A practical transgenesis method must allow the delivery of relatively large cargos of several kilobases. To test the efficiency of hyPBase, we generated a piggyBac donor plasmid with a 4.4 kb insert with both a transgene and a transgenesis marker (Figure 1B). The cassette consisted of the mScarlet red fluorescent protein flanked by promoter and 3′UTR regions of the nanos-O gene Plodia homolog (nos-O), a germline determinant selected on its apparent specificity to gonadic tissues (Nakao and Takasu, 2019; Xu et al., 2022). The nos-O_prom::mScarlet component was an attempt to drive a fluorescent marker into the germline, as an exploratory experiment for future driving of Cas9 in germ cells to facilitate genome editing (Xu et al., 2022)— this did not yield positive results for this current study (see Discussion). As a transgenesis marker, we used a 3xP3::EGFP marker that labeled ocular tissues during previous somatic piggyBac transformation attempts in Plodia (Bossin et al., 2007). First, we injected this plasmid without hyPBase mRNA to control for episomal expression of the 3xP3::EGFP driver. These injections showed strong EGFP expression in large internal cells 48 h post-injection, suggesting episomal expression from the embryo vitellophages (Figure 2A). However, this signal was lost in 72-h old embryos, which only showed background levels of fluorescence or external autofluorescence artifacts at injection sites (Figure 2B). Thus, episomal expression of injected plasmids dissipates by 72 h of embryonic development and should not interfere with the screening of successful integration events at this stage and onwards.

We then co-injected the donor plasmid pBac[3xP3::EGFP; nosO::mScarlet] with a hyPBase mRNA and monitored somatic transformation efficiencies throughout the G₀ generation. In order to facilitate the screening of fluorescence, all experiments were performed in the Pi_wFog white-eyed strain that is devoid of screening pigments in eye tissues and also shows increased larval translucency (Heryanto et al., 2022). Transformed embryos and first instar hatchlings showed ocellar and glial EGFP fluorescence (Figures 2C, D), with 23.7% of injected eggs showing EGFP in 72-h embryos (Table 1). Injections produced viable larvae with persistent ocellar fluorescence, as well as eye fluorescence in pupae and adults (Figures 2E, F). Over several replicated experiments, we found that 16% of injected eggs resulted in pupae, of which 18.6% were EGFP⁺. Taking into account occasional pupal failure observed in normal rearing conditions, we determined that 2.5–3% of injected eggs become viable and fertile G₀ somatic transformants.

Next, we tested germline transmission by back-crossing G₀ EGFP⁺ individuals to uninjected stock (Table 2). Out of six fertile pairs, 50% yielded EGFP⁺ G₁ progeny, suggesting a practical level of germline mobilization among G₀ founders. This result is mitigated by the fact that only six out a total of 16 single-pair matings (37.5%) were fertile and generated offspring in our conditions, suggesting that single-pair matings have limited success in our conditions (see Discussion). This establishes a germline efficiency rate of 0.94% (Figure 1D, GTE = 6/16 × 2.5% G₀ founders), meaning that for 1000 G₀ embryos injected, 9.4 embryos will survive as fertile founders passing the transgene to the G₁ generation.

As we wanted to assess whether hyPBase would allow the rapid isolation of single-insertion lines, we needed to test if transgenes were integrated into multiple copies per G₀ gamete, or if they could cause sterility. EGFP⁺ G₁ individuals (N = 3) were back-crossed (Table 2) and produced a mean of 61 EGFP⁺ adults out of 124 emerged G₂ per cross (49.3%), showing no statistical difference from an expected 50% ratio of a single insertion event (0.06< X ² < 0.46; df = 1; 0.10 < p < 0.80). Likewise, a total of five subsequent in-crosses (G₂ EGFP⁺ x G₂ EGFP⁺) each resulted in positive offspring ratios close to the expected 75% (0.06< X ² < 0.44; df = 1; 0.507 < p < 0.80). Of note, the Plodia-nosO:mScarlet transgene failed to drive red fluorescent signals detectable by epifluorescent and confocal microscopy in dissected ovaries, and we will explore the activity of alternative germline-driving promoters in the future (Nakao and Takasu, 2019; Xu et al., 2022). Overall, these data demonstrate that hyPBase provides practical transformation rates for a relatively large cargo insert, with at least 2.5% of injected zygotes yielding potential founders ready for isogenic line establishment after only one or two generations of backcrossing.

