Generation of a Widespread Drosophila Inversion by a Transposable Element

See allHide authors and affiliations

Science  16 Jul 1999:
Vol. 285, Issue 5426, pp. 415-418
DOI: 10.1126/science.285.5426.415


Although polymorphic inversions in Drosophila are very common, the origin of these chromosomal rearrangements is unclear. The breakpoints of the cosmopolitan inversion 2j of D. buzzatii were cloned and sequenced. Both breakpoints contain large insertions corresponding to a transposable element. It appears that the two pairs of target site duplications generated upon insertion were exchanged during the inversion event, and that the inversion arose by ectopic recombination between two copies of the transposon that were in opposite orientations. This is apparently the mechanism by which transposable elements generate natural inversions inDrosophila.

Most Drosophilaspecies are naturally polymorphic for inversions in one or more chromosomes (1, 2), but little is known about the molecular mechanisms underlying the generation of these inversions. Indirect evidence suggests that transposable elements (TEs) causeDrosophila inversions: TEs mediated chromosomal rearrangements in laboratory populations (3), and in situ hybridization studies detected the transposon hobo around the breakpoints of four endemic inversions of D. melanogaster (4). However, in D. subobscura and D. pseudoobscura there was no cytological association between middle repetitive sequences and inversion breakpoints (1, 2), and the two direct studies that sequenced the breakpoints of naturally occurring inversions did not detect any TE (5).

The species D. buzzatii belongs to the D. repleta group of the Drosophila subgenus (6). Two chromosomal arrangements are commonly observed in chromosome 2 of D. buzzatii: the ancestral one or 2 standard (2st), and the2j, which derived from the 2st by inversion2j and is distributed throughout the species range at high frequencies (7). Here, we cloned and sequenced the breakpoints of inversion 2j of D. buzzatii. First, the region of the proximal breakpoint in the 2stchromosome (designated as CD) was cloned from a line homozygous for the 2st arrangement (st-1). Then the proximal and distal breakpoints in the 2j chromosome (designated asBD and AC, respectively) were isolated from a line homozygous for the 2j arrangement (j-1). Finally, the region of the distal breakpoint in the 2stchromosome (named AB) was obtained from the st-1 line.

The proximal breakpoint of inversion 2j in the2st arrangement lies between thenAcRβ-96A and Pp1α-96Agenes (Fig. 1A), which are 4 kb apart in D. melanogaster (8). A cDNA clone of the D. melanogaster nAcRβ-96A gene (9) was used to screen a st-1 lambda genomic library (10). Three positive clones were identified, and one of these, λst9, contained the Pp1α-96A gene (Fig. 1A) and spanned the breakpoint (Fig. 1B). By sequential subcloning of λst9 and in situ hybridization of the different subclones to 2st and 2j chromosomes, theCD breakpoint was further located in a 2-kb Hind III fragment, pGPE102.2, and a 0.9-kb Dra I fragment, pGPE102.2.2 (Fig. 1A). The AC and BD breakpoints in the 2jarrangement were recovered by screening a j-1 lambda genomic library with pGPE102.2 used as a probe. Five positive phages were identified, and the two with the most divergent restriction patterns, λj3.1 and λj10, were selected. In situ hybridization of both phages to st-1 and j-1 chromosomes yielded multiple signals (Fig. 1B), showing that they include repetitive sequences. However, the stronger signals on 2j chromosomes at the proximal breakpoint with λj3.1 and at the distal breakpoint with λj10 indicated that the inserts of these phages came from the BD and AC regions, respectively. Fragments of λj3.1 containing D sequences (pGPE105.4.1) and of λj10 containing C sequences (pGPE107.2.1.1) were identified by hybridization of digested DNA from each phage with a probe of the CD breakpoint (Fig. 1A). Next, a series of contiguous fragments of λj3.1 adjacent to pGPE105.4.1 and of λj10 adjacent to pGPE107.2.1.1 were subcloned. Most showed a repetitive in situ hybridization pattern, similar to that of the entire phages. A 1.2-kb Pst I–Bam HI fragment of λj3.1 (pGPE109.2), which contained B sequences and gave a strong signal at the AB region of st-1 chromosomes, was found 3.3 kb away from pGPE105.4.1 (Fig. 1A). Likewise, a 2.4-kb Pst I–Xba I fragment of λj10 (pGPE107.3), containing A sequences and giving a unique signal in the AB region of st-1 chromosomes, was found 1.9 kb away from pGPE107.2.1.1 (Fig. 1A). All these fragments were sequenced. A 95% identity at the amino acid level was found between the proximal end of pGPE107.3 and the 44 COOH-terminal residues of the D. melanogaster Rox8 protein (11), thus revealing the location of the D. buzzatii rox8 gene (Fig. 1A). The AB region in the2st arrangement was finally obtained by polymerase chain reaction (PCR) amplification from st-1 genomic DNA using primers Rox1 and B0, located at the 3′ end of the rox8coding region and the distal end of pGPE109.2, respectively. In situ hybridization to st-1 and j-1 chromosomes confirmed that the resulting 2.1-kb PCR product (pGPE115) spanned the 2j distal breakpoint (Fig. 1B).

