The isolation of the medfly homologs of the cellularization genes sry α and nullo by degenerate primer PCR using an embryonic cDNA pool was not successful 9. Thus, we carried out PCR-based cDNA subtractions of different embryonic stages and identified several cellularization-specifically-expressed genes (Figure 1). The genes C.c.-slow as molasses (C.c.-slam; Figure 1A), C.c.-sub2_99 (Figure 1B), C.c.-CG2186 (Figure 1C), C.c.-serendipity α (C.c.-sry α; Figure 1D), C.c.-sub2_63 (Figure 1E), and C.c.-sub2_65 (Figure 1F) are expressed specifically during medfly blastoderm cellularization (Figure 1). None of the genes shows maternal expression or expression at later stages, except C.c.-sub2_63, which is additionally expressed during germ band elongation (Figure 1E3).

By inverse PCR, we isolated the P/Es from C.c.-slam, C.c.-sub2_99, C.c.-CG2186, C.c.-sry α, and C.c.-sub2_63 containing about 0.4 to 1.9 kb of the complete 5'UTR and upstream sequences. The isolated P/Es were fused to the tetracycline-controlled transactivator gene tTA and used to engineer different driver constructs (sl1-tTA, sl2-tTA, 99-tTA, CG2186-tTA, sryα 2-tTA, and 63-tTA) embedded into piggyBac vectors carrying polyubiquitin (PUb)-driven DsRed as a germline transformation marker 10. Additionally, three effector constructs were generated (TREp-hid^Ala5, TREhs43-hid^Ala5, and >TREp-hid^Ala5>) carrying the lethal factor hid^Ala5 under control of either P element 11 or hsp70 basal promoters 12 from D. melanogaster. In the >TREp-hid^Ala5> construct the lethality inducing transgene is flanked by gypsy insulator elements (> = gypsy element in 5'-3' orientation), which should reduce position effect-caused variable expression 13. Except for sl1-tTA, all constructs carry a minimal attachment P (attP) site 14, which potentially enables site-specific phiC31-integrase-mediated integration to modify transgenes at successfully evaluated genomic positions 15.

Five driver constructs (sl1-tTA, sl2-tTA, 99-tTA, CG2186-tTA, and sryα 2-tTA) and the three effector constructs were used for germline transformation of medfly. For each construct we obtained transgenes of which we further analyzed a maximum of three independent lines (Figure 2 and Figure 3).

We crossed 12 independent homozygous driver lines with five independent homozygous effector lines to generate 60 different combinations (a crossing scheme presenting data on all 60 trans-heterozygous combinations is shown in Additional File 1). From each combination, we collected eggs to visualize the early expressed tTA and proapoptotic gene hid^Ala5 by in situ hybridizations. The lethal activity of each combination was checked by a second egg collection, which was counted for eggs and progeny. To describe the dimension of lethality we henceforth use the term 'complete lethality' for 100% lethality in laboratory experiments. The lethality was checked in this trans-heterozygous condition, since the heterozygous condition must suffice to induce lethality when homozygous males are crossed with WT flies in intended applications. All combinations that showed detectably lower or no progeny under these trans-heterozygous conditions were then inbred to generate homozygous (for both driver and effector construct) lethality lines (LLs). In addition four combinations with non-reduced progeny were made homozygous for comparison (see Additional file 1).

All LLs expressed tTA specifically during cellularization. However, due to the different P/Es as well as integration sites, the tTA expression strength varied and resulted in different expression strengths of hid^Ala5 (Figure 2 and Figure 3). This resulted in variable efficiencies of the lethality system. LLs deriving from the same driver line showed similar expression levels of tTA. The P/Es sl1 (0.4 kb; Figure 2 A1–3) and 99 (Figure 2 C1–3) mediated only very weak expression of tTA, which subsequently could not induce detectable levels of hid^Ala5 expression. The longer P/E region of sl2 (1.9 kb) was able to drive tTA and, indirectly, hid^Ala5 (Figure 2 B1–6), but the level of hid^Ala5 expression was not high enough to drive complete lethality (Figure 2). With the P/E CG2186 we obtained a strong level of tTA expression during cellularization (Figure 2 D2), which started the expression of hid^Ala5 during the cellularization stage (Figure 2 D5) and led to complete pupal lethality of LL #68 (Figure 2).

