Improving genetic transformation rates in honeybees

Functional genetic studies in honeybees have been limited by transformation tools that lead to a high rate of transposon integration into the germline of the queens. A high transformation rate is required to reduce screening efforts because each treated queen needs to be maintained in a separate honeybee colony. Here, we report on further improvement of the transformation rate in honeybees by using a combination of different procedures. We employed a hyperactive transposase protein (hyPBaseapis), we tripled the amount of injected transposase mRNAs and we injected embryos into the first third (anterior part) of the embryo. These three improvements together doubled the transformation rate from 19% to 44%. We propose that the hyperactive transposase (hyPBaseapis) and the other steps used may also help to improve the transformation rates in other species in which screening and crossing procedures are laborious.


Results
We used a previously published injection and rearing procedure 23 to transform honeybee queens with transposon cassettes using the piggyBac (PB) transposon system. We first examined if we can directly target the germ cells and increase transformation rates in the offspring by injections in the posterior third of the egg. We suggested targeting this area because markers of the germ cells (the transcript of the Am-vasa and Am-nos) have been detected in cells that are arranged in a line from the 3 rd to the 6 th abdominal segment 27 . We obtained 9 queens out of 47 queens (19%; Fig. 1, Suppl. Fig. S2).
To possibly improve the transformation rate, we introduced the mutations of the hyperactive variant of the piggyBac transposase protein 24 by synthesizing a gene that encoded the hyperactive mutations and adjusted the codon usage to that of the honeybee (hyPBase apis , Table 1). We compared the activity of hyPBase apis to that of iPBase (insect piggyBac transposase from Trichoplusia ni, that is usually used in insects (reviewed in 28 )) using an excision assay in Sf21 cells, a permanent insect cell line which is routinely used in cell based assays 29 . We transfected cells with the piggyBac transposon provided by PB[Ac5C rubia] plasmid together with either the hyBPase apis or the iPBase gene under the control of the OplE2 promoter (pIZ/V5-His vector). After 72 hours, we isolated plasmid DNA and semi-quantified the amount of excised transposon using PCRs (Fig. 2). In this semi-quantitative analysis (n = 3), we detected a strong PCR product (indicating the excision of the piggyBac transposon) when the hyPBase apis gene was expressed and observed only a weak PCR product when the iPBase gene was expressed. These results provide a first indication that the hyPBase apis gene can possibly provide a higher transposition activity in insect cells compared with the iPBase gene. Next, we examined the effect of hyPBase apis on transformation rates in queens. We obtained 6 out of 27 queens (22%) (Fig. 1) with transgenic offspring.
To possibly further improve rates, we increased the amount of injected transposase mRNAs per embryo from 90 pg to 240 pg. This resulted in 31% queens with transgenic offspring (5 out of 16 queens, Fig. 1). Finally, we injected into the anterior region of the embryo (Fig. 1) expecting that this change would transform one of the first nuclei located in the anterior pole of the embryo 30 . This might be possible because our injections were performed 0 to 1.5 hours after egg deposition, while the first mitotic division occurs 3 to 4 hours after egg deposition. With the anterior injection site and the 240 pg of hyPBase apis mRNA, we detected that 25 of 58 queens (43%) had transgenic offspring (Fig. 1). Injection with 240 ng of the new hyPBase apis mRNAs at the anterior pole doubled the transformation rate as compared with injection with 90 pg of iPBase mRNAs in the posterior area (19 to 43%; d.f. = 1, fisher's exact test P < 0.05).
To find further evidence of whether the change from posterior to anterior injection site may increase the chance of targeting the germ cells by transformation of the first nuclei, we compared how often we were able to detect the transgene in the ovaries of the queens generated from posterior and anterior injections with 240 pg of  hyPBase apis mRNA. We extracted genomic DNA from ovaries and amplified the transgene using PCRs (Suppl. Fig. S2a). We observed that 53% of the 49 queens from the anterior injection and only 30% of the 50 queens from the posterior injections (Suppl . Table S1) possessed ovaries with the transgene (Fisher's exact test, P < 0.05). This result supports our hypothesis that anterior injections can more likely target the ovary tissue possibly by a transformation event in early embryogenesis. We next explored whether the improvements increase the percentage of offspring that possess the transgene. Although the proportion of transgenic offspring is in general quite large, we detected no evidence that the new procedures improved the proportion of transgenic offspring in the transformed queens (Fig. 3, cond. I-IV, Kruskal-Wallis-test, d.f. = 3, P = 0.7).

