- Top of page
- Materials and methods
Spatial and temporal alterations in intracellular calcium [Ca2+]i play a pivotal role in a wide array of neuronal functions. Disruption in Ca2+ homeostasis has been implicated in the decline in neuronal function in brain aging and in neurodegenerative disorders. The plasma membrane Ca2+-ATPase (PMCA) is a high affinity Ca2+ transporter that plays a crucial role in the termination of [Ca2+]i signals and in the maintenance of low [Ca2+]i essential for signaling. Recent evidence indicates that PMCA is uniquely sensitive to its lipid environment and is stimulated by lipids with ordered acyl chains. Here we show that both PMCA and its activator calmodulin (CaM) are partitioned into liquid-ordered, cholesterol-rich plasma membrane microdomains or ‘lipid rafts’ in primary cultured neurons. Association of PMCA with rafts was demonstrated in preparations isolated by sucrose density gradient centrifugation and in intact neurons by confocal microscopy. Total raft-associated PMCA activity was much higher than the PMCA activity excluded from these microdomains. Depletion of cellular cholesterol dramatically inhibited the activity of the raft-associated PMCA with no effect on the activity of the non-raft pool. We propose that association of PMCA with rafts represents a novel mechanism for its regulation and, consequently, of Ca2+ signaling in the central nervous system.
Spatial and temporal alterations in free intracellular Ca2+ concentration [Ca2+]i in neurons play a crucial role in key physiological processes such as neurotransmission, synaptic plasticity, signal transduction, and gene expression (Miller 1991; Berridge 1998, 2005). Disruption in the regulation of Ca2+ homeostasis has been implicated in neuronal dysfunction occurring in brain aging and in neurodegenerative disorders such as Alzheimer’s disease, Parkinson’s disease, epilepsy, and stroke (Choi 1995; Garcia and Strehler 1999; Sattler and Tymianski 2000; Mattson and Chan 2001). The plasma membrane Ca2+-ATPase (PMCA) is the primary Ca2+ extrusion system responsible for the transport of Ca2+ from the cytosol to the extracellular medium (Carafoli 1991; Garcia and Strehler 1999). In conjunction with the Na+/Ca2+ exchanger and sarco–endoplasmic reticulum Ca2+-ATPase, PMCA is responsible for the maintenance of tightly regulated levels of free intracellular Ca2+ (Brini and Carafoli 2000).
Mammalian PMCA is encoded by four genes giving rise to four different isoforms that differ quite significantly in their regulation by calmodulin (CaM), the primary Ca2+ sensor in cells (Strehler and Zacharias 2001). Binding of four Ca2+ ions to high affinity sites on apo-CaM triggers a conformational change in which the Ca2+-binding domains move farther apart, thus exposing hydrophobic residues on the central linker region (Babu et al. 1988; Ikura et al. 1992). CaM retains considerable structure in sodium dodecyl sulfate (SDS), and Ca2+-dependent conformational changes in the protein can be assessed by monitoring altered mobility in SDS–polyacrylamide gel electrophoresis (PAGE) (Klee et al. 1979; Zhang and Vogel 1994; Yao and Squier 1996). Calcium-activated CaM binds to an autoinhibitory domain on the C-terminal end of PMCA, dissociating it away from the active site, and resulting in several-fold stimulation of PMCA activity (Falchetto et al. 1992; Lushington et al. 2005). Recent in vitro studies have shown that PMCA is uniquely sensitive to the biophysical properties of its surrounding lipid environment (Pang et al. 2005; Duan et al. 2006; Tang et al. 2006). PMCA activity was found to be two- to threefold higher when the purified protein was reconstituted into ordered lipids such as eye lens fiber lipid, lens epithelial lipid, or the synthetic lipid dipalmitoylphosphatidylcholine compared with PMCA present in the disordered (fluid) lipid dioleoylphosphatidylcholine (Tang et al. 2006).
