Cells, cotransfected with the phEGFP plasmid (Clontech) and either PrP^c or Dpl, were loaded with fura-2 AM (5 μM, Molecular Probes) as in Malgaroli et al. (1987), and the coverslip was then placed on the stage of a Zeiss Axiovert 100 epifluorescence microscope (Göttingen, Germany), equipped with a 16-bit digital CCD videocamera (Micromax, Princeton Instruments, Trenton, NJ). Samples were alternately illuminated at 340 and 380 nm, and the emitted light (filtered with an interference filter centered at 510 nm) was collected by the camera. Images were acquired using the Metafluor software (Universal Imaging, West Chester, PA). The ratio values (1 ratio image/s) were calculated off-line, after background subtraction from each single image.

Statistical Analysis

Data are reported as means ± SD; n indicates the number of experiments. Statistical differences were evaluated by Student's two-tailed t test for impaired samples, with a p value lower than 0.05 being considered statistically significant.

RESULTS

Localization and Expression of Transiently Transfected PrP^c or Dpl

Before analyzing whether PrP^c, or Dpl, affects Ca²⁺ homeostasis, we ascertained their correct cell distribution by immunostaining CHO cells transiently transfected with either PrP^c or Dpl. As evident from the uniform labeling of the plasma membrane (Figure 1A), obtained by adding (4°C) anti-PrP-, or anti-Dpl-, mAbs to intact cells, PrP^c and Dpl are exposed to the exoplasmic space, in accord with the presence of a GPI anchor in both proteins, and with similar data reported previously (Massimino et al., 2004). Using dual-labeling immunocytochemistry, an identical localization of the proteins to the cell surface was observed in cells expressing PrP^c, or Dpl, together with the AEQ probes, and in those cotransfected with both PrP^c and Dpl (see also Massimino et al., 2004; unpublished data). Figure 1B (second lanes), reporting the immunoblots of PrP^c and Dpl in cells housing each protein alone, shows that the two cell types express similar protein amounts. Importantly, by applying immunocytochemistry and Western blot techniques to untransfected CHO cells, we found that no endogenous PrP^c, or Dpl, was detectable over the entire experimental time period.

Figure 1.

Localization (A) and expression level (B) of PrP^c and Dpl in transiently transfected CHO cells. (A) Distribution of PrP^c and Dpl in intact cells expressing each protein alone, after treatment (4°C) with anti-PrP mAb 8H4 (left), or anti-Dpl mAb Dpl-79 (right), followed by fluorescein isothiocyanate-conjugated secondary antibody. Both proteins are present on the cell surface, correctly exposed to the exoplasmic space. Bar, 20 μm. (B) Western blots of PrP^c (left) and Dpl (right) of cells expressing each protein alone (second lanes), or together (third lanes), show that both proteins exhibit high mass diffuse bands typical of their complex N-linked glycosylation (Massimino et al., 2004). Cell lysate SDS-PAGE separated proteins were immunoblotted with the above-described mAbs, and immunoreactive bands were then subjected to densitometric analysis. Comparison between the density of PrP^c- and Dpl-containing bands and the band density elicited by a known quantity of PrP_rec or Dpl_rec (0.5 ng, first lane of the left and right panel, respectively) yielded a value of 3.8 and 4.0 ng (in singly transfected cells, second lanes of both panels), and 2.1 and 2.3 ng (in doubly transfected cell, third lanes of both panels) for PrP^c and Dpl amounts, respectively.

ER Ca²⁺ in PrP^c- or Dpl-expressing Cells

The ER is the main Ca²⁺ storage of the cell as it houses Ca²⁺-binding proteins and SERCA pumps that transport Ca²⁺ against a high concentration gradient. To measure the ER luminal [Ca²⁺] ([Ca²⁺]_er), we used erAEQ that targets exclusively to this compartment (Montero et al., 1995). To function as a Ca²⁺ probe, recombinant AEQ needs to be reconstituted in the active complex with its prosthetic group, coelenterazine; added to cells, coelenterazine permeates the cell membranes and combines efficiently with AEQ. However, to avoid the instantaneous consumption of the holophotoprotein by the high luminal [Ca²⁺]_er, erAEQ-containing cells had to be first reduced in the ER Ca²⁺ content (by adding the Ca²⁺ ionophore ionomycin in the absence of external Ca²⁺) before reconstitution (in a Ca²⁺-free medium) with coelenterazine n, which is modified so as to reduce the Ca²⁺ affinity of AEQ (Barrero et al., 1997).

