Address correspondence and reprint requests to Dr. F. J. Barrantes, UNESCO Chair of Biophysics & Molec Neurobiol and Instituto de Investigaciones Bioquímicas de Bahía Blanca, C.C. 857, B8000FWB Bahía Blanca, Argentina. E-mail: email@example.com
The nicotinic acetylcholine receptor (AChR) is the prototype ligand-gated ion channel, and its function is dependent on its lipid environment. In order to study the involvement of sphingolipids (SL) in AChR trafficking, we used pharmacological approaches to dissect the SL biosynthetic pathway in CHO-K1/A5 cells heterologously expressing the muscle-type AChR. When SL biosynthesis was impaired, the cell surface targeting of AChR diminished with a concomitant increase in the intracellular receptor pool. The SL-inhibiting drugs increased unassembled AChR forms, which were retained at the endoplasmic reticulum (ER). These effects on AChR biogenesis and trafficking could be reversed by the addition of exogenous SL, such as sphingomyelin. On the basis of these effects we propose a ‘chaperone-like’ SL intervention at early stages of the AChR biosynthetic pathway, affecting both the efficiency of the assembly process and subsequent receptor trafficking to the cell surface.
green fluorescent protein-labeled vesicular stomatitis virus G protein.
At the cholinergic synapse, the chemical signal encoded in the neurotransmitter acetylcholine is transduced into an electrical signal in a very rapid process mediated by the nicotinic acetylcholine receptor (AChR). This transmembrane protein is the prototype of the superfamily of rapid ligand-gated ion channels. The AChR has five homologous, membrane-spanning subunits in stoichiometry α2βγδ (electric tissue and fetal muscle) or α2βεδ (adult muscle). Several studies have demonstrated that the AChR is quite sensitive to the membrane lipid environment (Barrantes 1993, 2004).
The receptor protein is assembled in the endoplasmic reticulum (ER) (Smith et al. 1987), purportedly by a stepwise addition of subunits in which the first stage is the formation of αδ and αγ heterodimers (Blount et al. 1990). In this model, these heterodimers interact with a β subunit and with each other to form a pentamer that leaves the ER and travels to the cell surface along the exocytic pathway (Gu et al. 1991; Kreienkamp et al. 1995). The assembly process is slow and inefficient (Merlie and Lindstrom 1983).
Sphingolipids (SL) are more abundant at the plasma membrane than in other intracellular membranes (Schwarzmann and Sandhoff 1990). Their distribution in the plane of the membrane bilayer appears not to be homogeneous, and it was postulated (Simons and Ikonen 1997) that SLs and cholesterol (Chol) occur in laterally segregated lipid microdomains termed ‘rafts’, which have been suggested to concentrate signaling molecules and receptors in particular regions of the cell surface (Simons and Toomre 2000; Edidin 2003).
The synthesis and sorting of lipids are important for several cellular events (Futerman and Hannun 2004). SL biosynthesis commences in the ER (see Supplementary material Scheme 1) and proceeds up to ceramide synthesis. Ceramide is translocated from the ER to the Golgi apparatus where the synthesis of sphingomyelin (SM) or glucosylceramide (GluCer) takes place (Holthuis et al. 2001). It has been suggested that transport of proteins through the secretory pathway is coupled to SL biosynthesis (Rosenwald et al. 1992). In hippocampal neurons, the viral glycoprotein hemagglutinin and Thy-1 interact with SL-Chol (‘classical’) lipid rafts and missorting of axonal Thy-1 occurs in SL-deprived cells (Ledesma et al. 1998). The targeting of the enzyme tyrosinase to melanosomes is glycosphingolipid (GSL)-dependent (Sprong et al. 2001). Oligomerization and trafficking of Pma1 to the plasma membrane in yeast is dependent on SL metabolism (Lee et al. 2002; Wang and Chang 2002). Reduced SL content also enhances solubility of glycosylphosphatidylinositol (GPI)-anchored proteins in the non-ionic detergent Triton X-100 (Hanada et al. 1995). Rafts have also been postulated to be involved in AChR clustering in ciliary neurons (Bruses et al. 2001) and to serve as platforms for trafficking of complexes of AChR with the non-receptor protein rapsyn to the plasma membrane (Marchand et al. 2002). More recently lipid rafts were postulated to regulate AChR clustering by facilitating the agrin/MuSK signaling and the interaction between AChR and rapsyn (Zhu et al. 2006) and to participate in the formation and subsequent stabilization of agrin-induced AChR clusters (Campagna and Fallon 2006; Stetzkowski-Marden et al. 2006). In this work we disclose a hitherto unknown function of SLs in AChR assembly and traffic from its site of synthesis at the ER to the plasma membrane.
