Introduction

The presence of active transposable elements (TEs) in the genome is an important factor responsible for genetic instability. As much as 3.8% of euchromatin (Kaminker et al., 2002) and up to 22% of total DNA (Kapitonov and Jurka, 2003) in Drosophila melanogaster is composed of TEs. The spontaneous rate of TE transposition in Drosophila is about 10−3–10−5 per site per generation depending on the genetic background and environmental conditions (Harada et al., 1990; Nuzhdin et al., 1996; Vieira and Biemont, 1997; Lampe et al., 1998; Maisonhaute et al., 2007).

P and hobo transposons invaded the D. melanogaster genome not earlier than 100 years ago (Blackman et al., 1987; Bazin et al., 1999). According to the hypothesis of Rouzic and Capy (2005), the transposition rate can vary by several orders of magnitude during the colonization process. They propose that the rate is high during the first generations following the horizontal transfer, but then decreases until an equilibrium value is reached corresponding to the known measured transposition rates (Rouzic and Capy, 2005).

TEs are now widely used to generate genome-wide mutant collections. P elements have been the vehicle most often employed to disrupt Drosophila genes because they transpose at the highest rates (Bellen et al., 2004). The minimal rate of transposition of another newcomer, hobo, was evaluated as 1.75 × 10−3 per site per generation in the strains studied by Harada et al. (1990). However, Aulard et al. (2004) show that a hobo copy artificially introduced into the fly genome previously free of hobo insertions transposes at the rate of 0.19–0.45 per site per generation.

The isogenic strain y cn bw sp of D. melanogaster is used by many researchers as the reference strain, in large part, because a complete genome sequence is available (Adams et al., 2000; Myers et al., 2000). This strain carries full-sized copies of hobo, Dm412 and mdg1 elements capable of inducing transpositions. It is, therefore, tempting to study the stability of TE patterns both for the newcomer hobo transposon and the ancient Dm412 and mdg1 retrotransposons in this strain, and determine transposition rates for those that are active. This may be of interest for understanding genome instability from an evolutionary perspective.

We found that the number of hobo hybridization sites on polytene chromosomes of the y cn bw sp genome is higher than that annotated in the database (http://flybase.bio.indiana.edu). In contrast, Dm412 and mdg1 hybridization sites largely coincided with those annotated in silico.

Materials and methods

Fluorescence in situ hybridization (FISH) experiments were performed with y[1] oc[R3.2]; cn[1] bw[1] sp[1]; LysC[1] MstProx[1] GstD5[1] Rh6[1] strain (hereafter referred to as y cn bw sp) obtained from Bloomington Drosophila Stock Center (USA) in summer 2004 and summer 2005. Flies of the isogenic strain y cn bw sp of D. melanogaster are suitable for the analysis of mobility because this stock has been made isogenic and labeled with recessive phenotypic markers. Thus, all heterogeneity should be due to novel instability as opposed to drift of preexisting polymorphism or contamination.

The cytological localization of hobo element, Dm412 and mdg1 in the nearly completely sequenced genome was performed with the use of FlyBase Insertions query results (http://flybase.net/transposons). Information concerning hobo (or H-element), Dm412 (or 412) and mdg1 (or mdg1) in sequenced genome is designated as H{}# (2003), hobo{2005}#, 412{}# and mdg1{}#, respectively in the database.

FISH was performed on the squashed preparations of larval salivary glands. We used 4–5 squashed preparations for every type of probe. Slides were heated at 60°C for 1 h and denatured in 0.07 M NaOH for 3 min. The completely cloned copies of hobo transposon (provided by J Lim), Dm 412 and mdg1 (provided by N Fedorova) were used as a probe. The probe was labeled by nick translation with biotinylated dUTP (Medigen, Novosibirsk, Russia), Cy3-UTP (Roche oligolabeling kit) or digoxigenin-dUTP (Molecular Probes, Invitrogen, Eugene, OR, USA). The in situ hybridization solution was as follows: 50% formamide, 10% dextransulfate, 4 × SSC, 1 × Denhardt solution, 0.1 M phosphate buffer (pH 7.6) and 10–20 ng labeled DNA per slide. After hybridization, the squashed preparations were washed with 2 × SSC three times for 5 min. The detection of biotin was performed with avidin-fluorescein isothiocyanate (avidin-FITC), digoxigenin was detected by antidigoxigenin-Cy3 (Molecular Probes). Vectashield with 4',6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Inc., Burlingame, CA, USA) was used as an antifading solution.

Following the FITC-DNA hybridization and FITC/Cy3-DNA hybridization, the preparations were examined with an Axioskop-2 Plus microscope equipped with a black-and-white CCD VC44 camera (PCO). The images were processed with the ISIS program (METSYSTEMS GmbH). The result of Cy3-labeled Dm412 DNA hybridization was analyzed on Zeiss Axiophot2 microscope equipped by Zeiss Axio Cam HRm. The images were processed with the AxioVision version 3.1.

