hobo episomes were recovered during in vivo hobo mobility assays performed in developing D. melanogaster embryos. They were obtained following the simultaneous introduction into developing D. melanogaster of 'donor' plasmids containing a hobo element carrying a kanamycin resistance gene, an origin of replication and the E. coli lacZ alpha peptide coding region and 'helper' plasmids containing the hobo transposase open reading frame under the control of a heat-inducible promoter. Following recovery from developed embryos approximately 18 hours post-injection, recovered DNA was digested with restriction endonucleases KpnI and XbaI. hobo episomes did not contain KpnI and XbaI restriction sites and were resistant to digestion by these restriction endonucleases while donor and helper plasmids were cleaved multiple times. Introduction of the digested DNA into E. coli permitted the recovery of only episomal forms of hobo. hobo episomes were readily recovered under these conditions and for approximately every 1000 donor plasmids recovered from injected embryos six hobo episomes were isolated ( = 0.0061, SEM = 0.0040, n = 7)). hobo excision products consisting of donor plasmids with an empty hobo integration site were also recovered during the same experiments at a rate of approximately three in every 1000 recovered donor plasmids ( = 0.0031, SEM = 0.0006, n = 5). Analysis of variance indicated that there was no significant statistical difference between the frequency of recovery of hobo episomes and empty hobo donor sites (F_1,10 = 0.3831, P = 0.5497).

A variety of specific forms of hobo episomes were recovered from plasmid-based mobility assays, differing in the sequence of the junction between the left and right terminal sequences (Table 1). Of the 38 episomes recovered from plasmid-injected embryos and whose sequence was determined, 15 (39%) contained intact left and right inverted terminal repeats and contained all of the information necessary for transposition. Three of these 15 'intact' episomes were perfect end-to-end joints of the left and right terminal inverted repeats. The remaining 12 'intact' episomes contained from 1 to 20 nucleotides between the inverted terminal repeats. Three of these episomes had intercalary DNA with sequences related to the inverted terminal repeats of hobo (ie. TTCTTCT, AGAACTTCTCTG, ATGCGGCTGCAGTTCTCTG). The remaining 23 (61%) 'defective' hobo episomes were missing one (19) or both (4) of the inverted terminal repeats and had variable amounts of their sub-terminal sequences deleted. These deletion-containing episomes are not expected to be transpositionally competent.

Hermes episomes were recovered from two transgenic D. melanogaster containing integrated, self-mobilizable Hermes elements (Hermes7011 and Hermes198H70-1) (Figure 1). Episomes in insects containing Hermes7011 were detected using a nested PCR strategy using two pairs of Hermes terminus-specific primers oriented so that PCR products will only arise when the termini are joined end-to-end (see Methods, Figure 1). The DNA sequence was determined for a sample of the cloned products arising from this PCR reaction (Table 2). Thirty-five percent (9/26) of the episomes recovered using this method had intact inverted repeats and all of the sequence information necessary for transposition. Only two of the recovered episomes were perfect end-to-end joins. Thirty-one percent (8/26) had only intact left inverted terminal repeats with variable amounts of the right end being deleted while 19% (5/26) had a reciprocal structure with intact right inverted terminal repeats and variable amounts of the left end being deleted. Fifteen percent of the recovered episomes were missing both inverted terminal repeats and adjacent sub-terminal sequences.