Lipid rafts are specialized microdomains in the plasma membrane known to serve as platforms for the assembly of protein complexes involved in signaling pathways (Simons and Toomre 2000; Grzybek et al. 2005). These membrane microdomains are characterized by their high content of cholesterol and sphingolipids (Brown and London 1998). It is believed that hydrogen bonding between the 3-OH group of cholesterol and the amide group of sphingolipids segregates these lipids away from the rest of the plasma membrane, resulting in the formation of tightly packed, ordered microdomains (Schroeder et al. 1994; Brown and London 1997, 1998). Given that PMCA activity is stimulated by the presence of ordered lipids in its surrounding membrane environment, it may be predicted that the naturally high degree of lipid order in neuronal rafts may be favorable for PMCA activity. Consistent with this possibility, PMCA was recently shown to be associated with lipid rafts isolated from synaptic membranes of porcine cerebellum (Sepulveda et al. 2006). The physiological significance of the raft-associated PMCA and its regulation in intact neurons, however, remain unclear.
The aim of the present study was twofold: (i) to determine and characterize the association of PMCA with lipid rafts in primary cultured neurons and (ii) to determine the effect of disruption of raft lipid environment on the properties of the raft-associated PMCA. Both biochemical and morphological evidence showed the partitioning of PMCA into rafts in preparations isolated from primary neurons and in intact cells. The raft-associated PMCA exhibited much higher activity than the non-raft PMCA, consistent with higher protein levels associated with these microdomains. Disruption of rafts by chronic depletion of cellular cholesterol dramatically decreased the activity of the raft-associated PMCA with no effect on the activity of PMCA excluded from these microdomains. The decrease in activity of the raft-PMCA was not due to alterations in its protein levels or the levels of its activator protein CaM. We propose that the ordered environment of rafts, maintained by high levels of cholesterol, enhance PMCA activity in vivo. The association of PMCA with lipid rafts therefore represents a novel CaM-independent mechanism for the regulation of its activity and, consequently, of Ca2+ signaling in the central nervous system.
Materials and methods
- Top of page
- Materials and methods
Lovastatin (LS), mevalonate, ouabain, thapsigargin, oligomycin, ATP, and horseradish peroxidase-coupled cholera toxin subunit B (CTXB) were purchased from Sigma Chemical Company (St. Louis, MO, USA). Bicinchoninic acid protein assay kit was obtained from Pierce (Rockford, IL, USA) and Protease Inhibitor Cocktail III was from Calbiochem (San Diego, CA, USA). Amplex Red kit for cholesterol measurement and gradient gels (8–16%) were purchased from Invitrogen (Carlsbad, CA, USA). The following sources were used for the various primary antibodies: anti-PMCA (Affinity Bioreagents, Golden, CO, USA), anti-CaM (Upstate Biotechnologies, Charlottesville, VA, USA), and anti-flotillin-1 and Thy-1 (BD Biosciences, San Jose, CA, USA). Isoform-specific PMCA antibodies were a gift from Dr E.E. Strehler, Mayo Clinic (Rochester, MN, USA) and NAP-22 antibodies were provided by Dr S. Maekawa, Kobe University (Kobe, Japan).
Dissociated cortical and hippocampal neuron cultures were established from 18-day-old Sprague–Dawley fetuses as described previously (Michaelis et al. 1994, 1998, 2005). Briefly, pups were delivered by cesarean section, while the dam was fully anesthetized with pentobarbital (0.1g/kg), and the brains recovered according to protocols approved by the institutional IACUC and in accordance with National Institutes of Health guidelines. The cortical lobes and hippocampi were dissected and cells dissociated by gentle trituration with trypsin. After the final precipitation step, neurons were re-suspended in fresh Dulbecco's modified Eagle's medium/F-12 with 10% fetal calf serum, and plated at densities ranging from 0.3–3 × 106 cells/dish on 35-mm tissue culture dishes coated with poly-d-lysine. After 24 h to allow cell attachment, the fetal calf serum-containing medium was replaced by a defined medium with Dulbecco's modified Eagle's medium/F-12 containing N2 supplements, potassium bicarbonate (15 mmol/L), and 20% glial conditioned medium. Cells were fed twice a week by replacing 1/3 of the medium with fresh medium. It was estimated that > 90% of the cells in the primary cultures were neurons, based on morphological characteristics and confirmation with immunofluorescent staining for glial fibrillary acidic protein and neuron-specific enolase.
