The human SKOV3 ovarian cell line was purchased from American Type Culture Collection (ATCC, Rockville, MD). Cells were maintained in RPMI1640 media (Mediatec, Inc., Herndon, VA) supplemented with 10% fetal bovine serum (FBS) at 37˚C in 5% CO₂. Generally, 2 × 10⁵ cells/well were plated in six-well plate and incubated overnight. The next day, 100 nM Taxol dissolved in dimethylsulfoxide (DMSO) were added into the indicated wells, and an equal amount of DMSO was added into control wells. The final concentration of DMSO was less than 0.1%. At 48 h after treatment, cells were collected for apoptosis analysis.

To clone the full-length GM3 synthase gene, we used the GM3 synthase sense primer 5’-GACATGGCTTCTGTTCCAAT-3’ and the GM3 synthase antisense primer 5’-TGTTCAAAATTCACGATCAATGCC-3’ to amplify the coding region of the GM3 synthase gene in SKOV3 cells by PCR. The full-length GM3 synthase DNA fragment was separated on 1.5% agarose gel and purified by a gel extraction kit (Qiagene #28704). The purified GM3 synthase DNA fragment was subcloned into the pEF6/ V5-His TOPO expression vector (Invitrogen Corp., Carlsbad, CA) following the protocol provided by the company. Four µL of gel-purified PCR product, 1 µL reaction salt buffer, and 1 µL TOPO vector were mixed gently and incubated at room temperature for 5 min. Two µL of the reaction mixture were added into a vial of ONE SHOT top 10 chemically competent E. coli and mixed gently. This mixture was kept on ice for 30 min, and then this mixture was heat shocked at 42˚C for 30 sec. After heat shock, the vial was immediately transferred onto ice. Then, 250 µL of room temperature SOC medium were added to the vial. The vial was capped tightly and shaken horizontally at 200 rpm for one h at 37˚C. Ten to 50 µL of transformed E. coli were spread onto an ampicillin selective plate. The obtained GM3 synthase clones were further confirmed by DNA sequence analysis.

SKOV3 cells were grown in a 6-well plate (2 × 10⁵ cells/well) overnight. The human ovarian SKOV3 cells were transfected with the GM3/V5-His plasmid or empty pEF/V5-His plasmid by using lipofectamine 2000 transfection reagent according to the manufacturer instructtions (Invitrogen). For cell transfection in each well, 5 µL lipofectamine 2000 were added into 250 µL OptiMEM serum reduced medium, and 4 µg plasmid DNA were added into another 250 µL Opti-MEM serum reduced medium. Each mixture was vortexed briefly and kept at room temperature for 5 min, then the mixtures were combined and kept at room temperature for 20 min. Five hundred µL of the mixture were then added evenly into a well of a 6-well plate; the plate was then shaken gently to thoroughly mix the transfection reagent and plasmid mix. Stable clones were obtained through screening transfected cells by adding 10 µg/mL Blastacidin (Invitrogen) to the growth medium at 24 h after transfection. After 2 - 4 weeks of growth, over-expression of GM3 synthase in the transfectants was confirmed by Western blot analysis using V5 antibody. The confirmed transfectants were expanded and stored for future studies.

Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Valencia, CA) as described by the manufacturer. Two µg of total RNA were used in reverse transcription reactions with M-MLV reverse transcriptase and oligo (dT) 15 primer (Promega, Madison, WI) as described by the manufacturer. One µL oligo-dT 15 primer, 1 µL dNTP, and 2 µg total RNA were mixed and the total reaction volume was brought to 18 µL by adding nuclease-free water. The reaction mixture was kept at 70˚C for 10 min, and then the mixture was immediately transferred onto ice. Then, 5 µL 5× reaction buffer, 1 µL M-MLV reverse transcriptase, and 0.5 µL nuclease inhibitor were added into the mixture, and the total 25 µL reaction volume was reached by adding nuclease-free water. This reaction mixture was kept at 42˚C for 1 h for cDNA synthesis, followed by 70˚C for 15 min to deactivate M-MLV reverse transcriptase. Four µL of the resulting total cDNA were then used as the template in PCR to measure the mRNA level of interest using the designed primers for GM3 synthase: sense, 5’-AAAGGAAACCCTGCCATTCT-3’; antisense, 5’-TCACCACTCCCTCTTTGACC-3’. β-actin mRNA levels were used as internal controls. The sense primer for actin amplification is 5’-GCCCCCCTGAACCCCAAGGCCAAC-3’. The antisense primer for actin amplification is 5’-CCGGTCGGCCGTGGTGGTGAAGCT-3’. Four µL total cDNA, 1 µL (5 pmol) primer, and 12.5 µL Gotaq Green Master Mix (Promega #M7122) were mixed and a total 25 µL reaction volume was reached by adding 6.5 µL nuclease-free water. The reactions were performed with denaturation at 95˚C for 55 sec, annealing at 55˚C for 55 sec, and extension at 72˚C for 2 min for a total of 30 cycles. The amplified fragments were separated on 1.5% agarose gel and visualized under ultraviolet light by ethidium bromide staining.

The cell survival assay was performed as we previously described [20]. Cells were treated with 10 - 500 nM of Taxol for 48 h. After harvesting the cells, survival was determined by cell survival assay. For each treatment, 1 × 10⁴ cells were seeded into a 96-well plate and treated with or without CMH. Cells were stained with 1% methylene blue in 50% methanol. After the plates were washed with PBS, the dye was dissolved in 100 μL of 0.5 M HCl. Absorbance was determined by an automated scanning photometer at a wavelength of 630 nM.

Following treatment as stated above, the cells (5 × 10⁵ cells/treatment) were used to determine the translocation of phosphatidylserine to the outer surface of the plasma membrane during apoptosis using the annexin V binding assay. For this assay, the cells were incubated with Taxol as described above, harvested, and stained with fluorescein isothiocyanate-labeled annexin V (BD Biosciences, San Jose, CA) and propidium iodide according to the manufacturer protocol. All harvested cells were washed two times with PBS then suspended in 300 µL annexin V binding buffer. For each sample, 100 µL cell suspension, 5 µL annexin V FITC-conjugated antibody and 1 µL 250 µM propidium iodide were mixed and kept in darkness place at room temperature for 30 min. Then, another 400 µL of annexin V binding buffer were added into each sample for final flow cytometry analysis. Apoptosis and necrosis were analyzed by quadrant statistics on the propidium iodide (PI)-negative, fluoresceinpositive cells, and PI-positive cells, respectively.

DAPI staining was performed on untreated and Taxoltreated cells as described previously [20]. Briefly, prior to staining, the cells were fixed with 4% paraformaldehyde for 30 min at room temperature. After washing with PBS by centrifugation at 650 × g, DAPI was added to the fixed cells for 5 min, then the cells were examined by fluorescence microscopy. Apoptotic cells were identified by condensation and fragmentation of nuclei. A minimum of 300 cells were counted for each treatment, and the percentage of apoptotic cells was calculated as the ratio of apoptotic cells to total cells counted ×100.

Clonogenic assay was carried out as we previously described [21]. SKOV3 cells were transfected with EF6 or GM3Syn-EF6 expression vector, treated with or without 5 or 10 nM Taxol for 48 h, and after washing the cells with growth media lacking Taxol, they were grown for 7 days. The colonies in control or Taxol-treated wells were then counted. Error bars represent the standard deviation for each point obtained from triplicate experiments.

In this study, the following primary antibodies were used: anti-caspase-3 (Cell Signaling Technology, Danvers, MA), anti-V5 antibody (Invitrogen Corp., Carlsbad, CA), anti-β-actin clone AC-74 (Sigma-Aldrich, St. Louis, MO). Protein concentrations were determined by using the BCA/Cu₂SO₄ protein assay (Sigma-Aldrich) as described by the manufacturer. For Western blot analysis, 50 - 80 µg protein/lane were separated by 12% SDS-PAGE, blotted onto a PVDF Immobilon membrane, and then protein levels were detected using peroxidase-conjugated secondary antibodies as described by the manufacturer. The primary antibody was used at a concentration of 1:2000 (v/v). The membranes were then exposed to Kodak X-Omat film for various times. The mouse monoclonal anti-β-actin (clone AC-74) was purchased from Sigma-Aldrich and used at 1:5000 (v/v).

