Segmental genomic replacement by Cre-mediated recombination: genotoxic stress activation of the p53 promoter in single-copy transformants
Segmental genomic replacement by Cre-mediated recombination: genotoxic stress activation of the p53 promoter in single-copy transformantsBruce Bethke+ and Brian Sauer*
National Institutes of Health, National Institute of Diabetes, Digestive and Kidney Disease, Bethesda,MD 2089-1800, USA
Received March 27, 1997;Revised and Accepted May 28, 1997
ABSTRACT
Genotoxic stress results in transcriptional activation of the p53 promoter. To gain more detailed information on genotoxic induction of the p53 promoter at a uniform genomic locus, we have developed an efficient strategy for replacing a defined genomic segment in mouse NIH 3T3 cells with exogenous transfected DNA using a `doublelox' targeting strategy mediated by Cre DNA recombinase. The strategy utilizes a pair of heterospecific lox sites engineered both into the genome and onto the targeting DNA. This allows direct replacement of genomic DNA by a Cre-catalyzed double crossover event. p53-CAT reporter constructs were site-specifically placed into the genomic target 20-fold more efficiently by doublelox recombination than by Cre-mediated single crossover insertional recombination, and the absolute frequency of site-specific doublelox targeting exceeded the frequency of transformation due to random illegitimate recombination of transfected DNA into the genome. Resulting targeted single-copy integrants of the p53-CAT reporter show strong genotoxic induction by mitomycin C, and a dynamic range of induction that exceeds that seen in transient transfection assays. The doublelox strategy is generally applicable to Cre-mediated genomic targeting in any cell and should be of particular utility in the site-specific targeting of DNA into embryonic stem (ES) cells for the production of gene-modified mice.
INTRODUCTION
Extreme variability of transgene expression is often observed between individual stable transformants of mammalian cells after gene transfer (1 ). This is due both to variation in copy number of the resulting integrated DNA and also to chromosome position effects on gene expression. For these reasons, and because the procedure is comparatively rapid, identification and characterization of important DNA elements controlling promoter function have been most conveniently achieved using transient transfection of mammalian cells and analysis. In certain situations, however, stable transformants are either necessary or would allow more in-depth analysis of gene expression. For example, a comprehensive understanding of inducible gene expression includes an appreciation for the remodeling of chromatin structure associated with induction (2 ,3 ). A second important example is the characterization of developmentally regulated, tissue-specific, or otherwise inducible promoters using transgenic animals, where incorporation of the transgene into the genome is a prerequisite for analysis.
Reproducible gene expression in stable mammalian cell transformants can be achieved by Cre-mediated site-specific integration of transfected DNA into a chromosomally positioned loxP site, the 34 bp recognition site for Cre DNA recombinase (4 ). Conservative site-specific DNA recombination by Cre results in the incorporation of a single copy of the entire circular loxP-containing targeting plasmid into the cellular genome by a precise single crossover event between the chromosomal and plasmid-borne loxP sites (5 -7 ). Although precise, the frequency of Cre-mediated targeting events is often less than the frequency of random illegitimate incorporation of DNA into the genome and may vary depending on the chromosomal position of the target loxP site (8 ).
We describe here an enhanced Cre-mediated site-specific targeting strategy and demonstrate its use to examine p53 induction by genotoxic agents. The p53 gene is one of the most commonly mutated genes in human tumors and plays an important role in normal cells in sensing DNA damage and regulating cell cycle progression. Exposure of non-transformed cells to DNA-damaging agents induces elevated levels of p53 protein (9 -11 ), extends the protein half-life (12 ) and results in enhanced nuclear accumulation (13 ). In addition, the p53 promoter contains a distinct DNA element that mediates transcriptional induction in response to genotoxic stress (14 ), as shown by transient transfection expression assays. We show here that stable cell lines with single-copy site-specific targeted genomic integration of a p53 promoter-CAT reporter construct show strong induction of the reporter gene after genotoxic stress.