Hermes episomes were also recovered from transgenic D. melanogaster using a plasmid rescue strategy. Autonomous Hermes elements containing a kanamycin resistance gene and an origin of replication permitted Hermes episomes to function as replicons in E. coli. Sixty Hermes episomes were recovered from adults and larvae of four independent 'lines' of transgenic D. melanogaster (Table 3). Approximately 50% (29/60) contained 'intact' Hermes elements with complete inverted terminal repeats. Only three (5%) were perfect end-to-end joins of the inverted terminal repeats while the remaining 26 contained variable amounts of intervening nucleotide information ranging from 1 bp to 1 kb. All other episomes contained 'defective' Hermes elements lacking either one or both inverted terminal repeats with variable amounts of sub-terminal sequences deleted; 20% had intact left inverted terminal repeats and 25% had intact right inverted terminal repeats. Hermes episomes were also detected in unfertilized D. melanogaster eggs by PCR (Table 3). The structure of episomes found in unfertilized eggs was generally similar to those found in other life stages although episomes with only the right inverted repeat (53%) were recovered more frequently in unfertilized eggs than from adults and larvae (25%). Approximately half (8) of the episomes contained complete elements including perfect end-to-end joints and those with a small number of nucleotides between the inverted terminal repeats.

Hermes episomes were also detected by PCR and recovered by plasmid rescue from transgenic Aedes aegypti containing the same autonomous Hermes element as was in the transgenic D. melanogaster described above (Hermes7011). Episomes were recovered from developing embryos, larvae and from adult ovarial tissue and resembled those recovered from D. melanogaster in their sequence (Table 4).

Hermes is a natural inhabitant of the genome of M. domestica and has been detected in all individuals (n = 65) sampled from 13 populations from four continents [36]. Because the Hermes element in these populations did not naturally contain a functional origin of replication, episomes could not be recovered by plasmid rescue and could only be detected using a PCR-based method that relied on the use of PCR primers specific to the right and left terminal sequences of the element that would result in PCR products only when the termini were joined end to end. Total genomic DNA isolated from individual M. domestica from three populations had evidence of episomal Hermes elements as indicated by the recovery of PCR products using end-specific primers (Figure 2B). The most abundant forms were perfect (434 bp) or near perfect (> 434 bp) end-to-end joints containing all of the information necessary for transposition (Figure 2B; Table 5). As in transgenic D. melanogaster, 'defective' forms were also recovered with variable amounts of terminal sequences deleted from PCR products less than 434 bp (Figure 2B). All 'defective' forms recovered were missing the right inverted terminal repeat and, in one case, both termini were absent (Table 5). Finally, episomal Hermes elements were detected in unfertilized eggs of M. domestica (Figure 2C).

The recombination potential of a variety of Hermes episomes was tested directly using a cell free Hermes transposition assay with purified Hermes transposase, episomal Hermes elements and a target plasmid. Hermes episomes with n nucleotides of intercalary DNA between the inverted repeats were tested where n was 0, 1, 4, 5, 17, 37,69 and 120 nucleotides. All episomal forms of Hermes tested were capable of transposition resulting in canonical 8 bp target site duplications with the sequence nTnnnnAn (Table 6 and Figure 3). The distribution of integration events within the target indicated that certain sites were preferred as we have previously described (Figure 3) although the apparent hot spots of integration in vitro are not the same as reported in vivo (Table 6) [37,38]. Integrations at nucleotide 94 of the target plasmid were recovered frequently however it is not known whether the primary nucleotide sequence of this target site (ATTGAGAT) is the major determinant of this site's preferred status.

Efforts to create transgenic D. melanogaster using Hermes gene vectors constructed from episomal forms of the element ('episomal Hermes gene vectors') resulted in the successful creation of transgenic insects in some cases. Germ-line transformation experiments were performed using well-established protocols (see Methods) involving the co-injection of vector and transposase-expressing helper plasmids into preblastoderm embryos followed by screening for expression of the dominant visible genetic marker present on the vector in the next generation. Episomal Hermes gene vectors containing 0, 17 and 37 bp of intercalary DNA separating the inverted terminal repeats did not result in the recovery of germ-line transformation events (n = 303; Table 7). On the other hand, identical episomal Hermes gene vectors with 80 bp and 160 bp of intercalary DNA separating the inverted terminal repeats did result in germ-line transformation in 22% (n = 23) and 28% (n = 54) of the germ-lines tested (Table 7). Integrated Hermes elements from these transgenic insects were analyzed and in all cases (n = 16) the Hermes element was found precisely inserted into D. melanogaster genomic DNA (Table 8). The 8 bp immediately flanking the left inverted terminal inverted repeat of the integrated Hermes elements conformed to the known target site consensus sequence of this element (Table 8) [37,38]. While we were unable to detect integration of episomal Hermes gene vectors with 0 bp or 17 bp of intercalary DNA, we discovered that these forms of the element actually inhibited transposition of canonical Hermes elements, i.e. elements with typical spacing of the inverted terminal repeats. Germ-line transformation experiments involving the co-injection of episomal Hermes gene vectors with 0 bp or 17 bp of intercalary DNA and canonical Hermes gene vectors (HermesActin5CEGFP or Hermes3 × P3EGFP which both contain over 2 kb of intercalary DNA and are effective gene vectors) failed to produce any transgenic insects (n = 306). Identical experiments using only the canonical Hermes gene vectors confirmed their functionality and transformation was observed at a frequency of approximately 10% (n = 49; Table 7).

