Efficient Mobilization of mariner in Vivo Requires Multiple Internal Sequences

Corresponding author: Daniel L. Hartl, Harvard University, 16 Divinity Ave., Cambridge, MA 02138., dhartl{at}oeb.harvard.edu (E-mail)

Communicating editor: J. A. BIRCHLER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Aberrant products of mariner excision that have an impaired ability to be mobilized often include internal deletions that do not encroach on either of the inverted repeats. Analysis of 13 such deletions, as well as 7 additional internal deletions obtained by various methods, has revealed at least three internal regions whose integrity is necessary for efficient mariner mobilization. Within the 1286-bp element, the essential regions are contained in the intervals bounded by coordinates 229–586, 735–765, and 939–1066, numbering in base pairs from the extreme 5' end of the element. These regions may contain sequences that are necessary for transposase binding or that are needed to maintain proper spacing between binding sites. The isolation of excision-defective elements with point mutations at nucleotide positions 993 and 161/179 supports the hypothesis of sequence requirements, but the reduced mobility of transformation vectors with insertions into the SacI site at position 790 supports the hypothesis of spacing requirements. The finding of multiple internal regions that are essential for efficient mariner mobilization in vivo contrasts with reports that mini-elements with as little as 43 bp of DNA between the inverted repeats can transpose efficiently in vitro.

MANY natural populations contain internally deleted copies of type II transposons (DNA transposons with inverted repeats). In Drosophila, the best known examples are the P element (BINGHAM et al. 1982 ; BLACK et al. 1987 ), the hobo element (PERIQUET et al. 1989 ; PASCUAL and PERIQUET 1991 ; KIM and KIM 1999 ), and mariner (MARUYAMA and HARTL 1991 ; BRUNET et al. 2001 ). In the case of the P element, the internal deletions can arise as a product of abortive transposition. Transposition of the P element is mediated by the P transposase encoded within the element that functions in DNA cleavage and strand transfer. Transposition proceeds by a "cut and paste" mechanism (KAUFMAN and RIO 1992 ), and the resulting gap in the donor molecule is repaired by template-directed gap repair (TDGR), using as template the sister chromatid, the homologous chromosome, or an ectopic site (ENGELS et al. 1990 ; GLOOR et al. 1991 ; NASSIF et al. 1994 ). Incomplete repair by TDGR results in a class of P-element mutations that are internally deleted, but the terminal 16 ± 1 bp of the inverted repeats are rarely included in the deletions (STAVELEY et al. 1995 ). In vitro studies have shown that P transposase makes a staggered cut within the 31-bp inverted repeats, leaving the extreme 17 nucleotides of each P-element end single stranded (BEALL and RIO 1997 ). The single-stranded DNA is bound by an inverted-repeat binding protein (IRBP; RIO and RUBIN 1988 ). It has been proposed that IRBP binding protects the P-element termini from exonuclease degradation (STAVELEY et al. 1995 ), thereby explaining the retention of the terminal inverted repeats in the deletion derivatives.

Most of the internal deletions of the P element are missing all or part of the transposase-coding region and hence are unable to produce functional transposase. They are, however, able to be mobilized in trans by a P element that does encode a functional transposase. It is the mobilization of such defective P elements that results in efficient P-element-mediated insertional mutagenesis (COOLEY et al. 1988A , COOLEY et al. 1988B ).

The generality of TDGR in transposition of inverted-repeat DNA transposons is supported by evidence from excision of the mariner element, a member of the mariner/Tc1 superfamily (HARTL et al. 1997 ; HARTL 2001 ). Homolog-dependent gap repair frequently accompanies mariner excision (LOHE et al. 2000 ). The ratio of excision events that leave a characteristic 5-bp footprint, compared to gap repair using a copy on the homologous chromosome, was estimated at 2.4:1 (LOHE et al. 2000 ). Therefore, at least 30% of all mariner excision events undergo TDGR.

On the other hand, recent experiments suggest that the details of the process of transposition leading to TDGR are quite different in the P-element and mariner systems. In a screen for aberrant products of excision of the mariner element denoted peach, we uncovered a high frequency of deletions near the termini of the inverted repeats (LOHE et al. 2000 ). The peach element is a mariner element that is inserted into the 5' untranslated region of the white gene in the wpch allele. The insertion changes the phenotype from bright red to peach (JACOBSON et al. 1986 ). The peach element does not produce a functional transposase, but it can be excised efficiently in response to active transposase supplied in trans. Somatic activity of the transposase in wpch flies results in a mosaic eye color consisting of red sectors on a peach background. Germline transposase activity results in excision of peach and restoration of the wildtype white gene function in the progeny (BRYAN et al. 1987 ). When under control of the dual hsp70::Mos1 promoter, the Mos1 mariner transposase is strongly expressed in the germline and soma even in the absence of heat shock. In flies carrying the wpch allele, hsp70::Mos1 induces high levels of germline reversion of peach and extreme mosaicism in the eyes (GARZA et al. 1991 ).

