| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Medical Chemistry, Biochemistry and Biotechnology and Center of Excellence on Neurodegenerative Diseases, University of Milan, Segrate, Italy;
Department of Biomedical Sciences and Biotechnology, University of Brescia, Brescia, Italy; and
Division of Biochemistry, Research Institute, Miyagi Prefectural Cancer Center, Natori, Miyagi, Japan
1Correspondence: Department of Medical Chemistry, Biochemistry and Biotechnology and Center of Excellence on Neurodegenerative Diseases, University of Milan, Via Fratelli Cervi 93, Segrate 20090 (Milan, Italy). E-mail: sandro.sonnino{at}unimi.it
ABSTRACT
Ceramide is a key lipid molecule necessary to regulate some cellular processes, including apoptosis and cell differentiation. In this context, its production has been shown to occur via sphingomyelin hydrolysis or sphingosine acylation. Here, we show that in human fibroblasts, plasma membrane ceramide is also produced from ganglioside GM3 by detachment of sugar units. Membrane-bound glycosylhydrolases have a role in this process. In fact, the production of ceramide from GM3 has been observed even under experimental conditions able to block endocytosis or lysosomal activity, and the overexpression of the plasma membrane ganglioside sialidase Neu3 corresponded to a higher production of ceramide in the plasma membrane. The increased activity of Neu3 was paralleled by an increase of GM3 synthase mRNA and GM3 synthase activity. Neu3-overexpressing fibroblasts were characterized by a reduced proliferation rate and higher basal number of apoptotic cells in comparison with wild-type cells. A similar behavior was observed when normal fibroblasts were treated with exogenous C2-ceramide.—Valaperta, R., Chigorno, V., Basso, L., Prinetti, A., Bresciani, R., Preti, A., Miyagi, T., and Sonnino, S. Plasma membrane production of ceramide from ganglioside GM3 in human fibroblasts.
Key Words: plasma membrane-associated sialidase • Neu3 • GM3 synthase • glycosylhydrolases • apoptosis
PLASMA MEMBRANE SPHINGOLIPIDS are a class of lipid mediators. Following various stimuli, the production of their catabolic fragments ceramide, sphingosine, and sphingosine-1-phosphate, leads to cell proliferation, cell differentiation, or apoptotic cell death, cell contraction, retraction, and migration (2)
. Sphingomyelin is considered the precursor of the sphingoid fragments (3)
. Ceramide glycanase, the enzyme that catalyzes the one-step release of the oligosaccharide chain from glycosphingolipids with the liberation of free ceramide, has been described in bacteria and low eukaryotes (4)
.
Gangliosides, sialic acid containing glycosphingolipids (1)
, have been described to participate to the cell-to-cell signaling process and to the process of signal transduction through the membrane (5)
. Sialidases remove sialic acid residues from sialocompounds (6)
. The membrane-associated sialidase Neu3 (7
8
9)
triggers selective ganglioside desialylation in neuroblastoma cells, thus modulating cell growth, differentiation, and neuritogenesis (10
11)
. Neu3 has been shown to release sialic acid from sialoglycolipids of neighboring cells through cell-to-cell interactions in COS-7 cells (9)
. A specific role of Neu3 has been proposed in cancer (12
13)
, in which high concentrations of the enzyme would maintain high levels of lactosylceramide considered an antiapoptotic compound, thus allowing cell proliferation.
Here, we report the first direct evidence that the plasma membrane-associated sialidase Neu3, in coordination with other cell surface glycosylhydrolases, modulates cell proliferation and cell apoptosis, by regulating the production of plasma membrane ceramide from gangliosides in cultured human fibroblasts.
MATERIALS AND METHODS
High-performance silica gel thin-layer plates (HPTLC Kieselgel 60, 10 x10 cm) were purchased from Merck GmbH (Milan, Italy). Vibrio cholerae sialidase was from Sigma (Milan, Italy). The methylumbelliferyl derivatives of 
-Gal and ß-glucose were from Sigma. [6-3H]-acetyl-D-mannosamine (20 Ci/mmole) was from American Radiolabeled Chemicals (St. Louis, MO) and [3H]thymidine (12 Ci/mmole) and [3-3H]sphingosine (23.1 Ci/mmole) were from Perkin Elmer (Norwalk, CT).