Isolation of lipid rafts from primary neurons
Rafts were isolated from primary cortical neurons at day in vitro 6 in culture by previously described methods (Sargiacomo et al. 1993). Briefly, the medium was removed from the cell culture dishes, cells were washed twice with 200 mmol/L Tris–Cl, pH 7.4, and lysed with 3 mmol/L Tris–Cl, pH 8.0, containing a cocktail of protease inhibitors. The neuronal lysate was solubilized in a buffer containing 25 mmol/L Tris–Cl, 75 mmol/L NaCl, pH 7.5 (TNB) plus 1% (v/v) Brij 98, and incubated on ice for 30 min with mild vortexing. The suspension was mixed with an equal volume of 90% (w/v) sucrose solution in TNB, overlayed with a two-step gradient consisting of 35% (w/v) and 5% (w/v) sucrose in TNB. Centrifugation was performed for 18 h at 135 000 g at 4°C in a TLS-55 rotor in an OptimaTM MAX Ultracentrifuge (Beckman Coulter, Fullerton, CA, USA). Ten fractions (0.5 mL) were collected from the top to the bottom of the tube. Protein estimation was performed by the bicinchoninic acid method according to the manufacturer’s instructions.
Cholesterol and GM1 measurement and cholesterol depletion
Cholesterol was measured using Amplex Red kit according to the manufacturer’s instructions. GM1 ganglioside levels were determined by dot-blot using CTXB, known to have a very high affinity for this lipid. Briefly, 1 µg protein from each fraction was applied onto a polyvinylidene fluoride (PVDF) membrane, blocked with 5% milk, and exposed to CTXB (10 µmol/L) for 2 h (Hering et al. 2003). Color was developed by incubating the dot-blot in a mixture of 1.4 mmol/L 3,3′-diaminobenzidine tetrahydrochloride, 200 mmol/L nickel chloride, and 6.2 mmol/L H2O2.
Chronic depletion of cellular cholesterol was achieved by treating neurons with LS, an inhibitor of the enzyme HMG CoA reductase, the rate limiting step in cholesterol biosynthesis. It has been shown that LS when used by itself causes apoptosis in cells, however, the toxic effects of LS are abrogated by the addition of mevalonate, the product of HMG CoA reductase (Wang and Macaulay 1999), presumably by allowing the synthesis of other downstream non-steroidal derivatives of this pathway important for cell survival. Briefly, neurons were treated for 4 days with LS (2.5 µmol/L) and mevalonate (200 µmol/L) added to the culture medium on day in vitro 2 (Hao et al. 2001; Hering et al. 2003). The vehicle (ethanol) -treated and LS-treated cells were collected and lysed on day in vitro 6, protein concentration was measured and equal amounts of protein solubilized and loaded on sucrose density gradients for raft isolation as described above.
Measurement of PMCA activity
The activity of PMCA was determined by monitoring the generation of Pi from ATP as described previously (Zaidi et al. 1998). The activity buffer contained 25 mmol/L Tris–Cl, pH 7.4, 1 mmol/L ATP, 50 mmol/L KCl, 0.1 mmol/L ouabain, 4 mg/L oligomycin, 0.1 µmol/L thapsigargin, 200 μmol/L EGTA, and CaCl2 added to yield a final free Ca2+ concentration of 0.5 µmol/L. The final free Ca2+ concentration was calculated using the software calcium.com that calculates the multiple equilibria between all ligands in solution. PMCA activity measured in the absence of added CaM is referred to as basal activity and that in the presence of 340 nmol/L CaM as CaM-stimulated activity. After a 5 min pre-incubation, the reaction was started by the addition of ATP, continued for 30 min at 37°C, and stopped by addition of the Malachite Green dye (Lanzetta et al. 1979). The PMCA activity was defined as the Ca2+-activated ATP hydrolysis and expressed as nanomoles of Pi liberated per milligram of protein per minute, based on values from a standard curve.
Proteins were separated by SDS–PAGE on 8–16% gradient gels and transferred to methanol-activated PVDF membranes as described previously (Zaidi and Michaelis 1999; Zaidi et al. 2003). Non-specific interactions were blocked with 5% (w/v) milk for 1 h at 25°C and the membranes incubated overnight with the indicated concentrations of primary antibodies. Alkaline phosphatase-conjugated secondary antibodies (1 : 1000) were added for 2 h at 25°C. Immunoreactive bands were visualized by developing the immunoblots using the substrate 5-bromo-4-chloro-3-indoyl phosphate and nitroblue tetrazolium. Blots were scanned using an HP Scanjet 4470 scanner (Hewlett Packard, Palo Alto, CA, USA) and relative intensities of the immunoreactive bands quantified using Adobe Photoshop 5.0 (Adobe Systems, San Jose, CA, USA).
