Electroporation-mediated somatic transgenesis for rapid functional analysis in insects

Reverse genetics and transgenic techniques are powerful methods for elucidating gene functions. However, such methods are available for only a few model organisms. Over the recent decades, because of its simplicity, RNA interference (RNAi) has been used as an alternative method for knocking out specific genes. RNAi has enabled in vivo functional analyses of specific genes in various non-model organisms (Tomoyasu et al., 2005; Moczek and Rose, 2009). However, several technical obstacles persist, such as low efficiency of the introducing dsRNA in some species (Tomoyasu et al., 2008) and the need for gain-of-function analysis when investigating gene function. To overcome such problems, we decided to directly introduce plasmid DNA into tissues in vivo using electroporation and insert exogenous DNA into host chromosomes using the piggyBac transposon. We designed a simple electroporator that costs under US$200 (supplementary material Figs S1, S2). Here, we demonstrate that exogenous DNA can be readily introduced into various tissues of the silkworm Bombyx mori after embryogenesis using electroporation, and that it can be stably maintained until adulthood. We also successfully induced RNAi using short-hairpin RNA (shRNA)-mediating DNA vectors. These methods were applicable to two other insect species: swallowtail butterfly Papilio xuthus and red flour beetle Tribolium castaneum. This suggests that our approach could be a powerful tool for rapid in vivo functional analysis in a broad range of insects.

We generated square voltage pulses (5 V) using an AVR microcontroller (ATTINY2313-20PU; Atmel, USA) and amplified it using an operational amplifier (op-amp) (OPA454; Texas Instruments, USA). The control program (C language, deposited at http://sourceforge.net/projects/vivoelec/files/) was compiled with WinAVR (http://winavr.sourceforge.net/) and written to the microcontroller using a USB-connection-type AVR programmer (AVRWRT-3; Kyoritsu Eleshop, Japan). Almost all electronic parts were manually soldered based on our designed electrical diagram (supplementary material Fig. S1). The op-amp was not directly soldered to the circuit board but was attached with a zero-insertion force socket (Kyoritsu Eleshop, Japan) such that only the op-amp chip needs to be exchanged when the output electrodes short-circuit. The power source for driving the op-amp (50-100 V constant voltage) was an electrophoresis power supply (model 500-200; Wakamori, Japan). To prevent the sudden current flux on connection to the power supply, a 10 kΩ variable resistor was inserted between the power supply and the circuit; the resistance was gradually reduced to 0 Ω after connection. The waveform of the amplified square pulses was confirmed using an oscilloscope (LBO-310A; Leader Electronics, Japan), and the voltage amplitude was calibrated using the oscilloscope and power supply.

B. mori (strain N4) was reared on an artificial diet (NOSAN, Japan) under long-day conditions (16-hour light/8-hour dark) at 25°C. P. xuthus was reared on tangerine (Citrus reticulata) leaves under long-day conditions at 25°C. T. castaneum (strain Ga2) was reared on wholewheat flour containing 5% yeast extract at 30°C.

DNA vectors (Fig. 1A) were based on pPIGA3GFP (Tamura et al., 2000) (supplementary material Figs S5, S6, Table S2). We first constructed the transitional vector pPIG-A3GG. The partial coding sequence of actin A3 in the A3 promoter (A3⁺ promoter) was eliminated to yield the A3 promoter. The fragment containing the entire vector sequence from pPIGA3EGFP and a fragment containing another EGFP expression cassette flanked by the IE2 promoter derived from pIZT-V5/His (Invitrogen, USA) from an unpublished plasmid [that includes the expression cassette between SalI and NotI sites of pBluescript KS– (Stratagene, USA)] were amplified by PCR (supplementary material Fig. S5). In-Fusion Enzyme (Takara, Japan) was used to ligate these fragments and for all other ligations, except self-ligations. An unnecessary multiple cloning site was eliminated by SmaI and SalI digestion and re-ligation. The redundant IE2 promoter was excised by BstBI digestion, and the 5′-sequence of the A3 promoter was amplified from pPIGA3EGFP and cloned into the vector after digestion with SalI. The A3⁺ promoter remaining in the plasmid was excised with XhoI and BamHI digestion and replaced with another A3 promoter fragment amplified from pPIGA3GFP to generate pPIG-A3GG.

pPIG-A3GR was obtained by cloning the DsRed2 ORF cassette from pBac3xP3DsRed2 (Inoue et al., 2005) into BamHI/NotI-digested pPIG-A3GG.

pPIG-A3GR-U6shG was obtained by isolating the U6 promoter fragment flanked with the sequence encoding shRNA against EGFP from the Bombyx genome (strain p50) using a previously described methods (Wakiyama et al., 2005; Ohtsuka et al., 2008) and cloning it downstream of the A3/DsRed2/SV40polyA reporter cassette using SfiI-digested pPIG-A3GR.

pPIG-A3G-U6shw3 was obtained by amplifying the U6 promoter and shRNA sequence against w-3 from pPIG-A3GR-U6shG and cloning it into XhoI/NotI-digested pPIG-A3GG.