To expand the toolkit of transgenesis markers, we sought to test the activity of a pBac[3xP3::DsRed] donor vector for the screening of red eye fluorescence. We used the pHD-DsRed plasmid available through Addgene (Gratz et al., 2014, 2015), which carries a 1,146 bp 3xP3::DsRed-SV40 cassette tightly flanked by piggyBac internal repeats. Control injections without hyPBase mRNA revealed weak episomal expression in vitellophages and red background fluorescence (Figure 3A). Injection sites, which show non-specific autofluorescence under EGFP filter sets, do not fluoresce in the red channel. HyPBase-mediated insertion of pBac[3xP3::DsRed] resulted in glial signals in 14.4% of injected 72-h AEL (after egg laying) embryos, but intriguingly, no signal in the head region. Likewise, larval transformants showed sporadic signals in abdominal regions, seemingly nervous ganglia, but these patterns were always mosaic (Figure 3C). About 25/30 G₀ DsRed⁺ pupae (83%) exhibited DsRed expression in the body (Figure 3D, G₀ DsRed^body). DsRed fluorescence in the head region was observed in only five G₀ pupae (Figure 3D, G₀ DsRed^eye), but its expression failed to reproduce the 3xP3::EGFP signal pattern in ocelli and eye tissues (Figure 2D).

The presence of DsRed in abdominal regions suggested successful integration of the donor plasmid including in tissues close to the germline. To evaluate the germline transmission in G₀ DsRed^eye individuals, we backcrossed DsRed^body individuals to the uninjected stock. Only two out of 14 G₀ DsRed^eye backcrossed pairs gave G₁ progeny (Table 2), and none inherited any DsRed fluorescence expression. In contrast, we recovered eggs from five G₀ DsRed^body individuals that were incrossed liberally in a container, and showed full embryonic 3xP3::DsRed signals (Figure 3E). This salvaged stock resulted in six G₁ pupae with DsRed expression in the eyes (Figure 3E) out of 52 isolated G₁ pupae (11.5%), with no body phenotype observed. These six G₁ DsRed⁺ Plodia were then individually crossed with Pi_wFog, three of which generated 83%, 70%, and 67% G₂ DsRed⁺ G₂ progeny (Figure 3F). As these ratios deviate from the 1:1 ratio expected in these crosses, we conclude that more than one insert occurred in the parental G₀ founder germline. Finally, we used Splinkerette PCR (Potter and Luo, 2010; Shao and Lok, 2014) to map the piggyBac insertions in the G₃ generation and found a single insertion site in the genome (Supplementary Table S1). We infer that a second insertion was either eliminated by chance when establishing the G₃ line, or undetected due to the set of restriction enzymes used for digesting the genome.

The 3xP3 activity in this DsRed donor plasmid showed inconsistent results not seen with the EGFP donor, including absence of activity in G₀ eye tissues, unusual abdominal fluorescent patches in G₀ pupae, and reduced activity in G₁ eyes. Intriguingly, full 3xP3::DsRed activity was recovered in G_2-3 pupae, suggesting possible epigenetic effects of transient nature in earlier generations. This unusual behavior may be due to minor differences in the cassette proximal promoter (Supplementary Figure S1), to the compact design of this cassette (Figure 1B), or to other sequence features making the insert prone to abnormal expression.

We co-injected the pBac[3XP3::EYFP; attP] plasmid (Stern et al., 2017) with hyPBase mRNA into Pi_wFog. This donor includes an attP docking site (Figure 1B), a feature that may facilitate genetic engineering using site-specific recombination, if successfully integrated into the Plodia genome. Control injections show little background autofluorescence and vitellophage signals under the EYFP filter set (Figure 4A). Transgenic G₀ embryos and larvae showed strong somatic 3xP3 activity consistent with ocular and glial expression (Figures 4B, C), with expected mosaic variations such as unilateral expression in one side of ocellus glia and ocelli-only expression. We recovered 14 pupae with mosaic G₀ EYFP expression (Figure 2D) from a total of 63 surviving pupae, out of 797 embryos injected over two trials (Table 1).

