The Sleeping Beauty transposon system (SB) has been shown to mediate nonviral integration of expression constructs resulting in long-term gene expression in several mammalian targets. Often, however, it is difficult to discern long-term expression resulting from transposition vs nonhomologous chromosomal recombination or maintenance of plasmid DNA in an extrachromosomal form. We have designed a system to silence expression from nontransposed sequences, making it possible to determine more specifically the amount of expression resulting from transposition. A transposon plasmid, pT2F/Cage (carrying a murine erythropoietin (Epo) gene transcriptionally regulated by the ubiquitously expressed CAGS promoter), was engineered to contain LoxP sites positioned so as to interrupt expression upon Cre-mediated recombination. Upon transposition these sites become segregated, thus protecting the expression construct from Cre-mediated recombination and subsequent silencing. Interferon-inducible Mx1Cre mice were administered pT2F/Cage with or without transposase-encoding plasmid. At 2 to 4 weeks postinjection, in the absence of SB transposase, Cre induction reduced Epo expression to about 1% of that seen in the group that was administered transposase-encoding plasmid, which maintained Epo levels near those of the uninduced groups. Southern hybridization analysis and plasmid rescue of transfected tissue supported the efficient Cre-mediated silencing of nontransposed sequences. These results indicate a substantial level of DNA-mediated expression not associated with transposition, but which can be quantitatively distinguished from transposition by its sensitivity to Cre recombinase. The results also provide additional evidence for the effectiveness of the Sleeping Beauty transposon system as an in vivo DNA-mediated gene transfer strategy for achieving long-term expression.