A single vector containing modified cre recombinase and LOX recombination sequences for inducible tissue-specific amplification of gene expression.

The selective alteration of the genome using Cre recombinase to target the rearrangement of genes flanked by LOX recognition sequences has required the use of two separate genetic constructs in trans, one containing cre and the other containing the gene of interest flanked by LOX sites. We have developed a strategy in which both the cre recombinase gene and LOX recombination sites may be cloned within a single vector in cis. This method uses a modified form of Cre (CREM) that contains alterations to the 5' region including the introduction of a Kozak consensus sequence and insertion of a functional intron. This system allows for the inducible, tissue-specific activation or inactivation of gene expression in a single vector and can be utilized for the 300-fold amplification of gene expression from a weak promoter. This approach can be applied to targeting strategies for generating genetically altered mice and gene therapy.


INTRODUCTION
Recent advances in transgenic technology provide a means to fundamentally alter the mouse genome by either overexpressing a gene of interest (1)(2)(3)(4), or deleting or modifying the endogenous allele of a gene in a tissue-specific and temporal manner. (5-7) Several of these approaches utilize technology based upon the Cre/LOX system where the P1 bacteriophage recombinase is used to act on target LOX sites (8) in order to modify a genetic locus or activate transgene expression in a tissue-specific manner. In this system, two distant LOX sites recombine resulting in the excision and loss of intervening nucleotide sequences. This technology has been particularly useful for conditionally altering genetic loci in mice where the expression of the Cre recombinase is under the transcriptional control of a tissue-specific and temporally regulated promoter (9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26). A major limitation of this approach, however, is that the genetically engineered constructs used to achieve these manipulations cannot contain both the sequences for cre recombinase and the LOX sites. Cre will be expressed when present in a plasmid transformed into bacteria which will lead to the recombination of LOX sites contained within the same plasmid and excision of genetic material contained between them. For this reason, two separate lines of transgenic animals must be generated and bred together in order to effect Cre recombination of genes flanked by LOX sites (fLOXed genes). Similarly, the use of Cre-LOX technology for gene therapy applications requires the use of two separate constructs, which makes its use inefficient for such purposes.
Another major limitation of transgenic technology and gene therapy approaches has been that many tissue-specific promoters are transcriptionally weak and lead to relatively low levels of heterologous gene expression. A method to maintain tissue-specificity but amplify levels of transcription of a gene of interest would be highly desirable.
In order to address these limitations, we have developed a novel system in which cre recombinase has been modified (CREM) such that it is not expressed in prokaryotic cells during the cloning process, but is actively expressed in eukaryotic cells. Based upon this strategy, we have generated single vectors containing both the modified cre recombinase and fLOXed gene of interest. In addition, we have developed novel reporter systems to assess Cre activity. By using Cre/LOX strategies in combination with very strong transcriptional promoters, we can achieve high levels of inducible, tissue-specific gene expression using a single vector. These strategies can be applied to manipulating the mouse genome and gene therapy approaches to greatly improve the targeted expression of genes of interest.

Construction of plasmids
Plasmid constructs were generated using standard cloning techniques as described below. For all PCR reactions, the following conditions were used: 10-100 ng of DNA template, 1 mM MgSO 4 , 0.3 mM dNTPs, 0.3 µM primers, and 1.25 U of PFX-polymerase (Life Technologies, Rockville, MD) in a total reaction volume of 50 µl. Thermocycle parameters were: 94°C for 2 min, 50°C for 30 s, 68°C for 1 min for 1 cycle, followed by 94°C for 15 s, 55°C for 30 s and 68°C for 1 min for a total of 34 cycles. *To whom correspondence should be addressed. Tel: +1 301 435 5193; Fax: +1 301 496 8395; Email: jegreen@nih.gov
Generation of plasmids containing a gene switch cassette under the transcriptional regulation of CMV. The StuI/SalI fragment from pSV-EGFP/β-gal containing the floxed EGFP was removed and ligated into the SmaI/SalI sites of of pCMVβ-gal (Clontech) resulting in the generation of pCMV-EGFP/ β-gal (p155) containing the floxed EGFP gene switch cassette driven by the CMV. A second construct utilizing the CMV enhancer/chicken β-actin promoter was also generated. The StuI/SalI fragment containing the floxed EGFP/β-gal from pSV-EGFP/β-gal (p144) was inserted into the XhoI site of pCAGGS (27) resulting in pCMV/βAc-EGFP/β-gal (p169).
Generation of an EGFP-RFP gene switch plasmid. The KpnI fragment from pSV-EGFP/β-gal (p144) containing the floxed EGFP cassette was cloned into the KpnI site of pDSRed1-N1 (Clontech) upstream of the coding sequence for red fluorescent protein resulting in the generation of pCMV-EGFP/RFP (p212) ( Fig. 2A).
Generation of a CMV-enhancer-βAc driven luciferase reporter gene containing a floxed stop-cassette. pCMVe-βAcluc was digested with BglII, blunt-ended, and digested with HindIII. pSV-STOP-luc was digested with StuI/HindIII and the fragment containing the floxed STOP cassette was isolated and cloned into the HindIII site of pCMVe-βAc-luc in which the BglII site had been previously removed by blunt-ending. This resulted in generation of pCMVe-βAc-STOP-luc (p 231) (Fig. 3A).

