Dear Editor,

Drosophila oogenesis is of great interest because it represents an excellent model system to study a number of fascinating biological processes, such as stem cell regulation, germ cell meiosis and oocyte determination, as well as signal interactions between germline and soma. A typical Drosophila ovary is composed of 16-20 ovarioles, each consisting of an anterior functional unit called a germarium and a linear string of differentiated egg chambers posterior to the germarium 1 (Figure 1A and 1B). Drosophila oogenesis initiates at the tip of the germarium, when a germline stem cell (GSC) divides asymmetrically to generate a daughter GSC and a cystoblast that eventually develops into a mature egg 2 (Figure 1C and 1E). During the last decades, much progress has been made on identifying the key regulators controlling oogenesis, using traditional forward genetic screening via ethyl methanesulfonate and P-element mutagenesis. Such approaches are undoubtedly fruitful. To date, many sterile mutations that disrupt genes necessary for oogenesis have been generated. However, systematic genetic screening and subsequent identification of the role of genes in oogenesis largely depend on phenotypic analysis; therefore, the effects of many unidentified genes on oogenesis are often masked by lethality. Thus, given the limitation of inherent bias in traditional forward genetics, it would be advantageous to develop alternative methods to remove a gene's function in a tissue-specific manner for screening and characterization of novel gene functions in oogenesis.

Figure 1
figure 1

Effective gene silencing in Drosophila ovarian germline by artificial microRNAs. (A) A schematic diagram of a Drosophila ovary. A single female fly has a pair of ovaries, each of which consists of 16-20 ovarioles. (B) Diagram of an ovariole containing an anterior functional unit called a germarium and a linear string of differentiated egg chambers posterior to the germarium. (C) Diagram of the germarium consisting of different cell types and organelles indicated as follows: terminal filaments (TFs, dark green) and niche cap cells (CPCs, light green) are in close contact with germline stem cells (GSCs, yellow) that divide in coordination with somatic escort stem cells (ESCs, blue). ESCs produce squamous daughter escort cells (ECs, light pink), which encyst the differentiating GSC daughter cystoblasts (CB, orange). As a CB divides, it develops into cyst cells (green), one of which is determined to form the oocyte (OC). A 16-cell cyst is surrounded by follicle cells (FCs), generated by follicle stem cells (FSCs, violet). GSCs and CBs carry a spectrosome (SP, red), whereas cysts have extending or branched fusomes (F, red). (D) Construction and biogenesis of artificial microRNAs. (E-J) Characterization of the function of the microRNAs in Drosophila ovary. Ovaries from wild-type w1118 (E), P{UASp-3XartmiR-bam}/P{nosP-gal4:vp16} (F), P{UASp-2XartmiR-mad}/P{nosP-gal6:vp16} (G), P{UASp-2XartmiR-ote}/P{nosP-gal4:vp16} (H, I) and P{UASp-artmiR-dpp}/P{bab1-gal4} (J) flies were stained with anti-Vasa (green) and anti-Hts (red) antibodies. Anti-Hts was used to outline the germarium and the morphology of the spectrosomes (as indicated by arrows or arrowheads) and fusomes, whereas the staining of anti-Vasa was used to visualize all germ cells in the germarium and egg chambers. (K) Quantitative RT-PCR experiments were used to analyze the effectiveness of target gene knockdown in the examined fly ovaries as indicated. (L) Western blot result showing the effectiveness of knockdown by 2XartmiR-ote in the ovary. (M) Quantitative analysis of the numbers of germ cells carrying spectrosomes in w1118 and artmiR transgenic flies.

Full size image

Small RNA-mediated gene silencing controls a wide range of processes during development. So far, three classes of small RNAs have been characterized in Drosophila, including small interfering RNAs (siRNAs), microRNAs (miRNAs) and Piwi-interacting RNAs 3. Double-stranded siRNA-mediated gene silencing, as a potent tool, has been widely used for knocking down gene expression in S2 cell cultures and in multiple tissues in flies. In Drosophila oogenesis, the siRNA pathway is apparently activated during egg activation, because the introduction of exogenous dsRNAs results in degradation of target mRNAs. However, in the early stage of germ cell development these cells fail to support dsRNA-mediated RNA interference because of unknown inhibitory mechanisms. To our knowledge no practical protocol for siRNA-mediated gene knockdown has yet been established for the Drosophila ovarian germline.

The miRNAs are structurally similar to siRNAs, but have a distinct biogenesis mechanism and execute gene silencing either through degradation of target mRNAs or by translation repression of target genes. There is increasing evidence that the miRNA pathway plays critical roles in the regulation of GSCs and germ cell development, suggesting that miRNA-mediated gene silencing functions as an important mechanism in regulating Drosophila germline development 4, 5. Haley et al. 6 have now successfully established a simplified gene silencing method by using artificial miRNAs in Drosophila S2 cell cultures and in wing imaginal discs. Therefore, we sought to test whether artificial miRNAs could also be applied in germ cells for knocking down genes of interest and for systematic reverse genetic screening to identify hitherto undiscovered genes involved in germline development.

