MCF-7 cells were transfected with the pcDNA3-GD3S expression vector containing the full-length cDNA of human GD3S or the empty pcDNA3 vector as control. Transfected cells were cultured 21 days in the presence of 1 mg/mL G418. Individual G418-resistant colonies were isolated by limiting dilution cloning. Forty-four clones were obtained and analyzed for the expression of GD3S. As previously shown [30], QPCR analysis of GD3S expression (Figure 2) indicates that GD3S mRNA is express at a very low level in wild-type and control (empty vector transfected) MCF-7 cells compared to SK-Mel 28 melanoma cells used as positive control [31]. Within the forty-four analyzed clones, three GD3S+ clones (clone #31, #41 and #44) were selected according to the high expression of GD3S compared to SK-Mel 28 (1.3-fold, 2.3-fold and 6.3-fold, respectively) (Figure 2).

The pattern of gangliosides was monitored in the three selected MCF-7 GD3S+ clones (clone #31, #41, #44) by flow cytometry using anti-G_D3 R24 and anti-G_T3 A2B5 mAbs. As shown in Figure 3, the three GD3S+ clones expressed G_D3 and G_T3 whereas wild-type and control (empty vector transfected) MCF-7 cells did not expressed complex gangliosides. G_T3 is expressed at a similar level in the three GD3S+ clones but a decrease of G_D3 is observed in clone #44 compared to clone #31 and #41 whereas the expression level of GD3S was 4.8-fold or 2.7-fold higher in clone #44 compared to clone #31 and #41, respectively (Figure 2). Control cells showed no change in the ganglioside profile compared with wild-type MCF-7 (data not shown).

Glycolipids were extracted from cells, purified by reverse phase chromatography and permethylated prior to MS analysis. Mass spectrometry analysis established that glycolipid profiles of MCF-7 WT and GD3S+ clones were characterized by complex patterns of neutral and sialylated glycosphingolipids from globo- and ganglio-series. Profiles of all cell lines were dominated by two signals at m/z 1460 and 1572 both identify based on their MALDI-TOF/TOF fragmentation patterns (data not shown) and in agreement with previously published analyses [33] as mixtures of G_b4 and G_A1 differing by the nature of their lipid moieties (d18:1-16:0 or d18:1-24:0). Along these two major components, MS and MS/MS analyses permitted us to identified other minor neutral GSLs including LacCer and G_b3 (Figure 4). The comparison of MS profiles did not show any significant difference in neutral GSLs content between MCF-7 WT and GD3+ clones. On the contrary, the content in sialylated glycolipids varied among the different cell lines. MCF-7 WT cells exhibited monosialylated gangliosides including G_M3, G_M2, and G_M1. The only disialylated GSL observed in MCF-7 WT cells was G_D1 at m/z 2182 and 2194. Its sequence analysis by MALDI-TOF/TOF typified it as G_D1a, thus lacking disialylated motif (data not shown). GD3+ clones did not show GSLs from the G2 and G1 families but synthesized instead a family of unusual highly sialylated lactosylceramide derivatives substituted by up to 5 Neu5Ac residues tentatively identified as G_D3, G_T3, G_Q3 and G_P3. The structure of these four compounds was confirmed by MALDI-TOF/TOF sequencing.

The Glycolipid profile of SK-Mel 28 cells was also analyzed in respect to the presence of polysialylated lactosylceramide derivatives. Contrarily to MCF-7 WT cells, GSLs extracted from SK-Mel 28 cells did not contain globo-series but were dominated by ganglio-series, which induces much higher overall sialic acid content (data not shown). In particular, disialylated G_D3 appears to be the major component along with monosialylated G_M3 and trisialylated G_T3. However, the tetra- and pentasialylated gangliosides synthesized by MCF-7 GD3+ clones were not detected in SK-Mel 28.