Sphingosine was prepared from cerebroside (14)
; C2-ceramide was prepared by acetylation of sphingosine using acetic anhydride in methanol; [1-3H]sphingosine (2.0 Ci/mmole) was prepared from sphingosine (15)
. Ganglioside GM3 was prepared from bovine brain GM1 (16)
. The preparation of isotopically labeled [3-3H(sphingosine)]GM3 (specific radioactivity, 2.3 Ci/mmole) and [3-3H(sphingosine)]GD1a (specific radioactivity, 1.3 Ci/mmole) have been described in detail (17)
. Bovine brain SM (Avanti Polar Lipids, Alabaster, AL) was labeled as for GM3. LacCer was prepared by mild acidic hydrolysis of GM1 (18)
. [3-3H(sphingosine)]LacCer was prepared from [3-3H(sphingosine)]GM3 (specific radioactivity, 2.3 Ci/mmole) by treatment with vibrio cholerae sialidase (18)
. [1-3H(sphingosine)]Cer (specific radioactivity, 2.2 Ci/mmole) was prepared by direct acylation of [1-3H(sphingosine)] (specific radioactivity, 2.0 Ci/mmole) with palmitic anhydride. Radioactive sphingolipids were extracted from cells fed with [1-3H]sphingosine, purified, characterized as described previously and used as chromatographic standards (19)
.
Normal human skin fibroblasts were cultured and propagated as described previously (20)
in 100-mm dishes (0.42±0.10 mg cell protein/dish), using 10% FCS-EMEM and used for the experiments when confluent. Mock and Neu3 cells were cultured in 100-mm plastic dishes containing 15% FCS-EMEM.
Expression of mouse ganglioside sialidase cDNA in normal human fibroblasts
The plasmid construct (PC) DNA-Neu3HA (21)
was digested with EcorI restriction enzyme. The product of the digestion was cloned into pIRES-neo eukaryotic expression vector. The presence, orientation and fidelity of the cDNA in the vector were confirmed by DNA sequencing. Human fibroblasts were transfected by nucleofection with Nucleofector System (Instrumentation Laboratory, Barcelona, Spain), with the empty expression vector or the expression vector containing the mouse ganglioside sialidase cDNA by the following procedures. Cells that were washed twice with PBS were resuspended in Nucleofector solution, according to manufacturer’s instructions, at a final concentration of 5 x 105 cells/100 µl. The suspension was added to sterile cuvettes with 5 µg of plasmid DNA. The cuvette was placed on ice for 10 min and electrophorated using specific program by Instrumentation Laboratory. The cuvette was placed on ice for an additional 10 min and cells were cultured in 4 ml of medium in 35-mM dishes. 24 h after nucleofection, selection started in growth medium containing 200 µg/ml neomycin. Following 5 wk of selection, few colonies were subcloned and expanded.
Semiquantitative real-time polymerase chain reaction for endogenous Neu3
Total RNA was isolated from mock and Neu3 transfected cells using TRIZOL reagent, followed by RNase free DNase I treatment according to the manufacturer’s instructions. cDNA was prepared from 1 µg total RNA using cDNA synthesis kit for real-time polymerase chain reaction (RT-PCR; Invitrogen, Carlsbad, CA), according to the manufacturer’s protocol. Fifty nanograms of each cDNA were diluted to a volume of 25 µl PCR mix (Premix Taq, Ex TaqTM R-polymerase chain reaction Version, Takara Bio Inc., Shiga, Japan) containing primers and probe.
The analysis of human endogenous Neu3 in mock and in Neu3-transfected cells was performed using Smart Cycler System (CEPHED).
The sequences and final concentrations of primers in the reaction mixture for human sialidase Neu3 were: 0.3 µM forward 5'-CCTGAAGCCACTGATGGAA -3', 0.3 µM reverse 5'-TTCCTGCCTGACACAATCTG-3' and 0.2 µM probe FAM-5'-CCACACTACCGGGGCATCGG-3'-tamra (GenBankTM accession no. NM_006656). For amplification, the initial denaturation at 95°C for 10 s (1 cycle) was followed by a second step at 95°C for 10 s, 60°C for 15 s (40 cycles). To normalize data obtained, the ß-actin expression was used as internal control. The sequences of forward primer was 5'-CGACAGGATGCAGAAGGAG-3' and reverse primer was 5'-ACATCTGCTGGAAGGTGGA-3'.
The relative expression of endogenous Neu3 in Neu3 transfected cells compared to mock cells was normalized to the expression of human ß-actin and was calculated by equation 2-
Ct where Ct is the cycle threshold and Ct = (Ctendogenous Neu3 – Ctß-actin)Neu3 cells – (Ctendogenous Neu3 – Ctß-actin)mock cells. To demonstrate that amplification efficiency of Neu3 and ß-actin were approximately equal, Ct were determined using 5, 10, 25, 50 and 100 ng of total RNA. After the amplification by real-time PCR the products were loaded on gel electrophoresis.
RNA isolation and RT-PCR analysis
Total RNA was isolated by the single-step acid-guanidine isothiocyanate-phenol-chloroform extraction method, according to the manufacturer’s instructions. One microgram of RNA was treated with 1 U of RNase-free DNase (Invitrogen) for 15 min at room temperature to remove any possible DNA contamination. Complementary DNA (cDNA) was synthesized with 4 µg of total RNA using SuperScriptTM III Reverse Transcriptase (RT; Invitrogen) in a final 20-µl reaction vol.