Besides the finding that different P/Es or P/E regions act differently on tTA and the dependent hid^Ala5 expression, also the integration site of the driver construct could influence the tTA expression (Figure 3). Three independent lines, carrying the driver construct with the sry α P/E at different integration sites, expressed the tTA specifically but with different strengths during cellularization (Figure 3 A2–C2). In LL #65, a weak expression of tTA led to a late expression of hid^Ala5 during germ band retraction, which was not able to drive complete lethality (Figure 3). In contrast, the LLs #66 and #67 expressed tTA strongly during cellularization (Figure 3 B2, C2), which activated hid^Ala5 expression at the cellularization stage (Figure 3 B5, C5) and led to complete L1 larval lethality for LL #66 and complete embryonic lethality for LL #67 (Figure 3). Thus, a strong tTA expression seems to be important to start the hid^Ala5 expression early enough to cause complete embryonic lethality.

In addition, the effector constructs with different basal D. melanogaster promoters or different integrations of the same effector construct influence the levels of hid^Ala5 expression and lethality. The effector constructs TREp-hid^Ala5 and >TREp-hid^Ala5>, carrying the p-basal promoter, were able to express hid^Ala5 in medfly after activation through the 12 independent driver lines, but did not cause complete lethality in 36 different combinations (Additional file 1). Interestingly, the effector construct TREhs43-hid^Ala5, which carries the basal promoter (43 bp) of D. melanogaster hsp70, showed differences in the expression strength of hid^Ala5 depending on the integration site of the construct. In comparison with the larval or embryonic LLs #66 or #67, which are derived from the effector line TREhs43-hid^Ala5_F1m2, the hid^Ala5 expression in LLs #29 and #72 deriving from TREhs43-hid^Ala5_F1m1 started at later stages (Figure 3 D6, E6) and was not sufficient to drive complete lethality at the embryonic, larval, or pupal stage. Thus, we can show that the choice of E/Ps as well as the integration site of both the driver and the effector is crucial to set up a successful embryonic lethality system.

Transgenic driver and effector lines were preliminarily screened for homozygous condition by fluorescence patterns and intensity. Southern blots indicated single copy integrations of driver and effector constructs of the lines sryα 2-tTA_F4m1, sryα 2-tTA_M2m1, TREhs43-hid^Ala5_F1m1, and TREhs43-hid^Ala5_F1m2 (Figure 4). Moreover, the correct piggyBac-mediated integrations at canonical TTAA target sites were verified by isolation of 5' and 3' insertion site sequences by inverse PCR (see Additional file 2). Therefore we know that differences in expression strength and functionality of the lethality system in different LLs are not a result of multiple insertions of the driver or effector constructs, but must be due to position effects. Furthermore, the integration sites of the driver and effector construct for LLs #66 and #67 were mapped by chromosome spreads. We found the driver and the effector of LL #66 or LL #67 located on chromosome 5 at the positions 74B and 70B or 63B and 70B, respectively (Figure 5). Thus, both LLs #66 and #67 have been results of recombinations. LL #66 was more difficult to establish than LL #67, corresponding to an expected lower recombination frequency between loci 74B and 70B than between loci 63B and 70B (Figure 5).