Figure 2A reports ER Ca²⁺ measurements in parallel batches of cells transiently expressing erAEQ alone (control cells, light gray trace) or together with PrP^c (black trace) or Dpl (dark gray trace). In agreement with previous reports (Brini et al., 2000), ∼1 min after adding CaCl₂ (1 mM), the [Ca²⁺]_er of control cells reached a steady state level close to 400 μM (390 ± 20 μM, n = 15). This value, however, is ∼15% lower than that observed in the presence of PrP^c, or Dpl, which was 455 ± 20 μM (n = 15) and 460 ± 20 μM (n = 15), respectively. Importantly, unless an ER Ca²⁺-discharging stimulus was applied, in all cases the shown Ca²⁺ plateau levels were stably maintained (unpublished data) and thus well represent the uppermost free Ca²⁺ amount that the three cell types can accumulate in the ER.

Figure 2.

Effects of the presence of PrP^c or Dpl (A), and of GFP-GPI_PrP (B), on the Ca²⁺ filling of the ER lumen and on the agonist-stimulated ER Ca²⁺ discharge, and effects of the presence of PrP^c, or/and Dpl, on the ER membrane Ca²⁺ passive efflux (C), and SERCA expression (D). (A–C) CHO cells, transiently expressing erAEQ alone (control, light gray trace), or together with PrP^c (black trace) or Dpl (dark gray trace; A and C), or GFP-GPI_PrP (dark gray trace; B), were first depleted of Ca²⁺ (see Materials and Methods and text) and then incubated with 5 μM coelenterazine n (in EGTA-supplemented KRB), and finally, after extensive washings, transferred to the thermostatted chamber of the luminometer and perfused with EGTA-containing KRB. As monitored by erAEQ, EGTA replacement by CaCl₂(1 mM) at the indicated time points induced the Ca²⁺ accumulation into the ER lumen, whereas addition of InsP₃-generating agonist ATP (100 μM; A and B), or of the SERCA inhibitor, tBuBHQ (20 μM; C), stimulated the ion discharge. Note that, although cells with either PrP^c or Dpl maintained a similar resting [Ca²⁺]_er (by ∼15% higher than in controls; A and C), ATP-induced Ca²⁺ efflux from the ER was more extensive in Dpl-containing cells (A), followed by a partial refilling of ER stores. Conversely, the levels of resting and discharged [Ca²⁺]_er in cells with GFP-GPI_PrP were as in the control (B). When cells were challenged with tBuBHQ, no difference was observed in the ER membrane passive permeability of the three cell types (C). Presented data are typical of at least seven independent experiments, which yielded equivalent results. (D) Western blots of the endogenous SERCA and β-actin in controls (first lane) or in cells expressing PrP^c (second lane) and Dpl (third lane) separately, or together (fourth lane), using mAbs against the ubiquitous SERCA 2 isoform and β-actin, respectively. By normalizing the density of each SERCA band to the corresponding β-actin band, cells with PrP^c resulted to express the highest, and those expressing Dpl (alone or together with PrP^c) the lowest, quantity of SERCA.