For PCR amplification of mouse Neu3-hemagglutinin cDNA we used a forward primer 5'-GACTTGGTGGCGTGTTTGTT-3' and reverse primer 5'-TTAAGCGTAATCCGGAACATC-3' (GenBankTM accession no. NM_016720). Reverse sequence primer is localized on tag hemagglutinin (HA); for the GM3 synthase cDNA, we used a forward primer 5'-AATGGCGCTGTTATTTGAGC-3' and reverse primer 5'-CTGGCAAGAGTTCCAAGAGG-3' (GenBankTM accession no. AY152815) and for the Gb3Cer synthase cDNA we used a forward primer 5'-TTCTCAAGAACCTGCGGAAC-3' and reverse primer 5'-GATCCAGCCGTTGTAGTGGT-3' (GenBank accession no. NM_017436). As in control, housekeeping gene ß-actin cDNA was measured at the same time. Thirty-five cycles were performed at 94°C (denaturation), 58°C (annealing) and 72°C (elongation). The PCR products were subjected to electrophoresis in 1,2% (w/v) agarose gel and were visualized by UV after ethidium bromide staining.
Indirect immunofluorescence
Mock and Neu3-transfected cells, grown in slide culture chamber, were briefly washed with PBS and fixed for 10 min with 4% paraformaldehyde in PBS. For permeabilization, cells were incubated for 15 min in the presence of 0.5% Triton X-100 in PBS for 2 min. Cells were then washed twice for 5 min with PBS and incubated for 1 h with mouse anti-hemagglutinin (1: 200, Sigma) or with a commercial monoclonal antibody for 6ß1-integrin (1:200, Serotec, Raleigh, NC), in PBS containing 1% BSA. Samples were then washed twice with PBS for 10 min and incubated with Cy2-conjugated donkey antimouse IgG (1:200, Jackson ImmunoResearch Laboratories, West Grove, PA) for detection of the Neu3-hemagglutinin and Cy3-conjugated donkey antimouse IgG (1:200, Jackson ImmunoResearch Laboratories) for detection of the 6ß1-integrin, in PBS containing 1% BSA. After two washes with PBS samples were embedded with fluorescence mounting medium (DAKO, Glostrup, Denmark). Laser confocal analysis was performed with Bio-Rad 1024 system (Bio-Rad Laboratories, Hercules, CA) and images were processed with Adobe Photoshop software.
Cell proliferation assay
5 x 105 cells were grown in 35-mm Falcon dishes and incubated with 1 µl medium containing 0.5 µCi of [3H]thymidine. After 1.5, 6, 20, and 48 h at 37°C, cells were harvested with 500 µl PBS, treated with 500 µl trichloroacetic acid 20%, shaken for 30 min at 4°C and centrifuged at 13,000 rpm for 15 min. After the removal of the supernatant, 500 µl of 20% trichloroacetic acid were added to the pellets, and mixtures treated as before. After 1 h at 37°C, 50 µl of 0.4 M HCl were added to the reaction mixtures that were analyzed for radioactivity content.
Determination of cell apoptosis markers
Cell homogenates from mock and stable transfected cells (35 µg of cell protein) were analyzed by SDS/PAGE gel electrophoresis, proteins were transferred to PVDF membranes, and the presence of Bcl-2 and caspase-3 were assessed by immunoblotting with specific antibodies, followed by reaction with peroxidase-conjugated secondary antibody (Ab) and enhanced chemiluminescence detection.
Expression of cytosolic cytochrome c was determined as described previously (22)
.
Cell treatments with C2-ceramide and DNA fragmentation assay
Mock and Neu3-transfected cells were plated in 96-well plastic plates (104 cells/well) and cultured for 48 h. Then the cells were incubated for 6, 12, and 24 h with 25 x 10–6 M C2-ceramide. After incubation, DNA fragmentation was investigated using a cell death detection ELISAPLUS kit (Roche Diagnostics, Munich, Germany), according to the manufacturer’s protocol. Each measurement was done in triplicate and the apoptotic index defined by the quantification of mono- and oligonucleosomes. The optical density405 is normalized to the total milligrams protein content of the sample used in the assay.
Metabolic tritium labeling of cell sphingolipids
Mock and transfected cells (48 h after seeding) were fed 3 x 10–8 M [3-3H]sphingosine (2 ml/dish). After a 2-h pulse, the medium was removed and replaced with fresh medium without radioactive sphingosine, and cells were incubated for 48 h (chase), allowing metabolic radiolabeling of all sphingolipids (including ceramide, SM, neutral glycolipids, and gangliosides) (19)
.
