Site- and Time-Specific Gene Targeting in the Mouse - Science Direct

METHODS 24, 71–80 (2001) doi:10.1006/meth.2001.1159, available online at http://www.idealibrary.com on

Site- and Time-Specific Gene Targeting in the Mouse Daniel Metzger and Pierre Chambon1 Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire, CNRS/INSERM/ULP, Colle`ge de France, BP 163 67404 Illkirch Cedex, C.U. de Strasbourg, France

The efficient introduction of somatic mutations in a given gene, at a given time, in a specific cell type, will facilitate studies of gene function and the generation of animal models for human diseases. We have established a conditional site-specific recombination system in mice using a new version of the Cre/lox system. The Cre recombinase has been fused to a mutated ligand binding domain of the human estrogen receptor (ER), resulting in a tamoxifen-dependent Cre recombinase, Cre-ERT, that is activated by tamoxifen, but not by estradiol. Transgenic mice were generated expressing Cre-ERT under the control of a cytomegalovirus promoter. Administration of tamoxifen to these transgenic mice induced excision of a chromosomally integrated gene flanked by loxP sites in a number of tissues, whereas no excision could be detected in untreated animals. However, the efficiency of excision varied between tissues, and the highest level (⬃40%) was obtained in the skin. To determine the efficiency of excision mediated by CreERT in a given cell type, Cre-ERT-expressing mice were crossed with reporter mice in which expression of Escherichia coli ␤-galactosidase can be induced through Cre-mediated recombination. The efficiency and kinetics of this recombination were analyzed at the cellular level in the epidermis of 6- to 8-week-old double transgenic mice. Sitespecific excision occurred within a few days of tamoxifen treatment in essentially all epidermis cells expressing Cre-ERT. These results indicate that cell-specific expression of Cre-ERT in transgenic mice can be used for efficient tamoxifen-dependent Cre-mediated recombination at loci containing loxP sites, to generate site-specific somatic mutations in a spatiotemporally controlled manner. This conditional site-specific recombination system should allow the analysis of knockout phenotypes that cannot be addressed by conventional gene targeting. 䉷 2001 Academic Press

1 To whom correspondence should be addressed. Fax: (33) 388.65.32.03.

1046-2023/01 $35.00 Copyright 䉷 2001 by Academic Press All rights of reproduction in any form reserved.

I. NEED FOR SPATIOTEMPORALLY CONTROLLED SOMATIC MUTAGENESIS IN THE MOUSE Gene targeting in the mouse has yielded remarkable advances in understanding the roles played by specific gene products in mammalian development and adult physiopathology. This strategy has, however, some inherent limitations, which are due mostly to the introduction of the mutation in the germ line. Indeed, the lack of a protein that serves essential functions in embryogenesis can result in early lethality, thus precluding analysis of its possible functions at subsequent stages (1–3). Furthermore, numerous genes exert multiple functions in distinct cell types during ontogeny and postnatally (pleiotropy). This may result in complex phenotypes and, therefore, in difficulties in distinguishing cellautonomous from more complex origins of abnormalities (4 and references therein). The effect of a germline mutation may also be compensated during development, thus preventing the appearance of an abnormal phenotype in the adult animal. In the case of closely related genes belonging to multigene families, it may be necessary to mutate several members of the family to prevent functional redundancies (possibly artifactual) which preclude identification of the function of a given member of the family. Defining the function of this member may become even more difficult when the family is involved in highly pleiotropic signalling pathways, such as those of nuclear 71

72

METZGER AND CHAMBON

receptors, e.g., retinoid receptors (4, 5). Other potential effects confounding conventional knockouts include the risk of impaired fertility and generalized, systemic disorders (6, 7). In many instances, these limitations prevent the determination of the function of a given gene product in a defined subset of cells at any given time during the animal’s life. Moreover, they prevent the engineering of mouse models for human diseases that are caused by somatic mutations, particularly when these diseases result from a combination of somatic mutations, such as in most forms of cancer (2, 8). Thus, methods to achieve conditional gene inactivation are highly needed. Strategies for conditional gene targeting in mice, based on cell type-specific or inducible expression of the bacteriophage P1 site-specific Cre recombinase have been developed (9–14). The Cre recombinase can efficiently excise a DNA segment flanked by two loxP sites (floxed DNA) in animal cells (9). Spatially or temporally controlled somatic mutations can be obtained by placing the Cre gene under the control of either a cell-specific or an inducible promoter, respectively. However, these conditional gene targeting systems have also a number of limitations, as they are either spatially or temporally controlled. Ideally one would like to have a system that allows generation of somatic mutations in a defined gene, at a given time in the life of the animal and in a specific cell type.