Next, we examined cells for the capacity to release Ca²⁺ after the addition of ATP (100 μM) at the indicated time point (Figure 2A). By activating plasma membrane P2Y receptors coupled to Gq proteins and phospholipase C, ATP induces the generation of inositol-1,4,5-triphosphate (InsP₃), which promotes the discharge of the ER Ca²⁺ stores through the channel activity of InsP₃ receptors (InsP₃Rs; Streb et al., 1983). Obtained results clearly demonstrate that the dynamics of Ca²⁺ efflux differs in the three cell types. The largest Ca²⁺ discharge occurs in the presence of Dpl (dark gray trace), so that the remaining [Ca²⁺]_er (130 ± 20 μM, n = 9) is less than in PrP^c-containing cells (black trace; 215 ± 20 μM, n = 7), whereas it equals that of controls (light gray trace; 110 ± 20 μM, n = 9; note, however, the lower starting [Ca²⁺]_er in the latter cells). Yet differences regard also the kinetics of the process, in that the fastest value is exhibited by Dpl-expressing cells, whereas the rate rapidly slows down in cells with PrP^c (velocity at half maximal Ca²⁺ discharge [V_1/2]: 27 ± 3 μM/s, n = 5, in cells with Dpl; 18 ± 1 μM/s, n = 5, in cells with PrP^c; 20 ± 2 μM/s, n = 5, in controls). Intriguingly, however, after ∼30–40 s from the maximal Ca²⁺ depletion and in the continuous presence of the agonist, a partial refilling of ER stores occurred in Dpl-containing cells, whereby the final [Ca²⁺]_er plateau settled to a value (210 ± 30 μM, n = 9), closely approximating that observed with PrP^c.

Notably, we could exclude that the peculiar ER Ca²⁺ handling in cells with PrP^c, or Dpl, is the consequence of the double transfection protocol (i.e., PrP^c, or Dpl, associated with erAEQ), given that cells coexpressing erAEQ and a non-natural GPI-linked protein (GFP-GPI_PrP, made of GFP fused to the GPI-attachment signal of PrP^c), have the same behavior of controls (Figure 2B). Nor did PrP^c, or Dpl, alter the passive permeability of the ER membrane, at least judging from the similar response (in terms of both kinetics and amplitude) in PrP^c- and Dpl-containing cells to the addition of 2,5-di-(tert-butyl)-1,4-benzohydroquinone (tBuBHQ), an inhibitor of the ER Ca²⁺ pump (Kass et al., 1989; Figure 2C). After the demonstration that IP₃-sensitive stores are fully depleted by inhibiting ER Ca²⁺ ATPases (Vanoevelen et al., 2004), these data also suggest that the larger Ca²⁺ amounts accumulated by cells with PrP^c (or Dpl), and the smaller IP₃-induced Ca²⁺ efflux by cells expressing PrP^c (shown in Figure 2A), are not due to the recruitment of IP₃-insensitive Ca²⁺ compartments. This conclusion is further corroborated by the finding that, when pretreating control, and PrP^c-expressing, cells with thapsigargin (an irreversible SERCA inhibitor; Thastrup et al., 1990), in both cases the uptake of Ca²⁺ in the ER was <10% of the maximal quantity observed in thapsigargin-untreated controls (unpublished data; see also Brini et al., 2000). We could also exclude that data with PrP^c are biased by the mode of transfection, or the protein quantity, in that the effects of the protein on the ER Ca²⁺ handling are reproduced in PrP^c-stably transfected CHO cells, which express the protein in lower amounts (unpublished data). Unfortunately, analysis of stably expressed Dpl molecules could not be carried out, given the impossibility, under our conditions, to establish viable clones with CHO cells.

We then examined whether the presence of PrP^c and Dpl affected the expression of ER proteins such as SERCA or calreticulin. The latter protein is not only crucial for the ER Ca²⁺-buffering capacity (Krause and Michalak, 1997), but can also regulate InsP₃Rs (Arnaudeau et al., 2002). We could not attribute to calreticulin the different modes of Ca²⁺ efflux from the ER, because the three cell types expressed the protein at similar levels (unpublished data). However, with respect to controls, we found higher and lower SERCA amounts in the presence of PrP^c and Dpl, respectively (Figure 2D, lanes 1–3). Possibly, PrP^c and Dpl trigger a signal that interferes with the expression of the protein in an opposed way.