Metabolic tritium labeling of cell sialoglycoproteins
Mock and transfected cells (48 h after seeding) were fed 2.25 x 10–7 M [6-3H]acetyl-D-mannosamine (5 ml/dish) in culture medium for 24 h and then chased for 24 and 48 h. After the chase, the cells were washed twice with PBS and resuspended in sterile water. 30 µg of cell proteins were separated by 10% SDS-PAGE and blotted to a PVDF membrane. Radioactive proteins were visualized by radioimaging.
Lipid extraction and analysis
Cells were harvested in ice-cold water (2 ml) by scraping with a rubber policeman, and lyophilized. Lipids were extracted twice with chloroform/methanol/water 2:1:01 by volume (first extraction, 1.5 ml; second extraction, 0.25 ml). Total lipid extracts were analyzed by HPTLC performed with the solvent system chloroform/methanol/0.2% aqueous CaCl2, 55:45:10 (v/v), followed by radioactivity imaging and radioactivity quantification. The identity of radioactive lipids was assessed by comparison with standard lipids.
Enzyme activities
Activity of GM3 synthase, ceramidase, ß-galactosidase, sialidase and sphingomyelinase were determined in cell homogenates. Negative controls were performed using heat-inactivated cell homogenates (100°C for 3 min). The enzymatic reactions were stopped by adding chloroform/methanol (2:1) and analyzed by HPTLC using the solvent system chloroform/methanol/water, 55:20:3 by volume or 55:45:10 chloroform/methanol/0.2% aqueous CaCl2. Separated radioactive lipids were detected and quantified by radioactivity imaging.
GM3 synthase activity was assayed as described previously in a cell-free assay using [3-3H(sphingosine)]LacCer as a substrate (23)
.
To perform ceramide assay, cell homogenates were prepared as described for the GM3 synthase assay (23)
. The activity of ceramidases was assayed as described previously using [1-3H(sphingosine)]Cer as a substrate (23
24)
.
ß-Galactosidase, ß-glucosidase and -galactosidase assays were performed on cells. Then cells washed twice with PBS, suspended in sterile water, and pelleted. Activities on artificial substrates were determined as reported (25)
.
ß-Galactosidase activity was determined on the natural substrate. In each reaction tube containing 0.1 ml of sodium taurocholate (1 mg/ml in C/M 2/1) and 50 µg of the substrate [3-3H(sphingosine)]LacCer (corresponding to 1,35 nCi), from a stock solution in chloroform/methanol 2/1 (v/v), were dried under nitrogen flow. To this mixture, 0.1 ml of sterile water was added and the tubes were sonicated. ß-Galactosidase activity was assayed by adding 0.2 ml of 0.2 M sodium citrate buffer (pH 4.8) containing 20 mM NaCl, 0.1 ml of enzyme source (containing 350 µg cellular protein) in a total reaction volume of 400 µl. Incubation was carried out at 37°C for 2 h.
Sialidase assays were performed on mock and stable Neu3 fibroblasts. Then, they were washed twice with PBS, suspended in sterile water, and pelleted. The membrane bound Neu3 and the lysosomal Neu1 were assayed as described previously (26)
, using both GD3 and GD1a gangliosides.
For the sphingomyelinase assays, cell homogenates were prepared in 150 mM sodium cacodylate-HCl buffer, pH 6.6 (20 mg of cell protein/ml) with protease inhibitors (2 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 0.0016 mM aprotinin, 0.044 mM leupeptin, 0.08 mM bestatin, 0.03 mM pepstatin A, 0.028 mM E-64) (Sigma) and 0.2% Triton-X 100. The activity of acid and neutral sphingomyelinase was assayed as follows. In each reaction tube, 25 µl of 0.2% Triton-X100 (v/v) in chloroform/ methanol (2:1) was mixed with 0.74 µM of [3H]sphingomyelin (corresponding to 13.5 nCi) from a stock solution in chloroform/ methanol (2:1) and dried under N2. To this mixture, 25 µl of 250 mM sodium acetate buffer, pH 5.2 (for the assay of the acidic enzyme) or 25 µl of 40 mM HEPES, 5 mM MgCl2, pH 7.4 (for the assay of neutral enzyme) and 25 µl of cell homogenate (containing 100 µg of protein for the acidic sphingomyelinase and 10 µg for the neutral enzyme) were added in a total reaction volume of 50 µl. The incubation was performed at 37°C for 2 h with continuous shaking.
