The aim of this study was to improve the generation of gain-of-function ES cell lines and mouse models by facilitating the use of exogenous promoters in the Rosa26 locus and accelerating the generation of mutant ES cells using RMCE. We therefore generated two ES cell lines with modified Rosa26 loci by introducing either heterospecific Lox511/LoxP sites (modRosa26LoxP
ES cells) or FRT3/FRT sites (modRosa26FRT
ES cells). ModRosa26LoxP
ES cells allow for RMCE using the Cre/LoxP system, while modRosa26FRT
ES cells can be used for RMCE using the Flp/FRT system. All transgenic ES cell lines were successfully used to generate highly chimeric transgenic mice that showed germline transmission within the first litter. Compared to homologous recombination, RMCE into the modRosa26 locus dramatically increased the targeting efficacy and therefore minimizes time, effort and costs for the generation of transgenic ES cells and mice thereof. Most previous studies used the endogenous Rosa26 promoter to achieve ubiquitous gene expression in vitro
and in vivo
at moderate levels 
. However, the pCAG promoter targeted to the Rosa26 locus offers 8- to 10-fold stronger transgene expression when compared to the Rosa26 promoter 
. Whether other ubiquitous promoters or even tissue-specific promoters can be used to drive transgene expression from the Rosa26 locus remained elusive. Therefore, we decided to test several ubiquitous and tissue-specific promoters in the modRosa26LoxP
locus in combination with rapid ES cell targeting by RMCE. Although the CMV, EF1α and pCAG promoters showed efficient and comparable transgene expression in vitro
in ES cells, only the pCAG promoter induced high transgene levels in vivo
in the modRosa26LoxP
locus. In contrast to the pCAG and EF1a promoters, the CMV promoter was not active in most tissues in vivo
, consistent with other studies showing that the CMV promoter is susceptible to silencing effects in transgenic mice 
. It is possible that the NeoR cassette interferes with the transgene expression, depending on the promoter used.
When targeted to the Rosa26 locus without shielding the Rosa26 promoter, the pCAG promoter showed mosaic expression in skeletal muscle, lung and liver 
. Since it was suggested that the Rosa26 promoter can influence the expression of inserted transgenes 
, we introduced a Stop cassette downstream of the Rosa26 sense promoter to shield our expression cassette. Indeed, transgenic ES cells with this modification showed no EGFP expression in vitro
when no exogenous promoter was inserted, indicating that the Rosa26 promoter was functionally silenced. In the modRosa26LoxP
locus, the pCAG promoter showed strong and ubiquitous EGFP expression in skeletal muscle and other organs analyzed. Only in the liver the EGFP expression remained mosaic, showing a clear improvement to previous pCAG driven transgene expression from the Rosa26 locus 
. Possibly, removing the NeoR cassette or improved shielding of the Rosa26 promoter elements could further improve ubiquitous transgene expression using the pCAG promoter targeted to the Rosa26 locus. A clear advantage of the pCAG promoter is that it yields very high expression levels. Thus, pCAG promoter mice enable tracing of live cells for differentiation studies, electrophysiology and intravital microscopy in which recombination events need to be monitored by direct EGFP fluorescence. This is in contrast to the endogenous, ubiquitously active Rosa26 promoter that results in low and often insufficient levels of transgene expression that is not detectable without further immunostaining 
Using Cre recombinase as a sensitive reporter that was monitored by ActB-EGFP reporter mice 
, we analysed specificity tissue-specific promoters in the modRosa26 locus. Although αSMA and VeCad promoters mostly retained their specificity in vivo
when targeted to the modRosa26 locus, rare ectopic transgene expression was observed. Col1a1 promoters entirely lost the tissue-specificity, resulting in ubiquitous transgene expression. Although we could show that the endogenous Rosa26 promoter was functionally silenced in the modRosa26 locus, it is possible that inserted tissue-specific promoters lose specificity due to positional influences. Therefore, using the modRosa26 locus in combination with tissue-specific promoters driving Cre requires careful characterization. However, unspecific or mosaic Cre expression is a general problem when using isolated promoter constructs and can often only be circumvented by a knock-in of Cre into the endogenous gene locus. When high amounts of non-conditional transgene expression from tissue-specific promoters are required, classical transgenic mice with multiple transgene insertions are still inevitable despite all their disadvantages, as mentioned above.
In order to facilitate high level transgene expression in a tissue-specific manner along with the advantages of single-copy integration into a defined locus, we now generated the modRosa26FRT locus. We combined Flp/FRT-mediated RMCE into the modRosa26FRT locus with conditional Cre/LoxP-mediated transgene expression from the pCAG promoter. This approach allows for reliable high-level transgene expression in the desired tissue using established Cre lines. Using this approach we generated two reporter mouse lines. The great advantage of mR26CS-EGFP reporter mice is the high EGFP fluorescence that allows monitoring of recombination events in live cells. Going one step further, mR26CS-Luc reporter mice even allow monitoring of recombination events in live animals upon luciferase injection. In particular, this allows initial screening of new Cre lines without extensive histological analysis, since promising candidates can be selected very rapidly by Xenoimaging. In addition, these mice could be used to monitor tumor formation in combination with Cre/LoxP conditional deletion of tumor suppressors or activation of oncogenes.
In summary, we modified the Rosa26 locus to facilitate the use of exogenous promoters and to accelerate the generation of transgenic ES cells and transgenic mice for gain-of-function studies. We characterized several ubiquitous and tissue-specific promoters in vivo. Using tissue-specific promoters for low-level transgene expression or combined Flp-RMCE and Cre/LoxP conditional transgene expression for high-level expression, we accelerated the generation of single-copy transgenic mice targeted to a defined locus as an alternative to classical transgenic mice produced by pronuclear microinjections. When well-characterized Cre lines are available for the tissue of interest, we believe that our system is preferable to classical transgenic mice, since it offers reliable and predictable expression profiles and rapid generation and therefore minimized costs and effort.