Discussion
In this study, we demonstrated a twofold increase in transformation rates for 3.8-to 5.0-kb sized transposon cassettes using the piggyBac transposon system. In a previous study, we were able to produce transgenic queens with on average less than 10% of transgenic offspring (shown in 23 , therein presented in Table 2) suggesting that this low rate is not very supportive for further genetic experiments in honeybees. In this study, we employed a codon-adjusted hyperactive transposase gene, tripled the amount of transposase mRNA in the injections and changed the injection site from the posterior to the anterior site of the egg to target the first nuclei of the embryo. The implementation of all three procedures doubled the transformation rate from 19% to 43% of the queens with transgenic offspring. These improvements suggest that the analysis of gene functions using transposon cassettes (consisting of an endogenous promotor and gene as well as the flanking pBacL and R sequences) is feasible for an average bee facility.
Among the transgenic queens of the improved condition IV a single queen had more than 70% transgenic offspring. In combination with the observation that 26% of those transgenic queens produce more than 10% transgenic offspring, our results indicate that we can produce, with the improved transformation rates, queens with a large fraction of transgenic worker or male offspring in the first generation following the injected generation. This will offer the perspective to perform genetic experiments with the first generation and to do crossing experiments that require only a limited number of resources and honeybee colonies. Contrary to our assumption, it was not possible to transform the embryo at least before the first two nuclei divisions.
Our Sf21 cell experiments and transgenic ovary studies provide support that the hyperactive transposase expressed from a codon usage-adjusted gene and that the choice of the anterior injection site contributed to the overall increase in the transformation rate in honeybees. The expression of hyPBase apis increased the number of plasmids in which the transposon was excised relative to the expression of the original iPBase. Further, the change from posterior to anterior injections increased the number of transgenic ovaries detected in the queens. We suggest that we have targeted with our anterior injections the first nuclei possibly located in that region 30 .
Similar improvements have been reported from experiments performed in other species, but not in this combination. Studies in mice demonstrated that the adjusted codon usage of the transposase gene from an insect to that of the mice increases the transposition efficiency up to 20-fold in mouse embryonic stem cells 25 , suggesting the importance of codon usage adjustments. The hyperactive transposase variant developed by Yusa et al. for mice increased the transformation rate by 10-fold 24 . Studies in the silk worm Bombyx mori reported that changes in the injection site and direction increased the transformation rates from 2.1% to 6.5%, suggesting the germ cells were better targeted 26 . We propose that any of the steps -the use of hyPBase apis , the choice of injection site, and the amount of the injected transposase mRNA -or a combination of all three procedures can enhance transformation rates in other insect species, which have been reported to vary from 0.1% to 15% (reviewed in 28,31 ). The further increased transformation rates would be especially rewarding for species in which the maintenance of females and the selection procedure in the offspring is as laborious as it is in honeybees.

Materials and Methods
Microinjection and rearing. Honeybee eggs (age: 1.5 hours after egg deposition) were injected and larvae reared to queens as previously described 23 . To induce egg laying, the treated queens were 9 d old when they were incubated with CO 2 for 10 min on two successive days. One operator injected 30 pg of plasmid DNA in the years 2013-2015 and 10-15 pg in the year 2017 together with transposase mRNAs as previously described 23 . The plasmids used in this study (Suppl. Fig. S1) had a similar structure to those previously reported ([6xP3-rubia; Am-actin5c-egfp], 23 ). The transposon cassettes (including pBacR and pBacL sequences) in this study ranged in size from 3.8 to 5.0 kb. Bee sources. The bees were feral colonies of the A. mellifera carnica strain. DNA Preparation, PCR, Nucleotide Analysis and mRNA Synthesis. Genomic DNA from honeybee drones was extracted from viable larvae at the 1-2 instar stage or from the dissected ovaries of the reared queens using the peqGOLD Tissue DNA Mini Kit (PEQLAB, Erlangen, Germany). Because in some queens more than 16 male (drone) offspring were analysed, it was necessary to standardize the percentages of transgenic offspring per queen to 16 for statistical analysis. Plasmid preparation, PCR reactions, use of restriction enzymes, sequencing and transposase mRNA synthesis were performed as previously described 23 .
Cloning of the hyPBase apis gene in pGEMT and pIZ/V5-His vector. The nucleotide sequence of hyP-Base apis (hyperactive and codon adjusted piggyBac transposase) was synthesized, and the DNA was cloned into the pUC57 vector (Centic Biotech, Weimar, Deutschland). We replaced the original transposase gene 32 in pGEM-T iPBase 23 with the hyPBase apis sequence (nucleotide changes are shown in Table 1) using the ApaI/NcoI restriction sites. The resulting plasmid was named pGEM-T hyPB apis . Transposase mRNA was synthesized following the instructions of the mMESSAGE mMACHINE kit (Thermo Fisher Scientific, Dreieich, Germany). For the expression of the unmodified (iPBase) and the hyperactive (hyPBase apis ) PB transposases under the control of the OplE2 promoter in Sf21 cells, we cloned the coding sequences into the pIZ/V5-His vector (Thermo Fisher Scientific, Dreieich, Germany) using the SacII/EcoRI restriction sites. Biochrom, Berlin, Germany) at 27 °C in six-well plates and maintained according to the manufacturer's instructions. We transiently transfected 1 µg of piggyBac [Ac5C rubia] plasmid into 10 6 cells using Roti Insectofect (Roth, Karlsruhe, Germany) in combination with 1 µg of either pIZ/V5-His iPBase or pIZ/V5-His hyPBase apis transposase-expressing plasmid. After 72 hours, we isolated the plasmid DNA using the GeneJET Plasmid Miniprep Kit (Thermo Fisher Scientific, Dreieich, Germany).
Excision assay. PCR oligonucleotide primers that amplified a 180-bp sequence if the transposon was excised ( Fig. 2) were P1: CAGACAGCGTTGAGATATAC and P2: CAATGTGGTTTTTGTCAAACGAAG. The PCR oligonucleotide primers that amplified a fragment upstream of the Amp R gene, which was not affected by the excision event, were bb1: CGACGTGTTGGCTAAAATTATTAAA and bb2: GCTGCAAGGCGATTAAGTTGGGTA.