This mutation was chosen because the corresponding mouse ER LBD mutant (Gly525Arg) binds the synthetic ligands tamoxifen (Tam) and 4-hydroxytamoxifen (OHT), but does not bind E2 (17). As expected, Cre-ERT recombinase activity was induced by OHT in a cell culture system, whereas no activity was detected in the absence of ligand addition or in the

II. GENERAL STRATEGY: TISSUE-SPECIFIC EXPRESSION OF LIGAND-INDUCIBLE CRE RECOMBINASES 1. Ligand-Inducible Cre Recombinases A number of studies have revealed that the activity of a variety of proteins (e.g., oncoproteins, transcription factors, RNA-binding protein, kinases) can be controlled in a ligand-dependent manner when fused to the ligand-binding domain (LBD) of steroid hormone receptors (for a review, see Ref. 15). Similarly, we have shown that the activity of a chimeric recombinase generated by the fusion of Cre to the LBD of the estrogen receptor (Cre-ER) is dependent on the presence of an estrogen agonist (e.g., 17␤-estradiol, E2) (Figs. 1A and 1B; (16)). To achieve tight control of the activity of such a chimeric protein in the presence of endogenous estradiol, we introduced a mutation in the LBD of the human ER (Gly521Arg), resulting in the chimeric protein Cre-ERT (Fig. 1A).

FIG. 1. Conditional site-specific recombination in mammalian cells using ligand-dependent chimeric Cre recombinases. (A) Schematic representation of Cre recombinase, the human estrogen receptor (ER), and the Cre-ER and Cre-ERT fusion proteins. Amino acid sequences of Cre and LBDs of ER are represented by hatched and gray boxes, respectively. Numbers refer to amino acid positions. The ER A/B, C, D, E, and F regions (36), the DNA binding domain (DBD), and the LBD, as well as the G400V and G521R mutations, are indicated. (B) Ligand-dependent activation of CreER and Cre-ERT in mouse F9 embryonal carcinoma cells. RXR␣⫹/ ⫺(LNL) F9 cells (37), containing a floxed tkneo gene, were transfected with a plasmid encoding Cre-ER (15) or with a plasmid encoding Cre-ERT (18) and treated with vehicle (no ligand), 100 nM E2, or 1 ␮M OHT. Excision of the floxed DNA segment was analyzed by PCR (18). O, no excision; ⫹, excision. (C) Schematic representation of ligand-activated site-specific recombination.

GENE TARGETING IN THE MOUSE

presence of E2 (Figs. 1B and 1C; (18)). These results indicated that Cre-ERT was of potential use for spatiotemporally controlled somatic mutagenesis in the mouse.

2. Ligand-Activated Site-Specific Recombination in Mice a. Establishment of CMV-Cre-ERT Transgenic Mice To test the activity of the tamoxifen-dependent Cre recombinase in the mouse, we generated transgenic animals expressing Cre-ERT under the control of the human cytomegalovirus (CMV) major IE gene enhancer/promoter, which is active in a number of cell types (18, 19). To this end, the 4.6-kb PvuII DNA fragment of pCMVCre-ERT (Fig. 2A) was injected into (C57BL/6XSJL) F1 zygotes according to established procedures (20). The Cre-ERT transgene was detected in mouse tail DNA by polymerase chain reaction (PCR), using primers (1) 5⬘-ATCCGAAAAGAAAACGTTGA-3⬘ and (2) 5⬘- ATCCAGGTTACGGATATAGT-3⬘ (Fig. 2A). PCR amplification was carried out in a buffer containing 10 mM Tris–HCl, pH 8.0, 50 mM KCl, 1.5 mM MgCl2, 0.2 mM dNTPs, 0.25 mM of each primer, and 2 U Taq polymerase using 1 ␮g of genomic DNA as template. After 35 cycles (30 s at 94⬚C, 30 s at 55⬚C) the products were analyzed on ethidium bromide-stained 2.5% agarose gels (18). In principle, immunohistochemistry analyses should reveal the expression pattern of the chimeric recombinase in transgenic lines. However, using various Cre antibodies, the Cre-ERT protein could not be readily detected in most of the tissues analyzed, except in the skin (see below and our unpublished results). We therefore estimated the level of Cre-ERT mRNA in various tissues. Total RNA was isolated by the LiCl/urea method (21). cDNA synthesized for 20 min at 50⬚C using 1 ␮g RNA was PCR-amplified through 35 cycles, using primers 3 (5⬘-TTGACCTCCATAGAAGACAC-3⬘) and 4 (5⬘-GGCGATCCCTGAACATGTCC-3⬘) (Fig. 2A), resulting in a 254-bp Cre-ERT cDNA fragment. As an internal control, a 177-bp cDNA fragment of the HPRT mRNA was coamplified in the same reaction using the primers 5⬘-GTAATGATCAGTCAACGGGGGAC-3⬘ and 5⬘CCAGCAAGCTTGCAACCTTAACCA-3⬘. Analysis of the PCR product on an agarose gel revealed that, in fact, Cre-ERT mRNA was expressed at various levels in all organs analyzed except in the thymus (Fig. 2C)