Adults were collected from natural populations (kindly provided by Dr. Elliott Krasfur, Iowa State University) and from laboratory colonies (the laboratory strain Cs was kindly provided by Dr. Jeffery Scott, Cornel University).

Using previously reported methods, transgenic lines were created using the Hermes gene vector 198H70-1 and the host strain w¹¹¹⁸ (Figure 1) [42]. The vector 198H70-1 contained, in addition to approximately 500 bp and 700 bp of the right and left ends of the Hermes transposable element, respectively [42], the transposase open reading frame under the regulatory control of the D. melanogaster heat shock 70 (hsp70) promoter and the Enhanced Green Fluorescent Protein (EGFP) open reading frame under the regulatory control of the D. melanogaster Actin 5C promoter [45], a kanamycin resistance gene, a ColE 1 origin of replication and the E. coli lac Z alpha peptide coding region [38]. Because these transgenic insects contained an autonomous (self-mobilizing) Hermes element, stable lines could not be established and maintained. Therefore transgenic populations were maintained by selecting EGFP-expressing individuals every other generation to ensure the transgene was at a high frequency within the laboratory population. The presence of an antibiotic resistance marker and a prokaryotic origin of replication enabled this vector to be used in plasmid rescue experiments from transgenic individuals.

Transgenic D. melanogaster (w¹¹¹¹⁸) were also constructed using the Hermes gene vector 7011 [46]. This vector was identical to 198H70-1 except it lacked the kanamycin resistance gene, the origin of replication and the lacZ alpha peptide-coding region (Figure 1). Because this vector also was a self-mobilizing (autonomous) Hermes element stable lines could not be established and this line was also maintained by periodic selection of EGFP-expressing individuals, as described above.

D. melanogaster (w¹¹¹¹⁸) were used as hosts to perform Hermes and hobo plasmid-based mobility assays in developing embryos.

Using reported methods the Orlando line (a wild-type laboratory strain) was transformed with the Hermes vector 198H70-1 [47]. Because of the low level of germ-line remobilization activity of Hermes in this species, maintenance did not require periodic selection for individuals expressing EGFP as was necessary for transgenic D. melanogaster containing the same vector [48].

Genomic DNA was extracted from adult insects as described (using protocol 48 in "Drosophila: A laboratory manual", [49]) or using Wizard Genomic DNA Purification Kits (Promega, Madison Wisconsin) according to the manufacturers recommendations. Low molecular weight DNA (plasmids or episomes) was isolated according to the method of Hirt [50].

One to five μg of undigested D. melanogaster genomic DNA was used in the electroporation of DH10-β E. coli (Invitrogen, Carlsbad, California) and the treated cells were grown in 1 ml of SOC [51] at 37°C for 1 hour. Cells were then concentrated by centrifugation, suspended in approximately 100 μl of SOC and spread on LB plates containing 25 μg/ml of kanamycin. Resistant colonies were isolated and plasmid DNA extracted and digested with restriction endonucleases EcoRI and BamHI. Hermes episomes originating from 198H70-1 excision are predicted to yield BamH1 and EcoR1 fragments approximately 2.7, 2.6, 1.3, 0.6 and 0.4 kb in length depending on the exact structure of the joined ends of the element. Final confirmation of the structure of the recovered episomes was made by DNA sequencing using primers HL216 (5' GCA GGC GAC TGA GTA ACA ACA AT AAC AAC 3') and HR2500 (5' CAA TGA GTA TAC AAC ACA ACA AAG AAG TGA G 3').