Cytosolic Ca²⁺ in PrP^c- or Dpl-expressing Cells

Having found that the presence of PrP^c, or Dpl, affects the Ca²⁺ handling of the main intracellular Ca²⁺ store, we then set out to evaluate if and how the proteins influenced cytosolic Ca²⁺ concentration ([Ca²⁺]_c) under resting conditions, or after cell stimulation. AEQ steep response curve precludes the accurate estimation of [Ca²⁺] lower than 200–300 nM (Brini et al., 1995), a concentration typical of the cytosol of resting cells. This parameter was therefore determined through the image analysis of single cells loaded with the fluorescent Ca²⁺ indicator fura-2, after identifying successfully transfected cells by the coexpression of GFP (Brini et al., 1995). We found that fura-2 signals (given by the ratio of fluorescence emitted by illuminating cells at 340 and 380 nm) detected in PrP^c- or Dpl-expressing cells have a magnitude similar to that of the control (Figure 3A), suggesting that no appreciable alteration in basal [Ca²⁺]_c is attributable to these proteins.

Figure 3.

Resting [Ca²⁺]_c (A) and ATP-induced effects on [Ca²⁺]_c (B and C), in CHO cells transiently expressing cytAEQ alone (control), or together with PrP^c, or Dpl. (A) Resting [Ca²⁺]_c, monitored by the Ca²⁺-indicator fura-2, is expressed as the ratio of the fluorescence emitted by fura-2 after cell excitation at 340 and 380 nm (see Materials and Methods). No statistically significant difference was found in the three cell types. (B) [Ca²⁺]_c transients, after ATP (100 μM) addition at the indicated time point, were monitored with cytAEQ reconstituted with wild-type coelenterazine (5 μM; see Materials and Methods). Although a similar peak was found in controls (light gray trace) and Dpl-containing cells (dark gray trace), cells with PrP^c (black trace) gave rise to a transient of smaller magnitude. (C) Using the above described cytAEQ, a protocol was applied so as to evaluate the contribution to the ATP-induced [Ca²⁺]_c movements shown in B, first of the InsP₃-induced Ca²⁺ mobilization (first peak, ATP added in the presence of 100 μM EGTA, i.e., with no external Ca²⁺) and then of the Ca²⁺ influx from the external medium (second peak, in the presence of 1 mM CaCl₂). On Ca²⁺ release from the ER (first peak), a transient of smaller magnitude was again observed in cells with PrP^c, whereas no difference in the second peak was evident in the three cell types. Presented data are typical of at least seven independent experiments, which yielded equivalent results.

Conversely, cytAEQ was used to monitor changes in [Ca²⁺]_c after the cell stimulation by ATP. As shown (Figure 3B), addition of the agonist to cells induced a similar rise of [Ca²⁺]_c in both control cells (light gray trace; 3.9 ± 0.4 μM, n = 19) and in those expressing Dpl (dark gray trace; 3.9 ± 0.3 μM, n = 17), whereas in cells with PrP^c a statistically significant reduction of the Ca²⁺ transient was observed (black trace; 3.4 ± 0.4 μM, n = 18). It is to be noted, however, that under the used conditions two processes contribute to elevating [Ca²⁺]_c; one is the InsP₃-induced discharge of ER Ca²⁺ stores; the other is the Ca²⁺ influx from the extracellular space through the so-called capacitative Ca²⁺ entry (CCE), which becomes activated by a retrograde signal arising from the ER Ca²⁺ depletion (Putney et al., 2001). Hence, to dissect the contribution of each individual process, cells were subjected to a time-based protocol. First, adding ATP to cells placed in a Ca²⁺-free medium induced Ca²⁺ efflux from the ER (Figure 3C). As shown, we found an identical peak transient in control cells (light gray trace; 2.7 ± 0.6 μM, n = 12) and in those with Dpl (dark gray trace; 2.7 ± 0.3 μM, n = 14), whereas a reduced value was again observed in cells with PrP^c (black trace; 2.1 ± 0.3 μM, n = 12). After the transients subsided, cells were exposed to a 1 mM CaCl₂-containing medium. This caused a second peak, as expected from CCE through activated store-operated Ca²⁺ channels (SOCC). In this case, however, no statistically significant difference was evident in the three cell types (0.92 ± 0.13 μM, n = 10 in control cells; 0.80 ± 0.09 μM, n = 7, in cells with Dpl; 0.89 ± 0.14 μM, n = 8, in cells with PrP^c). Similar results were obtained with fura-2 (unpublished data).