We illustrate the sequence analyses of unusual G_Q3 and G_P3 in Figure 5. These two molecules differing in the presence of a single Neu5Ac residue were structurally related. They shared the fragmentation pattern of a linear stretch of four sialic acid residues in terminal non-reducing position as [M+Na]⁺ B-ions at m/z 398, 759, 1120 and 1481 and a linear sequence of Sia₄Hex₂Cer at reducing end as [M+Na]⁺ Y-ions at m/z 1106, 1466, 1827 and 2188. G_P3 showed additional B and Y ions at m/z 1842 and 2551 typifying a linear Sia₅ sequence. Altogether, these data established that MCF-7 GD3+ clone #44 synthesizes an unusual family of oligosialylated lactosylceramide derivatives presented from 2 to 5 Neu5Ac residues. The disappearance of G_M2, G_M1 and G_D1a in MCF-7 GD3+ clone #44 could be explained by the depletion of G_M3 substrate caused by the over-expression of the GD3S+ that competes with β-4GalNAcT1 for the use of G_M3 substrate.

Because mass spectrometry does not provide reliable quantitative data, we quantified the extent of oligosialylation modifications in GSLs induced by the overexpression of G_D3 in MCF-7 by screening all three GD3S+ clones (#31, #41 and #44) (Figure 6). To do that, oligosialylated motifs were released from purified GSLs by mild hydrolysis according to optimized procedures [35] and labeled by DMB before separation and quantification by FL-HPLC [36]. The analysis of relative quantifications of sialic acid chains in GSLs of all MCF-7 GD3S+ clones compared to MCF-7 WT demonstrated a sharp increase of sialylation oligomerization degree, up to five residues, in accordance with mass spectrometry analysis. In MCF-7 WT, no oligosialylation could be observed, in accordance with the sole identification of G_D1a by MS. In contrast, all GD3S+ clones contained about 30% of Sia₂ and Sia₃ motifs, up to 10% of Sia₄ and up to 5% of Sia₅. Although clones presented similar GSL oligosialylation profiles, small scale quantitative differences could be observed between the three clones, with a prevalence of higher DP values for clone #44 compared to clones #31 and #41.

Anti-G_D3 R24 mAb was purchased from Abcam (Cambridge, UK) and anti-G_T3 A2B5 mAb was kindly provided by Pr. Jacques Portoukalian (Depart. of Transplantation and Clinical Immunology, Claude Bernard University and Edouard Herriot Hospital, Lyon, France). FITC-conjugated sheep anti-mouse IgG was from GE Healthcare (Templemars, France). FITC-conjugated anti-mouse IgM was purchased from Molecular Probes (Invitrogen, Carlsbad, CA, USA).

The breast cancer cell line MCF-7 and the melanoma cell line SK-Mel 28 were obtained from the American Type Cell Culture Collection. Cell culture reagents were purchased from Lonza (Levallois-Perret, France). Cells were routinely grown in monolayer and maintained at 37 °C in an atmosphere of 5% CO₂, in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, and 100 units/mL penicillin-streptomycin. GD3S positive (GD3S+) MCF-7 clones were obtained by stable transfection of the pcDNA3-GD3S expression vector encoding the full-length human G_D3 synthase [37] as previously described [33]. Individual resistant colonies were isolated by limit dilution. Three positive clones (#31, #41 and #44), expressing different levels of GD3S were used for further study. Control cells (empty vector transfected) and GD3S+ clones were cultured in the presence of 1 mg/mL G418 (Invitrogen, Cergy-Pontoise, France).

Total RNA was extracted using the Nucleospin RNA II kit (Macherey Nagel, Hoerdt, France), quantified using a NanoDrop spectrophotometer (Thermo Scientifics, Wilmington, USA) and the purity of the preparation was checked by ratio of the absorbance at 260 and 280 nm. The cDNA was synthesized using 2 µg of RNA (GE Healthcare). PCR primers for GD3S and Hypoxanthine PhosphoRibosylTransferase (HPRT) were previously described [30,38] and synthesized by Eurogentec (Seraing, Belgium). PCR reactions (25 µL) were performed using 2X SYBR^® Green Universal QPCR Master Mix (Stratagene, Amsterdam, The Netherlands), with 2 µL of cDNA solution and 300 nM final concentration of each primer. PCR conditions were as follows: 95 °C for 30 s, 51 °C for 45 s, 72 °C for 30 s (40 cycles). Assays were performed in triplicate and GD3S transcript expression level was normalized to HPRT using the 2^−ΔΔCt method described by Livak and Schmittgen [32]. Serial dilutions of the appropriate positive control cDNA sample were used to create standard curves for relative quantification and negative control reactions were performed by replacing cDNA templates by sterile water.