Our modified site-specific integration strategy is based on the observation that certain nucleotide changes in the spacer region of the lox site can alter the site selection specificity of Cre-mediated recombination with little or no effect on the efficiency of recombination (15 ). Because Cre can catalyze recombination independently at two such different (heterospecific) lox sites in eukaryotic cells (16 ), we have incorporated a second heterospecific lox site into both the targeting vector and the chromosomal target to allow genomic targeting by a double reciprocal crossover replacement event. We show that this doublelox strategy (i) greatly increases the recovery of site-specific targeted integrants; (ii) allows the rapid one-step precise replacement of a defined chromosomal segment with sequences from exogenous transfected DNA; and (iii) eliminates unnecessary incorporation of vector DNA sequences into the genome during targeting.
MATERIALS AND METHODS
Plasmids and DNA manipulations
Plasmids were constructed and prepared by standard techniques (17 ). The CMV-cre expression plasmid pBS185 (5 ), the fusion neo target plasmid pSF1 (6 ) and the lox2STOP plasmid pBS302 (18 ) have been described previously.
The plasmid pBS428 was used to generate the chromosomal target in cell line B-13 shown in Figure 1 . From left to right, plasmid pBS428 carries the PGK promoter (PstI fragment) from pPNT (19 ), an ATG translational start in-frame with a following loxP-neo fusion gene, the STOP region from pBS302 followed by an ATG-less loxP-lacZ-BGH polyadenylation fusion gene, derived from pMC1871 (Pharmacia) and pRc/CMV (Invitrogen), and the lox511 site. The loxP sites of the loxP-neo and loxP-lacZ fusion genes share the same reading frame so that excision of intervening DNA by Cre will result in activation of the lacZ gene.
The double crossover vector pBS397 (see Fig. 2 ) carries the STOP signal from pBS302, a synthetic ATG-loxP site (Fig. 1 ) fused to the neo gene from pSF1, followed by a synthetic polylinker: EcoRI-EcoRV-HindIII-lox511-NheI-NotI (underlined sites are unique). Deletion of the lox511 site between the HindIII and NheI sites of pBS397 yielded the single crossover vector pBS398.
Cell culture and cell line construction
Mouse NIH 3T3 cells (American Type Culture Collection, Rockville, MD) were maintained in DMEM + 10% calf serum (Life Technologies, Inc., Gaithersburg, MD). To construct cell line B-13 (Fig. 1 ), 3T3 cells were electroporated with the 7.2 kb NotI fragment of pBS428 carrying PGK loxP2neo-lacZ-lox511. Candidate G418R colonies were screened to be [beta]-gal- by histochemical staining of cells with X-gal (21 ) and low copy integrants were identified by Southern analysis. One of these, B-13, was chosen as a starting point for the generation of a Neo- derivative. B-13 cells were transfected with the cre expression plasmid pBS185 (to allow transient expression of Cre) and then plated nonselectively. Resulting colonies were screened to be G418S and [beta]-gal+, the predicted phenotype of cells that have undergone Cre-mediated excision of the neo gene. Such colonies were obtained, as expected, at a frequency of ~15%. This frequency corresponds to the DNA transfection efficiency observed in a parallel experiment with a green fluorescent protein (GFP) reporter plasmid (see below). One of the G418S derivatives, 13-1, was analyzed in greater detail. This cell line showed light blue X-gal staining in situ whereas the parental B13 cell line was white. Southern analysis with a variety of enzymes confirmed that Cre-mediated recombination had excised the neo gene, leaving a single copy of the reconstructed lacZ reporter with the structure shown in Figure 1 . Cell line 13-1 was therefore used for subsequent Cre-mediated targeting.
Genomic DNA from individual cell lines was analyzed by Southern blotting (22 ) using either a 1 kb neo structural gene probe or a 2 kb lacZ structural gene probe. Detection of hybridization was with a Fuji BAS 1500 PhosphorImager.