Figure 1

Cloning strategy and in situ hybridization of the regions spanning the breakpoints of inversion 2j.(A) Physical map of 2st and 2jchromosomes showing the breakpoint regions (AB andCD in 2st and AC and BD in2j) and different clones obtained (thin lines shown above map). Vertical arrows mark the limits of the A,B, C, and D regions (in 2jchromosomes there are insertions between A and Cand between B and D). Empty boxes below map represent the genes flanking the breakpoints according to the sizes of the respective cDNAs in D. melanogaster. The direction of transcription, when known, is indicated as 5′ and 3′. Cen, centromere; B, Bam HI; D, Dra I; E, Eco RI; H, Hind III; P, Pst I; S, Sal I; X, Xba I. (B) In situ hybridization to the salivary gland chromosomes of the st-1 and j-1 D. buzzatii lines of the four clones containing the breakpoint regions of inversion2j (arrowheads). Clones spanning the breakpoints in the2st arrangement (λst9 and pGPE115) give one signal in st-1 and hybridize to both breakpoints in j-1. Clones spanning the breakpoints in the 2j arrangement (λj3.1 and λj10) give two signals in st-1 and one signal in j-1.

Alignment of AB and CD sequences in st-1 (2086 nucleotides) and AC and BD sequences in j-1 (8668 nucleotides) showed that there are large insertions at both2j inversion breakpoints that are not present in the2st arrangement (Fig. 2). The insertion betweenA and C is 392 base pairs (bp) long and has inverted terminal repeats (ITRs) of 106 bp. The insertion betweenB and D is 4319 bp long, with a complex internal structure. It has ITRs like those of the AC insertion, of 106 and 47 bp, the latter corresponding to the beginning of the 106-bp repeat. The central 180 bp of the AC insertion and theBD sequence that follows the proximal ITR are homologous (95% identity) but are oppositely oriented (Fig. 2). Flanking each insertion are 7-bp sequences that resemble target site duplications that were produced during the insertion event. However, the duplicated sequences are separated and inverted (Fig. 2), which strongly suggests that inversion 2j appeared by intrachromosomal pairing and recombination between the two homologous sequences inserted in distant sites in opposite orientations (Fig. 3). Although the precise structure of the original insertions is not known, they were likely homologous at least over a span of 274 bp, and this relatively short sequence appears to be long enough to sustain ectopic recombination in Drosophila(12).

Figure 2

Features of the sequenced regions of the distal and proximal breakpoints of inversion 2j in the2st and 2j arrangements. Horizontal lines correspond to the A, B, C, andD sequences, and insertions within them are represented as colored boxes. The insertions at the breakpoints are delimited by black blocks corresponding to the ITRs. Orientation of the homologous sequences of the two insertions is indicated by angle bracket patterns (〈〈〈 〉〉〉). Internal repeats within the BD insertion are represented as sharp-pointed blocks with different colors.ISBu-1 is a nondescribed, highly dispersed repeated sequence inserted in the 2jarrangement in the A region, 256 bp away from the insertion at the breakpoint. Details of the breakpoint sequences in st-1 and j-1 are shown below each diagram (A and Bsequences in blue, C and D sequences in red, and insertions in black). The target site duplications generated during insertion are underlined and in boldface. The distal breakpoint in2st shows a structure with 7-bp imperfect (one nucleotide difference) direct repeats flanking two adjacent 9-bp inverted repeats, which resembles the footprints left by class II TEs after excision and incomplete repair of the double strand break (30). The A sequence in the 2jarrangement is that of the j-14 line, because this sequence is absent in the j-1 line as the result of a 17-bp deletion. Horizontal arrows represent different PCR primers; those used for sequencing are marked with an asterisk. (Sequences of all primers can be found atScience Online, Codes for restriction sites are as in Fig. 2.

