Genomic targeting of a bicistronic DNA fragment by Cre-mediated site-specific recombination

Gene 203 (1997) 209–216

Genomic targeting of a bicistronic DNA fragment by Cre-mediated site-specific recombination Andreas F. Kolb a,b,*, Stuart G. Siddell a a Institute of Virology and Immunology, University of Wu¨rzburg, Versbacherstr. 7, D-97078 Wu¨rzburg, Germany b Hannah Research Institute, Mauchline Road, Ayr, KA6 5HL, UK Received 20 May 1997; accepted 28 August 1997; Received by A.M. Campbell

Abstract The Cre-recombinase of bacteriophage P1 catalyses site-specific recombination between DNA fragments containing loxP sites. Targeting of predefined genomic loci can be achieved by Cre-mediated linkage of a promoterless resistance marker gene to a floxed promoter pre-existing in the genome. In order to avoid the introduction of plasmid sequences into the host genome, we have constructed a series of plasmids in which the DNA segment to be integrated is flanked by two loxP sites. We show here that this floxed targeting fragment is reliably and effectively separated from the vector backbone and integrated into genomic loxP sites by Cre-mediated site-specific recombination in mammalian cells. We also demonstrate that by using this approach two convergent, promoterless coding regions can simultaneously be linked to two independent promoter elements at a pre-existing genomic loxP site. This methodology will be particularly useful for genomic targeting experiments in transgenic animals. © 1997 Elsevier Science B.V. Keywords: Genetic engineering; Cloning vectors; loxP site; Transgenic animals; Transcriptional interference

1. Introduction The Cre-recombinase of bacteriophage P1 catalyses site-specific recombination between loxP sites consisting of two 13-bp inverted repeats flanking an 8-bp unidirectional spacer segment. Expression of the cre gene in eukaryotic cells has allowed the precise excision of DNA fragments flanked by loxP sites ( Kanegae et al., 1995; Bergemann et al., 1995) and the integration of DNA fragments into pre-existing loxP sites ( Fukushige and Sauer, 1992; Albert et al., 1995). To date, however, the integration of DNA fragments into loxP sites in mammalian cells has usually meant the integration of the entire plasmid. Our long term goal is to utilise Cremediated recombination for the targeted integration of * Corresponding author. Tel. +44 1292 674020; Fax: +44 1292 678797; e-mail: [email protected] Abbreviations: CRFK, Crandell feline kidney cells; floxed, flanked by loxP sites; Hy, hygromycin B; HyR, hygromycin resistant/resistance; PGK, phosphoglycerol-kinase; Luc, firefly luciferase; MT, metallothionein; NmR, neomycin-resistant/resistance; ORF, open reading frame; PBS, phosphate buffered saline; PolIk, Klenow fragment of E. coli DNA polymerase I; SV40, simian virus 40; WAP, whey acidic protein. 0378-1119/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 03 7 8 -1 1 1 9 ( 9 7 ) 0 0 5 15 - 5

foreign genes into specific genomic loci in transgenic animals. Bacterial vector sequences, which have been shown to have adverse effects on the expression of transgenes, are normally excluded from the constructs used for oocyte injection (Jaenisch, 1988 and references therein). To avoid the insertion of bacterial vector sequences into the mammalian genome when using Cremediated recombination, we have constructed plasmid vectors in which the sequences to be integrated are flanked by two loxP sites. This floxed DNA fragment is separated from the vector backbone in a first recombination reaction and integrated into a genomic loxP site by a second recombination event. Such a two-step reaction has been successfully used in yeast (Sauer and Henderson, 1990). We demonstrate here that this process is also feasible in mammalian cells and that it occurs with an efficacy comparable with the single-step recombination reaction ( Fukushige and Sauer, 1992; Baubonis and Sauer, 1993). We have used vectors carrying two loxP sites to integrate a bicistronic DNA fragment carrying two convergent, promoterless open reading frames (ORFs) into a singular genomic loxP site. We show here that this integration event can simultaneously link the two genes to two independent eukaryotic

210

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216

promoter elements which subsequently direct the expression of the respective adjacent genes. By using this strategy, mammalian cell lines or embryonic stem cell lines in which a loxP site has been integrated next to a suitable mammalian promoter by homologous recombination can be used for the repeated integration of foreign genes into a predefined chromatin structure with a predictable expression pattern.