Activity of Neu3 and membrane-associated sphingomyelinase on [3-3H(sphingosine)]GM3 and of [3-3H(sphingosine)]SM-administered to cells
[3-3H(sphingosine)]GM3 and [3-3H(sphingosine)]SM were administered to cells, and their fate was determined under conditions that prevent lysosomal catabolism. To do this, cells were kept at 36°C with 50 µM chloroquine or 10 mM ammonium chloride or kept at 4°C for 30 min in EMEM. After removal of the medium and rapid washing of cells with EMEM, 2 ml of the medium containing the radioactive lipid were added to each dish, and the cells were incubated at 36°C in the presence of 50 µM chloroquine or 10 mM ammonium chloride, or at 4°C for 12 h. Medium containing tritium-labeled GM3 or SM was prepared as follows. The tritiated sphingolipid, dissolved in propan-1-ol/water, 7:3 (v/v), was pipetted into a sterile tube and dried under a nitrogen stream. The residue of [3-3H(sphingosine)]GM3 was solubilized in Eagle’s minimum essential medium (EMEM) containing 10% FCS, 1% glutamine, 1% penicillin/streptomycin, to obtain a GM3 concentration of 4.5 x 10–6 M. The residue of [3-3H(sphingosine)]SM was solubilized in the above medium without FCS to obtain a final concentration of 6 x 10–6 M.
At the end of incubation, cells were washed four times with complete EMEM, twice with PBS and scraped off with water by mean of a rubber policeman. Samples were lyophilized and submitted to lipid extraction resulting in a delipidized pellet and a total lipid mixture. Radioactive lipids were separated by HPTLC (1000 dpm/lane), and analyzed by radioimaging (48 h acquisition) using a Beta-Imager 2000 Instrument (Biospace, Paris, France).
Other analytical methods
The radioactivity associated with lipids, with total lipid extracts and with DNA was determined by liquid scintillation counting.
The protein assays were carried out according to Lowry (27)
, using BSA as the reference standard or using the bicinchoninic acid protein assay kit.
RESULTS
Figure 1
, lane 1, shows the radioactive sphingolipid pattern of cells fed with radioactive GM3. After TLC separation, the catabolic products lactosylceramide and ceramide were identified, together with some SM. The presence of radioactive SM has been shown to be associated with the very rapid recycling of lysosomal sphingosine produced in the catabolism of administered ganglioside (19)
.
|
To verify whether a part of the radioactive ceramide formed from GM3 administered to cells was produced at the plasma membrane, we fed cells with GM3 also in the presence of chloroquine and ammonium chloride, which are known to block lysosomal activity (28
29
30)
, or at 4°C, a condition known to block endocytosis (31
32
33)
. Figure 1
, lanes 2 to 4, shows that ceramide and traces of lactosylceramide, but not SM, are produced from GM3 in fibroblasts under experimental conditions which are able to block lysosomal activity or endocytosis. The lack of SM production is good proof that GM3 in the presence of chloroquine or ammonium chloride does not actually reach lysosomes or is in any way metabolized in the lysosomal compartment. In absence of lysosomal catabolism of GM3, sphingosine cannot be formed and recycled for the biosynthesis of SM. These results suggest that the observed ceramide is produced in the membrane.
The above described set of experiments was repeated on cells overexpressing the plasma membrane sialidase Neu3 to examine the possible increase in production of ceramide, parallel to the increase of membrane bound sialidase activity.
A full-length cDNA encoding mouse sialidase Neu3 cDNA, having the HA tag at the C-terminal was cloned and inserted into a mammalian expression vector. Transfected fibroblasts did not show any morphological change with respect to control cells. Positive clones were determined on the basis of Neu3-hemagglutinin mRNA levels by RT-PCR. Figure 2
A shows the expression levels of mRNA of mouse Neu3-hemagglutinin in Neu3-hemagglutinin-transfected cells. Of course, as expected, in mock cells, we did not find any Neu3-hemagglutinin product. To analyze hypothetical changes of endogenous Neu3 after cell transfection with mouse Neu3-hemagglutinin mRNA, we performed semiquantitative real-time PCR. Figure 2B
and Table 1
shows that any change of the Neu3 content occurred in Neu3-hemagglutinin transfected cells with respect to mock cells.
|
|
Figure 3
shows that the protein was mainly localized on plasma membrane in transfected cells. These cells had sialidase activity 2 times higher than mock cells, as determined on natural substrates GM3 and GD1a (Table 2
).
|
|
Figure 1
, lane 1a, shows that transfected cells fed with GM3, produced lactosylceramide, ceramide, and SM. Transfected cells produced a high quantity of ceramide, about 3 times that produced by mock cells, and a similar quantity of lactosylceramide (Table 3
). Instead, ceramide, but not lactosylceramide, was formed blocking the cell lysosomal activity with chloroquine and ammonium chloride, or maintaining cells at 4°C to block endocytosis by transfected cells fed with GM3 (Fig. 1
, lanes 2a to 4a). These results obtained feeding Neu3 cells with GM3 confirm that some ceramide is produced at the plasma membrane, by involvement of Neu3, whereas the observed lactosylceramide is largely produced in lysosomes.