73

(18), suggesting that the protein could be expressed in most tissues, but not necessarily in all cell types. b. Tamoxifen-Inducible Recombinase Activity in CMV-Cre-ERT Transgenic Mice The efficiency and kinetics of excision of a floxed DNA segment can be determined at the cellular level by taking advantage of existing transgenic Cre reporter mouse lines in which LacZ expression is conditional on the removal of a floxed intervening segment. The mouse ACZL Cre reporter line carries the CMV enhancer/chicken ␤-actin promoter-LoxPchloramphenicol acetyltransferase (CAT) cassetteLoxP-LacZ cassette (13) (see Fig. 4A). ␤-Galactosidase cannot be expressed from this transgene, unless the CAT cassette is excised. However, the ACZL reporter line does not express the LacZ transgene in all cell types (13; our unpublished results). For example, in the skin, the reporter gene is expressed in the granular but not basal layer of the epidermis (see below and data not shown). We therefore analyzed at the DNA level the excision rate of a floxed gene segment in several organs by using a reporter line containing one WT allele of the RXR␣ gene and one modified RXR␣ allele carrying a floxed tkneo selection marker integrated by homologous recombination into the intron located between exon 8 and exon 9 [RXR␣⌬AF2(LNL)] (22) (Fig. 2B). The WT RXR␣ allele and the excised RXR␣⌬AF2(L) allele were simultaneously detected by PCR using one set of primers. The relative efficiency of excision was estimated by comparing the intensity of the band amplified from the deleted RXR␣⌬AF2(L) allele with that of the band amplified from the WT RXR␣ allele which differs in sequence only by absence of the loxP site. Offspring generated by crossing Cre-ERT and RXR␣ reporter mice, which harbored both the Cre-ERT transgene and the RXR␣⌬AF2(LNL) allele, were identified by genotyping of tail biopsies. The selection marker primers for detection of the RXR␣⌬AF2(LNL) allele were 5⬘-GGTTCTCCGGCCGCTTGGGT-3⬘ and 5⬘-GAAGGCGATGCGCTGCGAAT3⬘ (primers 5 and 6, respectively; Fig. 2B). Fourweek-old Cre-ERT/RXR␣⌬AF2(LNL)-positive littermates were injected intraperitoneally (ip) once a day with vehicle (oil) or with 1 mg OHT for 5 consecutive days. The OHT stock solution was prepared by addition of ethanol to 10 mg of OHT (Sigma) to obtain a 10 mg/100 ␮l OHT suspension. A 10 mg/ml OHT solution was prepared by addition of autoclaved sunflower oil, followed by 30 min sonication with a Branson ultrasonicator (Model 1210), and stored either

74

METZGER AND CHAMBON

at 4⬚C for a week or at ⫺20⬚C for months. The OHT stock solutions were sonicated just before use. One day before the first injection, DNA was prepared from tail biopsies of the animals (23). Two days after the last injection mice were killed and genomic DNA was isolated from various organs and analyzed for tkneo excision by PCR using 1 ␮g of genomic DNA as template and primers 7 (5⬘-CAAGGAGCCTCCTTTCTCTA-3⬘) and 8 (5⬘-CCTGCTCTACCTGGTGACTT-3⬘) (Fig. 2B). These primers amplify a 156bp fragment of the RXR␣ WT allele and a 190-bp fragment of the RXR␣⌬AF2(L) allele. After 35 cycles (30 s at 94⬚C, 30 s at 55⬚C) the products were analyzed on

ethidium bromide-stained 2.5% agarose gels. Excision of the floxed marker gene was undetectable in oil-treated control animals, whereas mice injected with OHT reproducibly showed some degree of excision of the floxed target gene in all organs tested except in the thymus (Fig. 2C). Importantly, the deleted RXR␣⌬AF2(L) allele was absent in the tail before OHT administration to the animal, whereas its presence was readily detected following OHT treatment (18) (Fig. 2C and data not shown). These results indicate that Cre-ERT is a tightly regulated recombinase that displays undetectable activity in the absence of its cognate ligand and can be activated in