To estimate the efficiency of germline integration from these mosaic founders, we individually backcrossed the 14 EYFP⁺ G₀ adults to single Pi_wFog individuals (Figure 4E; Table 2). Seven pairs gave progeny, among which only one cross generated progeny with 22 G₁ EYFP⁺ pupal phenotypes out of 37 total isolated pupae, a ratio statistically close to the 50% proportion expected from a germline tissue heterozygous for a single insertion in the G₀ founder (X ² = 1.32; df = 1, p = 0.25). To test if positive G₁ individuals were heterozygous carriers for a single insertion, we simultaneously backcrossed 12 G₁ EYFP⁺ to Pi_wFog and in-crossed five pairs of G₁ EYFP⁺. Three of these crosses resulted in G₂ EYFP⁺ progenies, with 59% and 48% positive ratios matching the 50% expected from backcrossing (0.18< X ² < 3.17; df = 1; 0.07 < p < 0.67), and a 71% positive ratio matching the expected 75% in the in-cross (X ² = 0.84, df = 1 p = 0.36). In summary, injection of pBac[3XP3::EYFP; attP] had a somatic transformation efficiency of 1.8%. The high level of mosaicism in G₀ resulted in only one out of 14 successful backcrosses, resulting in a germline transformation efficiency of 0.13% (Figure 1D, GTE = 7.1.% x 1.8% G₀ founders), but this event was successfully carried into a stable transgenic line. Similar to 3xP3::DsRed, Splinkerette PCR revealed a single piggyBac insertion into the Plodia genome in the G₃ generation (Supplementary Table S1).

In this study, we carried out somatic and stable germline transformation in Plodia interpunctella using the hyperactive piggyBac transposase (Yusa et al., 2011), and achieved high rates of somatic transformation with three independent piggyBac donor plasmids. We injected the transposase as a mRNA and used hyPB ^apis , a version of hyPBase codon-optimized for honeybees (Otte et al., 2018). Because Apis and Plodia both have an average GC content of around 35% (Jørgensen et al., 2007; Kawahara et al., 2022), we can reasonably expect compatibility in their codon usage biases.

Our study is the second to use an hyPBase mRNA as a transposase for transgenesis in a lepidopteran insect after the fall armyworm Spodoptera frugiperda (Chen and Palli, 2021). Plodia injections generated 15–40% of G₀ somatic transformants when observed in 72-h embryos (mean of 22%), suggesting highly efficient integration. Importantly, while we expect a higher efficiency of hyPBase based on a previous report in Diptera (Eckermann et al., 2018), we have not directly compared the efficiency of hyPBase compared to pBase using our injection set-up and conditions here.

Across different trials, a mean of 2.5% of injected eggs expressed the transgene marker as adults, representing 30% of surviving adults. However, somatic fluorescence in the injected generation does not guarantee that the transgene has transposed into the germline, or that transgenic gametes are fertile. To assess transgene inheritability into the G₁ generation, we backcrossed G₀ fluorescent founders to non-transgenic individuals. We obtained three independent G₁ lines expressing 3xP3::EGFP out of six fertile G₀ crosses, and one line expressing 3xP3::EYFP out of seven fertile G₀ crosses. Founders expressing 3xP3::DsRed showed unusual patterns of G₀ mosaicism, possibly due to epigenetic regulatory effects (see Results section), and failed to propagate the transgene when mated in single outcrossing pairs (N = 14), but we recovered a stable insertion from G₁ eggs that had been laid in a container where five G₀ founders had been left to mate randomly, meaning that one out of 19 G₀ transmitted 3xP3::DsRed.