Materials and methods
Nutridoma-SP medium (‘Nutridoma’) was purchased from Boehringer Mannheim, Germany. [125I]-α-bungarotoxin (α BTX, 120 μCi/μmol) was from New England Nuclear (Boston, MA, USA). Native αBTX, carbamoylcholine (Carb), and SM were from Sigma Chemical Co (St Louis, MO, USA). DL-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol.HCl (PDMP), fumonisin B-1 (FB-1) and myriocin (ISP-1) were obtained from Biomol Res Laboratories (Plymouth Meeting, PA, USA). Alexa Fluor488-conjugated αBTX (Alexa Fluor488-αBTX) and Alexa Fluor594-αBTX, goat anti-rabbit and goat-anti mouse IgG antibodies labeled with Alexa Fluor488 or Alexa Fluor546, ER-Tracker (ER-trackerTM Blue-White DPX), 6-((N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino) hexanoyl)-sphingosine (NBD C6-ceramide were all from Molecular Probes (Eugene, OR, USA). Rabbit anti-calnexin polyclonal antibody and anti-syntaxin 6 monoclonal antibody were from Stressgen Biotechnologies Corp., Victoria, Canada. The mAb 210 monoclonal antibody against the major immunogenic region (MIR) in the α subunit of the AChR was kindly provided by Dr. J. Lindstrom, University of Pennsylvania School of Medicine. The antibody against disulfide isomerase (PDI) was kindly provided by Dr. D. Ferrari, Department of Neurobiology, Max-Planck-Institute for Biophysical Chemistry, Göttingen.
Plasmids for VSVG-GFP were kindly provided by Dr. Alfonso Gonzalez, Pontificia Universidad Católica de Chile. CHO-K1/A5 and SPB-1/SPH− cells were grown in Ham’s F-12 medium supplemented with 10% bovine fetal serum (BFS) (complete medium) for 24 h and then transfected with 1.5 μg DNA/35-mm dish of VSV-GFP using the PolyFect transfection Reagent (QIAGEN, Hilden, Germany) for 24 h according to the manufacturer’s recommendations. In cells treated with PDMP, the inhibitor, the cDNA, and the transfection reagent were incubated together.
Temperature-sensitive sphingolipid-defective cells and inhibition of complex sphingolipid biosynthesis by FB-1, PDMP or ISP-1
SPB-1/SPH− cells were first grown in complete Ham’s F-12 medium for 48 h and were subsequently divided into different sets and further grown for 24 additional hours in Nutridoma, a lipid-depleted medium, at the semi-permissive (37°C) temperature. Complex SL biosynthesis was blocked in control CHO-K1/A5 cells, a clone expressing adult muscle AChR (Roccamo et al. 1999), by growing them in the presence of different metabolic inhibitors, as follows. CHOK1/A5 cells were first grown in complete Ham’s F-12 medium for 48 h and subsequently grown in Nutridoma medium containing 50 μmol/L FB-1, 10 or 50 μmol/L PDMP, or 5–40 μmol/L ISP-1 for 24 h. Solvent (ethanol) was kept below 0.1% in all cases.
Recovery experiments using exogenous sphingomyelin
SM at a final concentration of 5–10 μmol/L was added to cells essentially as in (Furuya et al. 1998) except that ethanol:decane (98:2 vol/vol) kept below 0.16% was used as a vehicle.
To observe the AChR at the plasma membrane, cells were grown on coverslips, labeled with Alexa488-αBTX for 1 h at 4°C, washed with PBS (150 mmol/L NaCl, 10 mmol/L Na2HPO4, 10 mmol/L NaH2PO4, pH 7.4) or Medium 1 (140 mmol/L NaCl, 20 mmol/L HEPES, 1 mmol/L CaCl2, 1 mmol/L MgCl2, 5 mmol/L KCl, pH 7.4), mounted on glass slides and observed in vivo. To visualize intracellular AChR, plasma membrane AChRs were blocked with excess native αBTX at 4°C. Cells were then fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained with Alexa594-αBTX or, alternatively, with the antibody mAb 210 followed by labeling with rat Alexa488-labeled secondary antibody.
To visualize the ER, cells were incubated with ER-Tracker for 40 min at 37°C, washed with PBS and incubated with excess native αBTX at 4°C to block surface AChR. Cells were then fixed, permeabilized with Triton X-100 and labeled with Alexa594-αBTX (or mAb 210 antibody) as above. Anti-calnexin rabbit polyclonal antibody, or anti-PDI rabbit antibody, were used to label the ER. Anti-syntaxin 6 mouse monoclonal antibody was used to label the trans-Golgi network (TGN). Secondary antibodies (Alexa488-labeled anti-rabbit IgG, Alexa488-labeled anti-mouse IgG) were applied together with Alexa594-αBTX. Alternatively, AChR was identified using the monoclonal antibody mAb 210 followed by Alexa488-labeled anti-rat IgG.