FIG. 2. Tamoxifen-dependent site-specific recombination in mice. (A) Structure of the Cre-ERT transgene and strategies for its detection. The DNA fragments used to generate transgenic mice contained the enhancer/promoter region of the major IE gene of the human CMV, a rabbit ␤-globin intron, the Cre-ERT, and a SV40 polyadenylation signal [poly (A)]. The positions of the RNA start site (arrow) and of the primers used for PCR (primers 1 and 2) and for RT-PCR (primers 3 and 4), as well as the Pvull restriction sites, are indicated. (B) Genomic structure of the RXR␣ WT allele, the RXR␣⌬AF2(LNL) target allele, and the deleted RXR␣⌬AF2(L) allele, and PCR strategies to genotype RXR␣⌬AF2(LNL) mice (primers 5 and 6) and to analyze Cre-ERT-mediated excision of the floxed tkneo marker (primers 7 and 8). (C) Pattern of Cre-ERT-mediated DNA excision and Cre-ERT mRNA expression in various organs. The level of floxed DNA marker excision after five intraperitoneal injections of OHT in the indicated organs (hatched bars), the level of DNA excision in the tail 1 day after the third injection (open bar), and the corresponding levels of Cre-ERT mRNA (shaded bars) are shown. No excision could be detected in untreated animals.

GENE TARGETING IN THE MOUSE

mice by OHT treatment. We did not observe any deleterious effects of OHT treatment during this study, in agreement with reports indicating that short-term tamoxifen treatments have very low acute toxicity and cause no severe abnormalities in mice (24). Interestingly, the relative level of Cre-ERT mRNA correlated well with the level of DNA excision in the various organs examined (Fig. 2C). Excision was the most efficient in tail, skin, kidney, and spleen where 40 to 50% of the floxed marker gene was excised. Importantly, the excision in the tail remained at the same level (⬃50%) after three or five injections of OHT (Fig. 2C) (18). These results indicated that CreERT expression was probably restricted to specific cell types in the tail, and presumably also in other tissues, and that the excision measured for a given organ most likely underestimated the actual level of excision in a subset of specific cell types. c. Cre-Mediated Excision of Floxed DNA Is Specifically Induced by Tamoxifen in Cells Expressing the Cre-ERT Transgene Skin was used as a model tissue to further analyze the activity of Cre-ERT. Skin is composed of the dermis and the epidermis (Fig. 3A). The epidermis is made up of several stratified layers of increasingly differentiated keratinocytes. The proliferative basal keratinocytes (basal layer) are attached to the basement membrane that separates the dermis from the epidermis. Differentiation of basal keratinocytes is coordinated with vertical migration into the next layer, known as the spinous layer; further differentiation accompanied by further vertical migration results in the granular layer, which, with the spinous layer, constitutes the suprabasal layers. Continued differentiation is accompanied by further vertical migration, yielding a terminally differentiated keratinocyte, the corneocyte or squame, which is present in the most superficial epidermal layer, the cornified layer or strateum corneum (Fig. 3A). As squames are lost daily from the surface of the skin, the epidermis is a very dynamic tissue (25, 26). To reveal the expression of Cre-ERT in the skin of CMV-Cre-ERT mice, we performed immunohistochemistry on tail sections, using an anti-Cre rabbit polyclonal antibody and confocal microscopy. Transgenic mice were injected for 5 consecutive days with tamoxifen (Tam, 1 mg/day), and tail biopsies were collected before the first Tam injection (Day 0), and at Days 1, 3, 7, and 10. The stock solution was prepared by dissolving 10 mg Tam (free base, ICN)