In the presence of hsp70::Mos1 transposase, the frequency of excision-defective mutations in peach is 0.2% per exposed wpch allele per generation. The most frequent aberrations (51/68 events) could be explained by exonuclease degradation of the inverted repeats during host-mediated repair of a double-stranded gap made by transposase at either the 5' junction or the 3' junction of the peach element and genomic DNA. An analogous class of mutations has not been reported to arise from P-element transposition (STAVELEY et al. 1995 ).

In this article we report on 20 internal deletions in peach obtained by a variety of methods, ranging in size from 31 to 899 bp. The deletions overlap in such a way as to include all sequences internal to the inverted repeats. The fact that peach elements with a variety of internal deletions have an impaired ability to excise in the presence of active transposase seems to imply that, in vivo, the mariner transposase, or some complex between the mariner transposase and one or more host factors, must interact with internal sequences to catalyze efficient excision. Furthermore, from the overlaps of the deletions we can identify three distinct internal regions that are required for efficient excision, either because they contain essential binding sites or because their presence is needed to maintain the correct spacing between binding sites. A model requiring correct spacing is supported by evidence in a companion article (LOZOVSKY et al. 2001), which shows that insertions of exogenous DNA at any of a number of internal sites also impair mariner mobility.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The peach/Mos1 two-element system:
Although the peach and Mos1 elements both have an uninterrupted open reading frame, they differ at 11 nucleotide positions out of 1286 bp, and only Mos1 encodes a functional transposase. Both elements can excise and transpose in response to Mos1 transposase, which is active in the soma as well as the germline. The insertion of peach into the 5' untranslated region of the white gene alters the eye color from red to peach (JACOBSON et al. 1986 ). With Mos1 in the genetic background, the peach element excises and white gene function is restored. Excision of peach in ommatidia results in red spots or sectors on a peach background in the eye, and peach excision in germline cells results in full restoration of the wild-type red eye color in the progeny (BRYAN et al. 1987 ). The degree of somatic mosaicism is a semi-quantitative assay for the rate of peach excision and correlates well with the rate of germline excision.

The wpch allele was originally discovered in Drosophila mauritiana (JACOBSON et al. 1986 ) but was transferred onto the X chromosome of D. melanogaster by cloning and P-element transformation (GARZA et al. 1991 ). Thus, in D. melanogaster, the peach element is part of a P-element construct that is a composite of sequences from the D. melanogaster white gene and the D. mauritiana wpch allele (GARZA et al. 1991 ).

The wpch eye-color mosaicism screen:
A mutagenesis screen for peach mutations utilized the two-element peach/Mos1 system with the wpch allele as a scorable marker for Mos1-induced peach excision leading to eye-color mosaicism (LOHE et al. 2000 ). The source of transposase was not Mos1 itself but an hsp70::Mos1 construct that is unable to be mobilized because it lacks the 5' inverted repeat (LOHE et al. 1995 ). In the absence of functional transposase, the wpch allele yields a uniform peach-colored phenotype and there is no somatic mosaicism. In the presence of hsp70::Mos1, even without heat shock, somatic excision of peach restores wild-type white gene function and the eye phenotype is mosaic with hundreds of red ommatidia per eye on a peach background (LOHE et al. 1995 ). The particular hsp70::Mos1 construct used is designated Mr182, which consists of a P-element transformation vector containing hsp70::Mos1 inserted in chromosome 2 (LOHE et al. 1995 ). New mutations in the wpch allele affecting the ability of the peach element to excise were identified by a decrease in the level of eye-color mosaicism. Germline reversion rates of wpch were measured to verify that peach excision was also reduced in the germline and that the eye color of the revertants was wild type. Mutagenized wpch males were crossed to 10–15 y w^-; Mr182/Mr182 females in bottles, and female progeny were scored for somatic mosaicism. Stocks of wpch mutants were each made from a single male using the FM7a balancer chromosome, and Mr182 was immediately removed from the genetic background.

Methods of mutagenesis:
Three mutagenesis schemes were employed.

1. Mos1 mutagenesis was carried using the Mr182 hsp70::Mos1 construct as described above, selecting for mutants of wpch that showed markedly reduced levels of peach excision.

