The murine X-linked steroid sulfatase gene (Sts) normally escapes X inactivation. However, we have observed that most long-term murine cell cultures are deficient in STS activity even though only the L cells are known to be derived from an STS- mouse strain. To investigate this phenomenon, we developed a selective system whereby STS+ cells could be selected from STS- populations. The system is based on making cells dependent on cholesterol-sulfate as the sole source of cholesterol, allowing only STS+ cells to grow. Two STS- cell lines, after treatment with either 5-azacytidine (5AC) or ethyl methane sulfonate (EMS), yielded STS+ revertants, suggesting that their STS- phenotype was due to hypermethylation. To study the evolution of STS- cell lines, we established XO and XX primary lines from STS+ strains; the XX cell line remained STS+ after more than 200 cell doublings whereas the XO became STS- after about 100 doublings. Treatment of this STS- XO cell line with 5AC produced clones with restored STS activity. All the revertants showed a growth disadvantage compared to their STS- counterparts. It would appear that aberrant methylation is the basis for much of the STS deficiency observed in established murine lines and that its propagation is due to the growth advantage of STS- over STS+ cells.