75

in 100 ␮l ethanol. Autoclaved sunflower oil was added to obtain a 10 mg/ml solution, which was stored at ⫺20⬚C, as described above. Tail biopsies were embedded in OCT medium (Sakura Finetek Europe B.V., Zoeterwoude, The Netherlands), immediately frozen on dry ice, and sectioned with a cryostat (Leica CM 3050). Ten-micrometer-thick longitudinal sections were mounted on gelatin-coated slides and incubated in PBT [0.1% Triton X-100 in phosphate-buffered saline (PBS)] containing 5% normal goat serum (Sigma) for 1 h at room temperature. A 1/10 dilution of purified rabbit polyclonal anti-Cre antibody (D.M. et al., unpublished results) was applied to the slides for 2 h at 21⬚C. After five washes in PBT (10 min each), sections were incubated for 2 h at 21⬚C with a donkey anti-rabbit antibody coupled to the CY3 fluorochrome (Jackson Immunoresearch, West Grove, PA) at a 1/500 dilution. Slides were washed 5 ⫻ 10 min in PBT, and medium for fluorescence (Vectashield, Vector Laboratories, Inc.) containing 0.01% DAPI (4⬘,6-diamidino-2-phenylindole dihydrochloride; Sigma) was applied. The analysis of pictures taken on a Leica TSD4D confocal microscope revealed that there was no detectable Cre-ERT expression in the proliferative keratinocyte basal layer of untreated Cre-ERT mice, and that essentially all Cre-positive cells were located in the granular component of the suprabasal layers, whereas no significant staining was observed on sections of wild-type mouse skin (Fig. 3B). Interestingly, DAPI staining of keratinocyte nuclei clearly showed that the CreERT protein was essentially cytoplasmic in the absence of tamoxifen treatment, whereas it became nuclear on tamoxifen administration to be almost completely nuclear after 3 days of tamoxifen treatment (see Fig. 3C. The translocation of Cre-ERT from the cytoplasm to the nucleus of granular cells was clearly tamoxifen-dependent, as Cre-ERT was again mostly cytoplasmic when skin sections were examined 3 and 6 days after the last tamoxifen administration (Fig. 3D). As similar results were obtained when either Tam or OHT was used (data not shown), and as Tam is about 40 times cheaper than OHT, Tam was used in the following experiments. The efficiency and kinetics of floxed DNA excision in the epidermis were analyzed in double transgenic mice which carry both the CMV promoter-Cre-ERT transgene and the Cre recombinase reporter transgene present in the ACZL transgenic line (see above) (13) (Fig. 4A). Six- to eight-week-old CMVCre-ERT/ACZL double transgenic mice were injected for 5 consecutive days (Days 0 to 4) with tamoxifen

76

METZGER AND CHAMBON

(1 mg per day), and tail biopsies taken at various days were embedded in OCT medium and immediately frozen on dry ice. Thirty-micrometer-thick longitudinal sections were stained with X-Gal (5-bromo-4chloro-3-indolyl-␤-D-galactoside) as described (27, and references therein), dehydrated in ethanol (70, 90, and 100%, 2 ⫻ 10 min each), embedded in epoxy (Epon 812), cut with an ultramicrotome (Reichert Ultracut’s, Leica) in 2-␮m semithin sections, and counterstained with safranine. Analysis of the sections revealed that no X-Gal staining could be detected in the skin of untreated mice, whereas some staining, which was restricted to the granular layer, was visible after 1 day of treatment (Fig. 4B). All cells of the granular layer were clearly stained after

2 to 3 days of tamoxifen treatment (Fig. 4B) (see also Ref. 28). Most importantly, ␤-galactosidase expression was restricted to the Cre-ERT-expressing granular layer. If excision of the floxed CAT cassette did not occur in the proliferative basal keratinocytes, X-Gal staining should progressively disappear from the granular layer after cessation of tamoxifen treatment, move into the cornified layer, and finally disappear. This prediction was fulfilled when the CMV-Cre-ERT/ ACZL mice, initially treated for 5 days (Days 0 to 4) with tamoxifen, were examined at later times. Indeed, the X-Gal stain migrated vertically into the upper part of the granular layer (Fig. 4B, Day 7), in which very little stain was left in the lower part at

FIG. 3. (A) Epidermal stratification of mouse skin. Toluidine blue-stained 2-␮m semithin section of adult tail skin. A diagrammatic representation is given on the right-hand side. Note that mitosis is restricted to the proliferative basal cells. (B–D) Pattern of Cre-ERT expression in the tail epidermis of transgenic mice. Immunohistochemistry with anti-Cre antibody was performed on sections (10 ␮m thick) of tail biopsies of 6- to 8-week-old WT and CMV-Cre-ERT transgenic mice untreated (B) and treated with tamoxifen (C, D). Mice were injected intraperitonally for 5 consecutive days (Days 0–4) with tamoxifen (1 mg per day) as indicated. Sections were stained as indicated with DAPI and anti-Cre antibody, before the first tamoxifen injection (B), after 1 and 3 days of tamoxifen treatment (C), and 3 and 6 days after the fifth tamoxifen injection (D). Arrows point to the basement membrane (BM). B, S, G, and C: basal, spinous, granular, and cornified layers, respectively.

