It was reported recently by our group that the transfection of GD3 synthase cDNA into Neuro2a cells, a neuroblastoma cell line, caused cell differentiation with neurite sprouting (Kojima et al., 1994; J. Biol. Chem., 269, 30451-30456). To further explore this phenomenon in detail, we applied tetracycline-regulated system to control the expression of GD3 synthase cDNA in Neuro2a cells. Under this system, the process of Neuro2a cell differentiation was rather slow, about 3 weeks of cell culturing in the absence of tetracycline was required for most cells to extend the neurite-like structures. The RNase protection assay indicated that the mRNA of GD3 synthase gene was first detected between 4 h and 8 h after the gene was activated and kept at approximately the same level through the process. Furthermore, time-course analysis of total ganglioside expressions has shown that GD3 and GT1b gangliosides appeared on the cell surface early in the process and reached the maximum level around day 6. We also found that the amounts of GD3 and GT1b on the cell surface started to decrease after day 6 and returned gradually to the basal values after 3 weeks. On the other hand, GQ1b and GD1b were started to be synthesized at early stage and the amounts were continuously to increase through the whole Neuro2a morphological change process. In addition, time-course analysis by flow cytometry method for GD3 and GQ1b suggested that the conversions of simple gangliosides to more complex gangliosides may be required to induce the Neuro2a differentiation. Our results indicated that the combination of cDNA transfection and regulated gene expression is a powerful tool to study the function of glycolipids and should have a general application to the glycobiology field.