Using approximately 1 μg of D. melanogaster or Ae. aegypti (each transgenic with Hermes vector 198H70-1) genomic DNA as a template, PCR was performed in 1× Taq polymerase buffer containing 2.5 mM MgCl₂, 0.8 mM dNTPs and 8 pmoles each of primers HL216 and HR2500. Following an initial step of 95°C for 3 min. 25 iterations of the following cycle were performed: 95°C/15 sec, 55°C/15 sec, 72°C/30 sec. To complete the reaction and insure that all products were fully double-stranded the mixture was incubated at 72°C for 5 min. Episomal Hermes elements are expected to yield a 470 bp reaction product if the left and right inverted terminal repeats were precisely joined end to end. PCR amplification products were purified using QIAquick PCR Purification Kit (Qiagen, Valencia, CA) and ligated to the plasmid pGEM-T Easy (Promega). Electroporation was used to introduce some of the plasmids into DH10-β by and the DNA sequences of the plasmid inserts was determined from the recovered clones.

This semi-nested-primer method was used to detect and analyze Hermes episomes in wild-type M. domestica and D. melanogaster (transgenic with Hermes vector 7011). The first reaction was performed in a volume of 50 μl with 1 mM dNTP's, 2.5 mM MgCl₂ with primers Hermes2432F (5' AAT ATA CTT ATG CTC TTT TCC 3') (for M. domestica) or Hermes2681F (AAA ATA CTT GCA CTC AAA AGG C 3') (for D. melanogaster) and Hermes333R (5' TCG GAA CAT TTT GCT GTG 3'), each at 0.6 μM and 5% of the genomic DNA from individual insects as template. The reaction conditions involved a preliminary denaturation step of 95°C for 3 minutes followed by 25 cycles of 95°C/15 sec, 56°C/1 min, 72°C/1 min. Following these 25 cycles the reactions were incubated 72°C for 5 minutes. The reaction conditions for the second reaction were identical to those of the initial reaction but primers Hermes2432F (for M. domestica) or Hermes2681F (for D. melanogaster) and Cy5-66R+ (5' Cy5-AAT GAA TTT TTT GTT CAA GTG GCA AAG CAC 3') were used with 5 μl of a 20× dilution of the initial reaction as template. Following the reaction, approximately 5 μl was size-fractionated using a high resolution electrophoresis system consisting of a 1 mm thick 6% polyacrylamide in Tris Borate buffer and 8 M urea. The resulting gel was dried onto filter paper and scanned using the 633 nM light source of a Storm 860 gel/blot imaging system (Molecular Dynamics, Piscataway, NJ) to directly visualize the Cy5-labelled PCR products.

Transposable element display permits all members of a transposable element family in the genomic DNA of individual insects to be detected and visualized as unique PCR products. Hermes transposable element display was performed here as described [46].

Pre-blastoderm D. melanogaster embryos (w¹¹¹⁸) were injected with a mixture of the plasmids pHobo8bpdrKanOriLacZ and pHspHobo (each plasmid was at a concentration of 250 μg/ml) as described [52]. Injected embryos were incubated at 25°C overnight and approximately 18 hrs post injection the embryos were placed at 37°C for 1 hr and allowed to recover for 1 hr at 25°C. After heat shock and recovery viable embryos were collected and low molecular weight DNA was recovered as described [50]. To recover episomes low molecular weight DNA was digested with restriction endonucleases Kpn I and Xba I and used to electroporate E. coli (DH 10β). Kpn I and Xba I cut pHobo8bpdrKanOriLacZ within the donor plasmid backbone but not within the hobo element. Any excised episomal forms of hobo present within the sample will be resistant to digestion and will transform E. coli to kanamycin resistance with β-galactosidase activity. Plasmids from E. coli transformants were confirmed as hobo episomes by digesting with the restriction endonuclease Sal I and finally by determining the sequence of DNA spanning the inverted terminal repeats.