2. P-element mutagenesis made use of the fact that the wpch allele is present in a P-element vector and that excision of a P transposon and aberrant template-directed gap repair can result in internal deletions in the newly synthesized copy (GLOOR et al. 1991 ). With the wpch allele, internal deletions or other mutations in the peach element can be recovered. In the present experiments, the stable source of P transposase was the P2–3(99B) element (ROBERTSON et al. 1988 ). Before commencing the experiments, we verified that the P transposase does not mobilize the peach element even at low levels (data not shown).

3. Chemical and radiation mutagenesis made use of EMS (ethyl methanesulfonate), DEB (diethylbutane), or X rays. EMS mutagenesis was carried out as described (LOHE et al. 1996 ). About 100 crosses were set up with 4–5 mutagenized wpch males and 10–15 y w^-; Mr182/Mr182 females per bottle, and 20,000 female progeny were scored for somatic mosaicism. From this screen, 13 wpch mutants were recovered with reduced somatic mosaicism and low germline reversion rates. DEB mutagenesis was carried out using a protocol similar to the EMS mutagenesis, except that the concentration of DEB was 10 mM (CROSBY and MEYEROWITZ 1986 ). This resulted in a high loss of males and only 4000 female progeny were screened, which yielded 4 wpch mutants. For X-ray mutagenesis, 300 males (4–5 days old) were exposed to 4000 R from a cesium-137 source kindly provided by the CSIRO Division of Entomology. Groups of 4–5 mutagenized wpch males were crossed with 10–15 y w^-; Mr182/Mr182 females per bottle. Approximately 6000 female progeny were screened for somatic mosaicism and 4 wpch mutants were recovered.

Molecular analysis:
Analysis of wpch mutants and DNA sequencing were carried out using primers and PCR amplification of genomic DNA from single flies, as described (LOHE et al. 2000 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Internally deleted peach elements produced by the mariner Mos1 transposase:
In the peach/Mos1 two-element system, the peach element is tested for its ability to excise by means of a phenotypic screen for the observed level of eye-color mosaicism in wpch flies. This screen appears to be relatively free from biases in phenotypic selection, because the selection assays the ability of the target element to excise, not the phenotypic effect of the excision on a nearby gene. In the analysis of mutant peach elements arising from aberrant effects of the Mos1 transposase, two major classes of mutants were recovered. One class consisted of relatively small deletions at the extreme termini of either the 5' or the 3' inverted repeat, though primarily affecting the 5' inverted repeat (LOHE et al. 2000 ), and the other class consisted of the internal deletions in the element reported here. Among 68 mutant peach elements recovered in the peach/Mos1 screen, 13 proved to have internal deletions.

The internal deletions resulting from aberrant effects of Mos1 transposase are listed in Table 1, along with internal deletions from other sources described below. Although most of the Mos1-associated aberrations are simple deletions, a few also had a few nucleotides of "filler" DNA sequence inserted between the breakpoints, similar to some of the terminal deletions reported earlier (LOHE et al. 2000 ). Among the internal deletions the length of the filler sequence was 1 bp (16R and 57R), 2 bp (49R and J2), 8 bp (72R), or 16 bp (35R); all are much shorter than the length of the deletion.

Among the 13 Mos1-associated deletions, only 99R contained a deletion endpoint that removed part of one of the inverted repeats. This deletion is 705 bp in length and extends from nucleotides 577–1281, leaving only the terminal 6 bp of the 3' inverted repeat intact. Not surprisingly, 99R shows no somatic mosaicism.

Unexpectedly, while none of the remaining 12 internal deletions encroaches on either of the terminal inverted repeats, their effects on peach excision are profound. The level of somatic mosaicism is reduced by at least a factor of 10, and in many cases somatic mosaicism is no longer observed.

Internally deleted peach elements produced by P-element transposase:
It is possible that the internal deletions produced by the mariner transposase may be nonrandom in such a way as to yield a high frequency of peach derivatives with a drastically impaired ability to excise. For example, the breakpoints may be in regions in which there are essential contacts between the target element and the transposase. To evaluate this possibility we also examined mutant peach elements generated by mechanisms unrelated to the mariner transposase, and once again we found internally deleted elements.

Table 1 lists five mutant peach elements generated from aberrant effects of the P-element transposase. The wpch allele is present in a P-element construct that contains the peach element inserted into the 5' untranslated region of the white gene in a P-element vector (GARZA et al. 1991 ). The action of P transposase on the wpch allele should induce template-directed gap repair and result, on rare occasions, in internal deletions within the transposon including all or part of the peach element. Accordingly, crosses were carried out with wpch in the same genome as a stable source of P-element transposase (P2–3), and the wpch progeny were tested for mosaicism by crossing to a strain carrying hsp70::Mos1. In this test, defective peach elements induced by the P transposase could be distinguished from the rare defective peach element produced by the Mos1 transposase, because in the former case (P element) all of the progeny show reduced mosaicism, whereas in the latter case (Mos1) only 1/500 progeny shows a mutant phenotype.