Cells were examined with Nikon Eclipse E-600 or E-300 fluorescence microscopes (Nikon, Melville, NY, USA). Images were captured using a K2E Apogee CCD camera thermostatically cooled at −10°C and driven by PMIS software (GKR Computer Consulting, Unterfoehring, Germany) or a Hamamatsu ORCA-ER CCD camera driven by Metamorph software (Meta Imaging Software, Downingtown, PA, USA). Confocal images were taken with a Leica TCS SP2 microscope (Leica Microsystems, Heidelberg, GmbH).
Fluorescence microscopy assay of nicotinic acetylcholine receptor assembly
To determine the proportion of assembled /unassembled intracellular AChR, we employed an assay based on agonist inhibition of αBTX binding. Cell-surface AChRs were blocked with native toxin, fixed, permeabilized and incubated with 10 mmol/L carbamoylcholine (Carb) for 1 h at 25°C and subsequently incubated with Alexa594-αBTX in the presence of agonist. As a control for total intracellular AChR, cells were incubated with PBS instead of Carb for the same period. Quantitative measurements of fluorescence intensities of intracellular AChR label in the presence and absence of agonist provided an estimation of the proportion of unassembled AChR.
Sucrose density gradient analysis
This was performed essentially as in Barrantes (1982) and Kreienkamp et al. (1995). Control and PDMP-treated cells were incubated for 1 h with native αBTX in order to block surface AChR, and then permeabilized with 0.5% saponin-containing buffer (10 mmol/L sodium phosphate, 10 mmol/L EDTA, 0.1% BSA). Cells were incubated with excess [125I]- αBTX for 1 h at 25°C, washed with Medium 1, and solubilized in 600 μL of homogenization buffer (1% Triton X-100, 150 mmol/L NaCl, 5 mmol/L EDTA, 50 mmol/L Tris, pH 7.5, containing a protease inhibitor cocktail) for 3 h at 25°C. Samples were centrifuged at 12 000 g to eliminate the particulate material and the supernatants were loaded atop a 3–20% sucrose gradient containing the same buffer and detergent, and centrifuged in a Beckman SW41 rotor for 22 h at 200 000 g. Fractions were collected from the top of the tube in 250 μL fractions and counted in a γ-counter.
Quantitative fluorescence microscopy analysis
Fluorescence images were analyzed with Scion Image version 4.0.2 (Scion Corp., Frederick, MD, USA). Fluorescence intensities were measured from 16-bit images by delimiting membrane areas (cell surface AChR and VSVG-GFP-labeled surface membrane) or by delimiting the whole cell (intracellular AChR, unassembled AChR and total VSVG-GFP) in 10–30 different fields for each experimental condition. Fluorescence intensity values were corrected for fluorescence background measured in areas adjacent to the cells. The average fluorescence intensity over distinct areas of the cell surface or whole cells was calculated for randomly chosen cells in phase contrast for each experimental condition. Results are expressed as average ± SD of three or more independent experiments. For illustration purposes, images were processed using Adobe Photoshop 7, scaled with identical parameters, and pseudo-colored according to the corresponding emission wavelength.
Equilibrium [125I]- αBTX binding studies
Surface AChR expression was determined by incubating 60–70% confluent cells with excess [125I]- αBTX in cell culture medium for 1 h at 4°C. At the end of the incubation period, dishes were washed four times with PBS and the cells were lysed with 0.1 N NaOH. Radioactivity was measured in a gamma counter with an efficiency of 80%. Non-specific binding was determined from the radioactivity remaining in the dishes of cells pre-incubated with 10 μmol/L native αBTX or 2 mmol/L Carb before addition of [125I]- αBTX.
Cells grown in complete medium were treated with 10–50 μmol/L PDMP (and only with vehicle in the case of control cells) in Nutridoma medium supplemented with 0.1% of BFS for 1 h. Cells were then incubated with NBD-C6-ceramide for 24 h in the presence of 10–50 μmol/L PDMP. Cells were scraped with a spatula and lipids were extracted by CHCl3/CH3OH (2:1, by vol.) (Folch et al. 1957). Lipid extracts were analyzed as in (Lipsky and Pagano 1983) with some modifications: Thin-layer chromatography (TLC) was developed with CHCl3/CH3OH/H2O (65:25:4 by vol.) and appropriate standards. Fluorescent lipids were visualized by UV illumination. Lipids spots were scraped, extracted from the silica as in (Arvidson 1968) and quantified in a 4800 fluorimeter (SLM instruments, Urbana, Illinois, USA).
