FLP-mediated DNA mobilization to specific target sites in Drosophila chromosomes
FLP-mediated DNA mobilization to specific target sites in Drosophila chromosomesMary M. Golic, Yikang S. Rong, Robert B. Petersen1, S. L. Lindquist2 and Kent G. Golic*
Department of Biology, University of Utah, Salt Lake City, UT 84112, USA, 1Institute of Pathology, Case Western Reserve University, Cleveland, OH 44106, USA and 2Howard Hughes Medical Institute, University of Chicago, Chicago, IL 60637, USA
Received May 15, 1997;Revised and Accepted July 22, 1997
ABSTRACT
The ability to place a series of gene constructs at a specific site in the genome opens new possibilities for the experimental examination of gene expression and chromosomal position effects. We report that the FLP-FRT site-specific recombination system of the yeast 2[mu] plasmid can be used to integrate DNA at a chromosomal FRT target site in Drosophila. The technique we used was to first integrate an FRT-flanked gene by standard P element-mediated transformation. FLP was then used to excise the FRT-flanked donor DNA and screen for FLP-mediated re-integration at an FRT target at a different chromosome location. Such events were recovered from up to 5% of the crosses used to screen for mobilization and are easily detectable by altered linkage of a white reporter gene or by the generation of a white+ gene upon integration.
INTRODUCTION
Current methods for transformation of Drosophila utilize transposable elements as vectors to carry DNA into the genome. Because these elements transpose into essentially random locations, transformed genes end up scattered throughout the genome. Their expression is subject to the influences of their immediate chromosomal environments. These position effects on gene expression are quite commonly observed and in some experiments are a serious nuisance. For instance, if a series of genetically engineered gene alterations are to be compared, position effects on the integrated constructs complicate the interpretation of results. The typical solution is to assay a number of independently derived transformants of each construct and average the results (1 ). If a single transformant of each construct could be analyzed the amount of work could be significantly reduced. The degree of confidence in such averaged results is also diminished, because a single extreme position effect can strongly influence the average.
One solution to this problem is to target the integration of all constructs to a single site, where they would all be subject to the same position effect. Conclusions derived from such comparisons are more compelling because a major source of variability is eliminated. This can be achieved in yeast and mice by using homologous recombination to place DNA at a specific site (2 ,3 ). In plants, where homologous recombination techniques have not been developed, site-specific recombination has been used to integrate DNAs at a target site for a recombinase that was previously integrated into the genome (4 ).
DNA placement by homologous recombination and by site-specific recombination have so far only proven useful in systems where chemical selection can be applied to single cells. With both techniques integration events are relatively rare (typically 10-4-10-6). Drosophila does not readily lend itself to the type of chemical selection that would be required to recover such rare events. No method of cell culture has been reported that would allow chemical selection on isolated cells with subsequent return of the selected cells to the germline of an intact animal. Transformants are obtained in Drosophila by injecting DNA into very young embryos, mating the adults that survive this procedure and screening their progeny for flies that carry the DNA that must have integrated into the germline cells of their injected parents (5 ). Although chemical selection has been used to recover transformants among the progeny of injected flies, the technique does not greatly increase the number of progeny that can be screened. In contrast to cell culture systems, in which selection can be applied to millions of cells in a single Petri dish, embryo injection is still required as the first step when chemical selection is utilized in Drosophila.
We set out to use site-specific recombination to target DNA integration to specific sites in the Drosophila genome. Because of the difficulties described above, it seemed that using site-specific recombination to effect the direct integration of injected DNAs would not be feasible as a first step. Instead, we chose to utilize a method that would allow us to place a copy of the DNA that we wished to integrate (the donor) into every germline cell of whole animals. We used standard P element transformation to integrate the donor DNA in the Drosophila genome. This P element carried direct repeats of the target site (FRT) for the FLP site-specific recombinase (6 ) flanking the donor DNA. FLP can excise the FRT-flanked DNA from the chromosome with nearly 100% efficiency (7 ,8 ) and the DNA is excised as an intact circular DNA that carries one FRT (9 ). Therefore, FLP-mediated excision can generate a single episomal copy of the donor DNA in virtually every cell in an animal. If such an animal also has an FRT integrated at another site (the target), then when the extrachromosomal donor and the chromosomal target come into contact, FLP can mediate a second round of recombination that will integrate the donor DNA at the target site. The method is conceptually simple and abundant evidence demonstrates that site-specific recombination can be used to effect this DNA integration in other organisms (10 -13 ). The problem has been to find experimental conditions that take into account the biology of Drosophila so that the targeted integration event will occur with sufficient frequency. We describe experimental parameters that can be used to recover targeted integrants using the FLP site-specific recombinase in Drosophila melanogaster.
MATERIALS AND METHODS
All P element constructs were transformed by standard techniques (5 ). Heat shocks were performed in a circulating water bath as previously described (7 ).
FLP and FRT constructs
The hsFLP gene has been previously described (7 ).