This experiment yielded five wpch mutants with reduced or absent somatic mosaicism, in which the wpch allele remained at its original chromosomal location. All five mutants had an internal deletion within peach (Table 1). The deletions cause a reduction in level of somatic mosaicism of wpch of at least a factor of 10, although two of the four deletions that have the most drastic effects on somatic mosaicism have one breakpoint near the 5' inverted repeat, namely C1 and D1. (In fact, the deletion in C1 includes the innermost nucleotide of the 5' inverted repeat.)

Defective peach elements produced by chemical mutagens and radiation:
To generate still more peach mutations with impairments in somatic excision, the wpch allele was mutagenized with three different mutagens: EMS, DEB, or X rays. These mutagens are expected to induce a different spectrum of mutations, ranging from point mutations (EMS) through small deletions (DEB) to large and small deletions (X rays).

Five additional peach deletions were recovered from these screens, three of which affected the inverted repeats. Two small deletions were recovered from EMS treatment: One was of 5 bp at the extreme 5' end of the element (including the TA duplication and the first three nucleotides of the 5' inverted repeat), and the other was of 1 bp near the extreme 3' end of the element (deleting nucleotide 1284). In both of these mutants the level of somatic mosaicism was reduced by a factor of at least 10. The third small deletion was induced by DEB and deleted nucleotides 5–6 (or perhaps 7–8) from the 5' inverted repeat; this deletion reduced somatic mosaicism to 0.2% of the control level.

The other two deletions recovered from the mutagen screens are shown in Table 1. One small one (31 bp) was induced by DEB and one relatively large one (461 bp) by X rays. Both of these reduce somatic mosaicism of wpch by a factor of 10.

EMS-induced internal point mutations affecting peach excision:
EMS mutagenesis of wpch also produced four mutations with one or two nucleotide changes in which the peach element had reduced somatic excision. Three of these (EMS-15, EMS-36, and EMS-118) had identical G-to-A transitions at nucleotide position 993. The fourth (EMS-9) was a double mutant with C-to-A transversions at both nucleotide positions 161 and 179. All four of these mutant peach elements showed a level of wpch excision that was reduced by more than an order of magnitude, suggesting that nucleotides 161 and/or 179, and nucleotide 993, are key nucleotides required in the excision reaction of peach. None of these sites is close to either of the inverted repeats.

Transposase-induced recombination between mutant peach alleles:
We previously reported that, in females, the rate of recombination at the site of a homozygous peach element is increased in the presence of active transposase by a factor of 200 (LOHE et al. 2000 ). This estimate assumes that the transposase-induced increase in recombination occurs across the region of the 1.3-kb peach element only, but not more uniformly across adjacent regions. To test this assumption, we examined rates of recombination in females carrying two different mutant peach elements, in the presence or absence of an hsp70::Mos1 source of transposase. Progeny were scored for high levels of somatic mosaicism. Because both mutant peach elements in the parental female have low levels of somatic mosaicism, restoration of a normal level requires recombination in the interval between the mutations. (We use the term recombination for any mechanisms of information exchange between homologs, including gene conversion associated with template-directed gap repair.)

Four mutant peach elements were tested individually in heteroallelic combinations with EMS-118, which has a single nucleotide substitution (C-to-A) at position 993 of peach. The phenotype of EMS-118 is a weak mosaic, showing 10–20 mosaic spots per eye in the presence of Mos1 transposase. The results are shown in Table 2. The peach mutations tested in combination with EMS-118 were B1, 65R, 66R, and 68R. B1 is a deletion of peach that includes nucleotides 229–586 (Table 1), 65R is a deletion/insertion in which the first 18 bp of the 5' inverted repeat has been deleted and replaced with 25 bp of unrelated sequence, D66R is another deletion/insertion in which 123 bp at the 5' end of peach has been deleted and replaced with 9 bp of unrelated sequence, and 68R is a 9-bp deletion that includes 6 bp of flanking sequence and the first 3 bp of the 5' inverted repeat (LOHE et al. 2000 ).

The highest level of recombination was obtained with B1/EMS-118 (0.17% in Table 2) and this recombination was dependent on the presence of Mos1 transposase. Two recombinant elements from this genotype were verified as authentic peach by DNA sequencing. Since the value 0.17% includes only one of the two recombinant classes, the recombination rate between the mutant alleles is actually 0.34%. Because B1 has a 357-bp deletion between nucleotides 229 and 586 inclusive, any recombination that restores somatic mosaicism recombination must have occurred in the 406-bp region between the 3' breakpoint of B1 and position 993.