Subplasma Membrane Ca²⁺ Pools in PrP^c- or Dpl-expressing Cells

pmAEQ, which targets to the cytosolic rim of the plasma membrane (Marsault et al., 1997), was used to monitoring the Ca²⁺ concentration in these restricted cytosolic domains ([Ca²⁺]_pm; Figure 4). ATP addition to cells in the presence of 1 mM CaCl (panel A) resulted in transient [Ca²⁺]_{2 pm} rises of similar magnitude in controls (light gray trace; 4.7 ± 0.8 μM, n = 12) and in Dpl-expressing cells (dark gray trace; 5.3 ± 1.5 μM, n = 9), in contrast to the almost eightfold higher [Ca²⁺]_pm peak found in those housing PrP^c (black trace; 42 ± 12 μM, n = 7). After the reestablishment of resting conditions, cells with PrP^c continued to maintain a statistically significant enhanced [Ca²⁺]_pm value (1.9 ± 0.6 μM, n = 6) than in controls (0.90 ± 0.20 μM, n = 10), although also in the presence of Dpl the basal [Ca²⁺]_pm remained at a higher level (1.5 ± 0.4 μM, n = 7; see inset to the figure). As reported in panel B, a similar influence of PrP^c and Dpl on [Ca²⁺]_pm was evident when CCE was alternatively activated by adding CaCl₂ (1 mM) to Ca²⁺-depleted cells (peak and steady state values were, respectively, 40 ± 12 μM, n = 11 and 6.2 ± 2.7 μM, n = 7, in cells with PrP^c; 24 ± 7 μM, n = 11 and 4.5 ± 0.9 μM, n = 8, in cells with Dpl; and 22 ± 8 μM, n = 7 and 2.9 ± 0.7 μM, n = 6, in controls). Altogether, these results suggest that the recombinant expression of PrP^c strongly increases Ca²⁺ entry from the extracellular space, whereas the attenuated effect by Dpl becomes evident only under steady state conditions.

Figure 4.

Monitoring the subplasma membrane Ca²⁺ concentration with pmAEQ. (A) Response to ATP (100 μM) of CHO cells transiently expressing pmAEQ alone (control, light gray trace), or together with PrP^c (black trace) or Dpl (dark gray trace), and maintained in 1 mM CaCl₂. pmAEQ was reconstituted by incubating cells with wild-type coelenterazine (5 μM; see Materials and Methods). Addition of ATP provoked a [Ca²⁺]_pm transient that in cells with PrP^c was by far more elevated than in the other two cell types. With respect to the control, however, the resting [Ca²⁺] level remained significantly higher in the presence of both PrP^{c pm}and Dpl (in the inset, resting [Ca²⁺]_pm are reported on an expanded scale). (B) Response to CaCl₂ addition (1 mM) in the same cell types described in A, but maintained in (Ca²⁺-free) KRB supplemented with EGTA (100 μM). After pmAEQ reconstitution, addition of 1 mM CaCl₂ induced a response that mimicked that shown in A with respect to both peak and resting [Ca²⁺]_pm levels. The latter are also shown in the inset on an expanded scale. Presented data are typical of at least six independent experiments, which yielded equivalent results.

Mitochondrial Ca²⁺ in PrP^c- or Dpl-expressing Cells

Compelling evidence indicates that in several cell types, including CHO cells, mitochondria behave as reporters of localized Ca²⁺ changes originating from the stimulation of the InsP₃Rs, in that the Ca²⁺ rise at the channels' mouth triggers the activity of the low-affinity Ca²⁺ uniporter of mitochondria juxtaposed to InsP₃Rs (Rizzuto et al., 1998). Such a spatial link between the two organelles, which according to recent data may also involve stable physical interactions (Filippin et al., 2003), ensures that mitochondria receive ER signals with maximal efficiency. To examine the influence of PrP^c and Dpl on such mechanism, we used mtAEQ targeted to the mitochondrial matrix (Rizzuto et al., 1992; Montero et al., 2000). Intriguingly, ATP addition resulted in a mitochondrial Ca²⁺ signal ([Ca²⁺]_m) that was specific for each case (Figure 5); cells with PrP^c showed the lowest value (black trace; 42 ± 11 μM, n = 9), the control ones the intermediate value (light gray trace; 62 ± 15 μM, n = 7), whereas the highest Ca²⁺ accumulation was detected in the presence of Dpl (dark gray trace; 110 ± 15 μM, n = 10). In the latter case, the average [Ca²⁺]_m value was more than twice that occurring with PrP^c. Therefore, in line with the close coupling between quantity of Ca²⁺ released from the ER and Ca²⁺ amounts accumulated by mitochondria, these data emphasize the findings on the ER Ca²⁺ discharge monitored by erAEQ (see Figure 2A).

