Central to an understanding of the in vivo function of genes is their analysis by mutation, that is, inactivation or modification of a gene by mutation and the study of the consequences of the mutation in the mutant organism. In mammals, before gene targeting, this approach was limited to the rare spontaneous mutations reflected in obvious phenotypes, as in the case of inheritable diseases in the human. The targeted mutagenesis of the mouse germline was thus a fundamental breakthrough in this area of research. In its original form, gene targeting involves the inactivation of a given gene in the genome of embryonic stem (ES) cells by homologous recombination (1–3).

ES cells derive from an early stage of mouse development and have retained their totipotency, thus they can participate in the generation of cell lineages of the mouse (including germ cells) if transferred into an early mouse embryo. Transfer of mutant ES cells into mouse embryos thus allows the transmission of the mutation in question into the mouse germline. In the classical experiment of Thomas and Capecchi (3) gene inactivation was achieved by replacing a predetermined gene segment with a mutant version of this segment, through homologous recombination. Since the latter is infrequent in mammalian cells, the isolation of the mutant ES cells requires stringent selection. This was achieved by placing a selectable gene (in that case the neomycin resistance gene) into the targeted locus in a manner that allows its expression (and hence cellular selection) while inactivating the target gene.