There is a critical need for new sources of hepatocytes, both clinically to provide support for patients with liver failure and in drug discovery for toxicity, metabolic and pharmacokinetic screening of new drug entities. We have reported previously a variety of methods for differentiating murine embryonic stem (ES) cells into hepatocyte-like cells. One major challenge of our work and others in the field has been the ability to selectively purify and enrich these cells from a heterogeneous population. Traditional approaches for inserting new genes (e.g., stable transfection, knock-in, retroviral transduction) involve permanent alterations in the genome. These approaches can lead to mutations and involve the extra costs and time of developing, validating and maintaining new cell lines. We have developed a transient gene delivery system that uses fluorescent gene reporters for purification of the cells. Following a transient transfection, the cells are purified through a fluorescence-activated cell sorter (FACS), re-plated in secondary culture and subsequent phenotypic analysis is performed. In an effort to test the ability of the reporters to work in a transient environment for our differentiation system, we engineered two non-viral plasmid reporters, the first driven by the mouse albumin enhancer/promoter and the second by the mouse cytochrome P450 7A1 (Cyp7A1) promoter. We optimized the transfection efficiency of delivering these genes into spontaneously differentiated ES cells and sorted independent fractions positive for each reporter 17 days after inducing differentiation. We found that cells sorted based on the Cyp7A1 promoter showed significant enrichment in terms of albumin secretion, urea secretion and cytochrome P450 1A2 detoxification activity as compared to enrichment garnered by the albumin promoter-based cell sort. Development of gene reporter systems that allow us to identify, purify and assess homogeneous populations of cells is important in better understanding stem cell differentiation pathways. And engineering cellular systems without making permanent gene changes will be critical for the generation of clinically acceptable cellular material in the future.