Figure 5.

ATP-induced effects on [Ca²⁺]_m in CHO cells transiently expressing mtAEQ alone (control, light gray trace), or together with PrP^c (black trace) or Dpl (dark gray trace). mtAEQ was reconstituted by incubating cells with wild-type coelenterazine (5 μM; see Materials and Methods). The addition of ATP (100 μM) induced a rise in the [Ca²⁺]_m that was higher in Dpl-expressing cells than in the other two cell types. Presented data are typical of at least seven independent experiments, which yielded equivalent results.

Coexpression of PrP^c and Dpl Abolishes the Divergent Effects on Ca²⁺ Signaling Observed in Cells Expressing PrP^c, or Dpl, Alone

In view of the opposite effects on local Ca²⁺ fluctuations elicited by PrP^c and Dpl reported here and of the functional interplay of the two proteins suggested by genetic approaches (Kuwahara et al., 1999; Moore et al., 2001; Rossi et al., 2001; Yamaguchi et al., 2004), one expects that the Ca²⁺ signals observed in cells expressing PrP^c and Dpl separately will be modified by the copresence of the proteins in the same cell. To verify this assumption, we monitored Ca²⁺ homeostasis in the various compartments of cells transiently cotransfected with PrP^c and Dpl, after controlling that the two proteins are expressed at levels comparable to those found in singly transfected cells (Figure 1B, third lanes; note that, irrespective of the used plasmid batch(es), the transfection efficiency was generally ∼35%). Indeed, we found that under this condition local Ca²⁺ signals are extremely close to those of control cells (Figure 6), but for the ER Ca²⁺ efflux rate (panel A) that after PrP^c and Dpl cotransfection is faster than in controls by around 30% (V_1/2 being 33 ± 3 μM/s, n = 6, in cotransfected cells; 23 ± 4 μM/s, n = 5, in controls). It is interesting to note that a similar percentage difference is also found when comparing the corresponding V_1/2 values of controls and Dpl-expressing cells (Figure 2A, and text), and that cells transfected with Dpl, either alone or together with PrP^c, present the lowest amounts of SERCA (third and fourth lanes of Figure 2D).

Figure 6.

Monitoring [Ca²⁺]_er (A), [Ca²⁺]_c (B), [Ca²⁺]_pm (C), and [Ca²⁺]_m (D) in CHO cells transiently expressing the corresponding AEQ alone (control, light gray trace), or together with both PrP^c and Dpl (black trace). Experimental conditions were as reported in the legend to Figures 2A, 3B, 4B, and 5, respectively. Clearly, the copresence of PrP^c and Dpl abolished the divergent Ca²⁺ signaling observed in the various compartments of cells expressing PrP^c, or Dpl, alone. Presented data are typical of at least nine independent experiments, which yielded equivalent results.

DISCUSSION

Several reports indicate that reduction of PrP^c functionality causes perturbations in Ca²⁺ signaling and that the provoked alteration in some plasma membrane currents may explain aspects of TSEs, including a few neurologic symptoms (Cathala and Baron, 1987) and loss of neurons (Gray et al., 1999). On this background, we examined whether PrP^c takes part in the complex mechanism that controls Ca²⁺ homeostasis by analyzing, for the first time, Ca²⁺ fluctuations in different compartments of cells transfected with PrP^c and recombinant targeted aequorins. We found that the way cells with PrP^c handle compartmentalized Ca²⁺ favors the cell protective role attributed to PrP^c, in contrast with the effects arising from the presence of Dpl. Such a divergent activity, and the capacity of PrP^c—cotransfected with Dpl—to suppress the effects elicited by Dpl alone, highlights the protective function of PrP^c N-domain (absent from Dpl), as also deduced from experiments with cells challenged by serum deprivation (Sakudo et al., 2003) or over expressing toxic PrP fragments (Shmerling et al., 1998), proapoptotic Bax (Bounhar et al., 2001), or Dpl itself (Atarashi et al., 2003).