Cells were washed in cold PBS and detached by EDTA 2 mM. Cells were incubated at 4 °C during 1 h with anti-G_D3 R24 (1:100) and anti-G_T3 A2B5 (1:10), diluted in phosphate buffered saline (PBS) containing 0.5% bovine serum albumin (PBS-BSA) (Sigma-Aldrich). After washing with PBS-BSA, cells were incubated on ice during 1 h with Alexa Fluor 488 anti-IgG or anti-IgM (1:500). After two washes in PBS-BSA, cells were analyzed by flow cytometry (FACScalibur, Becton Dickinson). Control experiments were performed using secondary antibody alone.

Twenty dishes (10 cm diameter) of cultured cells were washed twice with ice-cold PBS and cells were scraped and sonicated on ice in 200 µL of water. The resulting material was dried under vacuum and sequentially extracted by CHCl₃/CH₃OH (2:1, v/v), CHCl₃/CH₃OH (1:1, v/v) and CHCl₃/CH₃OH/H₂O (1:2:0.8, v/v/v). Supernatants were pooled, dried and subjected to a mild saponification in 0.1 M NaOH in CHCl₃/CH₃OH (1:1) at 37 °C for 2 h and then evaporated to dryness [39]. Samples were reconstituted in CH₃OH/H₂O (1:1, v/v) and applied to a reverse phase C₁₈cartridge (Waters, Milford, MA, USA) equilibrated in the same solvent. After washing with CH₃OH/H₂O (1:1, v/v), GSLs were eluted by CH₃OH, CHCl₃/CH₃OH (1:1, v/v) and CHCl₃/CH₃OH (2:1, v/v).

Prior to mass spectrometry analysis, GSL were permethylated according to Ciucanu and Kerek [40]. Briefly, compounds were incubated 2 h in a suspension of 200 mg/mL NaOH in dry DMSO (300 µL) and CH₃I (200 µL). The methylated derivatives were extracted in CHCl₃ and washed several times with water. The reagents were evaporated and the sample was dissolved in CHCl₃ in the appropriate dilution. MALDI-MS and MS/MS analyses of permethylated GSL were performed on 4800 Proteomics Analyzer (Applied Biosystems, Framingham, MA, USA) mass spectrometer, operated in the reflectron mode. For MS acquisition, 5 µL of diluted permethylated samples in CHCl₃ were mixed with 5 µL of 2,5-dihydroxybenzoic acid matrix solution (10 mg/mL dissolved in CHCl₃/CH₃OH (1:1, v/v)). The mixtures (2 µL) were then spotted on the target plate and air dried. MS survey data comprises a total of 50 sub-spectra of 1500 laser shots. Peaks observed in the MS spectra were selected for further MS/MS. CID MS/MS data comprises a total of 100 sub-spectra of 3000 laser shots. Two or more spectra can be combined post-acquisition with mass tolerance set at 0.1 Da to improve S/N ratio. The potential difference between the source acceleration voltage and the collision cell was set to 1 kV and argon was used as collision gas.

In order to minimize internal fragmentation of polysialylated sequences, sialylated glycan samples were directly coupled to 1,2-diamino-4,5-methylenedioxybenzene (DMB) without prior mild hydrolysis [35]. Samples were incubated for 2.5 h at 50 °C in 50 μL of a DMB reagent solution (2.7 mM DMB, 9 mM sodium hydrosulfite, and 0.5 mM β-mercaptoethanol in 20 mM TFA). 10 μL of 1 M NaOH was then added and the reaction mixtures further incubated in the dark at room temperature for 1 h. Samples were stored at 4 °C before analysis. DMB-derivatized sialic acid oligomers were separated on a HPLC apparatus fitted with a CarboPac PA-100 column (Dionex). CarboPac column was eluted at 1 mL/min with a concentration gradient of 2 to 32% of 1 M NaNO₃ in water. Elution was monitored by an on line fluorescence detector set at wavelengths of 373 nm for excitation and 448 nm for emission.