We have found that Taxol at 100 - 500 nM induces robust apoptosis in cell lines from different types of cancer [22]. As shown in Figure 1(a), the annexin V binding assay revealed that Taxol at 100 nM induced early apop-

tosis in 28% of SKOV3 cells, and RT-PCR analysis showed that while SKOV3 cells express little GM3 synthase at the mRNA level, it was significantly increased after Taxol treatment (Figure 1(b)). These data suggest that Taxol induces GM3 synthase expression.

To determine the role of GM3 synthase in Taxol-induced apoptosis, we constructed a GM3 synthase III-EF6 plasmid and transfected SKOV3 cells. Stable SKOV3 transfected populations harboring this plasmid contained elevated levels of GM3 synthase as determined by immunoblot analysis (Figure 2). As revealed by annexin V binding assay, treating SKOV3 cells with 100 nM Taxol for 48 h induced early apoptosis in 28% - 30% of the cells, as compared to 17% in transfectants overexpressing GM3 synthase (SKOV3/GS) (Figures 1(a) and (b)). These results show that the overexpression of GM3 synthase reduces early apoptosis triggered by Taxol in SKOV3/GS cells (P < 0.05). Next, we determined the levels of late apoptosis induced by Taxol in SKOV3 and SKOV3/GS cells by DAPI staining. Apoptotic cells were identified by condensation and fragmentation of nuclei. The data shown in Figures 1(c) and (d) clearly show that Taxol induced apoptosis as visualized by DAPI staining of these cells treated with 10 - 100 nM Taxol. Quantitation of apoptosis assessed by DAPI staining revealed that treating the cells with 10 - 100 μM Taxol for 48 h induced 60% - 80% apoptosis in SKOV3 cells, but only

triggered 35% - 38% apoptosis in SKOV3/GS cells, indicating that GM3 synthase overexpression attenuates late Taxol-induced apoptosis.

To determine whether overexpression of GM3 synthase in SKOV3 transfectants increases cell proliferation, we determined number of cells after 48 h of culturing as described in Materials and Methods. As shown in Figure 3(a), GM3 synthase-overexpressing cells displayed 20%

increased proliferation compared to SKOV3 cells. Moreover, to determine whether Taxol decreases cell survival to the same extent in SKOV3 and GM3 synthase-overexpressing SKOV3/GS cells, we treated these cells with increasing concentrations of Taxol and determined cell survival by methylene blue cell survival assay as described in Materials and Methods. The results shown in Figure 3(b) revealed that SKOV3/GS survived cells are about 2 - 3-fold more resistant to Taxol than SKOV3 cells.

To examine whether GM3 synthase-overexpressing cells exhibited resistance to Taxol, we executed a colony forming assay. Clonogenic assay was carried out as described in Materials and Methods. As shown in Figure 3(c), more colonies were survived in SKOV3/GS cells compared to SKOV3 cells following treatment with 5 and 10 nM Taxol.

We also examined the role of caspase-3 in Taxol-induced apoptosis in SKOV3 and SKOV3/GS cells, since it is a central caspase in the mitochondrial apoptotic signaling pathway. SKOV3 cells were treated with Taxol and caspsase-3 levels were analyzed by immunoblotting. As shown in Figure 2(a), Taxol induced the processing of caspase-3 to its active forms, revealing that caspase-3 is involved in mediating Taxol death signals. While the results shown in Figure 2 demonstrated that Taxol induced processing of procaspase-3 to its active forms, the ectopic expression of GM3 synthase inhibited this process, and barely active caspase-3 forms were detected expressing GM3 synthase. These data clearly show that GM3 synthase functions as an endogenous caspase-3 inhibitor in SKOV3 cells.