Generation of modified cre-recombinase
Alteration of the 5′ region and introduction of a Kozak consensus translational start site, LOX sites and intron. Plasmid pBS185 (Life Technologies) containing the CMV enhancer/promoter driving expression of wild-type cre was digested with EcoRI, partially digested with HindIII, bluntended and religated to remove a 1.0 kb fragment containing a KpnI site generating pCMV-cre-del (p205) (Fig. 4A). Plasmid p205 was digested with XhoI and BssHII to remove a 5′ region of cre. A 5′ portion of cre was PCR amplified using OLN C  (Fig. 4A). These modifications resulted in the significant shortening of the 5′ untranslated region from 484 to 19 bp, an optimized Kozak translational start site (28) and the introduction of an EcoRI site at nucleotide 509 and a PstI restriction site through a silent T→G mutation at nucleotide 718 within the coding region of cre to allow for additional alterations of cre. These alterations resulted in the changes in the amino acid sequence of Cre as depicted in Figure 4A. The second amino acid was changed from serine to alanine, the eighth amino acid was changed from histidine to proline and codons for glutamic acid and phenylalanine were inserted at amino acid positions 9 and 10. The human β-globin intron from plasmid pCI (Promega) was altered by PCR mutagenesis to place a PstI site at the 5′ end and a BssHII site at the 3′ end using PCR primers OLN E (5′-GCG-ATC-TGC-AGG-TAA-GTA-TCA-AGG-TTA-CAA-GAC-AGG-3′) and OLN F (5′-ATA-TGC-GCG-CCT-GTG-GAG-AGA-AAG-GCA-AAG-TGG-AT-3′). The 5′ primer E contains a PstI restriction site and the first 35 nt of the β-globin intron (position 953 to 881 in pCI) and the 3′ primer F contains a BssHII restriction site and 26 nt of the 3′ portion of the β-globin intron (position 890 to 965 in pCI). The modified human β-globin intron was cloned into the PstI/BssH II sites of pCMV-cre-K to generate pCMV-CREM (p210) (Fig. 4A).
Generation of a plasmid containing floxed RFP inserted into the coding sequence of cre recombinase. Plasmid pDsRed1-N1 (Clontech) was digested with DraIII. An aliquot of this digest was cut with KpnI and the 3.5 kb DraIII/KpnI fragment was isolated. Another aliquot was digested with AgeI and the 1.

Generation of cre expression vectors using the androgenresponsive, prostate-specific probasin promoter
Generation of probasin driven Cre recombinase. The 428 bp probasin promoter fragment was isolated from prPb-tva950 (kindly provided by Pam Schwartzenberg, NCI) following digestion and blunt-ending of the NdeI site, and digestion with KpnI. This fragment was cloned into the SmaI/KpnI site of pβgal-Basic (Clontech) resulting in pPr-β-gal (p150). pPr-β-gal was cut with ClaI, filled-in, and cut with XhoI in order to remove the coding sequence for β-gal. The MluI (bluntended)-XhoI fragment from pBS185 (Life Technologies) containing the coding sequence for cre recombinase was isolated and inserted into the XhoI/ClaI (blunt-ended) fragment of p150 resulting in the generation of pPr-cre (p153) (Fig. 5A).

