A possible cell source for a bio-artificial liver is the human hepatblastoma-derived cell line HepG2 as it confers many hepatocyte functions, however, the urea cycle is not maintained resulting in the lack of ammonia detoxification via this cycle. We investigated urea cycle activity in HepG2 cells at both a molecular and biochemical level to determine the causes for the lack of urea cycle expression, and subsequently addressed reinstatement of the cycle by gene transfer. Metabolic labelling studies showed that urea production from 15N-ammonium chloride was not detectable in HepG2 conditioned medium, nor could 14C-labelled urea cycle intermediates be detected. Gene expression data from HepG2 cells revealed that although expression of three urea cycle genes Carbamoyl Phosphate Synthase I, Arginosuccinate Synthetase and Arginosuccinate Lyase was evident, Ornithine Transcarbamylase and Arginase I expression were completely absent. These results were confirmed by Western blot for arginase I, where no protein was detected. Radiolabelled enzyme assays showed that Ornithine Transcarbamylase functional activity was missing but that Carbamoyl Phosphate Synthase I, Arginosuccinate Synthetase and Arginosuccinate Lyase were functionally expressed at levels comparable to cultured primary human hepatocytes. To restore the urea cycle, HepG2 cells were transfected with full length Ornithine Transcarbamylase and Arginase I cDNA constructs under a CMV promoter. Co-transfected HepG2 cells displayed complete urea cycle activity, producing both labelled urea and urea cycle intermediates. This strategy could provide a cell source capable of urea synthesis, and hence ammonia detoxificatory function, which would be useful in a bio-artificial liver.