2. Materials and methods 2.1. DNA cloning DNA cloning was performed according to standard procedures. Plasmids were grown in E. coli strain DH5a. The construction of the plasmids pB-lox1/2-hyg, pB-lox1-hyg and pB-lox1/2-hyg/luc was described previously ( Kolb and Siddell, 1996). The plasmid pB-lox1/2-hyg/luc2 was generated by insertion of a 3.7-kb XhoI–XmnI DNA fragment encompassing the firefly luciferase (Luc) coding region from the plasmid pT109luc(−)tk ( Kolb and Siddell, 1996) into the XhoI and EcoRV sites of pB-lox1/2-hyg. To generate the plasmid pMTloxSV, a 615-bp (KpnI–RsaI ) DNA fragment of the murine metallothionein I (MT ) promoter was excised from the plasmid pXGH (Selden et al., 1986) and ligated with the plasmid ploxSV-neo that had been digested with KpnI+XhoI. The XhoI restriction site was blunt ended by using PolIk. The plasmid ploxSV-neo was generated by PCR amplification of the SV40 promoter from the plasmid Lys17 (Orlandi et al., 1989), using the primer pair SV-low (5∞-CCATCGATCGACCCTGTGGAATGTGTGTC-3∞) and SV-up (5∞-GAACGTCGACATAACTTCGTATAGCATACATTATACGAAGTTATTTTGCAAAAGCCTGCCTCCA-3∞), and subsequent cloning of the ClaI+SalI-digested PCR product into the plasmid pBluescript-PGK-neo that had been digested with the same enzymes. The amplification was done at a final primer concentration of 5 mM with 30 cycles of 1 min at 94°C, 1 min at 62°C and 2 min at 72°C, using 10 ng of plasmid DNA as template. To generate the fusion plasmid pMTSVhyg/luc, the plasmids pMTloxSV and pB-lox1/2-hyg/luc were recombined in vitro by using an MBP–Cre fusion protein ( Kolb and Siddell, 1996). The recombination reaction was purified and transformed into competent bacteria. Bacterial colonies carrying the correct recombination product were identified by restriction digestion of plasmid DNA after selection on agar plates supplemented with ampicillin and hygromycin B (Hy; Boehringer-Mannheim, Mannheim, Germany). The fusion plasmid pMTSV-hyg/luc2 was generated correspondingly by MBP–Cre-mediated in vitro recombination of the plasmids pMTloxSV and pB-lox1/2-hyg/luc2.

2.2. Cell cultivation and transfection CRFK ( ECACC 86093002) feline kidney cells were grown at 37°C and 5% CO in minimal essential medium 2 (MEM ) with Earle’s salts, supplemented with 10% fetal bovine serum, non-essential amino acids, glutamine and antibiotics. Transfections involving a recombination reaction were performed by using an electroporation device ( Invitrogen, Leek, The Netherlands). 107 cells in 0.6 ml ice cold PBS were mixed with 40 mg of the Creexpression plasmid pMC1-Cre (Gu et al., 1993) and 20 mg of the plasmid carrying the loxP site(s) and incubated on ice for 10 min. The cells were subsequently pulsed at 250 V and 1000 mF. After incubating on ice for a further 10 min, the cells were seeded into a 10-cm Petri dish with fresh medium. After a 24-h incubation period at 37°C the medium was removed and replaced by medium supplemented with 100 mg/ml Hy. The number of living cells which had survived the electroporation procedure was scored at that point by trypan blue exclusion staining in cells transfected in a parallel experiment. Routinely, 10% of the cells remained viable after the electroporation procedure. Hygromycin-resistant ( HyR) cell clones were counted after 14 days. Individual clones were isolated and expanded in medium containing 400 mg/ml Hy. Cre-mediated integration of the floxed plasmid was assessed by Southern blotting and PCR. Control transfections of 10 mg of the plasmid pCMV-b (Clontech, Heidelberg, Germany) using the same number of cells and the same electroporation conditions as above resulted in a transfection efficiency of approx. 15%. All other transfections were performed as described by Chen and Okayama (1987). 5 mg of plasmid DNA was transfected into 5×105 CRFK cells plated in a 6-cm Petri dish. 12 h post-transfection, the medium was removed and fresh medium was added. After a further 24 h the transfected CRFK cells were transferred to a 10-cm Petri dish and selected in medium supplemented with 400 mg/ml G418 (Sigma, Deisenhofen, Germany). To generate pools of stably transfected cells, a minimum of 200 independent cell clones were mixed. 2.3. DNA analysis To isolate genomic DNA, CRFK feline kidney cells were trypsinised and washed twice with ice-cold PBS. The cells were pelleted and incubated in a buffer containing 100 mM Tris–HCl, pH 8.5, 5 mM EDTA pH 8.0, 0.2% SDS, 200 mM NaCl and 100 mg/ml Proteinase K at 37°C for 16 h. DNA was precipitated from the cellular lysates by the addition of 0.8 vol. of isopropanol. The precipitated DNA was washed twice with 70% ethanol, air-dried and resuspended in 1×TE buffer (10 mM Tris–HCl, pH 7.5, 1 mM EDTA). Southern blot analysis was done as described previously ( Kolb and Siddell, 1996). PCR amplifications were done with