Generation of a plasmid containing probasin-modified
Cre flanked by insulators. The vector backbone of plasmid   pPr-CREM (p236) was isolated following digestion with DraIII and partially digested with SalI. The DraIII and BamHI fragment of pPr-CREM (p210) containing the probasin promoter and modified cre recombinase was isolated separately. Plasmid pJC13.1 (kindly provided by G.Felsenfeld, NCI) was digested with SalI and BamHI and the fragment containing the pair of chicken globin insulators (29) was isolated. All three isolated fragments were ligated together to produce pPr-CREM/Ins (p240) (Fig. 5A).
Generation of single vector plasmids containing modified cre and a reporter gene conditionally activated using LOX recombination sites. The DraIII and EcoICRI fragment containing the CMV-STOP-luc cassette was isolated from pCMVe-βAc-STOP-luc (p231) and ligated to the DraIII and Eco47III fragment of plasmid pPr-CREM (p236) containing probasin-modified cre to create plasmid pPr-CREM/CMV-STOP-luc (p252) (Fig. 6A). A second construct was made in which chicken globin insulators (29) were placed between the Pr-CREM and CMV-STOP-luc cassettes. The region containing the probasinmodified cre with insulators was isolated from pPr-CREM/Ins following digestion with DraIII and Eco47III and ligated into the EcoRICI/DraIII fragment of pCMVe-βAc-STOP-luc (p231). This resulted in the generation of a single vector pPr-CREM/Ins/CMV-STOP-luc (p255) containing the modified cre under the transcriptional regulation of probasin separated from the Cre-dependent CMV-STOP-luc cassette by insulators (Fig. 6A).
pCMV-mAR. The cDNA for mouse androgen receptor was modified by PCR mutagenesis to place a BglII site 9 bp upstream of the ATG and the 2.8 kb BglII-PstI fragment was inserted into the BglII-PstI sites of pEGFP-C1 vector (Clontech) after which the EGFP coding sequences were removed by digestion with AgeI and BglII, which were blunt-ended and religated.
pTK-RL. PTK-RL (Promega) uses the TK promoter to direct expression of Renilla reniformis luciferase and was used to normalize for transfection efficiency.

Transfection of cells with DNA
Cells were plated at a density of 1.5 × 10 5 cells/well in 3 ml of culture media 24 h prior to transfection using 6-well plates (Costar, Corning, NY). For hormonal induction studies, PC-3 cells were trypsinized (Life Technologies), washed in culture media and centrifuged at 190 g for 5 min at room temperature. Cells pellets were resuspended in 1× PBS, centrifuged and resuspended in RPMI 1640 without phenol red and supplemented with charcoal/dextran-treated FCS (HyCLONE). On the morning of transfection, 2 ml of fresh media supplemented with 4 nM of mibolerone (NEN, Boston, MS) was added to plates. Transfections were performed 1 h later with a total of 2 µg of plasmid DNA per 6 µl of FuGENE TM 6 (Roche, Indianapolis, IN) in a total of 100 µl of transfection solution containing serum-free culture medium per well. Fresh medium with mibolerone was added daily. Transfections were performed in triplicate and experiments were repeated at least three times. Cells were visualized using a Zeiss-G436 fluorescent microscope. GFP was visualized using a 436 nm excitation, 510 nm emission filter. RFP was visualized using a 546 nm excitation, 580 nm emission filter. Dual fluorescence was observed using a 460-490 m excitation, 510 emission filter.

Luciferase assay
The Dual-Luciferase TM Reporter assay system (Promega) was used to determine activity of firefly luciferase (Photinus pyralisus) and sea pansy Renilla luciferase (R.reniformis) according to the manufacturer's instructions. Cells were washed in 1× PBS and 500 µl passive lysis solution was added per well. Lystate was collected after 15 min and centrifuged in an Eppendorf 5415C centrifuge at 13 000 r.p.m. for 30 s. An aliqout of 20 µl of supernatant was used for luciferase assays.

Statistical analyses
InStat R (GraphPad Software Inc, San Diego, CA) statistical package program was used for statistical analysis.