pPIG-A3GFF-UASG was obtained by cloning the Gal4FF ORF fragment from pT2KhspGFF (Asakawa et al., 2008) into BamHI/NotI-digested pPIG-A3GG, excising the A3 promoter upstream of the EGFP ORF with XhoI and SalI, and replacing it with the 5×UAS/HSP70Bb fragment from pUAST (Brand and Perrimon, 1993).

pPIG-A3⁺G-HSP70R was a gift from Dr J. Yamaguchi (The University of Tokyo, Japan).

siRNA against EGFP and w-3 was designed based on our siRNA design guidelines (Yamaguchi et al., 2011) using the siDirect program (Naito et al., 2009) (http://sidirect2.rnai.jp; supplementary material Table S3) and chemically synthesized and annealed (FASMAC, Japan).

Silkworms and swallowtail butterfly larvae were anesthetized at 4°C before electroporation. Red flour beetle larvae were anesthetized with ethyl ether. After immobilizing the insects with adhesives or forceps, DNA solution (1 μg/μl) was injected into the body cavity using a microinjector (FemtoJet; Eppendorf, Germany) and a broken-tip glass needle (GD-1; Narishige, Japan), which was prepared using a needle puller (PP-830; Narishige) (supplementary material Fig. S3). The larval epidermis and antennal primordium were injected with 0.5 μl of the DNA solution, and the pupal wing was injected with 2.0 μl of the DNA solution. Tribolium larvae were injected with the largest possible volume of the DNA solution. Immediately after injection, platinum electrodes were placed near the injection site; on the opposite side of each tissue, PBS droplets were placed nearby and appropriate voltage was applied.

All images, except for those of the antenna and wing, were collected from live insects under fluorescence stereomicroscope (Leica M165FC; Germany) using a digital camera (AxioCam MRc5; Carl Zeiss, Germany). The contrast and γ values were adjusted using AxioVision software (release 4.5, Carl Zeiss). Antenna and wing images were collected using fixed tissues (4% paraformaldehyde fixation for 30 minutes at room temperature) under a LASER confocal microscope (FluoView FV1000; Olympus, Japan).

The optimal voltage for introducing exogenous DNA into tissues in vivo of B. mori was established using a simple, low-cost electroporator. Unless otherwise stated, different voltages (10-45 V) were applied for the same duration (five 280 ms pulses per 5 seconds). The target tissues were larval epidermis, antennal primordium and pupal wing.

Plasmid DNA solution was injected into the body cavity near each tissue, and platinum electrodes and droplets of phosphate-buffered saline (PBS) were placed nearby before electroporation (supplementary material Fig. S3). To stably express exogenous genes, we constructed the pPIG-A3GR vector (Fig. 1A) based on the vector used for transgenesis (Tamura et al., 2000). The overexpression cassette of the vector was flanked with the target sequence of the piggyBac transposon to be inserted into the chromosome (Fig. 1A-E, ‘L’ and ‘R’). piggyBac transposase was supplied by a separate helper plasmid [Fig. 1F, pHA3PIG (helper)]. First, exogenous enhanced green and red fluorescence protein expression (EGFP and DsRed2) was examined in the larval epidermis (second instar) post-electroporation at 20 V (Fig. 2A). Although the expression was observed 3 days post-electroporation with and without the helper plasmid (Fig. 2A, 3 d), it was still present 12 days post-electroporation only when the vector and the helper plasmid were electroporated together (Fig. 2A, 12 d). Moreover, it was maintained until adulthood (Fig. 2A, + piggyBac, 18 d and 21 d). This implied that piggyBac facilitated stable expression by generating transgenic somatic cells. Thereafter, the donor and helper plasmids were co-injected in all experiments. Exogenous gene expression was also observed in the antennal primordium (early final instar) post-electroporation with 10 pulses at 45 V and in the pupal wing with five pulses at 20 V. The expression in the antennal primordium was maintained after pupation (12 days post-electroporation), and several cells differentiated into olfactory neurons (Fig. 2B, arrows). In the pupal wing, several cells differentiated into scale cells (Fig. 2C, arrows). Furthermore, several tissues beneath the targeted tissues (e.g. muscle, fat body and trachea) were also transfected in these electroporation treatments, suggesting that functional analyses of various tissue types could be possible using our approach. The survival rates and gene transfer efficiencies are summarized in supplementary material Table S1.

Next, we investigated whether RNAi was applicable to this electroporation-mediated system. The EGFP gene expressed from a plasmid in the larval epidermis was used as the RNAi target. To stably induce RNAi, we used an RNA polymerase III promoter (U6 promoter)-driven shRNA (Wakiyama et al., 2005; Ohtsuka et al., 2008). We constructed the pPIG-A3GR-U6shG vector that expressed shRNA against EGFP (Fig. 1B) and compared the fluorescence of EGFP with that of pPIG-A3GR. EGFP fluorescence in the DsRed2-positive cells was weaker in all silkworms (compare Fig. 3A and Fig. 2A, n=10), indicating that the shRNA-mediated RNAi system was applicable. We also evaluated the efficacy of direct transfer of siRNA by electroporation. Electroporation of siRNA (400 μM, 0.5 μl) against EGFP weakened the EGFP signal (Fig. 3B, n=10), whereas injection of siRNA did not attenuate this signal (Fig. 3C, n=12). These data indicate that RNAi can be induced by electroporation-mediated direct transfer of siRNA and that our electroporation approach can overcome the low efficiency of introducing siRNA reported in some insects (Tomoyasu et al., 2008; Terenius et al., 2011).

