The advent of conditional mutagenesis, such as the cre/loxP system, has allowed certain aspects of lethality induced by traditional mutagenesis to be circumvented (1,2). The Cre recombinase catalyzes excision of DNA flanked by oriented loxP sequences (3). When combined with promoter driven expression of cre, this is a powerful tool for cell specific or drug inducible mutagenesis (4,5,6).

In 2000, a heat-inducible Cre recombinase using an Hsp70-1 promoter, dubbed the HsCre6 line, was reported in mammals (7). Modeled after Drosophila heat shock approaches, it allowed precise temporal control of cre expression to bypass potential early lethal effects of gene deletion. This system is particularly advantageous because it eschews the complicated kinetics of drug administration/withdrawal and provides for elegant stage-specific mutagenesis while eliminating the potentially confounding toxic effects of drug treatments on embryos (6,8,9,10). One problem, however, was that occasional sporadic, un-induced recombination was reported in the initial study. To improve the usefulness of this system in mice, we sought to reduce un-induced, sporadic recombination and improve the efficiency of Cre-mediated deletion after heat shock in embryos.

To assay for recombination efficiency before and after heat shock, HsCre6 mice (carried on a mixed background of C57BL/6J and CBA) were crossed with the 129 Z/EG reporter strain. Cre recombinase excises a floxed stop codon in the Z/EG transgene, which activates expression of green fluorescent protein (GFP) (11). GFP fluorescence is first visible by standard microscopy 3–4 h after the heat shock and reaches its maximum intensity after 12 h. With this fluorescent reporter of recombination, it is possible to quickly sort high-versus low-percentage recombination in embryos (12) (Figure 1), where the percent recombination is defined as the percentage of GFP-positive cells in tissue sections.

Figure 1. Macroscopic view of recombination with a GFP reporter. (Click to enlarge)

First, we repeated the published protocol, which uses 17-min heat shocks at 42°C in a humidified environment for postnatal mice, and 13–15 min heat shocks for pregnant females to induce recombination in fetuses. We attained similar efficiency to the original published results with the HsCre6 line, which frequently generates high-percentage mosaic (>90%) animals (7; data not shown). In our preliminary experiments, we arranged some matings so that the female carried the HsCre6 transgene. However, this revealed that un-induced recombination was more frequent when the matings were arranged in this fashion and caused us to alter our strategy to only use male transmission, similar to the original report (7). Interestingly, un-induced recombination also occurred if the mice were exposed to even slightly elevated temperatures in our colony (>22°C) or when cages were moved between buildings when the ambient temperature was >29°C (30%; n=25 and 33%; n=30, respectively). The data we present only includes experiments where extraneous sources of heat were not present and where a young male transmitted the HsCre6 transgene.

We tested various temperatures and durations of heat treatments. We found that a 7-min heat shock at 42°C was the shortest treatment sufficient to induce recombination (Table 1). Furthermore, the pregnant females were lethargic for 3–5 min, as opposed to approximately 20 min with the 17-min heat treatment (7). This single 7-min treatment induces varying degrees of mosaicism that may be desirable for certain experiments. If a second 7-min heat shock was administered 1 h after the first, >70% of treated animals showed recombination efficiency of >90% of cells, an increase of >35% (7) (Table 1 and Figure 2), and the females were lethargic for only 4–5 min after the second heat treatment. No preference for cell type-specific recombination was observed in any samples, as all cells examined throughout the embryo, including germ cells, underwent recombination with similar efficiency. Furthermore, the viability of embryos was not affected by theheat shocks (pups per litter = 7±3.1, control; 8±1.9, one heat shock; 8±2.7, two heat shocks), and the animals were born in expected ratios.

Figure 2. Immunocytochemistry examples of mosaicism achieved in 12.5-dpc gonads. (Click to enlarge)

Table 1. Heat Shock Summary (Click to enlarge)

The original report described varying efficiency of un-induced recombination dependent upon the floxed target line. To eliminate the possibility that the improved efficiency of our protocol is limited to the Z/EG floxed target, we repeated these experiments with a second reporter, R26R. Cre recombinase excises a stop codon in R26R mice and allows transcription of lacZ (13). We achieved similar efficiency of recombination with this second reporter and found no un-induced recombination (Table 1). Therefore, we conclude that this dual-heat shock protocol is applicable to other floxed targets.

We also investigated the efficacy of administering heat shocks to explanted organs. The dual-heat shock can be combined with explants to consistently induce high-percentage mosaicorgans in culture (Table 1, Figure 1, B and C, and Figure 2D).

In summary, we have developed a protocol that significantly increases the efficiency of stage-specific heat-inducible recombination in two reporter strains of mice, as well as decreasing stress on the animals. By organizing our breeding schemes to transmit the HsCre6 transgene from the male and by eliminating extraneous sources of heat, un-induced recombination has been greatly minimized. Secondly, by using two shorter (7 min) heat shocks separated by 1 h, treated, pregnant females recovered in a quarter of the time. Finally, the incidence of near complete (>90%) recombination increased by over 35% from previous protocols.