GENE TARGETING IN THE MOUSE

Day 10 (Fig. 4B, Day 10) and almost none in the upper part at Day 15, as well as into the cornified layer (Fig. 4B, Days 10 and 15). The observation that the newly differentiated granular cells were not stained indicated that no Cre-mediated recombination had occurred in the proliferative basal keratinocytes or, alternatively, that for unknown reasons, the tamoxifen treatment could have silenced expression of the lacZ gene. This latter possibility was clearly ruled out by the reappearance of X-Gal staining in the granular layer of tail biopsies of CMV-Cre-ERT/ACZL mice to which tamoxifen was

77

readministered between Days 25 and 29 (Day 30, Fig. 4B). To further demonstrate the efficiency of the tamoxifen-inducible site-specific recombination system in the mouse, we have recently established transgenic mice expressing Cre-ERT under the control of the bovine keratin K5 promoter, which is active selectively in the epidermal proliferative basal keratinocytes (29). As expected, Cre-mediated excision of floxed DNA occurred in all basal cells after tamoxifen treatment of the transgenic mice, and no background activity was observed (our unpublished results).

FIG. 4. (A) Schematic representation of the reporter cassette present in the ACZL transgenic line. A floxed chloramphenicol acetyltransferase (CAT) gene, located between the CMV enhancer/chicken ␤-actin promoter and the lacZ gene, prevents ␤-galactosidase expression (␤-gal⫺). Cre-mediated deletion of the floxed CAT gene allows expression of ␤-galactosidase (␤-gal⫹). pA, polyadenylation signal. LoxP sites, the CMV enhancer, and the chicken ␤-actin promoter sequences are indicated by filled arrowheads and crosshatched and open boxes, respectively. (B) Kinetics of ␤-galactosidase expression in tail epidermis granular layer of CMV-Cre-ERT/ACZL double heterozygous mice. Two series of daily tamoxifen (Tam) injections from Day 0 to Day 4 and from Day 25 to Day 29, were administered to 6- to 8-weekold CMV-Cre-ERT/ACZL double heterozygous mice. Tail biopsies were collected just before the first tamoxifen injection (Day 0) and at different days following the first injection, as indicated, sectioned, and stained with X-Gal. Two-micrometer semithin sections counterstained with safranine are shown. Arrows point to the basement membrane (BM). B, S, G, and C, basal, spinous, granular, and cornified layers, respectively.

78

METZGER AND CHAMBON

DISCUSSION The present Cre-ERT system appears particularly attractive for conditional somatic mutagenesis. Indeed, in two transgenic lines expressing Cre-ERT in either the basal or the granular layer of the epidermis, tamoxifen administration induced recombination in all cells expressing the fusion protein, whereas none could be detected in the absence of tamoxifen. The observation that tamoxifen is required for translocation of the unliganded cytoplasmic Cre-ERT to the nucleus may account, at least in part, for the undetectable background of recombination in the absence of tamoxifen. The excision kinetics can only be indirectly estimated from ␤-galactosidase activity. At least some recombination must take place within the first 24 h following tamoxifen injection (see Fig. 4B). This is also the time at which some Cre-ERT has been translocated into the nucleus, although complete translocation may take 2–3

days of tamoxifen treatment (see Fig. 3). It appears therefore that the intracellular level of tamoxifen could be a more limiting factor than the amount of expressed Cre-ERT. In this respect, we note that the Cre-ERT2 recombinase, containing the G400V/ M543A/L544A triple mutation of the human ER LBD, is about three- to four-fold more sensitive in F9 cells to OHT when compared with Cre-ERT (30), and therefore might be also more potent than CreERT in mice. Temporally controlled generation of cell type-restricted somatic mutations will require careful evaluation of the pattern of expression and activity of CreERT in transgenic animals. Indeed, mosaic expression of Cre-ERT in the target tissue will result in incomplete gene inactivation in the cell population. On the other hand, expression of Cre-ERT outside of the target tissue will lead to mutation of the gene in undesired cell types, and thus will complicate analysis of the phenotype. Mosaic gene expression is not

FIG. 5. Engineering of germ line and site-directed cell-specific temporally controlled somatic mutageneses in the mouse. (A) Establishment of mouse lines harboring either an exon 2-deleted or an exon 2-floxed gene. (B) Conditional somatic mutation in the mouse.