Furthermore, we investigated whether endogenous gene function could be inhibited using these RNAi systems. We focused on the w-3 gene, which encodes an ABC transporter that regulates accumulation of white pigment in the leucophores in larval silkworm epidermal cells (Quan et al., 2002; Kômoto et al., 2009). Both shRNA-mediated system and direct transfer of siRNA (300 μM, 0.5 μl) mimicked the w-3 mutant phenotype of translucent skin in the electroporated epidermis (8 d in Fig. 3D,E, arrowheads; n=11 for each), which was not observed in the negative controls (Fig. 3F, n=16; supplementary material Fig. S4). However, these two systems resulted in different phenotypic patterns. shRNA-mediated RNAi induced mosaic translucent skin that was maintained until the final instar (Fig. 3D, arrows), whereas direct siRNA transfer induced a broader translucent region (Fig. 3E, 8 d, arrow), which reduced over time (Fig. 3E, 13 d). These data suggest that the transgenesis-based RNAi system is suitable for inducing RNAi for prolonged periods and that direct siRNA transfer could be useful for testing short-term effects.

Evaluating the effects of gene product dose on phenotype is also important for analyzing gene function. To increase transcriptional efficiency, we constructed the pPIG-A3GFF-UASG vector (Fig. 1D) that uses the strong transcriptional activity of the GAL4/UAS system (Brand and Perrimon, 1993). Using this vector, we observed an increase in the levels of the exogenous gene product (Fig. 3G; n=5). Further manipulation of transcriptional activity with the GAL4/UAS system, such as modifying the number of UAS repeat sequences, should enable fine-tuning of the dosage.

We investigated whether our electroporation procedure was applicable to a broader range of insect species. We used pPIG-A3⁺G-HSP70R (Fig. 1E), in which reporter genes are driven by two different promoters active in non-host insects: BmA3, honey bee (Ando et al., 2007); DmHSP70, beetle, butterfly and silk moth (Oppenheimer et al., 1999; Ramos et al., 2006; Uhlirova et al., 2002). With the HSP70 promoter, we observed a leaky promoter activity under standard rearing conditions. We first tested the swallowtail butterfly P. xuthus as a species relatively closely related to B. mori. As P. xuthus displays two different mimicry body patterns on the larval epidermis during development, it is suitable for analyzing molecular mechanisms of mimicry and its evolution (Futahashi and Fujiwara, 2008) (Fig. 4A, left). We conducted electroporation in the same manner as for B. mori larval epidermis. Exogenous gene expression was observed with both promoters (20 V, second instar treatment) and was maintained even after a change in body color (Fig. 4A, 14 d, 17 d; n=1). Simultaneous electroporation of siRNA against EGFP (30 V, third instar treatment) attenuated EGFP expression in P. xuthus (Fig. 4B, dotted square; n=2), suggesting that transposon-mediated stable expression and direct transfer of siRNA are applicable to P. xuthus.

We subsequently tested the red flour beetle T. castaneum as a species more distantly related to B. mori. We electroporated the larvae before the last instar in essentially the same manner as described above (20 V, 5 pulses). Seven days post-electroporation, EGFP fluorescence was primarily observed in muscles, and DsRed2 fluorescence was observed in a subset of the EGFP-positive tissues (Fig. 4C, 7 d). The lower DsRed2 signal may be attributed to the low core activity of DmHSP70 promoter reported in this insect (Schinko et al., 2010). However, both of the fluorescence signals were maintained even after pupation (Fig. 4C, 11 d, arrows; n=7). These data suggest that a stable expression of piggyBac can also be achieved in Tribolium. The survival rates and gene transfer efficiencies in these insects are listed in supplementary material Table S1.

Our electroporation-mediated approach enabled gain-of-function and loss-of-function analyses in various tissues of three insect species, including two non-model insects. Using the piggyBac transposon, stable transgenic somatic cells were generated within different tissues at various developmental stages. In this system, the time from experimental design to obtaining results is very short; therefore, functional analyses of many genes in non-model insects may become possible. Generation of genetic mosaics will allow functional analyses of embryonically lethal signaling molecules and the study of neighboring cell interactions in non-model insects. We believe that this method could be a powerful tool for developmental genetic analyses in insects.

We thank Dr T. Tamura (National Institute of Agrobiological Sciences, Japan) for providing us plasmids pHA3PIG, pPIGA3GFP and pBac3xP3DsRed2; Dr K. Kawakami (National Institute of Genetics, Japan) for plasmid pT2KhspGFF; and Drs T. Kojima and J. Yamaguchi (The University of Tokyo, Japan) for helpful discussion.