Development of LOX/frameshift vectors for selective gene expression
The regulation of gene expression using Cre/LOX technology has relied on the use of either a floxed 'STOP cassette' inserted into the 5′ untranslated region of the gene of interest (14,24) or a floxed gene cassette upstream of a second gene to be expressed (12,15,18). In the latter case, the second gene of interest is out of frame for proper translation and therefore not expressed. In the presence of Cre recombinase, the floxed gene or STOP cassette will be removed allowing for expression of the downstream gene of interest. However, due to occasional read-through of transcription and protein translation despite the STOP cassette, there may be detectable background levels of expression of the downstream gene in the absence of Cre (C.Couldrey, unpublished observations).
To address this issue, we designed a novel expression cassette where expression of gene A could be selectively turned off with the simultaneous activation of gene B by placing the floxed gene A within the coding sequence of gene B, whose nucleotide sequence is out of frame to gene A ( Fig. 1A and B). This strategy results in an undetectable background level of expression of gene B in the absence of Cre (Fig. 1C-F). Our expression vector pSV-EGFP/βgal uses the SV40 early region promoter to drive the expression of one of two genes depending upon whether cre is expressed or not. To demonstrate the feasibility of this approach, the coding sequence for EGFP flanked by LOX sites was inserted in frame into the 5′ region of the Escherichia coli gpt gene downstream of the translational start site but out of frame with the coding sequence for the β-gal gene placed downstream (Fig. 1A). Similar constructs were made using the CMV enhancer promoter (pCMV-EGFP/βgal) or the CMV enhance/β-actin promoter (pCMV/βAc-EGFP/βgal). In order to maintain the open reading frame between gtp and EGFP, but place the β-gal gene out of frame, an additional cytosine was added after the LOX I recognition sequence (Fig. 1B). The insertion of this additional cytosine also resulted in the generation of a translational stop codon upstream of the lacZ reading frame within the LOX II recognition sequence (Fig. 1B).
Results of transfection experiments demonstrated that this strategy of gene switching worked as predicted (Fig. 1C-F). Transfection of pSV-EGFP/β-gal alone resulted only in the expression of EGFP, whereas co-transfection with pCMV-cre led to the excision of EGFP and the functional expression of the LacZ gene (Fig. 1C-F). Similar constructs utilizing the CMV enhancer-promoter derived from pCMV β-gal (Clontech) or the CMV enhancer β-Actin promoter from pCAGGS (27) were also generated and gave similar results (data not shown). However, because of the high level of GFP expression under the transcriptional regulation of the CMV promoter/ enhancer, it took longer for all GFP protein to be eliminated compared to that using the weaker SV40 promoter (data not shown). Nonetheless, the gene-switch occurred using the CMV promoter/enhancer. These results demonstrate that the LOX/ frame-shift strategy is effective for the tight regulation of switching gene expression.
Since the assay for β-gal activity cannot easily be performed in vivo, we also developed the reporter construct pCMV-EGFP/RFP for Cre recombinase activity in which the EGFP gene is expressed without Cre, but the RFP gene is expressed following excision of the EGFP gene by Cre (Fig. 2). Both the EGFP and RFP reporter genes can be assayed in vivo.

Amplification of gene expression using alternate promoters in combination with Cre/LOX
Although tissue-specific promoters are expressed in a limited range of cell types, their level of transcription may often be relatively low and limiting (30). In order to increase the levels of expression but maintain tissue-specificity, we developed a strategy which utilized the tissue-specific expression of Cre in combination with the gene-switch approach as well as a gene activation approach under the transcriptional control of a strong, ubiquitously expressed promoter. The 428 bp tissuespecific, hormone-responsive probasin promoter was used to drive the expression of Cre (pPr-cre). The gene-switch expression vectors (pSV-EGFP/β-gal (Fig. 1B), pCMV-EGFP/β-gal and pCMVe/βAc-EGFP/β-gal direct the expression of a specific reporter gene only in the presence of Cre. This strategy maintains the tissue-specific expression of the desired reporter gene since Cre is produced using the tissue-specific probasin promoter, but greatly amplifies the expression of the desired reporter gene, which is under the transcriptional control of the CMV promoter.
In order to quantitatively assess the level of amplification by switching transcriptional regulation from the probasin promoter to the CMV enhancer/β-actin promoter, a series of luciferase reporter constructs were generated (Fig. 3A). PC-3 cells were co-transfected with pCMV-mAR to express the murine androgen receptor and the reporter construct probasinluciferase (Pr-luc) (Fig. 3A) in the presence of the androgen analog mibolerone (Fig. 3B). This was compared to the levels of luciferase activity generated by the second reporter construct pCMV-STOP-luc in the presence or absence of the Cre producing plasmid probasin-cre. The results of these experiments demonstrated that luciferase activity was amplified at least 15-fold when the CMV-STOP-luc was activated in the presence of Pr-cre compared to the level of luciferase activity generated directly under the transcriptional regulation of the probasin promoter by pPr-luc (Fig. 3B).