Previous results have also suggested that the Mos1 transposase preferentially attacks the 5' inverted repeat of peach. If this is the case, then we would also expect disruption of the 5' end to reduce the rate of transposase-induced recombination in heteroallelic mutant peach combinations. This prediction is borne out by the results with 65R, 66R, and 68R, all three of which have a mutant 5' inverted repeat. The total distance between the lesion in 68R and that in EMS-993 is 989 nucleotides, approximately threefold greater than that between the lesions in B1 and EMS-993, yet the rate of transposase-induced recombination is reduced by a factor of more than four. (The unexpected recombinant obtained in the control cross between 65R and EMS-118 is most easily explained by a maternal effect of Mos1 transposase present in the tested females, since their mothers were heterozygous for Mr182.)

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Hotspots of recombination:
The results in Table 2 confirm that, in the presence of transposase, mariner elements are hotspots of recombination in the Drosophila genome, as they also are, at least at some sites, in the human genome (KIYOSAWA and CHANCE 1996 ; REITER et al. 1996 ). It should be emphasized, however, that there is as yet no direct evidence for an active mariner transposase produced in human cells. The Drosophila results also demonstrate that, in the presence of transposase, high rates of recombination between allelic, defective elements can restore excision competency to mutant elements. In the genotype B1/EMS-118, the rate of recombination is 0.34 map units across the 406 bp between the mutational lesions, which implies an overall rate of recombination of 1 map unit per 1.2 kb. The wpch transgene is located in the region of the X chromosome between singed and lozenge, in which the normal rate of recombination is 1 map unit per 250 kb (HEINO et al. 1994 ). Hence the rate of recombination within mariner elements is increased by a factor of 200.

Most of the recombinant mariner elements probably result from template-directed gap repair rather than from the usual homologous recombination pathway. This inference is supported by the observation in Table 2 that mutant mariner elements missing all or part of the 5' inverted repeat show less recombination than mutant elements, like B1, with internal deletions. The difference is statistically significant (² = 12.5, P < 0.01). This finding also supports a previous inference, based on other evidence, that the mariner transposase makes its initial attack on the 5' inverted repeat (LOHE et al. 2000 ).

Differences between mariner transposition and P-element transposition:
Our results indicate many points of difference between mariner transposition and P-element transposition. For example, the major class of defective elements recovered as aberrant products of mariner transposition have alterations in the 5' inverted repeat or, to a lesser extent, in the 3' inverted repeat. These alterations are most easily interpreted as arising from exonuclease degradation initiated at the site of a double-stranded cleavage made by transposase near the junction of the element and genomic DNA, usually between nucleotides 2 and 3 in the 5' inverted repeat, which is repaired by host enzymes with or without insertion of filler sequences (LOHE et al. 2000 ). An analogous class of mutations affecting the extreme ends of the P-element inverted repeats has not been reported (STAVELEY et al. 1995 ). The P-element transposase generates asymmetrical double-stranded cleavage with 3' single-stranded overhangs of the terminal 17 bp of each of the inverted repeats. Aberrant deletion products almost always retain the terminal 15–17 bp of the inverted repeats, which are apparently protected following cleavage by an IRBP (RIO and RUBIN 1988 ; STAVELEY et al. 1995 ; BEALL and RIO 1997 ). Our results with mariner are most easily interpreted by supposing that the IRBP does not bind to the mariner inverted repeats. Alternatively, if IRBP does bind to the mariner inverted repeats, then there is a different interaction with the transposase such that the bound ends are not protected from exonuclease attack.

The characteristics of internal deletions also indicate a major difference between the mechanisms of mariner and P-element transposition. Internal deletions of mariner resulting from aberrant excision appear to have endpoints scattered at random throughout the element (Table 1). Among the 13 Mos1-associated internal deletions in Table 1, only one (99R) extends into one of the inverted repeats. In contrast, internal P-element deletions usually terminate within one of the inverted repeats, and sometimes both endpoints lie within the inverted repeats (STAVELEY et al. 1995 ).