Cre-mediated targeting
Cell line 13-1 cells (5 * 105 cells in a 10 cm dish) was transfected overnight with 20 [mu]g targeting DNA (either pBS398-p53cat or pBS397-p53cat) and with the indicated amount of the cre vector pBS185 using calcium phosphate (23 ). Transformants were selected 48 h later by replating cells in 400 [mu]g G418/ml and colonies were scored after 12 days. Illegitimate recombination was measured by transfection of cells with 20 [mu]g pBS428 in the absence of co-transfecting cre vector. Transfection efficiency was monitored by transfection with 20 [mu]g pBS377, a GFP reporter construct (S. Gagneten and B.S., unpublished), and examination of cells 48 h later with a Nikon Optiphot-2 microscope equipped for FITC epiflourescence. In all cases, pBluescript II KS- DNA (Stratagene) was used to bring the total amount of DNA per transfection to 30 [mu]g.
CAT assay of gene expression
For transient expression assays, 3T3 cells were transfected as above with 20 [mu]g of the indicated DNAs. One day later, cells were induced for 24 h with 15 [mu]g/ml mitomycin C in serum-free media (14 ) and analyzed for CAT expression by thin layer chromatography (17 ,24 ). CAT activity was quantitated using a Fuji BAS 1500 PhosphorImager with multiple exposures to ensure a linear relationship between the signal and image intensity. The BCA assay (Pierce Chemical Co.) was used for protein determination.
RESULTS
Strategy for replacement recombination
Cre-mediated DNA recombination can direct insertion of an intact circular targeting DNA molecule into a loxP target prepositioned in the mammalian genome by a single crossover recombination event between a loxP site on the targeting DNA and the chromosomal loxP site (5 ). Because the resulting integrant contains two directly repeated loxP sites flanking the inserted DNA (Fig. 2 ), stable integation is attained by allowing only a short burst of Cre expression.
One drawback of Cre-mediated single crossover integration is that vector sequences are necessarily integrated in addition to the desired transgene. To circumvent this limitation, we designed a chromosomal target that would allow integration of DNA by a Cre-mediated double crossover event (Fig. 1 ). Cell line 13-1 carries two lox sites: a loxP site fused in frame with the lacZ gene, and a lox511 site 3" to the lacZ gene. Because Cre does not catalyze recombination between two different heterospecific lox sites (15 ,16 ), no recombination can occur between lox511 and loxP, although there is no barrier to loxP * loxP or lox511 * lox511 recombination. Thus, a double crossover event between the chromosomal target and an exogenous targeting plasmid carrying both loxP and lox511 would permit a simple replacement of a chromosomal gene for one on the targeting DNA (Fig. 2 ), with no integration of vector DNA. We designed the double-lox targeting vector to carry both a lox511 site and a defective loxP-neo fusion gene (lacking both an ATG start and a promoter), thus permitting direct selection for Cre-mediated replacement. Correct targeting results in reconstruction of a functional loxP-neo fusion gene and concommitant resistance to G418.
Site-specific DNA integration at the genomic target
To evaluate Cre-mediated chromosomal replacement events, we constructed the doublelox plasmid pBS397-p53cat (Fig. 2 ) to serve as the targeting vector. Double crossover events catalyzed by Cre between the pairs of chromosomal and targeting vector lox sites is designed to replace the chromosomal loxP-lacZ gene in cell line 13-1 with a loxP-neo fusion gene. To further assess the utility of Cre-directed chromosomal placement of a transgene, an accompanying p53 promoter-CAT reporter gene was included on the targeting vector between the heterospecific lox sites. Only Cre-mediated integration into the genome efficiently generates G418R colonies since the pBS397-p53cat construct lacks both an ATG start and a promoter for neo expression. To further minimize the occurrence of adventitious non-site-specific transformants in which the neo gene might become activated by integration into an endogenous gene, we also included an expression STOP (18 ) signal upstream of the loxP-neo gene. In order to compare the efficiency of the replacement reaction with the single crossover Cre-mediated insertion strategy, a nearly identical plasmid was constructed, pBS398-p53cat, that differs from pBS397-p53cat only in that it lacks the lox511 site and, hence, can only insert into the genome by a single crossover event at the resident loxP site.