|
Neutral glycosphingolipids could not be observed blocking the lysosomal ganglioside catabolism. This suggests that in the membrane, detachment of neutral sugars is faster than that of sialic acid. In addition, intermediate catabolytes could not even be observed in Neu3 cells in which the sialidase activity is quite high. This would suggest that in Neu3 cells the plasma membrane galactosidase and glucosidase activities are also increased. There are no methods to determine plasma membrane activities of these two enzymes, but Table 2
shows that the total cell enzyme activities on the neutral sugars dramatically increased in transfected cells.
The activity of the three known ceramidases, the neutral, the acidic and the basic, were comparable in transfected and normal cells. Thus, ceramide accumulates in transfected cells.
The sphingolipid pattern of human fibroblasts and Neu3-transfected cells, as determined after feeding tritiated sphingosine to cells, lipid extraction, TLC separation of the total lipid mixture and radioimaging are reported in Figure 4
A. The glycosphingolipid pattern of fibroblasts is in agreement with previous information (34)
, suggesting that GM3 and Gb3Cer are the main components of the glycosphingolipid mixture. However, Fig. 4
shows also that sphingomyelin is one of the main sphingolipids, while ceramide is hardly detectable.
|
In Neu3 cells with comparison to normal fibroblasts, we observed a minor, but statistically significant, reduction in the GM3 content, a minor increase of LacCer, and a dramatic decrease of Gb3Cer. In addition to this, ceramide was clearly present at higher levels in Neu3 overexpressing cells, and we calculated a 6-fold increase with respect to mock cells. The sphingolipid percent distributions in normal and Neu3-transfected cells are reported in Table 4
.
|
Figure 4B
shows the radioactive sialoprotein pattern in normal and Neu3 fibroblasts fed for two days [6-3H]-acetyl-D-mannosamine, a specific precursor of sialic acid. Cells incorporated the same radioactivity, but it is clear that in Neu3-transduced cells, the radioactivity associated with proteins was lower than in normal cells, suggesting that Neu3 is capable of acting on sialoproteins.
Despite the enzyme overexpression and increase of activity, only a very minor decrease in GM3 content was observed in stably transfected fibroblasts. Then we investigated the activity of the GM3 synthase SAT1. Figure 5
A shows that in stably transfected cells the activity of SAT1 increased by 
150% with respect to normal cells, as determined in an in vitro assay performed with [3-3H(sphingosine)]LacCer and cold CMP-Neu5Ac. The increase of enzyme activity was parallel to an increase of SAT1 mRNA (Fig. 5B, C
). This suggests that SAT1 maximum activity did not change, but that more enzyme was available. The higher availability of SAT1 could explain why the Neu3 overexpression in transfected cells does not produce any change in GM3 content. Thus, in Neu3 cells, the high turnover of GM3 largely consumes LacCer, which is no more available as a substrate for the synthesis of Gb3Cer. In confirmation of this, we found that the Gb3Cer synthase mRNA content did not change in Neu3 cells (Fig. 5D, E
).
|
Cell surface SM is precursor of ceramide. Tritium-labeled SM was administered to cells cultured in the presence of chloroquine. We did not find any statistical difference in ceramide contents between control and Neu3 cells, suggesting that the sphingomyelinase activity is not responsible, or it is only to a small extent, for the increase of ceramide in Neu3-transfected cells. On the other hand, in both control cells and cells maintained in the presence of chloroquine, ceramide is produced in a similar amount, suggesting that a constant quantity of ceramide is produced out of the lysosomes from SM (Fig. 6
and Table 5
).
|
|
The rate of growth of Neu3-transfected cells was reduced. This was determined by measuring DNA synthesis through the incorporation of [3H]thymidine in cells. As shown in Fig. 7
, Neu3 overexpression caused a marked diminution of [3H]thymidine incorporation, indicating inhibition of DNA synthesis.
|
The effect of Neu3 overexpression on programmed cell death was also evaluated. In Neu3 stable transfected human fibroblasts, the expression of the apoptosis-suppressing protein Bcl-2 was markedly reduced respect to mock transfected cells (Fig. 8
A), cytosolic cytocrome c was detectable (Fig. 8B
) and caspase-3 was cleaved into its active fragment (Figure 8C
), suggesting that the apoptotic pathway is executing in these cells. Quantitative evaluation of the apoptotic cell death (Fig. 9
) indicated that the entity of the process was 8 times higher in Neu3 than in mock cells.
|
|
To correlate a specific role of high concentration of ceramide in Neu3 cells with the apoptotic process, we administered C2-ceramide, a synthetic cell-permeable ceramide analog (35)
, to cells. Figure 9
shows that after 12 h-incubation with ceramide, the rate of cell death in mock cells was very similar to that of nontreated Neu3 cells. In both mock and Neu3 cells, C2-treatment progressively induced cell death by apoptosis. Nevertheless, data shown in Fig. 9
suggest that the higher value of apoptotic index in Neu3 cells with respect to mock cells after C2-ceramide treatment, is probably due to the higher basal concentration of intracellular apoptotic ceramide in the transfected fibroblasts.