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216

Gold-Star Polymerase ( Eurogentec, Seraing, Belgium) ¨ berlingen, in a PE 480 thermocycler (Perkin-Elmer, U Germany). Amplifications with the primer pair C68 (5∞-CAGGAAACAGCTATGAC-3∞) and luc-int (5∞-CTGCAACTCCGATAAATAAC-3∞) were done at a final primer concentration of 1 mM with 40 cycles of 1 min at 94°C, 1 min at 56°C and 2 min at 72°C. 2 mg of genomic DNA or 50 ng of plasmid DNA were used as template. 2.4. Protein analysis Luciferase activity in cytoplasmic extracts of transfected cells was determined as described previously ( Kolb et al., 1994). Protein concentrations were determined by the method of Bradford (Bradford, 1976).

3. Results and discussion 3.1. Genomic targeting with a vector plasmid carrying two loxP sites The plasmids currently used for Cre-mediated targeting of DNA fragments in mammalian cells contain a single loxP site flanked by a promoter which becomes integrated next to a promoterless resistance gene upon site-specific recombination (Fukushige and Sauer, 1992; Baubonis and Sauer, 1993). This strategy leads to the incorporation of extensive, plasmid-derived vector sequences into the mammalian genome, which have been shown to exert adverse effects on eukaryotic gene expression, particularly in transgenic animals (Jaenisch, 1988). To exclude vector sequences from Cre-mediated integration we have constructed a plasmid containing two loxP recombination sites, which flank a promoterless hygromycin resistance (HyR) gene (pB-lox1/2-hyg, Fig. 1). pB-lox1/2-hyg is separated into an insert and a vector fragment ( Fig. 1) by Cre-mediated recombination in vitro ( Kolb and Siddell, 1996) and upon transfection into tissue culture cells (data not shown). Both of these fragments carry one loxP site and can be integrated into the genomic loxP site. However, only the linkage of the fragment carrying the promoterless HyR gene to the preexisting promoter element will render the cells resistant to Hy (Fig. 1). pB-lox1/2-hyg and a molar excess of the Creexpression plasmid pMC1-Cre (Gu et al., 1993) were transfected into the cell clone CRFK-pBR-WloxSV c-3 ( Kolb and Siddell, 1996), which carries a singular loxP site flanked by the SV40 and the murine WAP promoter (Fig. 1). Twelve of the 32 HyR colonies which could be recovered after selection of the transfected cells in medium containing 100 mg/ml Hy, were expanded and analysed by Southern blotting, PCR and DNA sequencing (data not shown). In 9 of the 12 clones analysed

