Full Text (PDF)
Post-transcriptional gene silencing has become the fastest and most frequently used approach to reduce gene function in cell culture. In an organism without an interferon response, such as Drosophila melanogaster, RNA interference (RNAi) can be evoked using long, double-stranded RNAs (1). In order to achieve a gene knockdown at a stage not amenable to dsRNA injection into early embryos, transgenic approaches have been used in Drosophila (2,3). Commonly, a 500 to 700 bp cDNA product is cloned as an inverted repeat with or without a spacer and expressed under the control of UAS-Gal4 sequences (4). To allow tissue-specific and/or temporally controlled expression of the RNAi transgene, the bimodular UAS-Gal4 system can be used, bypassing any potential toxicity of such transgenes.
Inverted repeats tend to be inherently difficult to clone in Escherichia coli due to recombination-mediated DNA repair that occurs during DNA replication (5). Lee and Carthew (6) designed pWIZ, a pUAST-based (4) fly transformation vector in which the two inverted repeats are separated by an intron of the white gene. After transcription, the intron is spliced out and the mRNA forms a hairpin in vivo. The cloning of inverted repeats into pWIZ is very difficult, even when sbcC− E. coli strains reducing recombination repair are used. A smaller shuttle vector in which the repeat/intron cassette is pre-assembled is easier to use, but requires at least 3 or 4 sub-cloning steps (7). Zhu and Stein used a strategy allowing directional cloning of inverted repeats in the germline UAS expression vector pUASp (8,9). pWIZ was also adapted for the GATEWAY in vitro recombineering system (Invitrogen, Carlsbad, CA, USA) (10). This elegant system, however, requires the purchase of expensive, special reagents.
In an attempt to simplify the generation of genomic RNAi constructs, we modified pWIZ to improve the cloning efficiency of dsRNA probes. The biggest problem during the cloning of inverted repeats is to obtain the inverted versus the direct orientation of the dsRNA target fragment during the second round of cloning. We therefore constructed pWIZdir by changing the polylinkers of pWIZ to allow the consecutive, directional insertion of a single PCR product using pairs of compatible restriction enzyme sets: SpeI/AvrII and NheI/XbaI adjacent to the white intron, and BglII and BamHI upstream and downstream, respectively ((Figure 1)). The directionality of the second cloning step greatly increases the frequency of obtaining correctly inverted repeats and obviates the need for the intermediate use of a shuttle plasmid. We have made eight dsRNAi constructs and obtained between 16% and 90% correctly inserted second-round cloning products, a significant improvement over previous attempts (not shown). All eight constructs were successfully used to generate transgenic fly lines. Examples of eye and wing phenotypes obtained in vivo are shown in (Figure 2).
Figure 1.

Target sequences are chosen using the Heidelberg eRNAi prediction site (www.dkfz.de/signaling2/e-rnai). Fragments should lack either a BglII, BamHI, or BclI site, and should not contain either a XbaI, AvrII, NheI, or SpeI site. Preferentially, we attach 3 to 5 random nucleotides followed by an AvrII and BamHI site at the 5′ ends of the forward and reverse primers, respectively. This strategy results in the first insert having the antisense orientation, avoiding the expression of any potential protein fragment ((Figure 1)B).
All restriction digests are performed in 20 µl reactions using 5 µl (about 2.5 µg) of a typical midiprep or 10 µl of a miniprep and 1 µL of each enzyme for 2 h to overnight. Gel purifications are done using the Qiaquick gel-elution kit (Qiagen, Valencia, CA, USA) and DNAs are eluted in 30 µl elution buffer. Ligations are performed in 10 µl volumes containing 7 µl insert, 1 µL vector, 1 µl 10× ligation buffer, and 1 µl ligase (NEB, Ipswich, MA, USA) for 2 h to overnight at room temperature.
The PCR product is ethanol-precipitated according to standard procedures (11) and digested with AvrII and BamHI in buffer 2 (NEB), gel-purified, and ligated into pWIZdir digested (and dephosphorylated) with BglII and AvrII. After transformation, minipreps are analyzed for the presence of the insert with EcoRI/AvrII (buffer 4, NEB).
A positive miniprep is then sequenced with pUAST_for (CTGAAATCTGCCAAGAAGT) to confirm the identity of the PCR product. 10 µl of the miniprep are digested with XbaI/BamHI and ethanol-precipitated. After dephos-phorylation, the second-round vector is gel-purified and ligated to the same PCR product used in the first round of cloning. The ligation is transformed into chemically competent (12) Sure cells (Stratagene, La Jolla, CA, USA). Recovery of the heatshock and growth of colonies, as well as the following liquid cultures for minipreps are grown at 30°C instead of 37°C. Positive minipreps are identified by digestion with BamHI/NheI. Positive clones are characterized by restriction analysis and compared with the parental pWIZdir to avoid clones with deletions in the insert or vector backbone. We typically test again for the presence of both inserts and assess the integrity of the vector backbone with AvaI and StyI ((Figure 1)A).
Vector constructionTo generate pWIZdir, the BamHI sites flanking the UAS cassette of pWIZ (6) were destroyed by opening, blunt-ending, and religating the fragments. Sequencing confirmed the absence of BamHI sites. Annealed oligos Wiz_dir_for and Wiz_dir_rev (CTAGCTCTAGAATATCAATTGGGATCCA; CTAGTGGATCCCAATTGATATTCTAGAG) were subsequently ligated into the NheI/XbaI sites of pWIZΔBam to produce pWIZdir ((Figure 1)A). We also generated a RNAi-based vector in pUASp (pUASPdir), an expression vector that, unlike pUAST, enables Gal4-mediated expression in the germline, as well as in somatic tissues (9). To assemble pUASPdir, pWIZdir was amplified by PCR with oligos UASP_WIZ_for and UASP_WIZ_rev (ATATAGGTACCGAATTCGTTAACAGATCTGCGG; TATATCCCGGGGGATCCCAATTGATATTCTAGAG), the product digested with KpnI and SmaI and inserted into the KpnI/XbaI (blunt) sites of pUASp. Although we have not tested this vector in vivo, it is available to individuals who are interested in using it. Compiled sequences of pWIZdir and pUASPdir can be obtained from GenBank (accession nos. EU178814 and EU178815, respectively).
In summary, we generated a new vector useful to knockdown gene functions in transgenic Drosophila melanogaster. In contrast to pWIZ, pWIZdir allows efficient, directional cloning of the hairpin stems. This greatly increases the frequency of successfully cloning the required inverted repeats without use of expensive reagents. Furthermore, our cloning strategy eliminates the need to determine the orientation of each fragment, a task often complicated by the lack of appropriate, asymmetric restriction sites in such short inserts. In addition to its use as a fly transformation vector, pWIZdir could also be used to knock down genes in cell culture in combination with a co-trans-fected Gal4 expressing plasmid.
We thank Richard Carthew for pWIZ and Katya Serysheva for help with sectioning eyes. We thank Steve Cohen for sdGal4 and enGal4 strains and Uwe Walldorf for eyGal4. We are grateful to David DelAlamo, Matias Simons, and Elisa Wurmbach for critically reading the manuscript. We thank the reviewers for bringing the publications of Zhu and Stein (8) and Kondo et al. (10) to our attention. The work has been supported by the National Institutes of Health (grant no. R03 DC007055 to A.J. and grant no. RO1 GM062917 to M.M.).