A second source of FLP used in these experiments was an hsp70-FLP fusion gene called 70FLP. It is carried in a P element marked with ry+, called P[ry+, 70FLP]. It was constructed by cloning the XbaI-SalI FLP fragment from pDM420-FLP (7 ) into the plasmid pVZ1. This was transformed into strain RZ1032 (dut-/ung-; provided by S.Henikoff, Fred Hutchinson Cancer Research Center). Single-stranded DNA was prepared from a 20 ml culture and mutagenized to add a XhoI site at the 3'-end of the gene by synthesizing the second strand using a phosphorylated primer containing the mutation. The primer sequence was 5'-GCTTAAATGCTCGAGCTTATATGCG-3'. The resulting plasmid was called pFLP-XhoI. The mutated FLP gene was cloned into the plasmid p70ATG-Bam in which the normal hsp70 translation start site was replaced by a BamHI site (14 ). This was accomplished by cutting p70ATG-Bam with BamHI and filling in the 5'-overhang and then cutting with SalI to remove the hsp70 sequences. The FLP coding sequence was obtained by cutting pFLP-XhoI with SalI and filling in the 5'-overhang and then cutting with XhoI to isolate the FLP sequence from the vector, which was then ligated into the prepared p70ATG-Bam vector. The resulting plasmid was called p70/Flp/70. To facilitate moving the hsp70-FLP fusion gene into a P element vector the partially cleaved fusion gene (EcoRI and HindIII) was cloned into EcoRI and HindIII-cleaved pHSS6 (15 ), which served as a shuttle vector by providing flanking NotI sites. After isolation of the NotI fragment from the shuttle vector it was cloned into NotI-digested pDM30 (1 ), yielding the plasmid pP[ry+, 70CFLP70], or simply 70FLP. 70FLP3F is an insertion of this gene on the X chromosome.
A third FLP gene used here is the [beta]2t-FLP gene. This was constructed by cloning the [beta]2tubulin promoter from pUMB2 (16 ) into pHSS6 as a 0.7 kb EcoRI-HindIII fragment. The coding sequence of FLP was amplified by PCR with primers 5'-GGGATCCAAGCTTGCGCAGCTGAACAAGCTAAAC-3' and 5'-GTCGACTCTAGAGCGCTTCCGAAAATGCAAC-3', using a plasmid carrying hsFLP as template. These add HindIII and XbaI sites to the FLP coding sequence and these sites were used to insert the coding sequence behind the [beta]2t promoter in pHSS6. The [beta]2t-FLP gene was removed as a NotI fragment and cloned into the NotI site of the P element vector pDm30.
Two donor constructs were used in this work. The P[>whs>] element has been described previously (7 ). P[X97] was constructed by placing a SpeI linker with stops in all three reading frames (NEB 1061) into the PvuII site in the second exon of whs in plasmid P[X96] (9 ). This plasmid has a unique NotI site between the two halves of the whs gene. An hsGFP fusion gene (unpublished) was placed into this NotI site to generate the P[X97, hsGFP] element that was used in these experiments.
Two FRT-containing target site constructs were used: RS5r and RS3r. Their construction has been described (17 ).
PCR and primers
To confirm that the P[X97]-derived donor DNA was integrated at the RS3r target sites as expected we used genomic DNA from the putative integrants as template for PCR. The primer pair used [5'-GATAGCCGAAGCTTACCGAAGT-3' and 5'-TCATCGCAGATCAGAAGCGG-3'] amplifies fragments of characteristic size from P[X97, hsGFP], RS3r and from the donor DNA integrated at RS3r. The sizes and locations of these fragments are indicated in Figure 2 .
RESULTS
The white+ gene construct used in these experiments is actually a mini-white gene (18 ) that we call whs. All flies used in these experiments carried the w1118 null mutation. In a w1118 background whs usually produces an orange colored eye; for the sake of simplicity we refer to this phenotype as white+ (w+).
DISCUSSION
We have shown that site-specific recombination can be used to place genes at specific, pre-selected target sites in the Drosophila genome. Targeted gene integration was detected in some experiments by altered linkage of the mobilized gene. In other experiments a donor vector was used (P[X97]) that generates a w+ gene when it recombines into a target site, allowing the easy detection of integrants without the necessity of following inheritance of marked chromosomes. We used a variety of protocols to recover targeted integrants. The highest rate of success was obtained when the [beta]2t-FLP gene was used to promote mobilization between heterologous chromosomes in the male germline. Targeted integrants were also recovered using a heat shock during the first 2 days of development to induce expression of hsp70-FLP fusion genes in the germline. Early developmental heat shocks were used to ensure that all cells of the male germline were in a stage when they were still susceptible to heat shock (8 ). In the heat shock experiments there was no significant difference in the rates of germline mobilization achieved with 36, 37 or 38oC heat shocks nor was there significant variation in the efficiency of integration when three different locations of the RS3r target site were used. There was a barely significant (P = 0.045) variation in efficiency between the two P[X97] donors. Although our experiments did not reveal strong differences in efficiency with different sites, it is certainly conceivable that different donor and target locations might exhibit varying efficiencies of mobilization. For instance, if a donor and target were located close to each other, on the same chromosome or on homologs, the efficiency of integration might be higher because the excised DNA molecule would be more likely to come into contact with the target site.