211

the HyR gene had been integrated at the pre-existing loxP site and at no other site in the genome ( Table 1). In all of these clones the plasmid backbone had not been co-integrated. As evidenced by S1 nuclease protection analysis, transcription of the HyR gene was directed by the SV40 promoter (data not shown). None of the clones examined contained the cre ORF integrated into the genome (data not shown), suggesting that a constant expression of the cre gene is detrimental to the survival of cells in the selection process. Three of the 12 clones analysed carried the HyR gene at a location other than the pre-existing loxP site ( Table 1). Two hygromycinresistant (HyR) colonies could also be recovered after a control transfection of pMC1-Cre and pB-lox1/2-hyg into unmodified CRFK cells, confirming that the HyR gene can be activated upon random integration into the cellular genome ( Table 1). When the transfection experiments were repeated with a second cell clone (CRFK-pBR-WloxSV c-2), which contains an identical but differently located genomic loxP site, we could only derive one HyR clone. Moreover, the HyR gene had not been integrated at the pre-existing loxP site in this clone, but rather was inserted at random ( Table 1). This indicates, in accordance with published data (Baubonis and Sauer, 1993), that the chromosomal location of the pre-existing loxP site is a key factor determining the efficiency with which site-specific recombinations occur. The frequency with which HyR cell clones could be recovered upon transfection of plasmids carrying two loxP sites ( Table 1) was comparable with the frequency reported for plasmids carrying a singular loxP site ( Fukushige and Sauer, 1992; Baubonis and Sauer, 1993). The proportion of HyR clones arising from random integration appears to be higher when the two loxP site plasmids were used ( Table 1; Fukushige and Sauer, 1992). These differences, however, may be related to the different cell lines and resistance marker genes used in the different studies. Surprisingly, only very few HyR descendants could be recovered when the plasmid pB-lox1-hyg ( Kolb and Siddell, 1996), which carries only a singular loxP site next to the promoterless HyR gene, was transfected into the cell line CRFK-pBRWloxSV c-3 ( Table 1). Additionally, these colonies might have arisen from a random integration event. This finding is in contrast with previous results obtained in yeast (Sauer and Henderson, 1990), where plasmids carrying a singular loxP site were more efficiently targeted than plasmids carrying two loxP sites. As the HyR expression cassette (HyR coding region and SV40 polyadenylation signal ) is identical in the plasmids pB-lox1-hyg and pB-lox1/2-hyg, the low efficiency with which HyR descendants can be recovered after transfection of pB-lox1-hyg is probably not due to an impaired expression of the HyR gene. The sequences which are absent from pB-lox1-hyg but are present in pB-lox1/

212

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216

Fig. 1. Schematic representation of the plasmid pB-lox1/2-hyg and its anticipated reaction products during the process of recombination with the pre-existing loxP site in the cell line CRFK-pBR-WloxSV c-3. The loxP recombination target site and the hygromycin phosphotransferase gene (HyR) are represented as striped boxes. WAP and SV40 promoter elements are marked as thick arrows. In the first recombination reaction the plasmid insert which is flanked by two directly repeated loxP sites is separated from the vector backbone. In the second recombination reaction the vector or the insert fragment can be integrated into the pre-existing genomic loxP site. However, only the fusion of the HyR gene to the SV40 promoter confers resistance to Hy selection.

Table 1 Cre-mediated integration of plasmids carrying one or two loxP sites DNA transfected

Cell line transfected

HyR coloniesa

Colonies analysedb

Rand. integrationc

HyR clones/ 106 cellsd

Experiment 1 pB-lox1/2-hyg + pMC1-Cre pB-lox1/2-hyg + pMC1-Cre pB-lox1/2-hyg + pMC1-Cre pB-lox1-hyg + pMC1-Cre

CRFK CRFK+WloxSV c-3 CRFK+WloxSV c-2 CRFK+WloxSV c-3

2 32 1 2

0 12 1 0

2e 3 1

~2 32 ~1

Experiment 2 pB-lox1/2-hyg + pMC1-Cre pB-lox1/2-hyg + pMC1-Cre pB-lox1-hyg + pMC1-Cre

CRFK+WloxSV c-3 CRFK+WloxSV c-2 CRFK+WloxSV c-3

12 0 0

7

0

12 ~0 ~0

Experiment 3 pB-lox1/2-hyg/luc + pMC1-Cre

CRFK+WloxSV c-3

22

8

0

22

aNumber of HyR colonies scored 14 days post-transfection. bNumber of colonies analysed further by Southern blotting and/or PCR. cNumber of colonies in which the floxed DNA fragment was integrated at random (where applicable). dNumbers of HyR colonies recovered per 106 viable cells post-transfection as scored by Trypan blue exclusion staining. eAlthough these two colonies have not been analysed, the HyR gene could not have integrated into a pre-existing loxP site, as no such site is present in the unmodified CRFK cells. The colonies have therefore (by definition) been classified as carrying random integrations.