Females feeding on adult food containing a suppressor, tetracycline (Tc) or doxycycline (Dox), should be able to transfer the suppressor molecules maternally into the oocytes. After egg laying, this maternally contributed suppressor should bind to the tTA protein produced during early embryogenesis, blocking the trans-activating potential of tTA and thus the switch of the lethality system. To identify the minimal concentrations of Tc needed to rear the LLs #29, #72, #66, #67, and #68, flies were bred on larval and adult media containing different concentrations of Tc. We defined the optimal Tc concentration in adult and larval medium for rearing as the lowest possible amount of Tc combined with the highest possible number of descendants. The LLs #72, #66, and #67 could be reared efficiently on adult medium containing 10 μg/ml Tc and LL #29 even on 1 μg/ml Tc. All LLs could be reared on larval medium containing 1 μg/ml Tc, except LL #68 (10 μg/ml Tc). LLs #66 and #67 were in addition tested to be efficiently rearable on larval medium containing only 0.1 μg/ml Tc or Dox. When reared on larval medium lacking Tc, all lines showed a reduction of progeny. Rearing LLs #66 and #67 on adult food containing 100 μg/ml Dox resulted in maternal suppression of the lethality, even without Tc or Dox in the larval food. Hatching and pupation rates were comparable to WT flies reared on adult and larval food without Dox. Thus, Dox is able to maternally suppress the embryonic lethality. However, eclosion rates were still reduced when Dox was used in adult food only (data not shown).

In addition all lines and WT showed delayed ovary development and 5–7 day postponed egg laying when reared on adult medium containing 100 μg/ml Tc and larval medium containing 300 μg/ml Tc. This indicates the importance of reducing the Tc concentrations to a minimum for the efficient rearing of medfly lines.

To test for reversible sterility of lines #66 and #67, adults were reared on Tc-containing medium (10 μg/ml) for two days (Figure 6A). After transfer to Tc-free medium the rate of progeny decreased in five days to 0%. The sterility could be reversed by retransfer of the adults to Tc-containing medium. The reduced rate of progeny after this procedure could be due to a slight irreversible effect of the lethal system or to the advanced age of flies, as shown in other studies 16.

During medfly SIT programs, irradiation sterilized males are released into affected areas and mate with WT females, which leads to infertile matings. Ideally all progeny die as embryos to exclude damage to fruits from larval feeding. To show the efficacy and time point of lethality for the newly generated LLs, transgenic males (homozygous for driver and effector) from LLs #29, #72, #66, #67, #68, or WT were crossed with WT females, respectively (Figure 6B and Additional file 3). For the LLs #29 and #72, about 20% of the eggs survived to become L1 larvae, whereas pupae and adult progeny were highly reduced. Crossings with males from LLs #66, #67, and #68 showed complete pupal lethality but varying larval and embryonic lethality. Only 0.8% of the laid eggs from the LL #66-crossing hatched and all of those died during L1 larval stage. The LL #67-crossing showed the desired complete embryonic lethality.

An ideal line for releasing purposes should be embryonic lethal, but should also be competitive. We therefore did competition tests with LLs #66 and #67 (Figure 6C). WT females were crossed with WT males and transgenic males in different ratios (1:1:1, 1:1:3, 1:1:5, and 1:1:9; these numbers represent the ratio of WT females: WT males: transgenic males). The reduction of progeny compared with WT-only controls showed that both lines are highly competitive. Remarkable is the higher fertilization success of LL #67 males compared with WT males starting from ratio 1:1:5. For the ratio 1:1:9 an overall progeny rate of only 0.4% was measured. At the same time a WT control at ratio 1:10:0 gave only little reduction of overall progeny (Figure 6C). Thus, transgenic males from LLs #66 and #67 performed in laboratory competition tests comparably or even better than WT males. Progeny from all competition tests were identified as non-transgenic individuals by fluorescent microscopy, which additionally indicated the complete lethality of LLs #66 and #67. Interestingly, all lines deriving from the effector line TREhs43-hid^Ala5_F1m2 (#66, #67, and #68) partially lacked anterior orbital bristles, which did not obviously interfere with the mating success of these transgenic males. In addition to laboratory tests, field cage tests 17 with LL #67 males showed a comparable or even better competitiveness than WT (Figure 7).

WT and transgenic medfly lines were maintained under standard rearing conditions 25. The WT strain Egypt-II (EgII), which has been reared in laboratories for more than 25 years, was obtained from the FAO/IAEA Agriculture and Biotechnology Laboratory (Entomology Unit, Seibersdorf, Austria). Laboratory tests for efficiency, competition, and reversibility of the lethality system were performed in 15 × 15 × 20 cm acrylic cages. The Argentinean WT strain (Arg; from Mendoza, Argentina) was reared for about 20 generations under relaxed artificial rearing conditions (Entomology Unit, Seibersdorf, Austria) and then used for the field cage tests.