Internal sequences and/or spacing of sequences required for efficient mariner excision:
The most unexpected result of our study is the apparent presence of internal sites, far removed from the inverted repeats, that appear to be necessary for efficient mariner excision. The locations of the deletions in Table 1 are represented graphically in Fig 1. Because each of these deletions reduces the level of excision by at least a factor of 10, either each deletion must be missing sites that are essential for efficient transposition or, alternatively, each deletion alters an essential spacing between binding sites that flank the deletion. The remarkable finding is that the deletions are scattered throughout the element. We emphasize that these deletions were selected on the basis of their reduced levels of transposition; hence there may be other deletions that do not markedly affect the rate of excision. Nevertheless, if internal deletions that affected excision were rare or were restricted to a particular region, we would neither have expected to recover so many of them (13/68 mutant peach elements recovered in the peach/Mos1 screen had internal deletions) nor have expected them to be scattered throughout the entire element (Fig 1).

View larger version (36K): In this window In a new window Download PPT slide   Figure 1. Location of internal deletions (gaps) within the peach element, with the nondeleted element across the top for comparison. The numbers at the right are the rates of excision of each deleted element (in percentages), relative to the nondeleted element. The rectangles indicate the critical regions I, II, and III defined by the deletions B1, DEB-342, and 4R, respectively.

Inspection of Fig 1 indicates that, generally speaking, the larger the deletion the smaller the rate of element excision. (The numbers on the right are the rates of excision, in percentages, relative to excision of the peach element.) On the assumption that there might be discrete internal regions that are required for efficient element excision, three regions can be identified on the basis of the three smallest deletions: B1, DEB-342, and 4R. The B1 deletion implicates the region 229–586 bp (which we denote as region I), DEB-342 implicates the region 735–765 bp (region II), and 4R implicates the region 939–1066 bp (region III). Deletion of any of these regions reduces the efficiency of element excision by more than an order of magnitude. Sequences flanking these regions may also be important. For example, the deletion 72R completely abolishes element excision, yet it is only slightly larger than the deletion B1. We emphasize that internal regions other than I, II, and III may also be essential for efficient excision but have escaped identification and also that the sequences required for recognition may be large enough to overlap regions I and II or regions II and III. A more extensive mutational analysis will be required to define the minimal regions required for efficient excision.

It is possible that regions I, II, and III are necessary for efficient excision not because they contain essential binding sites but because they maintain an essential spacing between binding sites. Evidence for length dependence in the case of the Escherichia coli insertion sequence IS50 has been reported by GORYSHIN et al. 1994 , who show that elements <64 bp are severely impaired in transposition, whereas those >200 bp exhibit efficient transposition with essentially no length dependence. In our experiments, the lengths of the deleted peach elements are >>200 bp, ranging from 386 to 1255 bp. The hypothesis of critical internal binding sites is also supported by the point mutations recovered in the EMS screen. EMS-118 is a single-nucleotide substitution at position 993, which is very near the middle of region III. Clearly a sequence in this region is necessary for efficient excision, but we cannot exclude the possibility that the region is also necessary for spacing. EMS-9 is a double mutant with single-nucleotide substitutions at positions 161 and 179. None of these lie in regions I, II, or III, but it may be of some interest that both of them lie in the region immediately to the left of region I in the interval defined by the left-hand breakpoints of 72R and B1. This result again suggests that 72R contains essential sequences not present in region I, and it is sequence, not spacing, that is important.

On the other hand, correct spacing between essential binding sites also appears to be important for efficient mariner mobilization. Evidence is presented in an accompanying article (LOZOVSKY et al. 2001), in which we report that mariner vectors with exogenous DNA fragments of 1–5 kb inserted into any of the unique internal restriction sites SalI, SphI, or ClaI are severely compromised in their ability to excise and transpose.

Comparison with mariner deletions found in natural populations:
BRUNET et al. 2001 recently studied deletions of Mos1-related mariner elements in natural populations of multiple species of Drosophilidae, which were detected by means of PCR amplification using the Mos1 inverted repeats as primers. They found 27 distinct deletions, 21 of which were >50 bp. Among these, 10 overlap our region I; 5 overlap regions I and II; 2 overlap regions II and III; and 4 overlap regions I, II, and III. Furthermore, among the 6 deletions <50 bp, 2 (of lengths 27 and 45 bp) are in the critical region identified between the left-hand breakpoints of deletions 72R and B1, and one (of length 17 bp) overlaps region II. Judging from the locations of these deletions and the impaired mobility of the deletions that overlap regions I, II, and III (Fig 1), it seems very likely that most of the deletions found in natural populations are unable to undergo excision or transposition at anywhere near normal levels. This finding supports the conclusion of BRUNET et al. 2001 that many of the deletions found in nature are created spontaneously by mechanisms unrelated to transposase activity. For example, one-third of the naturally occurring deletions occur between short direct repeats of five to eight nucleotides (BRUNET et al. 2001 ).

