Slicing and dicing Cre/lox: part 1

The Cre/lox system has become the cornerstone of modern mouse genetics because it enables sophisticated control over the timing and location of gene expression. This is the first in a series of posts that looks at some finer points that often go overlooked in experiments with the Cre/lox system.

1. Cre itself can produce a phenotype.

Cre is usually expressed from a randomly integrated transgene, and transgene insertion sites are rarely, if ever, mapped in mice. A Cre transgene can produce a phenotype of its own in one of several ways. For example, Cre might disrupt an endogenous mouse gene at its insertion site and cause unanticipated effects of its own, especially when the transgene is bred to homozygosity. Second, the expression of Cre itself could lead to unexpected results. This is known as “Cre toxicity,” and it can come in one of several forms. The expression levels from some promoters can be high enough to affect physiology of the cell. This has been documented in at least one prominent case (Ins2-Cre; see Pomplun, et al. Horm Metab Res. 2007 May;39(5):336-40).

There are also naturally occurring “cryptic” loxP sites interspersed throughout the mouse genome. Cryptic loxP sites resemble loxP sites but differ in a few nucleotides. Cre can recognize cryptic loxP sites, albeit with much lower efficiency. When this happens, Cre can cause DNA damage. This was first shown in mouse embryonic fibroblasts (Loonstra A, et al. Proc Natl Acad Sci U S A. 2001 Jul 31;98(16):9209-14), but similar DNA damage from Cre alone has been shown in other cell types as well (Jae Huh W, et al. Am J Physiol Gastrointest Liver Physiol. 2010 Aug;299(2):G368-80). For these reasons, it is very important to include the Cre transgenic mouse itself (without any floxed sequences) as a control animal.

2. Germline Cre can (and should) be bred away after making a complete knockout.

It is possible to convert a conditional knockout (also called “floxed” for “flanked by loxP”) allele to a complete null allele by crossing to a Cre line with activity in the germline. Mice with germline Cre activity delete the floxed sequence in the sperm and/or oocytes, and transmit the deleted allele to their progeny. When using Cre/lox to convert a floxed allele to a complete knockout, it is a good idea to breed away any Cre with germline activity once you have demonstrated that the complete knockout has been transmitted to the next generation. This is especially true when using a Cre transgene with widespread activity throughout the mouse. Remember that Cre is a DNA recombinase, and its expression can have unintended effects. Once Cre has done its job (creating a heritable null allele, in this case), there is no reason to keep the transgene around. In a future post I will introduce different strategies for using the Cre-lox system to convert a floxed allele to a knockout allele.

3. Cre can be active in unexpected sites.

Many Cre transgenes have expression in unexpected organs and tissues, even when the promoters are thought to be tissue-specific. In some cases, there is unexpected or ectopic activity in the germline. For example, the AQP2-Cre strain we distribute (B6.Cg-Tg(Aqp2-cre)1Dek/J) is active not only in expected places (kidney cells), but also the male germline. In this example, care must be taken not to breed a male mouse carrying the floxed allele and Cre, because recombination will likely occur in the germline and the deleted (knockout) allele will be transmitted to its progeny. Another example is the Ins2-Cre line commonly known as RIP-Cre (B6.Cg-Tg(Ins2-cre)25Mgn/J), which is used to delete floxed sequences in beta cells of the pancreas. Not only is Cre active in the pancreas, but a recent report showed that Cre is unexpectedly active throughout the brain as well (Wicksteed, et al. Diabetes. 2010 Dec;59(12):3090-8).

Although this idea is more theoretical than empirical, the use of “knock-in” Cre alleles might help to reduce the incidence of unexpected, ectopic activity. A knock in Cre is a type of mutation where the Cre coding sequence either replaces the coding sequence of an endogenous gene (for example,B6.129P2(C)-Cd19tm1(cre)Cgn/J) or is appended to the end of the RNA message using an “internal ribosomal entry site,” which produces a bicistronic message encoding both the endogenous gene and Cre. An example of this is B6.129-Leprtm2(cre)Rck/J. One advantage of using a knock-in Cre allele is that Cre is expressed from the endogenous promoter, using all of the mouse gene’s regulatory sequences. Using a knock-in Cre also allows you to evaluate the likelihood of any unexpected expression by first reviewing the expression patterns of the endogenous gene. One location where expression information for endogenous mouse genes is collected is the Mouse Genome Informatics database. A knock-in Cre is created by gene targeting in ES cells, instead of pronuclear injection of a transgene sequence. In a future post I will delve into more details about knock-in Cre and BAC transgenics.

To address this issue the JAX Cre Repository has developed a comprehensive pipeline for the characterization of Cre driver strains using a LacZ reporter strain in a wide range of tissues and at multiple time points, including both target and non-target tissues. The data are annotated to include 11 broad organ systems, 30 individual organs/structures and 89 substructures, all of which are consistent with the mouse Anatomical Dictionary. Slide-scanned images and associated annotations are published on a dedicated website ( and submitted to, a comprehensive database of Cre strain functionality. The results indicate the vast majority of Cre driver strains exhibit unexpected recombinase activity in a number of tissue types, highlighting the need for this extended analysis.

4. Cre activity can be mosaic.

Some researchers find that a tissue-specific Cre transgene can fail to completely delete floxed alleles throughout the tissue under study. This is called “Cre mosaicism” and can arise in one of at least two ways. First, Cre might not be expressed in every cell in the organ of interest. Second, Cre might be expressed but fail to recombine the floxed allele in every cell. Cre mosaicism can be assessed using reporter lines. Failure to observe robust reporter activity throughout the organ of interest is a good indication that the Cre line will also fail to recombine a floxed allele. If you are creating a new Cre transgenic mouse, remember that Cre mosaicism can vary among founder lines. For this reason it is wise to evaluate Cre expression in different founder lines.