To recover hobo excision events (empty donor sites) recovered plasmids were introduced directly into E. coli and selected for chloramphenicol resistance. Chloramphenicol resistant colonies that were negative for β-galactosidase activity and sensitive to kanamycin were confirmed as excision products by digestion with restriction endonuclease digestion with Kpn I and Xba I.

In vivo inter-plasmid transpositions assays have been used extensively to investigate insect hAT elements (e.g. Sarkar et al. [38]). Here the assay has been adapted to a cell-free system and consists of a donor element (episomal forms of Hermes), a target plasmid (pGDV1) and purified Hermes transposase [53]. The episomal forms of Hermes tested in this assay contained a kanamycin resistance gene, a ColE1 origin of replication and the E. coli lacZ alpha-peptide coding region [38]. The target plasmid, pGDV1, is a gram-positive chloramphenicol resistant plasmid that cannot replicate in E. coli without the addition of a species-compatible origin of replication [38,54]. Reactions were performed in 20 μl of 20 mM HEPES pH7.9, 25% glycerol, 5 mM MgCl₂, 4 μg bovine serum albumin, 2 mM DTT with 250 ng each of donor and target plasmids, ands 1 μg of purified Hermes transposase [9]. The reaction was incubated at 30°C for 1-2 hrs at which time 80 μl of stop solution (50 mM Tris pH 7.5, 0.5 mg/ml proteinase K, 10 mM EDTA and 0.1 μg/ml tRNA) was added and incubated at 37°C for 1 hour. The reaction was extracted with phenol/chloroform and the DNA was precipitated with sodium acetate and ethanol and the dried precipitate was dissolved in 20 μl of water. Fifteen microliters were introduced into 80 μl of E. coli (DH10-β) by electroporation. Following electroporation 900 μl of SOC was added and the cells were incubated at 37°C for 1 hour at which time 1 μl was plated on LB plates containing ampicilin (50 μg/ml) and X-gal (20 μg/ml) to assess plasmid recovery. The rest of the cells were divided and plated on 10 LB plates containing chloramphenicol (10 μg/ml), kanamycin (25 μg/ml) and X-gal (20 μg/ml) to select for donor transpositions into the target plasmid. Restriction mapping and DNA sequencing of putative transposition events confirmed the presence and of transposed Hermes elements.

Germ-line integration was used to assess the mobility properties of episomal forms of Hermes in vivo. A 1 kb fragment containing the coding region of the DsRed protein under the regulatory control of the 3× P3 promoter was inserted between the terminal sequences of episomal forms of Hermes that contained a kanamycin resistance gene, a plasmid origin of replication and the LacZ alpha peptide coding region. Hermes transformation vectors were created with episomal forms containing 0, 17, 34, 80 and 160 bp of intercalary DNA separating the terminal inverted repeats of the element. The Hermes gene vectors HermesActin5CEGFP and Hermes3 × P3EGFP were used as controls [45,55]. Hermes transposase was supplied by co-injecting the plasmid pHSHH1.9 along with the vector containing plasmids being tested using methods previously described [42].

Transgenic insects were subsequently analyzed by transposable element display to confirm the presence of Hermes. Integration sites were sequenced following isolation, re-amplification and sequencing of transposable element display bands as described [46]. The DNA sequence flanking the integrated Hermes elements were used in a BLAST search [56] of the D. melanogaster genome sequence in FlyBase [57]. A sequence logo [58] was created using the aligned 8 bp of DNA immediately flanking the left inverted repeat of each integrated Hermes element using WebLogo [59]