Reduced cell-surface nicotinic acetylcholine receptor expression in sphingolipid-deficient SPB-1/SPH− cells
SPB-1 is a temperature-sensitive CHO cell mutant with thermolabile serine palmitoyltransferase (SPT) activity (Hanada et al. 1990). The SPT activity of these cells at 37°C is about 8% of that of the parenteral cell line, CHO-K1. When these cells are grown at 37°C the ceramide and SM levels decrease to 9% and 50%, respectively, of those of the parent cell line (Hanada et al. 1990). SPB-1/SPH− is a cell line derived from the SPB-1 cell, produced in our laboratory, that is stably transfected with the cDNA of the four adult mouse AChR subunits (Roccamo et al. 1999). In a first series of experiments SPB-1/SPH− cells were employed to study the effect of SL deficiency on AChR trafficking to the cell surface. The AChR-expressing SPB-1/SPH− cells were grown in Ham’s F-12 complete medium (i.e., containing 10% BFS) at 36–37°C for 48 h, and subsequently grown for 24 h in Nutridoma medium (a lipid-depleted medium) containing 0.1% BFS (Hanada et al. 1992). When we analyzed the plasma membrane-AChR levels by quantitative fluorescence microscopy, the cell-surface Alexa488-αBTX signal was found to diminish by about 40% in the SPB-1/SPH− cell line, as compared with CHO-K1/A5 cells (Fig. 1a), in agreement with our previous results using [125I]-αBTX (Roccamo et al. 1999). In order to test the specificity of these phenomena, CHO-K1/A5 and SPB-1/SPH− cells were transfected with VSVG-GFP and analyzed by quantitative fluorescence microscopy. As shown in Fig. 1b, no apparent differences in total or plasma membrane VSVG-GFP were observed between CHO-K1/A5 and SPB-1/SPH− cells, and the cellular distribution of this protein was similar in the two cell lines (Fig. 1b).
Cell-surface αBTX sites upon inhibition of serine palmitoyltransferase in control CHO-K1/A5 cells
In order to verify that the diminished AChR cell-surface labeling in the temperature-sensitive SPB-1/SPH− cells was indeed caused by impaired SPT activity, we employed ISP-1, a potent inhibitor of this enzyme, on control CHO-K1/A5 cells having normal SL metabolism. CHO-K1/A5 cells were grown for 48 h in complete medium, further incubated for 24 h with 5 μmol/L ISP-1 in Nutridoma medium, and labeled with Alexa488-αBTX (or [125I]-αBTX, data not shown). The fluorescence signal of cell-surface AChR was found to diminish by ∼30% (Figs 2a and b), in agreement with the results obtained with the temperature-sensitive mutant SPB-1/SPH− cells (Fig. 1).
Downstream inhibition of sphingolipid biosynthesis
The next series of experiments was designed to inhibit downstream SL biosynthesis at the level of dihydroceramide production (see Supplementary material Scheme 1). This was accomplished by growing CHO-K1/A5 cells with FB-1, a toxic fungal metabolite of Fusarium moniliforme that impairs SL biosynthesis by inhibiting the enzyme (dihydro)ceramide synthase. The cell-surface fluorescence intensity of Alexa488-αBTX (or [125I]-αBTX signal, data not shown) diminished by about 30% upon incubation of CHO-K1/A5 cells with 50 μmol/L FB-1 in Nutridoma for 24 h (Figs 2a and b). Complex GSL biosynthesis can also be inhibited at the level of ceramide glycosylation by blockage of the enzyme GluCer synthase. Incubation with 10 μmol/L PDMP for 24 h resulted in a reduction in fluorescent αBTX sites (or [125I]-αBTX signal, data not shown) at the plasmalemma similar to that observed with ISP-1 or FB-1, i.e., ∼30% (Figs 2a and b).
Block of sphingolipid biosynthesis leads to intracellular accumulation of nicotinic acetylcholine receptor
The obvious question raised by the diminution of cell-surface Alexa488-αBTX was whether defective SL biosynthesis could affect receptor biosynthesis or alter its traffic to the plasma membrane, with a concomitant accumulation of intracellular receptor. In order to measure intracellular AChR, CHO-K1/A5 cells were treated with the drugs for 24 h as described in the preceding sections (5 μmol/L ISP-1, 50 μmol/L FB-1, or 10 μmol/L PDMP), the surface AChRs were blocked with native toxin, fixed, permeabilized and AChR stained with Alexa594-αBTX. As shown in Figs 2a and b, in all cases in which SL synthesis was impaired and cell-surface receptor levels diminished, intracellular AChR levels were concomitantly elevated (∼50%), thus discarding the possibility that the diminution of cell-surface AChR resulted from diminished AChR biosynthesis.
The distribution of AChR inside the cell did not appear to be restricted to a particular organelle, unlike what happens when CHO-K1/A5 cells are chronically depleted of Chol (Pediconi et al. 2004). The AChR, as labeled with Alexa488-αBTX (or with mAb 210 anti-AChR antibody, data not shown) was found to be distributed all across the cytoplasm, as was the ER marker, calnexin (see Fig. S1). A certain fraction of the AChR label was co-localized with the TGN marker syntaxin-6 (see Fig. S2).
