Date Published: December 16, 2009
Publisher: Public Library of Science
Author(s): Johannes Hirrlinger, Robert P. Requardt, Ulrike Winkler, Franziska Wilhelm, Christine Schulze, Petra G. Hirrlinger, Darren P. Martin. http://doi.org/10.1371/journal.pone.0008354
Abstract: DNA recombination technologies such as the Cre/LoxP system advance modern biological research by allowing conditional gene regulation in vivo. However, the precise targeting of a particular cell type at a given time point has remained challenging since spatial specificity has so far depended exclusively on the promoter driving Cre recombinase expression. We have recently established split-Cre that allows DNA recombination to be controlled by coincidental activity of two promoters, thereby increasing spatial specificity of Cre-mediated DNA recombination. To allow temporal control of split-Cre-mediated DNA recombination we have now extended split-Cre by fusing split-Cre proteins with the tamoxifen inducible ERT2 domain derived from CreERT2.
Partial Text: DNA recombination technologies such as the Cre/LoxP system have advanced and refined the analysis of gene and cell functions in mice –. In contrast to classical in vivo “knockout” strategies, which result in a complete deletion of gene function in the whole organism, this conditional gene targeting technology enables a cell type-specific deletion of genes by driving the expression of Cre recombinase under the control of a cell type-specific promoter , , . In addition, the Cre/LoxP system has been used for fate mapping and for cell ablation in vivo–.
Based on our previous results with functional complementation of split-Cre proteins , we constructed inducible versions of split-Cre proteins by fusing the tamoxifen inducible ERT2-domain to each of the termini of both NCre and CCre (Fig. 1). The “split-CreERT2” proteins obtained were termed NCre-ERT2, CCre-ERT2, ERT2-NCre or ERT2-CCre (Fig. 1). We refer to these constructs below using the abbreviations NE, CE, EN and EC, respectively (the original NCre and CCre  will be referred to as N and C, respectively; Fig. 1). In a first attempt to determine whether complementation of these fusion proteins can reconstitute Cre-dependent DNA recombination activity and whether this activity is dependent on tamoxifen application, CHO cells were transfected with combinations of split-Cre and split-CreERT2 proteins and incubated in the absence or presence of 4OHT. To detect DNA recombination a CMV-LoxP-STOP-LoxP-firefly luciferase reporter plasmid was cotransfected and luciferase activity was measured (Fig. 2A). Luciferase activity was normalized to luciferase activity in N+C transfected cells cultured in the absence of 4OHT, which was set as 100%. Cells transfected with full-length Cre recombinase showed 234±25% (n = 10) of the activity of N+C, confirming our previous results . In contrast, all cells transfected with ERT2-tagged proteins showed a lower luciferase activity even after application of 4OHT compared to N+C transfected cells and the different combinations of split-CreERT2 proteins showed unequal activity (Fig. 2A). Western Blot analysis revealed expression of all constructs (Fig. 3), suggesting that the differences in activity are due to functional differences and not due to lack of expression. Nevertheless, several combinations of split-CreERT2 proteins showed inducibility of DNA recombination by 4OHT (Fig. 2A) and an induction ratio was calculated as [luminescence in the presence of 4OHT/luminescence in the absence of 4OHT] (Fig. 2B). This analysis revealed that the combination of NE+CE, NE+EC and EN+CE can be induced best by 4OHT (Fig. 2B). However, while e.g. NE+CE shows the highest inducibility (Fig. 2B), the total activity even in the presence of 4OHT is low (Fig. 2A). Therefore, a functional index was defined as [induction ratio×luminescence in the presence of 4OHT] to asses which combination of split-CreERT2 proteins combines the favorable properties of high activity and high inducibility (Fig. 2C). This analysis indicated that NE+CE and NE+EC are the best combinations (Fig. 2C).
In the current work, split-CreERT2 is presented as a new tool adding temporal control of DNA recombination to the recently described split-Cre system . The ERT2-domain  was fused to the N-terminus or the C-terminus of the constitutively active NCre- and CCre-proteins . Three parameters were assessed to dissect the functional properties of split-CreERT2 proteins: 1) low recombination in the absence of tamoxifen (no leakage); 2) high activity of recombination in the presence of tamoxifen; 3) a high induction ratio. When the different combinations of split-CreERT2 proteins were analyzed, not all versions showed the same properties regarding these criteria. Generally, the reporter gene expression from split-CreERT2 proteins was lower than that from the non-inducible forms of split-Cre (N+C; Fig. 2A, 4A), indicating lower recombination activity of split-CreERT2 compared to split-Cre in cell culture, which might hamper the efficient targeting of cells in living mice. However, a similarly reduced recombination activity of split-Cre was observed when compared to full-Cre in vitro. Nevertheless, in transgenic mice split-Cre induced robust reporter gene activation, which was probably more limited by mosaic transgene expression rather than by split-Cre efficiency itself . These findings may indicate that the situation in vitro and in vivo differs e.g. in expression kinetics, expression levels, kinetics of activation and DNA recombination. Whether the lower DNA recombination activity of split-CreERT2 in cell culture will limit DNA recombination in mice in vivo remains to be determined experimentally by generating appropriate transgenic mice.