Lipid raft patching in neurons
Exposure of hippocampal neurons to anti-Thy-1 antibody was used to induce patching of rafts according to previously described methods (Wong and Schlichter 2004). This was followed by labeling with anti-PMCA2 antibody. Briefly, hippocampal neurons were chilled on ice for 15 min and washed with ice-cold PBS supplemented with 0.7 mmol/L CaCl2 and 0.48 mmol/L MgCl2 (PBS-CM). Cells were subsequently incubated with anti-Thy-1 antibody (1 : 100) in 0.1% (w/v) BSA in PBS-CM for 1 h at 37°C, followed by Alexa Fluor 568 conjugated anti-mouse secondary antibody (1 : 200) for 1 h at 37°C. Neurons were then fixed with 4% (w/v) p-formaldehyde for 30 min on ice. After blocking with 5% (w/v) BSA, cells were labeled with anti-PMCA 2 antibody (1 : 500) and Alexa Fluor 488 conjugated anti-rabbit secondary antibody (1 : 500). Images were acquired on a Zeiss LSM-510 confocal microscope (Carl Zeiss, Thornwood, NY, USA) using fixed laser strength, pinhole size, and detector gain for all samples within a single experiment.
All data are means ± SEM obtained from six experiments using raft preparations isolated from independent cultures. Statistical significance of differences between various samples was assessed using Student’s t-test for unpaired samples.
- Top of page
- Materials and methods
This study shows for the first time that both PMCA and CaM are partitioned into lipid rafts in primary neurons. By taking advantage of the detergent insolubility and low buoyant density of these microdomains, we isolated rafts from cultured cortical neurons and showed the co-localization of PMCA with raft lipid markers cholesterol and GM1 and protein markers flotillin-1 and NAP-22 (Bickel et al. 1997; Naslavsky et al. 1997; Madore et al. 1999; Maekawa et al. 1999). Using intact hippocampal neurons, we also obtained morphological evidence showing that a substantial amount of PMCA was spatially segregated into lipid rafts, suggesting that such an association occurs in vivo. The partitioning of PMCA into raft and non-raft domains was corroborated by functional assays. The enzyme exhibited much higher total PMCA activity in the detergent insoluble and low buoyant density fractions compared with the pool in the non-raft domains, consistent with greater amounts of PMCA protein partitioned into rafts. Addition of exogenous CaM, however, stimulated PMCA activity only in fraction 3 with only marginal effects on PMCA in the other fractions. Low sensitivity or even a lack of sensitivity of PMCA to CaM in excitable tissue has been consistently observed by us and others and is believed to be either due to interaction with CaM constitutively bound to neuronal membranes and/or activation by acidic phospholipids known to render PMCA insensitive to CaM (Niggli et al. 1979; Missiaen et al. 1989; Zaidi et al. 1998; Levitan 1999a,b; Sepulveda et al. 2006).
A recent report showed the distribution of PMCA in rafts isolated from porcine cerebellum synaptic membranes (Sepulveda et al. 2006). Unlike the results in synaptic membranes showing the presence only of the isoform PMCA 4, our immunoblot analysis data shows the raft localization of all four PMCA isoforms. The discrepancy between the two studies may be attributed to distinct differences in the methodology used for raft isolation, such as detergent concentration, time of centrifugation, and the type of density gradient used. It seems quite likely that there are differences in the PMCA composition in rafts isolated from brain synaptic membranes compared with the entire plasma membrane of neurons as well as possible regional differences.