The most remarkable features observed in PrP^c-containing cells regard [Ca²⁺]_pm pools and the ER-mitochondria Ca²⁺ coupling. In considering the highest [Ca²⁺]_pm found in the presence of PrP^c (Figure 4), the simplest explanation is that PrP^c activates SOCCs. This is important, in view of the suggestion that the capacitative entry of Ca²⁺ in neuronal cells may serve in signaling pathways, not only in the refilling of Ca²⁺ stores observed in nonexcitable cells, and that subplasma membrane Ca²⁺-rich domains may trigger key physiological cell events (Putney, 2003). For example, SOCCs can be functionally coupled with adenylyl cyclase (Fagan et al., 2000); PrP^c-elicited [Ca²⁺]_pm rises may thus represent the so-far missing intermediate linking the activation of cytosolic cAMP-dependent protein kinase A by cell surface-attached PrP^cs, which was shown to inhibit apoptosis in retinal explants (Chiarini et al., 2002). Also, in analogy with SOCCs modulation of transmitter release and synaptic plasticity in certain neurons (Emptage et al., 2001), PrP^c-elicited Ca²⁺ hotspots may recruit effectors in the immediate vicinity, e.g., Ca²⁺-activated K⁺ channels, thereby explaining why impairment of (Ca²⁺-dependent) K⁺ currents in PrP-null cells could not be attributed to specific alterations of voltage-gated Ca²⁺ channels (Herms et al., 2000; but see Whatley et al., 1995). In the light of all these considerations, it is thus tempting to speculate that, as observed in the used cell model system, PrP^c affects SOCC-dependent subplasma membrane Ca²⁺ pools also in neurons. If this were the case, impairment of SOCCs could play a major role in TSE neurologic manifestations (Cathala and Baron, 1987).

Equally important is the way by which PrP^c-expressing cells control ER and mitochondrial Ca²⁺ homeostases, either of which is crucial in cell survival. Indeed, [Ca²⁺]_er variations may trigger stress responses, whereas substantial ER Ca²⁺ discharges may increase mitochondrial Ca²⁺ levels that, by activating caspase-9, a member of the death-specific caspase family (Cryns and Yuan, 1998), determine the pathway leading to apoptotic death (Li et al., 1997). Hence, despite the potentially dangerous elevated [Ca²⁺]_er (Pinton et al., 2000, 2001), our demonstration that the presence of PrP^c limits the agonist-stimulated ER Ca²⁺ release (Figures 2A and 3C) and, most importantly, the ion accumulation by mitochondria (Figure 5), is not only consistent with a PrP^c protective role, but it also predicts an increased cell vulnerability upon reducing PrP^c molecules, and/or expression of TSE-associated PrP mutants that accumulate in the ER (Singh et al., 1997; Zanusso et al., 1999; Jin et al., 2000; Negro et al., 2001). Accordingly, one of the first events observed in neuronal cells exposed to purified PrP^Sc is a substantial ER Ca²⁺ discharge, followed by up-regulation of stress proteins and activation of the ER-resident caspase-12 that acts on other effector caspases to produce apoptosis (Hetz et al., 2003b).