The efficiency of this method is undoubtedly open to improvement. Altered FRTs might be used to prevent re-excision of an integrated DNA (4 ,13 ,24 ). An area that deserves further study is the role of donor-target homology in obtaining integrants. In experiments (not reported here) in which the only homology between the donor episome and chromosomal target was ~200-600 bp surrounding the FRT, we recovered targeted integrants only rarely (two from >2000 vials). The successful cases reported here were instances where the donor and target sites bore matching white gene sequence. In the experiments diagramed in Figure 1 the donor had ~1.1 kb of homology to the RS5r target site and ~4.1 kb of homology to the RS3r target and in these experiments (Tables 1 and 2 ) integration was more efficient at the RS3r target site. The experiments in Tables 3 and 4 were also cases where the donor episome and chromosomal target had ~4.1 kb of homology. Homology between donor and target sites might contribute to efficiency by stimulating the maintenance of physical proximity. Drosophila pair homologous sequences in mitotic cells (25 ); this pairing mechanism might bind the episome to the donor when they come into contact, prolonging the opportunity for integrative recombination.
If pairing of homologous sequences does facilitate integration, then manipulations that improve the efficiency of that pairing might improve the efficiency of integration. One way to increase pairing might be to increase the length of time available for pairing of the episomal donor and chromosomal target FRTs (8 ). Primary spermatocytes have a very long meiotic prophase and if FLP were synthesized at this stage, pairing of the episome and target might be more frequent. The particular [beta]2t-FLP construct that we used is probably transcribed in primary spermatocytes, but not translated until after meiosis. By altering the 5'- and 3'-untranslated regions of the construct so that it more precisely resembles the [beta]2tubulin gene, it may be possible to synthesize FLP in primary spermatocytes, possibly leading to an improved efficiency of targeting.
When targeted integrants are recovered two simple tests typically suffice to confirm that the expected integration event has occurred. First, the putative integrants can be crossed to flies with a heat-inducible FLP gene and the progeny heat shocked to produce the somatic mosaicism that indicates excision of the w+ reporter gene (7 ). The vast majority of putative integrants recovered in the work reported here were re-excised by FLP (79/83 tested). As a second test, PCR can provide rapid molecular confirmation of the integration event. Forty four of the integrants that produced white mosaicism with FLP were tested by PCR and the expected junction fragment was detected in all 44 cases. The exceptional flies, in which FLP-mediated integration was not confirmed by genetic or molecular tests, are potentially quite interesting. Some may represent cases of FLP-mediated integration at genomic sites that resemble FRTs. Sequences that are used as low efficiency target sites by the Cre recombinase have been found in the yeast genome (26 ). It is also possible that occasional rare integrations occur by mechanisms unrelated to FLP. In other cases such flies might arise from mutational events occurring at an RS3r target site. The 3'-portion of white carried by RS3r is sufficient to encode a functional white protein if expressed and similar effects have been observed previously, also at low frequency (17 ).
The ability to place a variety of gene constructs at a specific site and in a specific orientation allows the design of experiments that were previously impractical or impossible. For instance, it should be possible to investigate subtle effects on gene expression when a series of variants are placed at the same site, because such effects will not be swamped by the effects of chromosomal position. This technique also provides the potential for the examination of particularly interesting position effects, such as variegated effects (27 ) or those that confer an unusual pattern of expression (28 ). A variety of potential insulators, enhancers and promoters could be tested for their resistance or susceptibility to position effects by placing them, one by one, at a specific chromosomal FRT. One method for placing two genes at a site has been previously described (29 ). FLP-mediated mobilization allows for the placement of an unlimited number of constructs at a given site.
Finally, the demonstration that FLP can mediate DNA excision and re-integration suggests the possibility that it may be used as a method of directly integrating, at a chromosomal FRT target site, DNAs injected into embryos. This would bypass the steps involved in constructing and transforming P elements with the donor DNA. This might also allow the transformation of DNA segments that are too large to transform in transposon vectors. Reports that injected DNAs can be used as templates for P transposase-induced gap repair show that the injected DNAs can pair with homologous chromosomal sequences, supporting the possibility that this technique is achievable (30 ,31 ).
ACKNOWLEDGEMENTS
We thank Kami Ahmad and Majid Golafshani for technical assistance. This work was supported by grant HD28694 from the National Institutes of Health.
REFERENCES
1 Mismer,D. and Rubin,G.M. (1987) Genetics, 116, 565-578.MEDLINE Abstract
22 Lindsley,D.L. and Zimm,G.G. (1992) The Genome of Drosophiila melanogaster. Academic Press, San Diego, CA.
23 Fuller,M.T. (1993) In Bate,M. and Martinez Arias,A. (eds), The Development of Drosophila melanogaster. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 71-147.
24 Schlake,T. and Bode,J. (1994) Biochemistry, 33, 12746-12751.MEDLINE Abstract
27 Spofford,J.B. (1976) In Ashburner,M. and Novitski,E. (eds), The Genetics and Biology of Drosophila. Academic Press, London, UK, Vol. 1C, pp. 955-1018.