Results

We compared the in silico localization sites of hobo, Dm412 and mdg1 TEs in the y cn bw sp strain of D. melanogaster to our FISH data. Since the correspondence between genome sequence and the cytological map is known (with a polytene chromosome bandwidth accuracy) in silico data can be used to localize a mobile element on the cytological map, to allow this comparison.

By 2003, as many as 24 hobo sequences had been annotated in the database for y cn bw sp strain. One of these annotated sequences represents the full-sized copy (2959 bp) encoding the active transposase (accession number M69216), while the other hobo copies carry deletions in the central part (http://flybase/). A 1406-bp-long variant is a predominantly truncated version with high homology to a full-sized hobo. According to the database, the hobo element is distributed non-uniformly on the chromosomes. Thus, the X chromosome contains five copies while 2L, 2R, 3L, and 3R carry 10, 2, 1 and 6 hobo hybridization sites, respectively.

In comparison to the database localizations, the number of FISH hobo sites is almost twofold higher (46 sites against 24); X chromosome has 7; 2L, 13; 2R, 3; 3L, 6; 3R, 17 hobo hybridization sites (Figure 1 and Table 1). For example, out of five annotated X-linked sites, two sites from chromocenter region were lost, while we were able to determine four novel sites in the regions 12EF, 13EF, 16A and 18EF (Figure 2). Some sites from 3R chromosome are polymorphic (Table 1).

Figure 1
figure 1

Fluorescence in situ hybridization (2004) of fluorescein isothiocyanate-labeled hobo DNA on the salivary gland polytene chromosomes from y cn bw sp strain. Arrows show chromosome arms.

Full size image
Table 1 Distribution of hobo transposable element on y cn bw sp polytene chromosomes according to FISH (2004 and 2005) and in silico data
Full size table
Figure 2
figure 2

Fluorescence in situ hybridization (2004) of fluorescein isothiocyanate-labeled hobo DNA on the salivary gland X chromosome from y cn bw sp strain. Black arrows show annotated positions; gray arrows, new sites; double arrows, lost sites.

Full size image

The same FISH experiments performed a year later (in 2005) revealed the additional hobo pattern changes in X and 3R chromosomes, namely, the loss of 18EF site and generation of a de novo X-linked site in the 16EF region. Novel 9A and 14C hobo sites were observed in a single slide out of five analyzed (Figure 3 and Table 1). The loss of four annotated sites and generation of one new site occurred in 3R chromosome (Figure 4 and Table 1). In addition, we observed the case of hobo localization polymorphism in 93 region of 3R chromosome (Figure 5). Gray arrows show hobo positions in different homologs on the different sides of mdg1 in the 93rd region. No hobo-pattern changes were observed in the other chromosomes in 2005 experiments as compared to 2004 (Table 1). In 2005, the database was supplemented with 39 additional defective hobo elements that had less homology with full-sized hobo and were much shorter than those annotated in 2003 (most of them being 100–300 bp derivatives). However, those additional potential hobo hybridization sites on X chromosome were annotated to the regions 11B, 20B, 20D1 and 20D2, whereas in our experiments, hybridization was observed in the other locations indicated above. Only three out of six annotated hobo sites (in regions 84B, 84D and 90D) are present in the 3R chromosome in our FISH data. Fourteen novel hobo hybridization sites also do not colocalize with those annotated additionally in silico in 2005 (Tables 1 and 2). In addition, those six sequences annotated in silico in 2005 represent short hobo derivatives with a low homology, thus, probably, being below FISH sensitivity (Table 2). We, therefore, believe that the novel hobo hybridization sites detected by FISH most likely appeared as a result of de novo transpositions.

Figure 3
figure 3

Fluorescence in situ hybridization (FISH; 2005) of fluorescein isothiocyanate-labeled hobo DNA (black arrows) and Cy3-labeled mdg1 DNA (3CF and 11C) on the X chromosome from y cn bw sp strain. Gray arrow shows the position of the new hobo sites in comparison with FISH (2004).

Full size image
Figure 4
figure 4

Fluorescence in situ hybridization (2005) of fluorescein isothiocyanate-labeled hobo DNA and Cy3-labeled mdg1 DNA on the salivary gland polytene chromosomes from y cn bw sp strain. Black arrows show chromosome arms, gray arrows show mdg1 hybridization sites.

Full size image
Figure 5
figure 5

Fluorescence in situ hybridization (2005) of fluorescein isothiocyanate-labeled hobo DNA and Cy3-labeled mdg1 DNA on 3R chromosome from y cn bw sp strain. Arrows show positions of mdg1 hybridization sites (85E, 93F, 98C). The polymorphic hobo hybridization pattern is shown in 93 regions.

Full size image
Table 2 Size and cytological localization of hobo sequences in 3R chromosome of y cn bw sp strain according to in silico data
Full size table

We were able to detect 20 mdg1 hybridization sites as compared to 29 annotated sites (Figure 4 and Table 3). It is noteworthy that mdg1 sites are found more often then hobo in the same band or neighboring bands of the polytene chromosome. In FISH experiments, this results in a single visible hybridization site instead of two sites actually present. In such twin cases, the loss of one copy might go unnoticed. We observed only one additional mdg1 site compared to the in silico data, 11C (Figure 3), while one site from the 27C7 region was lost (Table 3).