Comparison with mariner transposition in vitro:
Our results appear to conflict with those of TOSI and BEVERLEY 2000 , who reported that a Mos1 element with as little as 43 bp of DNA between the inverted repeats could transpose in vitro as efficiently as the complete element. However, the mini-element carried a kanamycin-resistance gene inserted between the inverted repeats, and the complete element carried a kanamycin-resistance gene inserted into the SacI site (TOSI and BEVERLEY 2000 ). Because the complete element forms the baseline for the comparison, the result implies that only the mini-element transposes in vitro as efficiently as an element with an insertion into the SacI site.

It may be that vectors with insertions into the SacI site undergo relatively efficient transposition in vitro. However, we have shown that, in vivo, Mos1 constructs with insertions into the SacI site are highly refractory to either excision or transposition by Mos1 transposase (LOHE and HARTL 1996 ). These results caution against extrapolation from data obtained in vitro to estimate the rate of transposition in vivo. We have shown in this article that the continuity of several internal regions of mariner is essential for efficient transposition in vivo and that excision of peach is abolished by many internal deletions > 400 bp (Fig 1). Furthermore, in vitro reactions are carried out with purified DNA, whereas in vivo the transposase protein must interact with chromatin.

	ACKNOWLEDGMENTS

We thank Nikki Lee and Courtney Griffin (née Courtney Timmons) for their help in carrying out some of the experiments and Pierre Capy for providing information on the location of naturally occurring mariner deletions. This work was supported by National Institutes of Health grants GM33741 and GM58423.

Manuscript received July 24, 2001; Accepted for publication November 8, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BEALL, E. L. and D. C. RIO, 1997  Drosophila P-element transposase is a novel site-specific endonuclease. Genes Dev. 11:2137-2151[Abstract/Free Full Text].

BINGHAM, P. M., M. G. KIDWELL, and G. M. RUBIN, 1982  The molecular basis of P-M hybrid dysgenesis: the role of the P element, a P-strain specific transposon family. Cell 29:995-1004[Medline].

BLACK, D. M., M. S. JACKSON, M. G. KIDWELL, and G. A. DOVER, 1987  KP elements repress P-induced hybrid dysgenesis in D. melanogaster. EMBO J. 6:4125-4135[Medline].

BRUNET, F., T. GIRAUD, F. GODIN, and P. CAPY, 2001  Do deletions of the Mos1-like elements occur randomly in the Drosophilidae family? J. Mol. Evol. in press.

BRYAN, G. J., J. W. JACOBSON, and D. L. HARTL, 1987  Heritable somatic excision of a Drosophila transposon. Science 235:1636-1638[Abstract/Free Full Text].

COOLEY, L., C. BERG, and A. SPRADLING, 1988a  Controlling P element insertional mutagenesis. Trends Genet. 4:254-258[Medline].

COOLEY, L., R. KELLEY, and A. SPRADLING, 1988b  Insertional mutagenesis of the Drosophila genome with single P elements. Science 239:1121-1128[Abstract/Free Full Text].

CROSBY, M. A. and E. M. MEYEROWITZ, 1986  Lethal mutations flanking the 68C glue gene cluster on chromosome 3 of Drosophila melanogaster. Genetics 112:785-802[Abstract/Free Full Text].

ENGELS, W. R., D. M. JOHNSON-SCHLITZ, W. B. EGGLESTON, and J. SVED, 1990  High-frequency P element loss in Drosophila is homologue dependent. Cell 62:515-525[Medline].

GARZA, D., M. MEDHORA, A. KOGA, and D. L. HARTL, 1991  Introduction of the transposable element mariner into the germline of Drosophila melanogaster. Genetics 128:303-310[Abstract].

GLOOR, G. B., N. A. NASSIF, D. M. JOHNSON-SCHLITZ, C. R. PRESTON, and W. R. ENGELS, 1991  Targeted gene replacement in Drosophila via P-element-induced gap repair. Science 253:1110-1117[Abstract/Free Full Text].

GORYSHIN, I. Y., Y. V. KIL, and W. S. REZNIKOFF, 1994  DNA length, bending, and twisting constraints on IS50 transposition. Proc. Natl. Acad. Sci. USA 91:10834-10838[Abstract/Free Full Text].

HARTL, D. L., 2001  Discovery of the transposable element mariner. Genetics 157:471-476[Free Full Text].

HARTL, D. L., A. R. LOHE, and E. R. LOZOVSKAYA, 1997  Modern thoughts on an ancyent marinere: function, evolution, regulation. Annu. Rev. Genet. 31:337-358[Medline].

HEINO, T. I., A. O. SAURA, and V. SORSA, 1994  Maps of the salivary gland chromosomes of Drosophila melanogaster. Dros. Inf. Serv. 73:619-738.