GENE TARGETING IN THE MOUSE

uncommon in transgenic animals and is likely to be affected both by the susceptibility to position effects of the promoter element chosen to express Cre-ERT and by its site of integration (31). Therefore, promoters displaying both position-independent gene expression and the correct spatial specificity should be selected. The use of chromatin insulator elements may help in achieving these goals (32–34). Alternatively, the desired expression specificity might be obtained by knocking in the Cre-ERT transgene at a chromosomal locus displaying the desired expression pattern. In any event, reporter mouse lines that can visualize Cre activity at the cellular level, by activation or suppression of a ubiquitously expressed cell marker on Cre-mediated excision, are required, but unfortunately not yet available. As the production of mice carrying a floxed allele represents little additional work when compared with classic knockout, it should clearly be considered when gene targeting experiments are being planned. To obtain a conditional allele (L2⫹) in place of a constitutive null mutant allele (⫺) (resulting from insertion of the selectable marker gene in an exon essential for gene function), a floxed marker gene and a loxP site have to be inserted into the introns flanking the exon, resulting in an L3 floxed targeted allele (Fig. 5A). Cre-mediated recombination will result in either the excision of the marker gene only, resulting in a floxed allele (L2⫹), or the excision of the marker gene and/or the floxed exon(s), resulting in null alleles (L- and L2- alleles, respectively). These various alleles can be obtained either by transient expression of Cre in ES cells or by crossing mice harboring the targeted L3 allele with a CMV-Cre mouse that we have established (35; our unpublished results) (see Fig. 5A). Indeed, this transgenic Cre line mediates deletion in the germ cells of all the sequences located between the three loxP sites, as well as deletion between two loxP sites only. Therefore, with a single step of gene modification by homologous recombination in ES cells and transmission of the mutation through the germ line, one can obtain gene inactivation either in the germ line or selectively in various tissues, at different times (spatiotemporally controlled somatic mutation), by simply mating mice harboring a floxed gene with the CMVCre mice or with a corresponding Cre-ERT transgenic that displayed the desired tissue specificity of expression, respectively (Fig. 5). Thus, the same genetically modified animal can be used to answer a variety of different questions related to the function of the targeted gene.

79

ACKNOWLEDGMENTS We are grateful to Professor A. Berns and to Dr. M. Giovannini for the generous gift of the ACZL mice. We thank J. M. Bornert, J. Brocard, S. Bronner, N. Chartoire, H. Chiba, J. Clifford, R. Feil, M. Gendron, B. Mascrez, N. Messaddeq, B. Schuhbaur, J. L. Vonesch, X. Warot, and O. Wendling for enthusiastic participation in the studies presented in this review, as well as M. LeMeur and the animal facility staff for animal care, the secretarial staff for typing, and the illustration staff for preparing the figures. This work was supported by funds from the Centre National de la Recherche Scientifique, the Institut National de la Sante´ et de la Recherche Me´dicale, the Colle`ge de France, the Hoˆpital Universitaire de Strasbourg, the Association pour la Recherche sur le Cancer, the Fondation pour la Recherche Me´dicale, the Human ´ ducation NatioFrontier Science Program, and the Ministe`re de I’E nale de la Recherche et de la Technologie De´c.97.C.0115.

