A reverse genetic screen in Drosophila using a deletion-inducing mutagen

Genome Biology 0 Institut fur Molekularbiologie, Universitat Zurich , Winterthurerstrasse 190, Zurich CH-8057 , Switzerland 1 Massachusetts General Hospital , Jackson 1402, 55 Fruit St, Boston, MA 02114 , USA 2 Zoologisches Institut, Universitat Zurich , Winterthurerstrasse 190, Zurich CH-8057 , Switzerland We report the use of the cross-linking drug hexamethylphosphoramide (HMPA), which introduces small deletions, as a mutagen suitable for reverse genetics in the model organism Drosophila melanogaster. A compatible mutation-detection method based on resolution of PCR fragmentlength polymorphisms on standard DNA sequencers is implemented. As the spectrum of HMPAinduced mutations is similar in a variety of organisms, it should be possible to transfer this mutagenesis and detection procedure to other model systems. - Background The fruitfly Drosophila melanogaster has been the prime genetic model organism for almost a century. This success story is mainly founded on countless so-called forward genetic screens designed to elucidate gene functions on the basis of their mutant phenotypes. Many of those screens reached a scale that has been termed 'saturating' as they identify all nonredundant genes involved in a certain phenotypic trait. However, forward genetic screens are limited in that they are only capable of uncovering functions that are easily measurable or visible. Furthermore, genes having a redundant or nonessential role are less likely to be found by forward genetics. The reverse genetic approach to unravel gene function starts with the DNA sequence. Mutations within the gene are induced and identified by various techniques and only subsequently is the mutant phenotype analyzed [1]. Reverse genetics may be undirected or directed, the undirected approach involving random mutagenesis, commonly by transposable elements or by chemicals, the establishment of mutant collections, and the identification of mutations in the gene of interest [2-5]. In contrast, directed reverse genetics is based on techniques that allow for specific inactivation of a gene. These include specific knockdown of gene activities through RNAmediated interference (RNAi) [6,7] and targeted gene disruption [8,9]. Both undirected and directed reverse genetic techniques have certain advantages and drawbacks. Transposon-based mutagenesis tends to be nonrandom because of the occurrence of hotspots for transposon integration. The use of transposable elements of different origin, such as P-elements and piggyBac, which exhibit a different insertion bias, can partly circumvent this problem. However despite large-scale efforts, the ultimate goal of covering the whole Drosophila genome by insertion mutagenesis is far from being achieved [10,11]. Moreover, while null mutants of P-element-tagged genes (Pelements have the tendency to integrate 5' to a gene) can easily be generated by imprecise excision, piggyBac transposons only excise precisely [10]. RNAi and small interfering RNA (siRNA) screens provide a powerful tool to dissect the function of genes at a genomewide scale [12-14], but the technique is most easily applied to cell cultures and is thus limited to cell-biological problems. Large-scale RNAi screens in multicellular organisms have been done only in C. elegans [15] and for technical reasons a similar approach in Drosophila is not feasible. Targeted gene knockout in Drosophila allows for generation of both null as well as hypomorphic mutations [16]. However, the technique is time-consuming and technically challenging and hence not applicable on a large scale. Random mutagenesis in reverse genetics generally relies on well-established techniques and commonly used mutagens, such as ethylmethansulfonate (EMS) [5,17] and N-ethyl-Nnitrosourea (ENU) [18]. Those chemicals primarily induce single-nucleotide polymorphisms, which can most efficiently be detected by sequencing [19], by denaturing high-pressure liquid chromatography (DHPLC) [5,17], or by enzymatic cleavage of heteroduplex DNA with single-strand-specific endonucleases such as Cel-I [18,20-22]. Mismatch-cleavage analysis and DHPLC require special machinery and DHPLC is not very well suited for high-throughput analysis. Fast neutrons have also been used to introduce small DNA lesions, which can simply be resolved by agarose electrophoresis after PCR amplification [23]. This kind of mutagenesis may be limited to seeds or to labs in the vicinity of a reactor. We reasoned that it would be worthwhile to establish a generally applicable reverse genetic technique based on an unbiased and practicable random mutagenesis and an efficient mutation-detection performed on standard laboratory equipment. Here we introduce a novel mutagenesis protocol utilizing the cross-linking drug hexamethylphosphoramide (HMPA) [24], streamlined fly genetics and high-throughput fragment analysis on sequencers to demonstrate the feasibility of our reverse genetics approach. Results and discussion Fly genetics There are two ways to handle mutagenized progeny. Either large collections are established and maintained, which then are systematically and continuously screened for mutations of interest, or mutagenized progeny are screened directly and only animals exhibiting a desired trait are kept. The first method is in practice an F3 screen, which requires balancing of mutagenized chromosomes and maintenance of many stocks. This approach is far more labor-intensive than a simple F1 screen of progeny and thus is more suited to stock centers. Moreover, balancer chromosomes have many DNA sequence polymorphisms to wild-type chromosomes (our unpublished data), which will interfere with detection of mutagen-induced sequence polymorphisms. To circumvent the inherent problems with balancers, we devised an alternative genetic strategy, which had to fulfill the following criteria. First, mutagenized chromosomes have to be passed on in an unrecombined form such that mutations cannot be lost. Second, the mutagenized chromosomes should be brought into an isogenic background for mutation detection. Third, for economic reasons stock-keeping should be kept at an absolute minimum. We generated a fly strain (KNF306) isogenic to our yw wildtype laboratory strain but containing the same dominant marker on the two major autosomes. Both chromosome 2 and chromosome 3 are carrying white+ marked P-element insertions, which were chosen because white+ expression is restricted to different subregions of the eye (Figure 1a). Chromosome 2 is marked by an insertion in the CG31666 locus, which results in white+ expression only in the posterior part of the eye. Chromosome 3 harbors an insertion in the promoter of CG32111, and this transgene causes dorsal white+ expression. The combined expression patterns of both show a 'pie-slice' eye-color appearance (Figure 1a). Thus, the same marker permits us to distinguish between linkage on chromosomes 2 or 3. Neither of the transgenes affects viability, a (...truncated)