Given that the liquid ordered phase of rafts arises from the tight packing of interstitial spaces by cholesterol molecules (Brown and London 1997, 1998), cholesterol depletion is commonly used as a tool to disrupt the raft microenvironment. Rather than using acute treatments such as cholesterol binding or extracting agents, which we found to be damaging to cells, we chose to inhibit cholesterol biosynthesis. Additionally, chronic depletion of cellular cholesterol over the course of several days would allow cells to adapt to the altered environment and hence be a better indicator of its effects as they might occur in vivo. Reduced levels of cellular cholesterol are known to induce alterations in the lipid and protein composition of rafts, depending upon the cell type and the experimental conditions used for depletion (Ilangumaran and Hoessli 1998; Isshiki and Anderson 1999; Hao et al. 2004). In our hands, inhibition of cholesterol biosynthesis in cortical neurons lowered total protein levels, but had no significant effect on PMCA levels, suggesting that PMCA may not be associated with rafts by direct interaction with cholesterol. However, it cannot be overruled that the ∼40% residual cholesterol may be sufficient to anchor PMCA to rafts. Although raft-PMCA protein levels did not change with cholesterol depletion, a dramatic loss in enzyme activity was observed, suggesting that the cholesterol-rich, liquid ordered raft environment favored PMCA activity, in close agreement with in vitro studies reported earlier (Zhao et al. 2004; Pang et al. 2005; Tang et al. 2006). Our results are consistent with earlier studies showing a significant reduction in PMCA activity upon depletion of synaptic plasma membrane cholesterol by cholesterol oxidase, a manipulation that converts cholesterol to its oxidation product cholestenone (Wood et al. 1995). Cholesterol oxidation was accompanied by a decrease in synaptic plasma membrane interdigitation and increased membrane fluidity and biophysical changes that may reduce PMCA activity by lowering its mobility and/or causing protein aggregation (Cornea and Thomas 1994). Another interesting study showed that pharmacological tools used to lower cholesterol such as oxidosqualene cyclase inhibitors (for example compound A18666A) can directly intercalate into the membrane bliayer thus causing perturbation of the membrane structure and function (Cenedella et al. 2004). Although the exact mechanism of PMCA inhibition by LS-induced cholesterol depletion is not clear from the present study, a number of possibilities, such as perturbation of the membrane bilayer and consequent change in its molecular structure, alterations in thickness and fluidity, disruption in the order of raft lipid environment, loss of acidic phospholipids, and/or up-regulation of GM1 ganglioside, a known inhibitor of PMCA activity (Zhao et al. 2004), may underlie these observations.
There has been a long-standing interest in the role of rafts and caveolae (specialized forms of rafts with a distinct cave-like morphology and the presence of caveolin) in Ca2+ signaling (Isshiki and Anderson 1999, 2003). Not surprisingly, a number of Ca2+-regulating proteins, including an inositol trisphosphate receptor-like protein, growth associated protein, heterotrimeric G proteins, transient receptor potential-like protein channels, PMCA, and CaM have been found in caveolae in several non-neuronal cell types (Fujimoto 1993; Fujimoto et al. 1995; Schnitzer et al. 1995; Arni et al. 1998; Keshet et al. 1999; Ambudkar 2004). Caveolae are believed to be local sites for the orchestration of Ca2+-dependent signal transduction events (Isshiki et al. 1998, 2002a,b). The role of neuronal lipid rafts in Ca2+ signaling has been less appreciated given the lack of a distinct morphology, smaller size precluding detection by optical microscopy, and the transient nature of these microdomains. The findings that both PMCA and CaM are localized in neuronal lipid rafts suggests that raft domains may play an important role in local Ca2+ signaling in neurons.
The association of Ca2+ signaling proteins such as the PMCA and CaM with lipid rafts may have potential implications on the spatiotemporal distribution of these proteins in the plasma membrane. Given the high lateral mobility of rafts, association of PMCA and CaM with these microdomains may serve as a mechanism for their trafficking to various parts of the membrane in response to specific stimuli. Indeed, a recent report showed that ∼60% of PMCA 2 in hair stereocilia exhibits high mobility and is rapidly delivered to the apical cell border from where it diffuses to the entire stereocilia surface (diffusion constant 0.1–0.2 µm2/s), whereas the remaining pool was present in a relatively immobile fraction (Grati et al. 2006). On the other hand, many raft-associated proteins including the PMCA may be anchored to the underlying membrane cytoskeleton. Such an association may help stabilize the protein complexes of PMCA and its binding partners and consequently restrict lateral mobility. It is tempting to speculate that the differential distribution of PMCA in raft versus non-raft domains may serve to regulate its mobility, trafficking, and protein–protein interactions with its binding partners. Association with rafts may therefore serve as a novel mechanism not only in the regulation of PMCA activity by modulation of the cholesterol content but also in controlling its spatiotemporal distribution, consequently allowing the fine-tuning of local Ca2+ signaling.