The Clontech PCR-Select cDNA Subtraction Kit (BD Biosciences, Heidelberg) was used to isolate fragments of the following genes expressed specifically during cellularization as described 9: C.c.-slam, C.c.-sub2_99, C.c.-CG2186, C.c.-sub2_63, and C.c.-sub2_65. An EST fragment of the medfly cellularization gene serendipity α (C.c.-sry α) was received from Dr. Ludvik Gomulski, Pavia. By RACE, 5' and 3' ends of cellularization-specific genes were isolated using the BD SMART RACE cDNA Amplification Kit (BD Biosciences, Heidelberg) and gene-specific primers. Complete cDNA sequences are deposited to GenBank [accession numbers: FJ460697 (Cc_slam), FJ460699 (Cc_sub2–99), FJ460701 (Cc_CG2186), FJ460705 (Cc_sub2–63), FJ460706 (Cc_sub2–65), FJ460703 (Cc_sry α)].

Inverse PCR was performed to get the 5' regions of genes specifically expressed during cellularization: 1.5 μg of medfly WT genomic DNA was digested for 24 h; restriction fragments were precipitated and self-ligated in a volume of 500 μl at 16°C for 24 h; PCR was performed on circularized fragments by using primer sequences (for a complete list of primers see Additional file 4) in opposite orientation within the 5' UTR or ORF of the genes. First, PCRs (1 min at 95°C; 6 cycles of 30 sec at 94°C, 45 sec at 66°C (-2°C each cycle), 6 min at 68°C; 25 cycles of 30 sec at 94°C, 45 sec at 54°C, 6 min at 68°C; and 6 min at 68°C) for C.c.-slam, C.c.-sub2_99, C.c.-CG2186, C.c.-sry α or C.c.-sub2_63 were performed on FspBI, NdeI, CviAII, PvuI or AclI cut genomic DNA with the primer pairs mfs-77/-79, mfs-85/-108, mfs-170/-172, mfs-159/-161 or mfs-83/-104, respectively, using BD Advantage 2 PCR (BD Biosciences, Heidelberg). Second, the obtained PCR products were diluted 1:50 with ddH₂0 and nested PCRs with primer pairs mfs-78/-80 (C.c.-slam), mfs-160/-162 (C.c.-sryα) or mfs-171/-173 (C.c.-CG2186) were performed (1 min at 95°C; 22 cycles of 30 sec at 94°C, 45 sec at 54°C, 6 min at 68°C; and 6 min at 68°C) using 5 μl of the dilution and the BD Advantage 2 PCR Kit (BD Biosciences, Heidelberg). PCR products from first (C.c.-sub2_99 and C.c.-sub2_63) and nested PCRs (C.c.-slam, C.c.-sry α and C.c.-CG2186) were cloned into pCRII vectors (Invitrogen, Karlsruhe) and sequenced. Upstream sequences are deposited to GenBank [accession numbers: FJ460696 (upstream_Cc_slam), FJ460698 (upstream_Cc_sub2–99), FJ460700 (upstream_Cc_CG2186), FJ460704 (upstream_Cc_sub2–63), FJ460702 (upstream_Cc_sryα)].

To localize the integration sites of piggyBac vectors, inverse PCR was performed with primers and protocols as described 26. Sequences flanking piggyBac insertions are shown in the Additional file 2.

Generally we compose our constructs in the cloning shuttle vector pSLfa1180fa. From the shuttle vectors the constructs can be easily placed in transformation vectors which carry FseI and AscI sites (fa-sites 27).