In summary, our hyPBase mRNA-based injections in Plodia resulted in germline transformation efficiency rates of 0.18% (DsRed), 0.25% (EYFP), and 0.94% (EGFP). For comparison, Plutella transgenic experiments using pBase have efficiency rates of 0.43–0.65% (Gregory et al., 2016). Our Plodia injection protocol has a median pharate survival of 9%, much lower than the published Plutella adult survival rate of 27.8% (Gregory et al., 2016). Indeed, our injection methods favor speed and quantity over precision, using relatively wide-open needle bores that avoid clogging during injections, as well as a rapid but aggressive glue-based egg sealing procedure (Heryanto et al., 2022). Only 10–25% of eggs injected with piggyBac reagents hatched across trials in our conditions—as opposed to 21–60% in a previous Plodia microinjection report conducted by another group (Bossin et al., 2007) — but this is balanced by the fact that a single experimenter can inject about 400 pre-blastoderm embryos in a 2 h session with our procedure. Overall, the germline efficiency rates reported here mean that one fertile G₀ founder was obtained for every 106 (EGFP), 555 (DsRed), and 777 (EYFP) injected embryos, making a 2–4 h injection effort (400–800 eggs) reasonably well suited for initiating each transgenic line attempt. Ultimately, practicality boils down to a trade-off between the number of injected embryos and their survival, and our data suggest that the high efficiency of hyPBase (Yusa et al., 2011; Eckermann et al., 2018) can make transgenesis feasible if one of these two factors is not optimal.

Mendelian segregation patterns observed at the G₂ generations indicate that all four out of five stable lines originated as single-insertion events, with G₀ founders likely carrying a single copy (Table 2). This feature can be used by experimenters to use various crossing strategies in the future, but we must highlight that single-mating strategies and crossing conditions resulted in few successful pairings in our initial attempts (e.g. 11–50% of G₀ crosses, Table 2). This artificially lowered germline transmission rates, likely due to founders failing to mate in small containers in suboptimal conditions. As we gained experience with Plodia husbandry during these experiments, we increased mating success rates to 66–78% in subsequent generations (see Methods for the optimized procedure). Furthermore we recommend to mix one transgene carrier with two to three wild-type unmated adults of the opposite sex instead of one, as this maximizes the likelihood of successful mating in this system (Brower, 1975; Huang and Subramanyam, 2003).

Each of the three constructs we tested provided complementary information. The EGFP construct was the largest and delivered the highest germline transformation rate. Of note, the compact DsRed construct resulted in unusual G₀ fluorescent patterns. While circumstantial, these observations bode well for using large inserts, and we caution that the pBac[3xP3::DsRed] (pHD-DsRed, Addgene #64703) has a more compact minimal promoter that might also explain its weaker expression (Supplementary Figure S1). The candidate germline driver of mScarlet, consisting of the proximal promoter and 3′UTR of nanos-O (Nakao and Takasu, 2019; Xu et al., 2019) cloned from the Plodia genome, failed to drive detectable fluorescence in ovarian tissues. We will investigate alternative germline promoters in the future (Xu et al., 2022), for instance by testing the PhiC31 site-specific integrase at the attP docking site from our new EYFP transgenic line (Yonemura et al., 2013; Haghighat-Khah et al., 2015; Stern et al., 2017; Stern, 2022). Both EGFP and EYFP showed robust and strong 3xP3-driven expression at all generations, without noticeable decrease over time in adult eyes (Das Gupta et al., 2015), with EYFP benefiting from lower autofluorescence effects than EGFP at various stages. We strategically used a white mutant strain deficient for eye-screening pigment, as routinely done in other insects to facilitate the screening of 3xP3-driven fluorescence (Stern et al., 2017; Klingler and Bucher, 2022), and this mutation also increases the translucency of Plodia larvae (Shirk, 2021; Heryanto et al., 2022). Of note, white mutations can be recessive-lethal in some lepidopteran species (Khan et al., 2017). Until alternative ways to generate depigmented eyes are found, this may limit the usefulness of 3xP3 drivers, especially in species where eggs and larvae are opaque and where screening becomes limited to narrow developmental windows (Das Gupta et al., 2015; Özsu et al., 2017). In such species, we suggest that stronger, more ubiquitous promoters of viral origin such as Op-ie2 and Hr5-ie1 may be more practical for transgenic screening (Martins et al., 2012; Xu et al., 2019, 2022). In future hyPBase transgenesis experiments, we intend to test the activity of viral promoters, potential ubiquitous promoters, and putative tissue-specific drivers for marking tissues such as the nervous system, wing epithelia, silk glands, and hemocytes. With these tools in hand, Plodia is well positioned to complement other organisms with functional genomics capacity like Bombyx and Plutella, due to its suitability for mass rearing, synchronized egg collection, and long-term maintenance of inbred lines.