Reversal of altered nicotinic acetylcholine receptor surface/intracellular ratio by exogenous sphingomyelin addition
We studied next whether exogenously supplied SM could modify the effect exerted by the inhibitors of SL biosynthesis. We selected SM because this lipid has a higher affinity for the AChR than GSL (Bonini et al. 2002; Mantipragada et al. 2003). When CHO-K1/A5 cells were treated for 24 h with FB-1, PDMP or ISP-1 in the presence of exogenous (10 μmol/L) SM, the amount of AChR reaching the plasma membrane, as assayed using [125I]-αBTX binding, was restored to the levels observed in control, intact cells. We observed that the amount of cell-surface [125I]αBTX-AChR complexes was not only brought back in all cases, but was significantly higher than that of control cells grown in complete medium (Fig. 3a). The effect of the inhibitors is thus abolished by supplying exogenous SM to cells having pharmacologically induced reduction of endogenous SL biosynthesis. The fate of intracellular AChR was analyzed in parallel in cells treated with inhibitors plus exogenous SM using quantitative fluorescence microscopy. As shown in Fig. 3b, the intensity of Alexa594-αBTX staining differed from that observed in cells treated with inhibitors alone (see Fig. 2b); fluorescence was slightly lower than that of control cells not treated with the SL inhibiting drugs. These results indicate that SM, or its sphingoid backbone (see Supplementary material Scheme 1), bypassed the drug effects, restoring the normal surface/intracellular AChR ratio when CHO-K1/A5 cells were co-incubated with SM and the inhibitory drugs.
The intracellular accumulation of nicotinic acetylcholine receptor is associated with defective nicotinic acetylcholine receptor assembly
We next asked whether the retention of AChR at the ER could be because of defective AChR pentamer assembly. To answer this we developed a new assay, which exploits the pharmacological properties of the competitive antagonist αBTX binding in the presence or absence of the full agonist Carb. The basis of the assay is as follows: Carb recognizes and binds to AChR α subunits associated with the ε (or γ in the embryonic AChR) or δ subunits in the form of dimers or trimers with β subunits, but does not bind to unassembled α subunits (Blount and Merlie 1988). In contrast, αBTX recognizes both unassembled and assembled α subunits. Therefore, exposure of cells to Carb followed by the addition of αBTX enables one to distinguish whether the α subunits are assembled or not (Keller et al. 2001). We treated CHO-K1/A5 cells with FB-1, PDMP or ISP-1 for 24 h in Nutridoma medium, and stained the intracellular AChR with Alexa594-αBTX in the presence of excess (10 mmol/L) Carb. Under these conditions, we observed a 30–40% increase in intracellular Alexa594-αBTX fluorescence intensity (Fig. 4a), corresponding to the pool of unpaired AChR α subunits, i.e., unassembled receptors, thus indicating that impairment of SL biosynthesis affects the efficiency of AChR assembly.
We next studied whether exogenous SM addition, which restored cell-surface/intracellular AChR levels in cells with impaired SL biosynthesis (see Fig. 3), could affect receptor assembly. CHO-K1/A5 cells were co-incubated with the SL inhibiting drugs (FB-1, PDMP or ISP-1) together with 5 μmol/L SM for 24 h in Nutridoma medium, stained with Alexa594-αBTX in the presence or absence of 10 mmol/L Carb, and imaged by fluorescence microscopy. Quantitative analysis of the fluorescence micrographs showed that exogenous SM was effective in restoring the levels of assembled AChR in SL-inhibited cells to those characteristic of control CHO-K1/A5 cells (ca. 40%, Fig. 4b).
Effect of high concentrations of PDMP on nicotinic acetylcholine receptor trafficking to the plasma membrane
In a next series of experiments, we used PDMP at high concentrations as in (Rosenwald et al. 1992), under which conditions the drug inhibits SM as well as GSL biosynthesis (see Supplementary material Scheme 1). Lipid analysis of CHO-K1/A5 cells treated with 10 and 50 μmol/L PDMP (Fig. 5a) confirmed this to be the case; GluCer synthesis decreased ∼65 and 90% with respect to untreated cells, at 10 and 50 μmol/L PDMP, respectively. SM synthesis increased at low PDMP concentration and decreased more than 50% at the high PDMP concentration. On the other hand, ceramide, precursor of SL synthesis, increased as the SL inhibition augmented (200 and 700% with respect to control cells at 10 and 50 μmol/L PDMP, respectively). When the amount of AChR in the plasma membrane was measured by radioligand binding assays with [125I]αBTX (Fig. 5b) or using fluorescence microscopy with Alexa488-αBTX (Figs. 2b and 5c,d, respectively) the AChR was found to diminish to 80% and 70%, respectively, at low PDMP concentrations, and to 50% and 35%, respectively, at high PDMP concentrations. Addition of exogenous SM (5 μmol/L) together with a high (50 μmol/L) PDMP dose did not prevent the effect of the drug (Figs 5c and d), at variance with what was observed for low PDMP concentrations, at which the SM biosynthetic pathway is not affected (Fig. 3). Thus the sphingoid backbone may not be available in the SL-salvage pathway (see Supplementary material Scheme 1 and Discussion) when the cells are incubated with high PDMP concentration. Figs 5a and b show that the amount of AChR at the plasma membrane decreases as the PDMP concentration increases; notably, the decrease in cell-surface AChR is accompanied by a concomitant decrease in GluCer levels. Taken together, these results suggest that GluCer plays a critical role in AChR trafficking.