The pSLaf_attP-sl2-tTA_af (#1231), pSLaf_attP-63-tTA_af (#1232), pSLaf_attP-99-tTA_af (#1234), pSLaf_attP-sryα 2-tTA_af (#1236), and pSLaf_attP-CG2186-tTA_af (#1237) carry a 52 bp attP site 28. #1231, #1232, or #1234 was created by ligating annealed attP primers (mfs-201/-202) in the EcoRI cut pSLaf_sl2-tTA_af (#1210), pSLaf_63-tTA_af (#1211) or pSLaf_99-tTA_af (#1212), respectively. #1236 or #1237 was created by ligating annealed attP primers (mfs-203/-204) in the NcoI cut pSLaf_sryα 2-tTA_af (#1225) or pSLaf_CG2186-tTA_af (#1226), respectively.

#1210, #1211 or #1212 was created by ligating the EcoRI-XbaI cut sl2 fragment (a 1.9 kb 5'-region of the gene C.c.-slam), the EcoRI-Eco31I cut 63 fragment (a 1.2 kb 5'-region of the gene C.c.-sub2_63) or the EcoRI-XbaI cut 99 fragment (a 0.7 kb 5'-region of the gene C.c.-sub2_99), amplified by PCR on genomic DNA with primer pairs mfs-141/-113, mfs-142/-143 or mfs-131/-133, in the EcoRI-XbaI cut pSLaf_tTA_af (#1215), respectively. #1225 or #1226 was created by cloning the NcoI-XbaI cut sryα 2 fragment (a 1.6 kb 5'-region of the gene C.c.-sry α) or the NcoI-Eco31I cut CG2186 fragment (a 1.2 kb 5'-region of the gene C.c.-CG2186), amplified with primer pair mfs-189/-188 or mfs-190/-191, in the NcoI-XbaI cut #1215, respectively. #1215 was generated by cloning a 1.5 kb XbaI-HindIII cut tTA-SV40 fragment from pTetOff (Clontech, CA) in the XbaI-HindIII cut pSLfa1180fa 27.

The driver construct pBac{sl1-tTA_PUb-DsRed} (sl1-tTA) was generated by ligating the BglII/XbaI cut sl1 (a 0.4 kb 5'-region of the gene C.c.-slam amplified with primer pair mfs-112/-113 from genomic DNA) and the XbaI/BglII cut tTA-SV40 (a 1.5 kb region amplified with primer pair mfs-110/-111 from pTetOff) in the BglII site of pB [PUbDsRed1] 29.

The driver constructs pBac{f_attP-sl2-tTA_a_PUb-DsRed} (sl2-tTA), pBac{f_attP-63-tTA_a_PUb-DsRed} (63-tTA), pBac{f_attP-99-tTA_a_PUb-DsRed} (99-tTA), pBac{f_attP-sryα 2-tTA_a_PUb-DsRed} (sryα 2-tTA) or pBac{f_attP-CG2186-tTA_a_PUb-DsRed} (CG2186-tTA) were generated by ligating the FseI-AscI fragment attP-sl2-tTA, attP-63-tTA, attP-99-tTA, attP-sryα 2-tTA or attP-CG2186-tTA from #1231, #1232, #1234, #1236 or #1237 in the FseI-AscI cut pBac{fa_PUb-DsRed} (#1200 16), respectively.

The effector constructs pBac{fa_attP_f_TREp-hid^Ala5_a_PUb-EGFP} (TREp-hid^Ala5) or pBac{fa_attP_f_TREhs43-hid^Ala5_a_PUb-EGFP} (TREhs43-hid^Ala5) were generated by cloning the hybridized primers mfs-211/-212 in the XmaJI site of pBac{faf_TREp-hid^Ala5_a_PUb-EGFP} (#1207) or pBac{faf_TREhs43-hid^Ala5_a_PUb-EGFP} (#1208), respectively. #1207 or #1208 were created by ligating the AscI fragments TREp-hid^Ala5 (5.0 kb) or TREhs43-hid^Ala5 (4.9 kb) from pSLfa_TREp-hid^Ala5_fa or pSLfa_TREhs43-hid^Ala5_fa 6 in the AscI site of pBac{fa_PUb-EGFP} #1201 16, respectively. The effector construct pBac{>fa_attP_f_TREp-hid^Ala5_a>_PUb-EGFP} (>TREp-hid^Ala5>) was generated by ligating the AscI-fragment attP_f_TREp-hid^Ala5 from TREp-hid^Ala5 in the AscI-site of pBac{>fa>_PUb-EGFP} 16.