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216

2-hyg may rather alter the efficiency with which the two plasmids are used in the Cre-mediated recombination reaction. 3.2. Integration of a bicistronic DNA fragment into a genomic loxP site With the strategy described above ( Fig. 1) it should also be possible to integrate two ORFs into a preexisting loxP site, thereby linking them to the two promoter elements flanking the pre-existing loxP site. To confirm this prediction we generated the plasmid pMTloxSV, in which the loxP site is flanked by the strong constitutive murine MT promoter and the SV40 promoter ( Fig. 2A). This plasmid was transfected into CRFK cells and two cell clones carrying pMTloxSV were established (CRFK-MTloxSV c-6 and CRFKMTloxSV c-8; Fig. 2A). The plasmid pB-lox1/2-hyg/luc ( Kolb and Siddell, 1996) which contains a Luc coding region in reverse orientation to the HyR gene (Fig. 2A) was subsequently transfected into these cell clones together with the Cre-expression plasmid pMC1-Cre. Eighteen and 21 HyR cell clones were obtained upon transfection of CRFK-MTloxSV c-6 and CRFKMTloxSV c-8 cells, respectively. DNA isolated from the parental cell lines CRFKMTloxSV c-6 and CRFK-MTloxSV c-8 and their HyR descendants was analysed by PCR using a primer pair encompassing the junction between the MT promoter and the Luc coding region ( Fig. 2A). The expected product of 1074 bp could be detected in the HyR cells, but not in the parental cell lines (Fig. 2B), indicating that the recombination process had linked the Luc gene to the MT promoter element. The same PCR product could also be detected when an in vitro recombination product of pMTloxSV and pB-lox1/2-hyg/luc (pMTSVhyg/luc in Fig. 2B) or genomic DNA from CRFK cells stably transfected with this in vitro recombination product (CRFK-MTSV-hyg/luc in Fig. 2B) was used as template. Sequence analysis of the cloned PCR product of 1074 bp confirmed the correct juxtaposition of the MT-promoter and the Luc gene (data not shown). Southern blot analysis with genomic DNA isolated from the HyR cell clone, CRFK-MTloxSV c-6-hyg/luc1, confirmed that the Luc coding region was only integrated into the genome at the pre-existing loxP site ( Fig. 2C ). Cytoplasmic extracts of the parental cells and the HyR cell clones CRFK-MTloxSV c-6-hyg/luc1 and CRFK-MTloxSV c-8-hyg/luc1 were analysed for Luc expression. The cell clones in which the MT promoter had been linked to the Luc gene by Cre-mediated sitespecific recombination expressed high levels of luciferase (C-6 HL1, C-8 HL1 in Fig. 2D). These results corroborate that it is possible to simultaneously link two promoterless coding regions to two independent promoter elements by Cre-mediated site-specific recombination.