Conversely, stimulation of Dpl-expressing cells leads to a more pronounced and faster efflux of ER Ca²⁺ (Figure 2A) and to a large rise in [Ca²⁺]_m (Figure 5). These results, collected in the absence of apoptotic stimuli, disclose a possible mechanism by which Dpl sensitizes cells to environmental insults, whereas the difficulty in obtaining neuroblastoma (Massimino et al., 2004), or CHO (this article), clones with Dpl suggests that the intrinsic toxicity of Dpl may manifest in cells other than Purkinje cells, the prime target of Dpl pathogenicity (Sakaguchi et al., 1996). Interestingly, biochemical analyses of Dpl-over-expressing brain tissues have proposed that the protein toxicity relates to an increased oxidative damage, possibly linked to impaired copper metabolism (Wong et al., 2001). These data well agree with the demonstration that, despite the capacity of Dpl to bind copper like the related PrP^c C-terminus, different structural and functional roles pertain to Dpl- and PrP^c-bound coppers (Cereghetti et al., 2004; see also Sakudo et al., 2004). However, it seems also consistent with the influence of Dpl on mitochondrial Ca²⁺ handling reported here, given that deregulation of the organelle Ca²⁺ homeostasis can lead to the overproduction of toxic reactive oxygen species (Toescu and Verkhratsky, 2003).

As for the divergent ER Ca²⁺ handling, we found that cells with PrP^c and Dpl express higher and lower amounts of SERCA, respectively, than control cells (Figure 2D). This data may explain why, irrespective of an identical Ca²⁺ load, the balance between InsP₃R-mediated Ca²⁺ efflux from, and SERCA-mediated Ca²⁺ (re-)loading in, the ER results in a less pronounced net Ca²⁺ discharge with PrP^c than with Dpl (Figure 2A). ER Ca²⁺ load, however, depends on both the amount and activity of the SERCA pumps. We have shown that in cells with Dpl CCE generates steady state [Ca²⁺]_pm levels higher than in controls (Figure 4). It seems therefore possible to hypothesize that, in line with other observations (Mogami et al., 1997), [Ca²⁺]_pm in Dpl-expressing cells stimulates the SERCA activity, thus explaining why, despite lower SERCA amounts, also these cells maintain a higher resting [Ca²⁺]_er than in controls (Figure 2A). Likewise, [Ca²⁺]_pm may also account for the slow refilling of ER stores occurring after the agonist-induced maximal Ca²⁺ depletion (Figure 2A). An important support to the key role of [Ca²⁺]_pm magnitude in regulating ER Ca²⁺ handling comes from cells housing PrP^c and Dpl together; they display low SERCA levels as well as low [Ca²⁺]_pm and are unable to exhibit the same [Ca²⁺]_er features observed in cells with Dpl alone (Figure 6).

In accord with the suggested functional interrelationship of the proteins, PrP^c and Dpl copresence abolishes Dpl effects on local Ca²⁺ movements (Figure 6). Several hypotheses have been suggested to explain PrP^c protection against Dpl (Behrens and Aguzzi, 2002): that PrP^c and Dpl compete for a common ligand (in the absence of PrP^c, the binding of Dpl to the ligand would elicit an apoptotic signal); that PrP^c antagonizes an independent action of Dpl that sensitizes neurons to apoptosis (PrP-null cells would no longer control oxidative stress and Dpl deleterious function would predominate); and that PrP^c interferes with unselective multimeric pores formed by Dpl in the ER and/or plasma membrane. With respect to the latter hypothesis, the similar passive ER Ca²⁺ efflux in PrP^c- or Dpl-expressing cells (Figure 2C) implies that Dpl is unlikely to form (Ca²⁺) leak pathways in the ER membrane. On the other hand, the Ca²⁺ handling in cells expressing PrP^c and Dpl together closely matches that of control cells (Figure 6), not that observed in the presence of PrP^c alone. A possible interpretation for such a balanced effect entails that, at least in CHO cells, PrP^c and Dpl have independent, opposite activities, rather than the capacity to competing for a common ligand.

In conclusion, we have shown that the recombinantly expressed PrP^c governs the cell Ca²⁺ metabolism in a cellprotective manner, by limiting the Ca²⁺ accumulation into mitochondria. Yet, PrP^c consistently increases [Ca²⁺]_pm, suggesting that these elevated Ca²⁺ pools may contribute to regulating the (Ca²⁺-dependent) cell excitability and/or PrP^c-mediated signaling, either of which is critical for neuronal survival.