Our best success to date with ZFN-induced gene targeting has been with the fruit fly, Drosophila melanogaster. We chose this organism for our first test of the technology because of its relatively small, sequenced genome, the ease of genetic analysis, and assistance we received at the outset from Kent Golic’s lab. Yikang Rong and Kent Golic developed the first effective gene targeting strategy for Drosophila, using FLP and I-SceI to generate an extrachromosomal, linear donor DNA in situ. We added our target cleavage approach to their procedure and developed a highly efficient method.
We have now successfully targeted 3 genes in Drosophila: yellow, brown and rosy. They were chosen because of their easily scored phenotypes. In each case we searched the gene for a sequence of the form (XXC)3N6(GXX)3 and readily found one or more. The corresponding ZFNs were produced and used to generate new mutants in two formats. First, we simply expressed the ZFN pair from a heat-shock promoter during larval development and screened progeny of induced flies for new mutations. These are produced by nonhomologous end joining (NHEJ) following targeted ZFN cleavage and are readily recovered. We determined DNA sequences of a number of these and confirmed that the mutations were located precisely at the site of targeted cleavage and had the form expected for NHEJ – mostly small deletions and insertions, with occasional larger deletions. In the best case (ry) we recovered new mutations in 14% of all offspring of induced females.
Second, we included a marked donor DNA and used the Rong & Golic method to make it most effective as a template for repair after the ZFN-induced break. We found that an extrachromosomal linear donor was more efficient than a circular one, and both were much better than an integrated (i.e., not excised) donor. Again the best target was ry, where we recovered new mutations in 25% of offspring of induced females: 15% as targeted gene replacements (products of homologous recombination), 10% as targeted NHEJ products. Efficiencies nearly as high have been achieved in males and at the y locus.
Methods (briefly)
To summarize our published method for targeting with ZFNs, we begin by identifying a plausible target in any gene of interest (see the section on Designing ZFNs for Novel Targets for a description of how to go about this). We construct coding sequences for the two required ZFNs and introduce them into the Drosophila genome on P element vectors, where they are under the control of a heat-shock promoter. We inject the ZFN transgenes separately, so they can be tested individually for potential lethality. Using marked balancer chromosomes, we map each transgene to one of the 4 chromosomes. For NHEJ-based mutagenesis, we simply cross flies carrying the individual ZFN transgenes. When we include a donor, we also introduce this into the genome on a P element vector. For this purpose we use one in which the cloning site is surrounded by recognition sites for I-SceI and, outside these, for FLP. The complete scheme requires that the donor, the two ZFN genes, and genes for FLP and I-SceI all be brought together in the same flies. This requires some mapping and crossing, but is basically straightforward. The mutagenesis and/or targeting events are induced by heat-shocking larvae carrying the necessary components (the FLP and I-SceI genes are also under heat-shock control). Eclosing adults are crossed to partners that will reveal germline mutations in the locus of interest. In cases where the mutant phenotype is unpredictable, molecular screens (by PCR) for the products of interest are also feasible.
Very recently we have succeeded in simplifying our procedures by injecting synthetic mRNAs for the ZFNs into early embryos. We have recovered NHEJ mutations at the ry locus at very high frequencies and donor DNA incorporation at lower efficiency.
References (Carroll Lab)
References (Golic procedure)