DISCUSSION
Glycosphingolipids are components of the cell plasma membranes where, through interactions with different proteins, they can play an important role in modulating several aspects of the signal transduction processes (5)
. Plasma membranes glycosylhydrolases (9
, 36
37
38)
could be the natural candidate for modifications of the cell surface glycolipids, participating to the modulation of the signal transduction processes. Within the membrane associated glycosylhydrolases, the membrane-bound sialidase Neu3 has been characterized (7
8
9)
.
In this paper, we show the involvement of Neu3 and other membrane-associated glycosylhydrolases in processing gangliosides belonging to the plasma membranes of human fibroblasts in culture, namely GM3 (34)
.
The production at the cell surface of normal fibroblasts of a small amount of ceramide was clearly shown by feeding cells with GM3 ganglioside-containing tritiated sphingosine. GM3 was administered to cells under experimental conditions that allowed gangliosides to enter into the cell plasma membranes and to become indistinguishable from the endogenous compounds (39
40
41)
. A part of the radioactive GM3 entered in the metabolic pathway, and we observed the production of some LacCer, Cer and SM (Fig. 1)
. Radioactive SM can be formed only by recycling of sphingosine, and this confirms that GM3 reaches lysosomes where sphingosine is produced. To understand whether a part of catabolic process occurred out of the lysosomes, we treated cells with ammonium chloride or chloroquine. Both ammonium chloride and chloroquine have been widely used to inhibit lysosomal activity and in studies aimed to clarify the glycolipid intracellular trafficking and metabolism (28
29
30)
. Figure 1
shows that, under these two experimental conditions, no lactosylceramide and SM could be observed, while ceramide was produced. Neu3 overexpressing fibroblasts with double Neu3 activity subjected to the above treatments gave similar qualitative results but produced ceramide levels almost 3 times more(Table 3)
. These results demonstrate that the experimental conditions used to block lysosomal activity are effective, as indicated by the complete lack of products deriving by the recycling of [3H]sphingosine that under our experimental conditions is generated in the active lysosomes from GM3 (19)
. Thus the large portion of cell ceramide observed is produced out of lysosomes.
To understand whether GM3 is converted to ceramide at the plasma membrane, experimental conditions known to block internalization of plasma membrane components via endocytosis were also used. Incubation of cells at low temperature has been widely used in the past for this purpose (31
32
33)
. When human skin fibroblasts were incubated with fluorescent sphingolipid analogs for 30 min at 10°C and observed by fluorescence microscopy, only plasma membrane labeling was observed (31)
. The use of temperatures between 2 and 10°C is a well-established method to distinguish endocytosis from other mechanisms of internalization of membrane components, as endocytosis does not occur in mammalian cells to any significant extent at temperatures below 11°C (32
, 42)
. As shown in Fig. 1
, when GM3 was administered to cells at 4°C, no lactosylceramide and SM could be observed, but still ceramide was produced (Table 3)
. These results, taken together, indicate that ceramide is produced in an extralysosomal compartment and strongly suggest that the site of ceramide production is the plasma membrane.
The production of ceramide, at the plasma membranes of mammal cells, has been always associated to the activity of sphingomyelinase on sphingomyelin. For the first time, we show that ceramide is produced at the plasma membrane starting from glycosphingolipids.
We did not observe the production of LacCer and GlcCer, blocking cell lysosomal activity and endocytosis, after administration of cells of radioactive GM3. This suggests that both galactosidase and glucosidase are available at the plasma membrane, and the rates of detachment of neutral sugars are higher than that of sialic acid. Surprisingly, in Neu3 cells, where the production of Cer is very high after administration of GM3, no radioactive neutral glycolipids could be observed blocking cell lysosomal activity and endocytosis. This, together with the increase of sialidase activity, could suggest that an increase of the activity of membrane-associated galactosidase and glucosidase could occur. Unfortunately, no assays are available to discriminate plasma membrane galactosidase and glucosidase activities from the lysosomal ones. Nevertheless, we found that the total cell galactosidase and glucosidase activities were highly increased in Neu3 overexpressing cells (Table 1)
.