Modification of cre recombinase for eukaryotic but not prokaryotic expression
In order to clone in bacteria a plasmid construct containing both the coding sequence for cre recombinase and the LOX target sequences, we designed modifications to the cre recombinase that prevented its translation in bacteria, but allowed for its translation in eukaryotic cells. This required the modification of the 5′ region of cre and insertion of an artificial intron within the cre coding sequence (Fig. 4A). Using mutagenesis strategies, the 5′ untranslated region of cre was removed, the translational start site was optimized for eukaryotic translation by using a Kozak consensus sequence (28) and unique KpnI and EcoRI restriction sites were introduced downstream of the ATG. As further described in the Materials and Methods, an artificial intron based upon a β-actin intron was inserted into a newly generated restriction site within the cre gene to create the plasmid pCMV-CREM (Fig. 4A).
To determine whether recombinase activity was maintained following these extensive modifications to cre, pCMV-CREM was co-transfected with the reporter plasmid pCMV-EGFP/ RFP. As demonstrated in Figure 4B-G, the functional Cre recombinase activity was maintained with these alterations of cre.

Addition of a second cre amplification step further enhances reporter gene expression
In order to further amplify the expression of the reporter gene, we developed an additional construct in which cre expression would also be conditionally amplified using a strong promoter. This would occur following the initial expression of Cre by the relatively weak, tissue-specific and hormone-inducible probasin promoter using the plasmid Pr-cre (Fig. 5A). An intermediate construct pCMV-CREM-L was first generated by placing a LOX 511 sequence (31) into the 5′ region of cre in pCMV-CREM. Cre activity was maintained despite the presence of the LOX 511 sequence within the cre coding region as demonstrated in co-transfection experiments with pCMV-EGFP/RFP (data not shown). The generation of the construct pCMV-RFP-CREM (p224) is depicted in Figure 4A. This construct uses the same frame-shift principles as those described previously for the EGFP/β-gal reporter constructs except that the LOX 511 target sequences flank the RFP gene. LOX 511 sequences were chosen to avoid potential recombination with LOX P sequences used in the other constructs. Without the expression of Cre, RFP is expressed and the modified cre is out of frame and not expressed. In the presence of Cre, RFP is excised and the modified cre is correctly translated.
Transfection experiments using PC-3 cells demonstrated that the presence of the second conditionally expressed modified cre-containing plasmid pCMV-RFP/CREM resulted in a 7-fold increase in luciferase activity compared to transfections in the absence of the second conditionally expressed Cre (Fig. 5B). In order to more clearly demonstrate the effect of Cre amplification by pCMV-RFP-CREM on luciferase expression using pCMV-STOP-luc, transfections were performed over several time points using varying amounts of Pr-cre (from 0.01 to 0.001 pmol) alone or with the additional inducible Cre recombinase using pCMV-RFP-CREM (Fig. 5C). A 32-fold increase in luciferase activity was observed using the gene expression amplification approach (Pr-cre and CMV-STOPluc) compared to an equimolar amount of Pr-luc alone (Fig. 5C, 3 versus 1). The addition of pCMV-RFP-CREM further increased the level of luciferase activity by 9-fold (Fig. 5C, 4 versus 3). Thus, there was an ∼300-fold amplification of luciferase activity using the gene amplification approach with the second conditional cre construct compared to luciferase activity from Pr-cre alone. These results strongly suggest that the addition of a second conditionally active Cre could significantly increase the effective expression from other weak promoters.