DNA analysis of targeting events
Southern analysis confirmed that Cre had correctly integrated the loxP-neo fusion gene and the accompanying p53-CAT reporter into the target locus. Southern blotting rather than PCR was used for analysis to allow detection of rare double insertion events (5 ) or of illegitimate DNA rearrangements that have occasionally been associated with Cre-mediated targeting in mammalian cells (7 ). As shown in Figure 2 , replacement recombination predicts that lacZ sequences are evicted from the genome and replaced with neo sequences, whereas insertion disrupts but does not supplant lacZ. A 4.5 kb SpeI-NheI band is diagnostic for correct Cre-mediated incorporation of neo sequences at the target site. Figure 4 shows Southern analysis of six representative transformants derived from each of the two targeting strategies. As expected, none of the doublelox replacement transformants retained the lacZ gene, whereas the insertion transformants exhibited a disrupted lacZ band of the predicted size. Both the replacement and insertion Cre-mediated transformants exhibited the correctly sized PGK-neo band at 4.5 kb. These results show that Cre precisely targeted the chromosomal target in cell line 13-1 to either insert the entire targeting vector or to instead replace a designated chromosomal gene segment with another carried on the targeting vector, depending simply on the choice of the targeting DNA vector.
Genotoxic induction of the p53 promoter
The p53 promoter contains a DNA element that is important in responding to DNA damage by mediating increased p53 transcription (14 ). We confirmed that the p53 promoter fragment responds to genotoxic induction in 3T3 cells by monitoring CAT expression after transient transfection (Table 1 ). Treatment of transfected cells with mitomycin C resulted in a 5-fold induction of CAT activity, similar to previously reported results. Note that even when induced, the p53 promoter fragment is far less powerful than the uninduced RSV promoter.
We next examined the expression of a single copy integrant of the p53-CAT construct after Cre-mediated site-specific targeting into the 3T3 cell genome. Two representative insertion integrants (pBS398-p53cat transformants I-1 asnd I-2) and two chromosomal replacement integrants (pBS397-p53cat transformants R-1 and R-2) were assayed for CAT activity with and without genotoxic induction (Fig. 5 ). In the absence of induction, all four cell lines exhibited a low, but detectable, amount of CAT activity compared with the parental 13-1 cell line. Upon treatment with mitomycin C, both of the insertion transformants showed an ~12-fold induction of CAT activity. Similarly, both of the replacement integrants displayed a >20-fold induction. These results demonstrate that the p53 promoter fragment responds appropriately to genotoxic induction when present at single copy in the genome. Moreover, they suggest that greater dynamic range of the induction response can be attained by single copy placement of the reporter into the genome compared with transient expression assays. Curiously, the plasmid ori and ApR sequences (present only in the insertion integrants) may have a slight inhibitory effect on the induction response seen here. Further work will be required to determine how general this effect is on reporter gene expression.
. Genotoxic induction of the p53 promoter during transient expression
Promoter
CAT activity (% acetylation)
No induction
+ Mitomycin C
p53 (ref. 20)
2.1
10.3
RSV (ref. 31)
75
n. d.
DISCUSSION
We show here the efficient targeted replacement of a defined chromosomal segment in the mammalian genome by a heterologous transgene using the Cre site-specific DNA recombinase. Targeting depends on a doublelox recombination strategy with heterospecific lox sites flanking the chromosomal segment targeted for replacement. In the experiments presented here, the doublelox strategy is at least 20-fold more efficient than the highest level attainable by Cre-mediated single crossover insertional recombination and exceeds the frequency of random, illegitimate recombination events. Moreover, this difference in efficiency can not be attributable to differences in chromosomal target site availability (8 ), as integration was designed in both strategies to occur at the same chromosomal target locus. In additional experiments, not shown here, we have routinely achieved integration frequencies of 0.5-1.2% with the doublelox vector carrying a variety of different inserts. Cre-mediated doublelox integration results in simple replacement of a defined chromosomal segment with a single, unrearranged copy of the desired DNA fragment from the targeting vector. In particular, we have not observed anomalous rearrangements of insert DNA, as has occasionally been obtained with Cre-mediated insertion using vectors having a single loxP site (7 ).