In order to discard the possibility that the PDMP effect on AChR traffic is a consequence of ceramide accumulation (see Fig. 5a) we treated CHO-K1/A5 cells with 50 μmol/L PDMP in the presence or absence of 40 μmol/L ISP-1 as in Maceyka and Machamer (1997). When the AChR expression levels in the plasma membrane were analyzed by quantitative fluorescence microscopy we did not observe an attenuation of the PDMP effect; on the contrary, the cell-surface AChR levels decreased even more (data not shown). These results indicate that ceramide accumulation does not mediate the observed effects of PDMP on AChR traffic.
The amount of intracellular AChR in cells treated with the high (50 μmol/L) PDMP concentration amounted to ∼70% (Fig. 5e) of control intracellular AChR levels. This is similar to the levels of AChR reported by Quiram et al. (1999) as a result of impaired interaction of β and δ AChR subunits. Interestingly, PDMP treatment increased the levels of unassembled AChR by ∼30% (Fig. 5e). Because the absolute amount of intracellular AChR diminished, (Fig. 5e), the actual relative increase in unassembled AChR was even more pronounced, rising from 40% of the total intracellular pool in control cells to 80% in PDMP-treated cells, i.e., ∼100% (Fig. 5e).
An independent series of experiments was used next to verify the observed changes in AChR intermediate pools. These were assayed using sucrose density gradient analysis as reported by Barrantes (1982) and Kreienkamp et al. (1995). Briefly, control and PDMP-treated cells were incubated with native αBTX in order to block the surface AChR, permeabilized with saponin and incubated with excess [125I]- αBTX for 1 h at 25°C. Cells were solubilized in Triton X-100-containing buffer and samples were subjected to discontinuous sucrose gradient centrifugation. As shown in Fig. 5f, in control cells only the 1.7S peak was apparent at the top of the gradient, corresponding to free [125I]-αBTX. PDMP-treated cells showed a broader band corresponding to the free toxin together with the 3.3S α subunit band, reflecting the relative, albeit small, increase in unassembled α subunits upon impairing SL synthesis.
Sphingolipids depletion affects nicotinic acetylcholine receptor export at the early secretory pathway.
The previous results indicate that in SL-depleted cells AChR assembly is more inefficient than in control cells. In order to analyze the intracellular localization of the AChR, we tagged fixed and permeabilized cells with (i) fluorescent αBTX to label intracellular AChR, (ii) ER-tracker to label the ER, and (iii) syntaxin 6 to identify the TGN in control and PDMP (50 μmol/L, 24 h)-treated cells. As shown in Fig. 6a, the AChR was found predominantly in the ER, both in control (as described Ross et al. 1991) and treated cells (see Fig. S1). To find out whether AChR could further traffic along the exocytic pathway, cells were submitted to a temperature arrest at 20°C, which results in AChR accumulation in the TGN (Pediconi et al. 2004). As shown in Fig. 6b, there was a clear buildup of AChR in the TGN of control cells, whereas no accumulation occurred in PDMP-treated cells, the AChR being retained in the ER. To further ascertain the ER localization of the AChR, CHO-K1/A5 cells were labeled with anti-calnexin or anti-PDI to localize these resident ER proteins, and Alexa488-αBTX to label the AChR. As shown in Fig. 6c, the two proteins displayed a similar intracellular distribution in cells treated with high PDMP concentration.
Inefficient nicotinic acetylcholine receptor assembly is not a consequence of diminished availability of the chaperone calnexin
Calnexin is a chaperone protein that has been shown to contribute to AChR biogenesis (Chang et al. 1997) and to be associated with the unassembled receptor subunits in the ER (Keller et al. 1996). To test whether the observed effects of high PDPM concentration were in fact an indirect consequence of, e.g., diminution in calnexin levels, we treated CHO-K1/A5 cells with 50 μmol/L PDMP for 24 h and measured both intracellular AChR and calnexin levels by fluorescence microscopy. Calnexin levels were not diminished (Figs 7a and b), thus discarding the possibility that inefficient AChR assembly is a consequence of the diminished availability of the chaperone calnexin to assist AChR subunit assembly in the ER.