Together with the increased production of ceramide, many biochemical events were strongly modified in Neu3 cells. The increased sialidase activity was not followed by a parallel decrease of the cell GM3 content. As shown in Fig. 4
, the GM3 cellular contents in control and Neu3 cells are only slightly different. This appears in contrast with the high content of ceramide in Neu3 cells. Surprisingly, we found that the increase of sialidase content and activity occurred together with a correspondent increase of the content and activity of SAT1 (the sialyltransferase known also as GM3 synthase) (see Fig. 5
). Thus, some plasma membrane GM3 disappears, as it becomes a substrate for the cell surface hydrolytic process GM3
LacCer GlcCer Cer, and a similar quantity is substituted by neosynthesized GM3. Thus, to support the GM3 turnover, a large amount of LacCer is necessary. LacCer is substrate for both SAT1 and Gb3Cer synthase (Gb3Cer is formed by addition of an -galactose to LacCer). But we did not find any increase in Gb3Cer synthase (-galactosyltransferase) expression. Thus, LacCer is not available to the -galactosyltransferase, as it is used to produce GM3 by SAT1, and this explains the loss of Gb3Cer in Neu3 overexpressing cells.
A direct correlation between the increase of membrane bound sialidase Neu3 and the increase of other enzymes of the glycosphingolipid metabolism is not so evident. At least as far as it concerns SAT1, we found that both the protein and the activity increased (Fig. 5)
. Thus a signal was necessary to synthesize more mRNA SAT1. Glycosphingolipids, together with cholesterol, are components of plasma membrane lipid domains, which are believed to contain the switch of several functional events. There is solid information that suggests that changes of the composition and organization of these domains can modulate the functional events. We had dramatic changes of the glycosphingolipid pattern after Neu3 cell overexpression (Fig. 4)
. This could be responsible for signals that can modify the contents of the enzymes for the glycosphingolipid catabolism. In addition to this, the new glycosphingolipid pattern could be responsible for the activation of the membrane-associated galactosidase and glucosidase throughout direct glycolipid-protein interactions. Glycosphingolipid ability to modulate membrane-bound enzymes has been well studied and described in detail in the past (43
44
45
46)
.
The biochemical events occurring with the increase of Neu3 activity lead to cell death. This is not surprising, because of the increase of ceramide. Ceramide, released from plasma membrane sphingomyelin by a plasma membrane-associated sphingomyelinase, with neutral or acidic optimal pH, has been shown to participate in some way in the activation of the apoptotic process (47)
. We have effectively shown that ceramide promotes apoptosis when added to normal fibroblasts and that it increases the existing apoptotic process when added to Neu3 cells (Fig. 9)
. This supports a correlation between the high content of ceramide produced by the high expression of Neu3 and the high rate of death by apoptosis in Neu3 cells. Of course, it is now necessary to look for physiological ligands able to increase directly or indirectly Neu3 activity/quantity. Some experiments in this direction are now in progress. In addition to this, we would like to stress that our results have been obtained using nontransformed cells and that for the first time, they associate Neu3 overexpression to increase apoptosis. The role of Neu3 was previously studied by overexpressing the enzyme in tumor cells. Results from these studies were very different from what we report here. In fact, Neu3 overexpression in tumor cells led to an increase of the LacCer cell content and enhanced an antiapoptotic effect (12)
. The different results could be related to the fact that in tumor cells the glycosylation process is aberrant.
Figure 6
shows that feeding of sphingomyelin, isotopically tritium labeled at position 3 of sphingosine, to human fibroblasts leads to the formation of some ceramide, as expected. Nevertheless, in Neu3 overexpressing cells, a similar amount of ceramide was produced. This would confirm that ceramide produced in Neu3 cells comes mainly from the degradation of GM3 but do not exclude that the minor increase of neutral sphingomyelinase activity can produce a minor quantity of it from sphingomyelin, in vivo.
Twenty-five years ago, it was suggested (48)
that polysialogangliosides, after biosynthesis in the Golgi and transport to the plasma membranes, could be substrates for a membrane-associated sialidase able to regulate the correct ratio between gangliosides at different sialic acid content and neutral glycolipids. Our results suggest that the role of Neu3, the membrane-associated sialidase, is to correctly maintain in the plasma membranes the ganglioside pattern and the content necessary for the cell communications.
ACKNOWLEDGMENTS
This work was supported by COFIN-PRIN (2002), Consiglio Nazionale delle Ricerche (PF Biotechnology) and FIRB; Italy.
Received for publication September 13, 2005. Accepted for publication January 3, 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  M. Valsecchi, L. Mauri, R. Casellato, S. Prioni, N. Loberto, A. Prinetti, V. Chigorno, and S. Sonnino Ceramide and sphingomyelin species of fibroblasts and neurons in culture J. Lipid Res., February 1, 2007; 48(2): 417 - 424. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||