213

3.3. Interference between the convergent ORFs As the two coding regions in the plasmid pB-lox1/2-hyg/luc are transcribed in opposite directions, we wondered whether they might have a suppressive influence on each other. To answer this question we generated the plasmid pB-lox1/2-hyg/luc2, which is identical to pB-lox1/2-hyg/luc but carries a 1.6-kb spacer fragment inserted between the HyR and Luc ORFs. pB-lox1/2-hyg/luc and pB-lox1/2-hyg/luc2 were recombined with the plasmid pMTloxSV in vitro by using an MBP–Cre fusion protein ( Kolb and Siddell, 1996). The in vitro recombination products pMTSV-hyg/luc and pMTSV-hyg/luc2, which carry the genes encoding NmR, HyR and Luc were transfected into CRFK cells. More than 200 independent stably transfected cell clones were pooled after selection in medium containing G418. The levels of Luc expression in cells transfected with pMTSV-hyg/luc2 (HL2-n; Fig. 3) were significantly higher than in cells transfected with pMTSVhyg/luc (HL-n; Fig. 3) indicating that the inserted spacer segment was beneficial for the Luc expression directed by the MT promoter. As evidenced by Southern blotting, the cell pools contained equivalent amounts of the Luc gene (data not shown), suggesting that the expression differences are not due to a different number of Luc gene copies. The cells were subsequently cultivated in medium containing 400 mg/ml Hy to eliminate cell clones only carrying the NmR but not the HyR gene. This reselection not only increased the number of HyR gene copies per cell, but also resulted in a 10-fold enrichment of the Luc gene copy number per cell in both cell pools (data not shown) and a concomitant increase in Luc expression ( Fig. 3). These results indicate that although transcription of the HyR gene is not refractory to Luc expression it is able to reduce it. The selection of suitable DNA spacer fragments which are able to buffer the interference between the two transcriptional units therefore remains an important goal for future experiments. 3.4. Conclusions (1) We have demonstrated that it is possible to avoid the introduction of bacterial vector sequences from Cre-mediated integration into the genome of mammalian cells by incorporating a second loxP site into the targeting plasmids. The floxed insert is excised from the plasmid backbone in an initial recombination reaction and subsequently integrated into a pre-existing genomic loxP site by a second recombination. As is obvious from Fig. 2(A), the genomic site at which the Cre-mediated integration occurs still contains the two resistance marker genes. Should these sequences interfere with the expression of an introduced transgene, they could, however, be eliminated during another round of site-specific

214

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216

Fig. 2. (A) Schematic representation of the plasmid pB-lox1/2-hyg/luc and its anticipated recombination products upon recombination with the pre-existing loxP site in the cell lines CRFK-MTloxSV c-6 and c-8. Labelling is identical to Fig. 1. The neomycin phosphotransferase gene (NmR) and the firefly luciferase gene (Luc) are represented as a dotted and a heavily dotted box, respectively. MT, SV40 and PGK promoter elements are marked as thick arrows. The position of the probe used for Southern blot analysis is marked as a solid box. Relevant restriction sites and PCR primer positions are indicated. (B) PCR analysis using the primer pair C68/luc-int. Aliquots (5 ml of a total volume of 100 ml ) of the PCR reactions were separated on a 1% agarose gel and the DNA visualised with ethidium bromide. The sizes of the DNA fragments in the marker lane (l-DNA digested with HindIII/EcoRI ) are indicated. (C ) Southern blot analysis. Genomic DNA isolated from CRFK-MTloxSV c-6 (20 mg), CRFK-MTloxSV c-6-hyg/luc1 (20 mg) and CRFK-MTSV-hyg/luc cells (5 mg) and 100 pg of the in vitro recombination product pMTSV-hyg/luc were separated on a 1% agarose gel and blotted to a Hybond N+ nylon membrane (Amersham, Braunschweig, Germany). The blot was hybridised with the indicated probe (A) encompassing the Luc gene. (D) Analysis of Luc expression in CRFK-MTloxSV c-6 cells (C-6), CRFK-MTloxSV c-8 cells (C-8) and the cell clones CRFK-MTloxSV c-6-hyg/luc1 (C-6 HL1) and CRFK-MTloxSV c-8-hyg/luc1 (C-8 HL1). Means of triplicate experiments are expressed as relative light units (rlu) per mg of protein.

recombination by using a different recombinase (e.g. Flp). (2) We have shown that two convergent promoterless genes on the floxed targeting fragment can be simultaneously activated by site-specific integration at a genomic target loxP site flanked by correctly oriented promoters. (3) In conclusion, our results expand the versatility and

potential of the Cre recombination system to introduce defined mutations into the eukaryotic genome. The strategy outlined above may be of particular benefit for the generation of transgenic animals expressing pharmaceutically relevant proteins. A loxP site could be integrated into the locus of an abundantly expressed gene (e.g., a milk protein or immunoglobulin gene) by homologous recombina-

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216 (b)

215

(c)

(d)

Fig. 2. (continued )

tion in embryonic stem cells (Smith et al., 1995). The loxP site should be positioned so that it is flanked by an introduced constitutive promoter (analogous to the SV40 promoter used in this study) and the promoter of the endogenous gene (analogous to the MT promoter used here). Subsequently, any foreign gene of interest may be placed under the control of the endogenous promoter by Cre-mediated recombination and appropriate selection. This approach will automatically supply the inserted foreign gene with the appropriate chromatin structure and regulatory elements required for its abundant expression and overcome the difficulties often encountered with transgenes

that are randomly integrated into chromatin structures which are inhibitory to transgene expression.