Germline transformation experiments were performed by microinjection of piggyBac constructs (500 ng/μl) together with the phspBac transposase helper plasmid (200 ng/μl) 30 into WT embryos as described by Handler and James 31 with the following exceptions: injected eggs were covered with Voltalef 10S oil (Lehmann & Voss, Hamburg, Germany), placed at 28°C in parafilm closed Petri dishes with watered Whatman paper in the lid; neither eggs, larvae nor pupae were heat shocked.

The vectors sl1-tTA, sl2-tTA, 99-tTA, sryα 2-tTA, CG2186-tTA, TREp-hid^Ala5, TREhs43-hid^Ala5, or >TREp-hid^Ala5> were injected into 600 embryos of which 260, 140, 160, 54, 83, 28, 63, or 52 survived to adulthood, respectively. Four female crossings (two to 25 G₀ females crossed with 15 WT males; F1–F4) and four male crossings (two to 25 G₀ males crossed with 15 WT females; M1–M4) were set up for each construct. G1 progeny were screened by epifluorescence for the expression of the PUb-DsRed or PUb-EGFP. Fluorescent progeny with different red or green patterns were backcrossed twice to WT to recognize possible multi-insertions and brought to homozygous conditions by inbreeding and checking fluorescence intensity. For screening of flies we used the fluorescence stereomicroscope Leica MZ16 FA with the filters DsRedwide (Ext. 546/12; Emm. 605/75) and EYFP (Ext. 500/20; Emm. 535/30).

To generate LLs, we crossed 12 homozygous driver lines and five homozygous effector lines in all possible combinations. We inbred those heterozygous combinations, which produced detectably lower or no progeny, screened the progeny by fluorescent intensity for homozygous individuals, and subsequently inbred these homozygous individuals to generate LLs homozygous for driver and effector construct.

Genomic DNA (~3–10 μg) from adult flies of different transgenic lines and the WT strain was digested with BamHI (Roche, Mannheim) and separated on 1% agarose gels. DNA was transferred to nylon membranes (Hybond-N⁺; GE Healthcare/Amersham, Little Chalfont) and immobilized by UV irradiation. Probe labeling and membrane hybridizations were performed according to the AlkPhos Direct kit (GE Healthcare, Little Chalfont). Signal detection was performed using CDP-star (GE Healthcare, Little Chalfont) followed by exposure for approximately 30 min on Kodak Biomax ML film.

The two probes for detecting DsRed or EGFP were amplified by PCR (2 min at 94°C; 30 cycles of 30 sec at 94°C, 30 sec at 53°C, 1 min at 72°C; 5 min at 72°C) from the constructs #1200 or #1201 with the primers mfs-333 and mfs-334 or mfs-335 and mfs-336, respectively.

WMISH with RNA probes to embryos were performed as described 32. RNA antisense probes were prepared by in vitro transcription with the DIG-RNA-Labeling Kit (Roche, Mannheim) from pCRII vectors (Invitrogen, Karlsruhe) containing subtraction cDNA fragments (p_slam, p_99, p_CG2186, p_63, p_65), an EST fragment (p_sryα), and the plasmids pBSK-hid^Ala5 or pBSK-tTA 6. By PCR using the primer pair mfs-41/-42, cDNA fragments were amplified and transcribed with Sp6 polymerase. The plasmids pBSK-hid^Ala5 or pBSK-tTA were linearized with ClaI or EcoRI and transcribed with T3 or T7 RNA polymerase, respectively.

Chromosome in situ hybridizations were performed with slight modifications as described 33. Instead of horseradish peroxidase, the Biotin/Avidin system VECTASTAIN Elite ABC was used (Vector laboratories, Peterborough). Hybridization sites were identified and photographed using 60× oil objectives (Olympus phase contrast microscope) with reference to medfly salivary gland chromosome maps 34. Squash preparations of salivary gland polytene chromosomes were made as described 33. A DNA-probe recognizing piggyBac insertions was prepared by PCR on genomic DNA from flies carrying a DsRed-marked piggyBac insertion 29 with the primers DsRed_F and DsRed_R (1813 bp) using the Biotin High-Prime kit (Roche Diagnostics, Mannheim).