Generation of a single vector containing a modified cre regulated by an inducible, tissue-specific promoter and conditionally expressed reporter gene
Having established that the modified cre maintained full activity for recombination of LOX target sequences, the modified cre, placed under the transcriptional control of the tissuespecific, androgen-inducible probasin promoter, was placed in cis with the expression reporter construct to generate pPr-CREM/CMV-STOP-luc (Fig. 6A). In order to prevent the strong enhancer activity of the CMV regulatory region from activating the upstream probasin promoter, an additional construct was made by placing a set of chicken globin insulators (pPr-CREM/Ins/CMV-STOP-luc). Both vectors were assayed in transient transfection experiments to determine whether the expression of the reporter genes was inducible by mibolerone and whether the insulators reduced potential background activity that might otherwise occur through the action of the CMV enhancer on the probasin promoter (Fig. 6B).
In order to determine whether chicken globin insulators might alter the level of expression from Pr-cre, insulators were added to flank Pr-cre to generate pPr-CREM/Ins (Fig. 5A) and co-transfected with pCMV-STOP-luc. No difference in luciferase activity was observed whether insulators did or did not flank Pr-cre (Fig. 6B, columns 4 and 6; P = 0.161, Mann-Whitney test). When the single construct containing pPr-CREM/CMV-STOP-luc was transfected in the absence of mibolerone, a significant basal level of luciferase activity was observed compared to co-transfection of the separate constructs Pr-CREM and CMV-STOP-luc (Fig. 6B, columns 3  and 7). This was presumably due to the strong transcriptional activating effect of the CMV enhancer on the probasin promoter. However, when the set of insulators was placed between the Pr-CREM and CMV-STOP-luc cassettes (pPr-CREM/Ins/CMV-STOP-luc), the uninduced level of luciferase activity was reduced ∼5-fold compared to the single construct without insulators (Fig. 6B, columns 7 and 9; P = 0.0001, Mann-Whitney test). The levels of mibolerone-induced luciferase activities were similar whether or not the insulators were present (Fig. 6B, columns 8 and 10; P = 0.297, Mann-Whitney test). No significant difference in inducible luciferase activities were observed using the single Pr-CREM/CMV-STOP-luc or Pr-CREM/Ins/CMV-STOP-luc vectors comparing to the cotransfection of the individual constructs (Fig. 6B, columns 4, 8 and 10; P = 0.681, Kruskal-Wallis test). These results demonstrate that the single vector containing the modified cre under an inducible promoter and the reporter cassette in cis worked equally as well as separate constructs in trans.

DISCUSSION
The results of this work demonstrate several novel modifications of Cre-LOX technology that can be applied to both human gene therapy and transgenic animal manipulations including the Cre-inducible activation or switching of gene expression, tissue-specific gene expression amplification and the development of single vectors containing both a modified cre recombinase and LOX recognition sequences.
Although the tissue-specific expression of Cre activation of gene transcription has been previously reported (11,20,(24)(25)(26) these approaches have generally relied upon the use of a STOP cassette to prevent transcriptional read-through flanked by LOX recombination sites (14,24). This method is adequate for certain experimental approaches such as β-gal expression, but is greatly limited by the fact that the STOP cassette does not absolutely prevent transcriptional read-through (C.Couldrey, unpublished data). Thus, highly transforming genes or genes whose biological activity only require very low levels of expression may still be transcribed despite the presence of the STOP cassette leading to unwanted biological activities. In order to overcome this technical limitation, we developed a strategy in which the switch in expression from one gene to another relied upon the design of altering the translational reading frame upon recombination of LOX sites in the presence of Cre (Fig. 1A). We have demonstrated the utility of this approach using several reporter genes including lacZ, GFP, RFP and luciferase.
Although numerous tissue-specific expression vectors exist, levels of transcriptional activity are often much lower than is desired to effect a biological response when applied to transgenic or gene therapy approaches. In order to overcome this technical limitation, we have demonstrated a method in which the level of gene expression from a relatively weak but tissuespecific promoter can be amplified through the Cre-induced transcriptional activation using a very strong promoter. Using this technical approach, we have shown that the level of transcription using the prostate-specific probasin promoter can be amplified 30-fold. When the use of a second, conditionally expressed Cre was employed, the amplification of the reporter increased an additional 10-fold resulting in a total amplification of reporter gene expression of 300-fold. This method can readily be applied to other tissue-specific but relatively weak promoters in transgenic applications and gene therapy approaches when high levels of expression are warranted. However, depending upon the experimental questions, low levels of tissue-specific expression of a particular gene may be more appropriate using existing approaches.
A limiting feature of Cre-LOX technology has been the fact that it has not been possible to develop single constructs in which an active Cre recombinase could be contained in cis with a pair of LOX recombination sites since the LOX sites would be combined when grown in bacteria. This has necessitated the use of two separate genetic constructs in which cre and LOX sites exist in trans. This requires the generation of two separate lines of transgenic animals or the less efficient introduction of two separate constructs in in vitro or gene therapy strategies. We have demonstrated that our modifications to cre recombinase allow for the placement of a modified cre in the same construct containing LOX recombination sites. The addition of chicken globin insulators reduces potential enhancer effects from a strong transcriptional regulatory unit on the relatively weak but tissue-specific promoter driving the expression of Cre. Although background levels of expression remain somewhat elevated, the addition of flanking insulators may overcome this problem.
The strategies and modifications to cre recombinase presented in this work will have broad applications in developing tissuespecific inducible gene expression systems.