Cellular levels and expression at the plasma membrane of VSVG-GFP is not affected by inhibition of sphingolipid biosynthesis
We used VSVG-GFP to analyze the effect of impaired SL biosynthesis on cellular traffic of a protein other than the AChR. Rosenwald et al. (1992) showed that PDMP treatment in CHO-K1 cells did not affect the arrival of VSV-G at the cis-Golgi compartment, but that the subsequent steps in the secretory pathway were more affected. We transiently transfected CHO-K1/A5 cells with VSVG-GFP cDNA and grew these cells in complete Ham’s F-12 medium, Nutridoma, and Nutridoma medium containing 50 μmol/L PDMP for 24 h (Fig. 8a). No statistically significant differences were observed in the fluorescence intensity of total VSVG-GFP in CHO-K1/A5 cells grown in Nutridoma medium, with or without PDMP (Fig. 8b). Similarly, the levels of plasma membrane-associated VSVG-GFP were not statistically different between PDMP-treated cells and Nutridoma grown control cells. These results fully concur with those of Fig. 1b and indicate that the effect on AChR trafficking brought about by altered SL biosynthesis is not a general phenomenon.
In the present work we have studied the effect of SL deprivation on the cell-surface targeting of the AChR in non-polarized mammalian cell systems. We demonstrate that the inhibition of SL biosynthesis impairs the normal transport of AChR to the plasma membrane and promotes its accumulation inside the cell. We further show that SLs intervene at an early stage of the exocytic pathway and that their deficit provokes the retention and accumulation of unassembled receptor in the ER. On the basis of these results we hypothesize that SLs display a chaperone-like activity in the assembly and trafficking of AChRs.
Complex SLs comprise GSLs and SM, which differ in the occurrence of carbohydrates and phosphocholine in their polar head, respectively. Complex SLs contain the hydrophobic residue ceramide, which consists in turn of the base sphingosine and fatty acids. SL biosynthesis starts at the cytoplasmic face of the ER and is initiated by condensation of l-serine with palmitoyl CoA, catalyzed by SPT, followed by ceramide biosynthesis. Ceramide is subsequently translocated from the ER to the Golgi apparatus by both vesicular and non-vesicular transport (Fukasawa et al. 1999; Hanada et al. 2003), and then either converted to SM by the enzyme SM synthase on the lumenal face of the Golgi apparatus, or to GluCer by the enzyme GluCer synthase, on the cytosolic surface of the Golgi complex. After translocation into the Golgi lumen, GluCer is further converted to lactosylceramide and more complex GSL (van Meer and Lisman 2002).
Earlier studies from our laboratory demonstrated that a class of GSL, the ganglioside GM1, co-localizes with the AChR at the innervated ventral face of the electrocyte (Marcheselli et al. 1993). More recently, we showed that SMs are asymmetrically distributed in AChR-rich membranes from Torpedo electrocytes, are predominantly made up of saturated fatty acid chains, and exhibit moderate selectivity for the AChR protein (Bonini et al. 2002; Mantipragada et al. 2003). Like the GPI-anchored proteins, SLs have been reported to reside in lipid domains termed ‘rafts’ (Simons and Ikonen 1997).
In the present work we first determined that cell-surface AChR levels drastically diminished in CHO SPB-1/SPH− cells deficient in SL biosynthesis grown under semi-permissive temperature (37°C and SL-deficient medium).
The same effect on cell-surface AChR plasma membrane expression and intracellular accumulation was observed when the three drugs were used independently (Fig. 2). Both FB-1 and PDMP can cause the accumulation of bioactive compounds like sphinganine, sphinganine-1P, or ceramide (Radin et al. 1993; Merrill et al. 1997, 2001). The fact that ISP-1 inhibited SL biosynthesis without bioactive intermediate accumulation (Miyake et al. 1995; Hanada et al. 2000) excludes the possibility that metabolic intermediates affect the normal trafficking of AChR to the plasma membrane. The observed effect is more likely to be due, therefore, to final products of the SL biosynthetic pathway, SM and GSLs.
When we added an exogenous SL such as SM together with FB-1, PDMP (at low concentrations) or ISP-1, we observed that the amount of AChR in the plasma membrane was similar to or even higher than that in control cells; intracellular receptor levels were also similar. The effects of inhibiting SL biosynthesis are therefore reversible.
Comparison of the experiments with low and high concentrations of PDMP threw further light on the mechanism of SL inhibition of AChR trafficking. Low PDMP concentrations blocked GluCer synthase enzymatic activity only (see Supplementary material Scheme 1 and Fig. 5a); hence, the restoration of AChR trafficking to the plasmalemma observed upon addition of exogenous SM can be accounted for by an excess of SM generated by the downstream restoration of the SM biosynthesis by reutilization of the sphingoid backbone in the SL metabolic pathway (Tettamanti et al. 2003). According to this hypothesis, exogenous SM is first degraded in lysosomal compartments (Chigorno et al. 2005) and its backbone shuttled to the ER where it is reutilized for de novo SM biosynthesis. In contrast, high PDMP concentrations blocked both SM synthase and GluCer synthase (see Supplementary material Scheme 1), and exogenous SM addition was unable to restore AChR trafficking (Fig. 5). The explanation for this phenomenon is that under such conditions the sphingoid backbone in the ER cannot be utilized along the SL biosynthetic pathway.