Upon reflection, it is not that surprising that doublelox integration is more efficient than with single loxP vectors. In both strategies Cre can mediate a single crossover event that integrates a circular plasmid into the genome to give an integrant containing two directly repeated loxP sites. Such a structure is a good substrate for excisive recombination and, as is the case using a single loxP-containing vector, allows efficient reversal of the integration reaction by a second round of Cre-mediated recombination. Stable insertion using a single loxP vector can only occur in the few percent of cells in which insertion into the genome has been trapped after the level of Cre recombinase in the cell has fallen below the amount required for recombination. In the case of the doublelox strategy, however, after such a single crossover event there is a choice: excisive recombination can occur either between the pair of directly repeated loxP sites or between the pair of directly repeated lox511 sites. Hence, although there is a 50% chance that insertion is reversed by excisive recombination, there is also a 50% chance that the other pair of lox sites are used, thus formally giving replacement of the chromosomal segment with that on the targeting plasmid. Alternatively, double reciprocal crossover recombination may proceed relatively directly. In both cases the result is the same: the original chromosomal segment is evicted from the chromosome and must compete with a vast excess of targeting DNA to reverse the incorporation of targeting DNA into the genome. This scenario underscores the likelihood that doublelox recombination is reciprocal, that is, not only is the targeting fragment transferred to the chromosome, but the chromosomal fragement is very likely transferred intact to the targeting vector. Because this may provide the basis for an attractive strategy for recovering predefined DNA segments from the genome for subsequent analysis, we are investigating this possibility further.
The most striking feature of Cre-mediated replacement, however, is the much reduced requirement for Cre recombinase compared with integration by single crossover insertion. Since equivalent high levels of recombinase are not required for replacement recombination, it is unlikely that high levels of Cre are inherently necessary for efficient recombination between plasmid DNA and the chromosome. Although we do not completely understand this observation, we suspect it indicates that efficient insertion (single crossover) recombination must require that Cre persist for a substantial amount of time in the cell. Such a requirement for greater Cre persistance for insertional recombination compared with replacement recombination could result if there is a reduced probability of productive insertional integration early after transfection compared with later. Exogenous naked DNA is subject to rearrangement and concatemerization after introduction into cells: Cre-mediated recombination of a linear concatemer with the chromosomal target would be expected to be nonproductive in generating a stable transformant, or could lead to aberrant DNA rearrangements. Later after transfection, high levels of Cre would act to resolve extrachromosomal linear concatemers into monomeric circles that could then productively insert into the chromosomal target. However, double reciprocal crossover recombination between incoming DNA and the chromosomal target, which is not dependent on circularity of the targeting DNA, would avoid such complications immediately after transfection. Alternatively, the conversion of naked DNA into a chromatin structure may act to slow excisive recombination and thus increase the probability of productive insertional recombination later after transfection.
Using Cre-mediated replacement recombination we demonstrate genotoxic stress induced gene expression from a single-copy chromosomal human p53 promoter reporter construct. A variety of transcription factor binding sites lie within the p53 promoter region, including a stress-responsive element that binds NF-[kappa]B (25 ) and also a distinct UV-stress response element (14 ). Because precise base-specific changes at these and other DNA elements within the p53 promoter can be rapidly installed directly into the genome by doublelox targeting, we anticipate that this strategy will permit a more detailed understanding of these elements' contribution to genotoxic induction at a uniform single-copy genomic locus. This will especially be useful in the examination of the role of chromatin remodeling during induction. The greater dynamic range that we observed with the chromosomal p53-CAT replacement transformants compared with transient expression may reflect a chromatin component to induction. Alternatively, transfected DNA itself may stimulate a slight gentoxic induction response, or factors affecting induction could be titrated by the multiple copies per cell of reporter plasmid necessarily present in transient expression experiments. Interestingly, the slight decrease in induction, compared with replacements, that we observed in the p53-CAT whole plasmid insertion transformants may be related to prior observations that prokaryotic vector sequences can adversely affect the expression of eukaryotic genes (26 ).