To explore whether ectopic expression of the artificial miRNAs could affect germline development, we chose bam, mad, ote and dpp as target genes (Supplementary information, Data S1). These genes have been demonstrated to play critical roles in determining the fate of GSCs and germ cell development. Using an earlier described method 6, we used the precursor of miR-1, which is enriched in Drosophila ovaries, as the backbone to design the pre-miRNA stem–loop base of artificial miRNAs (Figure 1D and Supplementary information, Figure S1). The sequence of pre-artmiRNA, 71 nucleotides (nt) long, contains 21-nt oligos that are complementary to the mRNA sequence of target genes, and the detailed information is shown in Supplementary information, Figure S1B. This design allows pre-artmiRNA to mimic the natural structure of pre-miR-1 that ensures the accurate processing and efficient production of the mature artmiRNAs by Dicer-1/Ago1 complexes.

To construct the artificial miRNA expression vectors, we used a modified UASp vector as the backbone to generate P{UASp-artmiRNA}, in which one or two or three hairpin sequences were placed under the control of the UASp promoter. The UASp promoter is the only active UAS promoter in germ cells when it is combined with P{nosP-gal4:vp16}, a germ cell-specific driver 7. To explore whether the artificial miRNAs would efficiently knock down the expression of target genes in germ cells, we first chose bam as a target gene and generated bam knockdown transgenic lines, P{UASp-3XartmiR-bam}, in which three different hairpins targeting bam mRNA were under the control of the UASp promoter. This gene has been shown to play key roles in germline development, given that loss of bam blocks germ cell differentiation resulting in GSC hyperplasia, while ectopic expression of bam in GSCs results in their elimination 8. To assess the effectiveness of expression of artmiR-bam in germ cells, we employed the Gal4-UAS system to express it by combining P{UASp-3XartmiR-bam} with P{nosP-gal4:vp16}. Notably, as shown in Figure 1F and 1K, expression of 3XartmiR-bam significantly knocked down the endogenous bam and blocked germ stem cell differentiation leading to germ cell hyperplasia, which simply phenocopied Drosophila with a loss of bam background, suggesting that artificial miRNAs could potentially silence the target genes of interest in germ cells. Additionally, we found that expression of one copy of bam hairpin (1XartmiR-bam) resulted in a weaker bam-like phenotype (data not shown), compared to 3XartmiR-bam, suggesting that artificial miRNAs act in a dose-dependent manner to down-regulate target gene expression. To test whether artificial miRNAs also knock down expression of other genes in germ cells, we then generated two more artificial miRNA transgenes, P{UASp-2XartmiR-mad} and P{UASp-2XartmiR-ote}, for knocking down the endogenous mad and ote genes, respectively. Both mad and ote genes have been shown to function as key regulators in the bone morphogenetic protein/decapentaplegic (BMP/Dpp) pathway and are essential for the maintenance of germline cells 7, 9. We first examined whether the artificial miRNAs of mad would affect adult GSC function by combining P{UASp-2XartmiR-mad} with P{nosP-gal4:vp16} to express artmiR-mad in adult germ cells. We took advantage of the temperature-dependent activity of Gal4:Vp16 to control the expression level of 2XartmiR-mad. According to the method described previously 10, the crosses were performed and the resulting progenies were initially raised at 18 °C until adult eclosion, then they were shifted to the standard culture temperature (25 °C). The numbers of GSCs per germarium from the controls and from female flies expressing 2XartmiR-mad were quantified at days 1, 3, 5 and 7 after eclosion. As shown in Figure 1G, 1K and 1M, compared with controls, the numbers of GSCs per germarium reduced progressively in the ovaries from the females expressing 2XartmiR-mad during the testing period. Together with quantitative RT-PCR analysis, our findings further demonstrated that expression of 2XartmiR-mad efficiently knocked down expression of the endogenous mad, resulting in loss of GSCs. Similar results were obtained when we knocked down the ote gene by expression of the artificial miRNAs of 2XartmiR-ote in germ cells (Figure 1H, 1I, 1K-1M).

Previous studies have revealed that maintenance of GSCs requires both extrinsic and intrinsic mechanisms. Dpp functions as an extrinsic niche factor to maintain GSCs by silencing bam transcription. To test whether artificial miRNA-mediated gene silencing could also be applied to somatic niche cells in the ovary we constructed dpp knockdown vectors, P{UASp-artmiR-dpp} and P{bab1-gal4}. As shown in Figure 1J, 1M and Supplementary information, Figure S2, specific knockdown of dpp expression in the cap cells of germaria resulted in loss of GSCs, which mimics the phenotype in dpp mutants 9, suggesting that artificial miRNAs could also interfere with target gene functions in somatic cells in ovaries.

In summary, our results support the idea that artificial miRNAs can effectively down-regulate endogenous target genes of interest in Drosophila germline, as well as in somatic cells in Drosophila ovary. We believe that the technique established in this study is not only useful for characterization of the functions of genes of interest, but also provides a powerful tool for systematic reverse genetic screening for identifying hitherto undiscovered genes involved in GSC regulation and germline development.