Since inhibition of SL biosynthesis resulted in intracellular accumulation of AChR, we also analyzed the possibility that this might be due to the inability of AChR subunits to assemble into a mature AChR oligomer. This possibility was studied by an assay based on the protection of binding sites with the agonist Carb (Blount and Merlie 1988; Kreienkamp et al. 1995; Chang et al. 1997; Keller et al. 2001). In cells with impaired SL biosynthesis the binding of intracellular Alexa594-αBTX was higher in Carb-exposed cells (cf. Fig. 4), indicating a lower efficiency in the subunit assembly process. This result is in agreement with the fluorescence microscopy experiments showing colocalization of intracellular AChR and ER probes: AChR subunits and unassembled receptors are retained in the ER (cf. Fig 6 and Fig. S1). This suggests that the quality control mechanisms somehow recognize trafficking signals in unassembled subunits which might be hidden in fully assembled receptors (Keller et al. 2001; Wang et al. 2002).
When we added SM to the cells together with the FB-1, PDMP (at low concentrations) or ISP-1, no differences were observed in the amount of assembled AChR with respect to untreated cells (Fig. 4). These results eliminate the possibility that the drugs affected AChR protein synthesis and the rate of receptor subunit folding and degradation in the ER, since if this were the case the addition of SM to cells with impaired SL biosynthesis would not have affected these parameters the way it did. In the case of high PDMP concentrations, we tested the amount of cellular calnexin in order to explore the possibility that the inefficient assembly process might be due to the diminished availability of this chaperone protein. This was found not to be the case (Fig. 7).
How can impaired SL biosynthesis affect AChR assembly? One possibility is that the optimal lipid environment of the AChR (or AChR subunits) in the ER membrane is disturbed, and this in turn affects correct inter-subunit association for normal assembly of the AChR oligomer. Alternatively, SLs could act as molecular chaperones in protein folding, as suggested by Fantini (2003) when analyzing the influence of the SLs on the conformation of HIV gp-120, PrP and the β-amyloid peptide. These proteins undergo significant conformational transitions following binding to SLs (Fantini et al. 2002; Mahfoud et al. 2002) and they share a motif termed the SL binding domain, which is also present in the AChR (Fantini, personal communication).
Phospholipids have also been proposed to act as molecular chaperones. The presence of phosphatidylethanolamine in membranes of E. coli is necessary for the correct assembly of lactose permease (Bogdanov et al. 1996; Bogdanov and Dowhan 1999). Recently, phospholipid chaperones were shown to be important in the pathogenesis of cystic fibrosis by perturbing CFTR trafficking (Eidelman et al. 2002).
Several factors intervene in the AChR folding and assembly processes (Wanamaker et al. 2003). These processes are ER chaperone-dependent, and AChR–chaperone interactions could be, in part, responsible for the slow kinetics of AChR assembly. The majority of the synthesized AChR subunits are normally degraded, and the ER-associated degradation (ERAD) systems play a key role in this process, as is apparent from the enhancement in mature AChR expression when the ERAD systems are blocked (Wanamaker et al. 2003). Cyclic AMP levels are also important in the assembly process, since increased cytoplasmic cAMP concentrations increase the surface expression of the AChR (Ross et al. 1991).
Lipid gradients occur along the secretory pathway: SLs are enriched in the plasma membrane and endocytic membranes, whereas only low amounts of SLs are found in the ER (Holthuis et al. 2001). Two GPI anchored proteins, Cwp2 and Gas1/Gpg1, are not transported normally from the ER to the Golgi when SL biosynthesis is inhibited (Skrzypek et al. 1997). GPI-anchored proteins were found to be associated with detergent-resistant membranes in the ER of HeLa cells (Sevlever et al. 1999), challenging previous work reporting that this type of protein acquires detergent insolubility upon reaching the Golgi apparatus (Brown and Rose 1992). More recently, the occurrence of protein-stabilizing associations was found between the prion protein PrPC and lipid rafts in the ER (Sarnataro et al. 2004). Impairment of raft association by cholesterol depletion during PrP biosynthesis led to protein misfolding (Sawamura et al. 2004).
SL levels in the ER, though low, could play an important role in stabilizing mature proteins exiting from this organelle. The detailed mechanisms are yet to be elucidated, but one could hypothesize that SLs participate as lipid chaperones in protein–lipid microdomain associations – the microdomains acting as sorting platforms- and/or in ER transport vesicle formation. Deficient intracellular levels of SLs appear to lead to the retention of unassembled subunits of a transmembrane protein, the AChR, in the ER, and to significantly reduce the trafficking of the mature oligomers to the cell surface in the non-polarized cell paradigm used in the present work.
This work was supported by grants from FONCYT and Universidad Nacional del Sur to FJB.