The facile genomic targeting by Cre-mediated replacement recombination will be of particular utility in embryonic stem (ES) cells. Because ES cells can contribute to the mouse germ line after injection into blastocysts, `subtle' mutations generated in ES cells can be evaluated in the resulting gene modified mice (27 ,28 ). After installation of heterospecific lox sites at a desired locus by homologous recombination, DNA could be quickly shuttled into that locus by Cre-mediated double lox replacement recombination, permitting rapid allelic exchange, reducing the time required for ES cell manipulation in vitro, and thus expediting the appraisement of targeted point mutations in the developing mouse.
ACKNOWLEDGEMENTS
We are grateful to L. Crawford and S. Tuck for the gift of pRXBCAT, and to M. Brennan for thoughtful comments on the manuscript.
REFERENCES
1 Lacy, E., Roberts, S., Evans, E. P., Burtenshaw, M. D. and Constantini, F. D. (1983) Cell, 34,343-358.MEDLINE Abstract
19 Tybulewicz, V. L., Crawford, C. E., Jackson, P. K., Bronson, R. T. and Mulligan, R. C. (1991) Cell, 65,1153-1163.
20 Tuck, S. P. and Crawford, L. (1989) Mol. Cell. Biol., 9,2163-2172.
21 Sanes, J. R., Rubenstein, J. L. R. and Nicolas, J.-F. (1986) EMBO. J., 5,3133-3142.
22 Nakamura, N., Sugino, H., Takahara, K., Fukushige, S. and Matsubara, K. (1991) Cytogenet. Cell Genet., 57,18-22.MEDLINE Abstract
23 Chen, C. and Okayama, H. (1987) Mol. Cell. Biol., 7,2745-2752.MEDLINE Abstract
24 Heard, J. M., Herbomel, P., Ott, M. O., Mottura-Rollier, A., Weiss, M. and Yaniv, M. (1987) Mol. Cell. Biol., 7,2425-2434.
25 Wu, H. and Lozano, G. (1994) J. Biol. Chem., 269,20067-20074.MEDLINE Abstract
26 Simonsen, C. C., Shepard, H. M., Gray, P. W., Leung, D. W., Pennica, D., Yelverton, E., Derynck, R., Sherwood, P. J., Levinson, A. D. and Goeddel, D. V. (1982) UCLA Symp. Mol. Cell. Biol., 25,1-6.
27 Valancius, V. and Smithies, O. (1991) Mol. Cell. Biol., 11,1402-1408.MEDLINE Abstract
28 Hasty, P., Ramirez-Solis, R. and Bradley, A. (1991) Nature, 350,243-246.MEDLINE Abstract
29 Lakso, M., Sauer, B., Mosinger, J., B., Lee, E. J., Manning, R. W., Yu, S.-H., Mulder, K. L. and Westphal, H. (1992) Proc. Natl. Acad. Sci. USA, 89,6232-6236.MEDLINE Abstract
30 Sauer, B., Whealy, M., Robbins, A. and Enquist, L. (1987) Proc. Natl. Acad. Sci. USA, 84,9108-9112.MEDLINE Abstract
31 Gorman, C. M., Moffat, L. F. and Howard, B. H. (1982) Mol. Cell. Biol., 2,1041-1051.
*To whom correspondence should be addressed. Tel: +1 301 402 4567; Fax: +1 301 496 0839; Email: sauerb@helix.nih.gov
+Present address: Department of Biology, St Vincent College, Latrobe, PA 15650, USA