Acknowledgement We wish to thank Dr. W. Mu¨ller (Institute of Genetics, University of Cologne, Germany) for plasmid pMC1-Cre, Dr. J. Heyer (Institute of Pathology, University of Wu¨rzburg, Germany) for plasmid PGKneo, Dr. Howard Goodman (Harvard Medical School, Massachusetts General Hospital, Boston, MA, USA) for plasmid pXGH and Dr. Greg Winter (MRC, Cambridge, UK ) for plasmid Lys17. Additionally, we

216

A. F. Kolb, S.G. Siddell / Gene 203 (1997) 209–216

Fig. 3. Luc expression of CRFK cells and CRFK cells transfected with pMTSV-hyg/luc (HL) or pMTSV-hyg/luc2 (HL2) as measured in cytoplasmic extracts of cell pools after G418 selection (HL-n; HL2-n) and after additional Hy selection (HL-nh; HL2-nh). Means of triplicate experiments are expressed as relative light units (rlu) per mg of protein.

gratefully acknowledge the technical assistance of Angelien Heister.

References Albert, H., Dale, E.C., Lee, E., Ow, D.W., 1995. Site-specific integration of DNA into wild-type and mutant lox sites placed in the plant genome. Plant J. 7, 649–659. Baubonis, W., Sauer, B., 1993. Genomic targeting with purified Cre recombinase. Nucleic Acids Res. 21, 2025–2029. Bergemann, J., Ku¨hlcke, K., Fehse, B., Ratz, I., Ostertag, W., Lother,

H., 1995. Excision of specific DNA-sequences from integrated retroviral vectors via site-specific recombination. Nucleic Acids Res. 23, 4451–4456. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. Chen, C., Okayama, H., 1987. High efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol. 7, 2745–2752. Fukushige, S., Sauer, B., 1992. Genomic targeting with a positiveselection lox integration vector allows highly reproducible gene expression in mammalian cells. Proc. Natl. Acad. Sci. USA 89, 7905–7909. Gu, H., Zou, Y.R., Rajewsky, K., 1993. Independent control of immunoglobulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting. Cell 73, 1155–1164. Jaenisch, R., 1988. Transgenic animals. Science 240, 1468–1474. Kanegae, Y., Lee, G., Sato, Y., Tanaka, M., Nakai, M., Sakaki, T., Sugano, S., Saito, I., 1995. Efficient gene activation in mammalian cells by using recombinant adenovirus expressing site-specific Cre recombinase. Nucleic Acids Res. 23, 3816–3821. Kolb, A.F., Gunzburg, W.H., Albang, R., Brem, G., Erfle, V., Salmons, B., 1994. Negative regulatory element in the mammary specific whey acidic protein promoter. J. Cell. Biochem. 56, 245–261. Kolb, A.F., Siddell, S.G., 1996. Genomic targeting with an MBP– Cre fusion protein. Gene 183, 53–60. Orlandi, R., Gu¨ssow, D.H., Jones, P.T., Winter, G., 1989. Cloning immunoglobulin domains for expression by the polymerase chain reaction. Proc. Natl. Acad. Sci. USA 86, 3833–3837. Sauer, B., Henderson, N., 1990. Targeted insertion of exogenous DNA into the eukaryotic genome by the Cre recombinase. New Biol. 2, 441–449. Selden, R.F., Howie, K.B., Rowe, M.E., Goodman, H.M., Moore, D.D., 1986. Human growth hormone as a reporter gene in regulation studies employing transient gene expression. Mol. Cell. Biol. 6, 3173–3179. Smith, A.J., De Sousa, M.A., Kwabi-Addo, B., Heppel-Parton, A., Impey, H., Rabbits, P., 1995. A site-directed chromosomal translocation induced in embryonic stem cells by Cre-loxP recombination. Nature Genet. 9, 376–385.