REFERENCES 1. Marth, J. D. (1996) J. Clin. Invest. 97, 1999–2002. 2. Rajewsky, K., Gu, H., Ku¨hn, R., Betz, U.A.K., Mu¨ller, W., Roes, J., and Schwenk, F. (1996) J. Clin. Invest. 98, 600–603. 3. Lobe, C., and Nagy, A. (1998) BioEssays 20, 200–208. 4. Kastner, P., Mark, M., and Chambon, P. (1995) Cell 83, 859–869. 5. Chambon, P. (1996). FASEB J. 10, 940–954. 6. Letterio, J. J., Geiser, A. G., Kulkarni, A. B., Roche, N. S., Sporn, M. B., and Roberts, A. B. (1994) Science 264, 1936– 1938. 7. Diebold, R. J., Eis, M. J., Yin, M., Boivin, G. P., Darrow, B. J., Saffitz, J. E., and Doetschman, T. (1995) Proc. Natl. Acad. Sci. USA 92, 12215–12219. 8. Olson, E. N., Arnold, H. H., Rigby, P. W. J., and Wold, B. J. (1996) Cell 85, 1–4. 9. Sauer, B. and Henderson, N. (1989) Nucleic Acids Res. 17, 147–161. 10. Gu, H., Marth, J., Orban, P. C., and Rajewsky, K. (1994) Science 265, 103–106. 11. Ku¨hn, R., Schwenk, F., Aguet, M., and Rajewsky, K. (1995) Science 269, 1427–1429. 12. Tsien, J. Z., Feng Chen, D., Gerber, D., Tom, C., Mercer, E. H., Anderson, D. J., Mayford, M., Kandel, E. R., and Tonegawa, S. (1996) Cell 87, 1317–1326. 13. Akagi, K., Sandig, V., Vooijs, M., Van der Valk, M., Giovannini, M., Strauss, M. and Berns, A. (1997) Nucleic Acids Res. 25, 1766–1773. 14. Barlow, C., Schroeder, M., Lekstrom-Himes, J., Kylefjord, H., Deng, C. X., Wynshaw-Boris, A., Spiegelman, B. M., and Xanthopoulos, K. G. (1997) Nucleic Acids Res. 25, 2543–2545. 15. Metzger, D., Clifford, J., Chiba, H., and Chambon, P. (1995) Proc. Natl. Acad. Sci. USA 92, 6991–6995. 16. Picard, D. (1994) Curr. Biol. 5, 511–515. 17. Danielian, P. S., White, R., Hoare, S. A., Fawell, S. E., and Parker, M. G. (1993) Mol. Endocrinol. 7, 232–240. 18. Feil, R., Brocard, J., Mascrez, B., LeMeur, M., Metzger, D., and Chambon, P. (1996) Proc. Natl. Acad. Sci. USA 93, 10887– 10890.

80

METZGER AND CHAMBON

19. Schmidt, E. V., Christoph, G., Zeller, R., and Leder, P. (1990) Mol. Cell. Biol. 10, 4406–4411. 20. Hogan, B., Costantini, F., and Lacy, E. (1986) Manipulating the Mouse Embryo: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 21. Auffray, C., and Rougeon, F. (1980) Eur. J. Biochem. 107, 303–314. 22. Mascrez, B., Mark, M., Dierich, A., Ghyselinck, N., Kastner, P., and Chambon, P. (1998) Development 125, 4691–4707. 23. Lufkin, T., Dierich, A., LeMeur, M., Mark, M., and Chambon, P. (1991) Cell 66, 1105–1119. 24. Furr, B. J. A., and Jordan, V. C. (1984) Pharmacol Ther. 25, 127–205. 25. Fuchs, E. (1990) J. Cell Biol. 111, 2807–2814. 26. Leigh, I. M., Lane, E. B., and Watt, F. M. (1994) The Keratinocyte Handbook, Cambridge Univ. Press, London/New York. 27. Takahashi, K., and Coulombe, P. A. (1996) Proc. Natl. Acad. Sci. USA 93, 14776–14781. 28. Brocard, J., Warot, X., Wendling, O., Messaddeq, N., Vonesch, J. L., Chambon, P., and Metzger, D. (1997) Proc. Natl. Acad. Sci. USA 94, 14559–14563.

29. Ramirez, A., Bravo, A., Jorcano, J. L., and Vidal, M. (1994) Differentiation 58, 53–64. 30. Feil, R., Wagner, J., Metzger, D., and Chambon, P. (1997) Biochem. Biophys. Res. Commun. 237, 752–757. 31. Dobie, K., Mehtali, M., McClenaghan, M., and Lathe, R. (1997) Trends Genet. 13, 127–130. 32. Chung, J. H., Whiteley, M., and Felsenfeld, G. (1993) Cell 74, 505–514. 33. Festenstein, R., Tolaini, M., Corbella, P., Mamalaki, C., Parrington, J., Fox, M., Miliou, A., Jones, M., and Kioussis, D. (1996) Science 271, 1123–1125. 34. Milot, E., Strouboulis, J., Trimborn, T., Wijgerde, M., de Boer, E., Langeveld, A., Tan-Un, K., Vergeer, W., Yannoutsos, N., Grosveld, F., and Fraser, P. (1996), Cell 87, 105–114. 35. Dupe´, V., Davenne, M., Brocard, J., Dolle´, P., Mark, M., Dierich, A., Chambon, P., and Rijli, F. (1997) Development 124, 399–410. 36. Krust, A., Green, S., Argos, P., Kumar, V., Walter P., Bornert, J. M., and Chambon P. (1986) EMBO J. 5, 891–897. 37. Clifford, J., Chiba, H., Sobieszczuk, D., Metzger, D., and Chambon, P. (1996) EMBO J. 15, 4142–4155.