Table 3 Distribution of mdg1 transposable element on the y cn bw sp polytene chromosomes according to FISH and in silico data
Full size table

FISH analysis of Dm412 retrotransposon distribution also yielded a better correspondence to in silico data (Figures 6a and b) than the hobo case. Out of 30 annotated insertions of Dm412, four sites (1E1, 41E3, 41E5 and 77D4) were lost, whereas three novel sites (21CD, 39CD and 59B) were detected (Figures 6a and b). In region 60 of 2R chromosome, we sometimes observe only three Dm412 sites instead of four annotated, which may be explained by a close localization of 60C and 60D Dm412 sites, poorly resolvable by FISH. No changes in Dm412 pattern were detected by FISH performed in 2005 as compared to the FISH results obtained a year earlier (not shown), thus testifying a lack of its transposition activity over this period.

Figure 6
figure 6

Fluorescence in situ hybridization (2004) of Cy3-labeled Dm412 DNA on the salivary gland X chromosome and third chromosome (a) and X and second chromosome (b) from y cn bw sp strain (black arrows—annotated positions; gray arrows—new sites; double arrows—lost sites). Two closely located lost sites are annotated to region 41 (41E3 and 41E5).

Full size image

Discussion

We observed clear differences in the distribution of the three TEs between the in silico records and our FISH data for the y cn bw sp strain. This difference is most prominent for the hobo element, apparently indicating a high transposition activity. The frequency of hobo insertions is higher than that of excisions, as confirmed by the numbers of hybridization sites appearing de novo and disappearing (Table 1).

Assuming that the DNA used for sequencing the complete D. melanogaster genome was isolated 10–15 years ago (Smoller et al., 1991; Tamkun et al., 1992; Shaffer et al., 1994; Adams et al., 2000) and that the fly life span under laboratory conditions is approximately 1 month, at least 100–150 generations have passed over that period. During that time, 10 hobo sites (out of the 24 annotated) have been lost, while 36 new sites have been generated (Table 1).

Since we are unable to follow all intermediate events of appearance/disappearance of sites, the minimal rate of hobo element transposition is 3 × 10−3 per site per generation for excision (10/24 × 150) and 10 × 10−3 per site per generation for insertion (36/24 × 150). The transposition frequency calculated for both excisions and insertions is 13 × 10−3 per site per generation (46/24 × 150). We cannot distinguish closely located hybridization sites by FISH, which means that we have probably underestimated the hobo transposition rate.

Assuming that all hobo elements transpose at the same rate, the data suggest that at least 1 hobo copy out of nearly 50 found in y cn bw sp strain will transpose in every second generation; this rate translates into six transpositions per genome per year. Over the period of 12 months (2004–2005), we have observed 4 novel hobo sites out of 46, while 5 sites were lost. The average frequency of transposition over this period is 16 × 10−3 (9/46 × 12) or 12 × 10−3 (6.4/46 × 12) taking into account the polymorphic nature of some sites. Comparing the average transposition rate over 1-year and the 10-year period, we can conclude that hobo transposition rate did not diminish with time.

Dm412 and mdg1 transposition rates are an order of magnitude lower than that of hobo element in y cn bw sp strain, being about 2 × 10−3 (7/30 × 150) per site per generation for Dm412 and 5 × 10−4 (2/29 × 150) per site per generation for mdg1.

D. melanogaster has nearly 100 different TEs in its genome (Kaminker et al., 2002). Assuming that every type of transposable element has an average of 10 copies per genome (Rouzic and Capy, 2005), and that the average transposition rate is in the range 10−3–10−5 (Nuzhdin et al., 1996; Vieira and Biemont, 1997), we can conclude that the total rate of TE transposition will be close to the transposition rate of hobo obtained in our study.

In comparing the transposition rates of transposons (hobo) and retrotransposons (Dm412 and mdg1), we should also consider the difference in the mechanism of transposition. Retroelements transpose via an RNA intermediate by a ‘copy and paste’ mechanism while transposons transpose by ‘cut and paste’ process in the DNA form. The ‘copy and paste’ process might be more time-consuming than the ‘cut and paste’, since it involves the reverse transcription step in cytoplasm, while the ‘cut and paste’ mechanism is completely localized to the nucleus. Possibly, retroelements are less active than the tranposons simply due to the different mechanism of transposition.

Isogenization does not rescue the Drosophila strains from instability. Moreover, during the process of isogenization the genome instability might be increased. One cannot rule out the possibility of hobo-mediated hybrid dysgenesis, when crossing the flies of hobo-carrying strains with the individuals free of this element, thus inducing the genome instability (Blackman et al., 1987; Bazin et al., 1999). Strain y cn bw sp of D. melanogaster was isogenized before DNA separation for sequencing, which may account for the high hobo instability.