Figure 3

Model for the generation of inversion2j. First, the two homologous sequences became inserted between the A and B sequences and between the C and D sequences of a 2stchromosome, generating direct duplications of seven nucleotides (boxed sequences). Then, ectopic recombination between the homologous sequences in opposite orientation generated the inversion. As a result of this process, target site duplications were exchanged between the insertions and acquired their present orientation in2j chromosomes, with each insertion flanked by its original target site and an inverted target site of the other element. Dashed lines refer to sequences not drawn to scale.

The duplication of a 7-bp sequence upon insertion, the presence of ITRs, and the moderately repetitive in situ hybridization pattern identify the breakpoint insertions as copies of a class II TE (13) that we designated as Galileo(14). Furthermore, we have observed variation among D. buzzatii lines both in the number of in situ hybridization signals and in their chromosomal location, which suggests recent transpositional activity of this TE (15). TheAC insertion is probably a defective copy that derived from a complete element by an internal deletion but retained the terminal sequences needed for transposition (16). TheBD copy may have degenerated partially, but it contains open reading frames (ORFs) that could code for a protein involved in transposition. The largest ORF has a predicted sequence of 172 amino acids, displaying 32% identity with eight gaps (53% similarity) to the putative transposase of the transposon Gandalf ofD. koepferae, a sibling species of D. buzzatii(17). Ectopic recombination between repeated DNA sequences has been implicated in inversion generation in diverse organisms, such as yeast (18) and humans (19), and TE activity could stimulate this recombination (20). However, precedents of TE involvement in the generation of naturally occurring inversions were restricted to bacteria (21). A TE-like sequence was found at one breakpoint of a polymorphic inversion of Anopheles arabiensis, but its implication in the origin of the inversion was not clear (22). Evidence from inversion 2j confirms the role of TEs inDrosophila chromosomal evolution.

The previous results were corroborated by PCR amplification of the breakpoint regions and sequencing of the PCR products in2st lines and 2j lines of different origin (23). Primer pairs A1-B1 and C1-D2 (Fig. 2) were used with genomic DNA of four 2st lines, whereas primer pairs A1-C1, B1-Gal2, and Gal1-D2 (Fig. 2) were used with DNA from 21 2jlines. In each case PCR products of the appropriate sizes were obtained, indicating that the breakpoint insertions are present in all2j lines and absent in all 2st lines (in the amplification of the AC segment only, j-9 and jq7-2 showed larger PCR products than expected, resulting apparently from additional insertions, and j-10 yielded no PCR product). We sequenced 521 bp of the A, B,C, and D regions in three 2st lines and five2j lines and 800 bp of the Galileo insertions in the 2j lines (24). All the 2jchromosomes in this geographically diverse set had the same insertions and identical breakpoints with regard to 2st chromosomes. Nucleotide diversity (25) over the 489 aligned sites in the eight sequences (excluding gaps) was 0 and 0.0279 for the 2jand 2st arrangements, respectively, and 2jchromosomes shared 22 fixed differences when compared to 2stchromosomes (including six small insertions of 1 to 4 bp).Galileo sequences showed more variation among the2j lines, with a nucleotide diversity over the 794 aligned sites of 0.0053. Thus, it appears that despite the transposon-mediated origin, all 2j arrangements descend from the product of a single ectopic recombination event: They are monophyletic.

Finally, we tested the effect of the inversion on the nearby genes rox8 and Pp1α-96A (Fig. 1A). The Pp1α-96A transcriptional startpoint (26) is just 156 bp away from the proximal breakpoint in the 2st arrangement, and a semiquantitative RT-PCR was carried out to compare Pp1α-96Aexpression between the st-1 and j-1 lines at four developmental stages (27). No apparent differences in expression between st-1 and j-1 were seen. As in D. melanogaster(26), the Pp1α-96Atranscript was present in embryos, was detected with a lower expression in adults, and was barely visible in larvae and pupae. We also performed Northern (RNA) analyses (28) using probes of theAB (pGPE115.1) and CD regions (pGPE102.2). Both probes hybridized to the transcripts ofPp1α-96A [2.3 kb (26)] androx8 [3 to 3.3 kb (11)] only, and no detectable differences in the expression of these genes between st-1 and j-1 were observed. It seems that inversion 2j neither disrupted any transcriptional unit nor affected the expression of the closest genes, ruling out a relation between the mutational effect of the inversion and its adaptive value (29).

  • * To whom correspondence should be addressed (in Spain). E-mail: mariocs{at}


View Abstract

Navigate This Article