Tc-free adult and larval food was used. In four independent repetitions, virgin WT females were crossed directly after eclosion with homozygous males from LLs #29, #72, #66, #67, #68, or WT, respectively. Four days later a 24 h egg collection was taken from each crossing. For statistical analyses on the reversibility, efficiency, and competition tests please see Additional file 5.

15 WT females and 15 WT males were placed together with different numbers of males of LLs #66 or #67 (15 (1:1:1) – 135 (1:1:9)). For control matings, 15 virgin WT females were crossed with either 15 WT males (1:1:0) or 150 WT males (1:10:0). Two independent crossings were performed for each ratio. Six 24 h egg collections were obtained from each crossing and the number of adult progeny was recorded. Adult progeny were verified by fluorescence light microscopy as WT or transgenic offspring.

Males from LL #67 (non-irradiated or irradiated with 120 Gy 48 hours before adult emergence) were competed against non-irradiated WT Argentinean males for mating with Argentinean WT females in a field cage (field cage size (W-D-H): 3 m-3 m-2.3 m; 17). Pupae from the different strains/treatment were placed in emergence cages, and every 24 h adults were removed, sorted by sex, and placed in cages with adult food (3:1, sugar:hydrolyzed yeast) and water for 6 days. Two days before the tests, flies were marked with a dot of water-based paint on the thorax (DEKA^®, Unterhaching, Germany). In each field cage, three potted Citrus aurantius trees, 1.6 m in height with 1.5 m diameter canopy, were used as a mating arena. To follow the quarantine protocol, tests were performed in a greenhouse with controlled temperature (24–26°C) and humidity (60–80%). On the day of the test, 20 sexually mature non-irradiated Argentinean males, 20 non-irradiated, and 20 irradiated males from line LL #67 were released into the cage around 08:30. Approximately 20 min later, 20 virgin and sexually mature Argentinean females were released in the cage. Tests lasted 3 hours. Mating pairs were collected as they formed by allowing the pair to walk into a small vial. The type of mating couple was recorded and the proportion of mating was calculated for each mating type (see Figure 7). After the couples separated, the males and females were identified and the mated females were grouped together depending on the type of mated male and transferred to small egging cages. Eggs were collected for five consecutive days and transferred to small Petri dishes with moist black filter paper. After four days of incubation, hatched larvae and un-hatched eggs were counted to determine the egg hatch for each mating type. Twelve replications of this test were carried out.

Starting from larval and adult media Tc-concentrations of 100 μg/ml, the minimal Tc concentrations for the LLs #29, #66, #72, #67, and #68 were tested with WT as a control. First, flies were reared on adult medium containing 100 μg/ml Tc and eggs were collected on larval medium containing 0, 1, 3, 10, 30, 100, or 300 μg/ml Tc. Hatching, pupation and eclosion rates were recorded. Second, the adult medium Tc concentrations (1, 3, 10, 30, 100, or 300 μg/ml) were tested over three generations by using the optimized larval media concentrations (1 μg/ml for LLs #29, #66, #72, and #67; 10 μg/ml for LL #68) in between the adult stages.

Additionally, LL #66 and LL #67 were reared on adult media containing 10 μg/ml Tc and eggs were collected on larval medium containing 0.1, 0.3, and 1 μg/ml Tc or Dox. Hatching, pupation, eclosion rates, and viability of progeny from the eclosed flies were recorded.

Moreover, freshly eclosed flies from the LLs #66 and #67 were placed on adult food containing 100 μg/ml Dox. Four and six days after eclosion an egg collection was taken and the eggs reared on Dox-free larval medium. Hatching, pupation and eclosion rates were recorded and compared with data from WT (EgII) flies reared on Dox-free adult and larval food.