The Pi_wFog strain (Heryanto et al., 2022) consists of an introgression of the recessive w-mutation (Shirk, 2021) from the Pi ^w- strain (origin: USA, kind gift of Paul Shirk), into the genetic background of the “Dundee” strain (origin: United Kingdom, kind gift of Mike Boots). Genome assemblies of both Pi ^w- and Pi_Dundee parental strains are available (Roberts et al., 2020; Kawahara et al., 2022). The resulting hybrid Pi_wFog strain has been maintained in inbred state for 3 years and used throughout this study. All rearing used previously published methods (Heryanto et al., 2022), using special containers and a wheat bran-sucrose-glycerol diet (Silhacek and Miller, 1972). A rearing temperature of 28°C resulted in a generation time of 28 days.

The pBac[3xP3::EGFP; Tc’hsp5′-Gal4Delta-3′UTR] (Addgene plasmid # 86449) was used as a donor plasmid with piggyBac insertion repeats and the 3xP3::EGFP reporter (Schinko et al., 2010). To generate pBac[3xP3::EGFP; nosO_prom::mScarlet-nosO_3′UTR], an mScarlet cassette preceded by 2 kb of promoter sequence immediately upstream of the Plodia nanos-O start codon, was synthesized in the pUC-GW-Amp backbone by Genewiz and sub-cloned into the FseI and AscI restriction sites of pBac[3xP3::EGFP; Tc’hsp5′-Gal4Delta-3′UTR]. The pBac[3xP3::DsRed] (pHD-DsRed) and pBac[3XP3::EYFP; attP] plasmids were obtained from Addgene (#64703, and #86860) and used without modification (Gratz et al., 2014, 2015; Stern et al., 2017). All the 3xP3-driven fluorophore genes included an SV40 termination sequence.

The pGEM-T_hyPB ^apis plasmid encodes a hyPBase that was codon-optimized for honeybees (Otte et al., 2018). The source plasmid was purified using the QIAprep Spin Miniprep Kit (Qiagen), linearized with NcoI-HF (New England Biolabs) and concentrated using acetate/ethanol precipitation. Around 500 ng of linearized template were transcribed using the mMESSAGE mMACHINE™ T7 ULTRA Transcription Kit (Invitrogen) and purified using the MEGAclear Transcription Clean-Up Kit (Invitrogen). After quantification with Nanodrop (Thermofisher), the solution was divided into 1,050 ng/μL one-time use aliquots and stored at −80°C.

Injection mixes consisted of 400 ng/μL hyPBase mRNA, 200 ng/μL donor plasmid, and 0.05% cell-culture grade Phenol Red (Sigma-Aldrich). Donor plasmids without hyPBase mRNA were injected in separate experiments as controls. Microinjection procedures (Figure 1A) followed a previously described procedure (Heryanto et al., 2022), with all embryo injections performed within 40 min after egg laying (AEL). Injected embryos were counted and kept in a rearing container with a small damp Kimwipe at 28°C. For the first 72 h, the container vent was covered with tape in order to maintain humidity saturation, a parameter that prevents egg desiccation. After 72 h, the vent was opened and the Kimwipe removed, and about five flakes of Plodia food added next to the eggs, in order to keep the emerging larvae within the injection dish. Mean emergence time of the Pi_wFog strain is 83 h AEL at 28°C for uninjected eggs, and is delayed by injection stress to 100–115 h AEL. Because of this variability, we report times of observation after injection in hours rather than in relative percentages.

Larvae and adult Plodia were anesthetized in tissue culture dishes positioned over a cold metal block during microscopy observation. All pictures were taken under the Olympus SZX16 stereomicroscope equipped with a Lumencor SOLA Light Engine SM 5-LCR-VA lightsource or standard stereomicroscope brightfield lamp, and with a trinocular tube connected to an Olympus DP73 digital color camera. Separation of fluorescent channels was performed using Chroma Technology filter sets ET-EGFP 470/40 × 510/20 m, ET-EYFP 500/20 × 535/30 m, and AT-TRICT-REDSHFT 540/25x, 620/60 m.

Embryonic survival rates (“egg hatching” rates) were determined by the ratio of hatched eggs at 120 h AEL over the number of injected eggs (N _inj ). Empty egg shells were counted for this purpose instead of first-instar hatchlings, which are difficult to count accurately in the presence of food. Pharate survival rates were determined by the ratio of pupae obtained from a given injection experiment, divided by N _inj , and thus accounts for mortality occurring at embryonic and larval stages. Pupal mortality was negligible, making pharate survival rates a reasonable proxy for overall adult survival, and is more convenient to couple to fluorescent screening than in mobile adults. G₀ transformation rates were independently measured in embryos and in pupae. For embryos, eggs with bright, internal fluorescent signals consistent with an ocellar or glial expression were counted as positive (fluorescent, F⁺ in Table 1) around 72 h AEL, and non-fluorescent eggs were counted as negative (F^‒). To isolate individual pupae, cardboard strips that are preferentially used as pupation sites (“hotels”) were added into containers containing fifth instar larvae, allowing a convenient isolation of individual Plodia pupae. Pupae were then extracted from these lodges and aligned on double-sided tape for fluorescence screening. Pupae with any glial or eye signal were counted as positive, while others were counted as negative. G₀ somatic transformation efficiency rate was determined as the number of healthy adult individuals emerged from fluorescent pupae, and normalized by N _inj .

Germline transformation efficiency rates factored the somatic transformation efficiency rate by the proportion of attempted G₀ backcrosses yielding transgenic offspring. G₀ transgenic adults or late pupae exhibiting positive fluorescent signals (G₀ F⁺) were crossed to a single unmated Pi_wFog adult of the opposite sex, by mixing in a 1.25 oz Plastic Souffle Cup (Solo) containing ∼0.2 g of diet and ∼1 cm² of paper towel. Pi_wFog outcrossing mates were replaced if found dead before any visible egg laying. These cups were monitored for up to 2 weeks for any larval emergence, after which they were transferred to a vented rearing container with a bed of Plodia food (modified LocknLock containers described in Heryanto et al., 2022; 177 ml and 350 ml formats). At the wandering L5 stage, cardboard “hotels” were added into the containers for pupal isolation. G₁ pupae with positive fluorescent signal were counted and backcrossed to an unmated Pi_wFog with the same procedure stated above. The resulting G₂ pupae were in-crossed as sib-matings and maintained as isogenic stock in the G₃ generations and henceforth.

The Genomic DNA of 3 separate transgenic G₃ Plodia adults carrying [3xP3::DsRed] or [3XP3::EYFP; attP] was isolated using the Quick-DNA Tissue/Insect Kits (Zymo Research) in 20 µL DNA Elution Buffer, RNAse-treated, and quantified by fluorimetry. Following the Splinkerette PCR protocols (Potter and Luo, 2010; Shao and Lok, 2014), 100 ng of the isolated DNA was digested with BfuCI (New Englands Biolabs). Two rounds of Splinkerette PCR were done using Q5 High-Fidelity DNA Polymerase (New England Biolabs) with the following cycle parameters: 98°C for 1 min; 30 cycles of 98°C for 20 s, 67°C (Tm of the primers targeting the insertion 5’ end) for 20 s, and 72°C for 2 min; and 72°C for 10 min. The PCR products were purified using the PureLink PCR Purification Kit (Thermofisher) prior to Sanger sequencing (Supplementary Table S1).