JACOBSON, J. W., M. M. MEDHORA, and D. L. HARTL, 1986  Molecular structure of a somatically unstable transposable element in Drosophila. Proc. Natl. Acad. Sci. USA 83:8684-8688[Abstract/Free Full Text].

KAUFMAN, P. D. and D. C. RIO, 1992  P element transposition in vitro proceeds by a cut-and-paste mechanism and uses GTP as a cofactor. Cell 69:27-39[Medline].

KIM, J. M. and W. KIM, 1999  Identification of a full-sized hobo element and deletion derivatives in Korean populations of Drosophila melanogaster. Mol. Cells 9:127-132[Medline].

KIYOSAWA, H. and P. F. CHANCE, 1996  Primate origin of the CMT1A-REP repeat and analysis of a putative transposon-associated recombinational hotspot. Hum. Mol. Genet. 5:745-753[Abstract/Free Full Text].

LOHE, A. R. and D. L. HARTL, 1996  Reduced germline mobility of a mariner vector containing exogenous DNA: Effect of size or site? Genetics 143:1299-1306[Abstract].

LOHE, A. R., D.-A. LIDHOLM, and D. L. HARTL, 1995  Genotypic effects, maternal effects and grand-maternal effects of immobilized derivatives of the transposable element mariner. Genetics 140:183-192[Abstract].

LOHE, A. R., D. T. SULLIVAN, and D. L. HARTL, 1996  Genetic evidence for subunit interactions in the transposase of the transposable element mariner. Genetics 144:1087-1095[Abstract].

LOHE, A. R., C. TIMMONS, I. BEERMAN, E. R. LOZOVSKAYA, and D. L. HARTL, 2000  Self-inflicted wounds, template-directed gap repair, and a recombination hotspot: effects of the mariner transposase. Genetics 154:647-656[Abstract/Free Full Text].

LOZOVSKY, E. R., D. NURMINSKY, E. A. WIMMER, and D. L. HARTL, 2002  Unexpected stability of mariner transgenes in Drosophila. Genetics 160:527-535[Abstract/Free Full Text].

MARUYAMA, K. and D. L. HARTL, 1991  Evolution of the transposable element mariner in Drosophila species. Genetics 128:319-329[Abstract].

NASSIF, N. A., J. PENNEY, S. PAL, W. R. ENGELS, and G. B. GLOOR, 1994  Efficient copying of nonhomologous sequences from ectopic sites via P element-induced gap repair. Mol. Cell. Biol. 14:1613-1625[Abstract/Free Full Text].

PASCUAL, L. and G. PERIQUET, 1991  Distribution of hobo transposable elements in natural populations of Drosophila melanogaster. Mol. Biol. Evol. 8:282-296[Abstract].

PERIQUET, G., M. H. HAMELIN, Y. BIGOT, and K. HU, 1989  Presence of the deleted hobo element Th in Eurasian populations of Drosophila melanogaster. Genet. Sel. Evol. 21:107-111.

REITER, L. T., T. MARUKAMI, T. KOEUTH, L. PENTAO, and D. M. MUZNY et al., 1996  A recombination hotspot responsible for two inherited peripheral neuropathies is located near a mariner transposon-like element. Nat. Genet. 12:288-297[Medline].

RIO, D. C. and G. M. RUBIN, 1988  Identification and purification of a Drosophila protein that binds to the terminal 31-base-pair inverted repeats of the P transposable element. Proc. Natl. Acad. Sci. USA 85:8929-8933[Abstract/Free Full Text].

ROBERTSON, H. M., C. R. PRESTON, R. W. PHILLIS, D. JOHNSON-SCHLITZ, and W. K. BENZ et al., 1988  A stable genomic source of P element transposase in Drosophila melanogaster. Genetics 119:75-83[Abstract/Free Full Text].

STAVELEY, B. E., T. R. HESLIP, R. B. HODGETTS, and J. B. BELL, 1995  Protected P-element termini suggest a role for inverted-repeat-binding protein in transposase-induced gap repair in Drosophila melanogaster. Genetics 139:1321-1329[Abstract].

TOSI, L. R. O. and S. M. BEVERLEY, 2000  Cis and trans factors affecting Mos1 mariner evolution and transposition in vitro, and its potential for functional genomics. Nucleic Acids Res. 28:784-790[Abstract/Free Full Text].

This article has been cited by other articles:

				D. C. Williams, T. Boulin, A.-F. Ruaud, E. M. Jorgensen, and J.-L. Bessereau Characterization of Mos1-Mediated Mutagenesis in Caenorhabditis elegans: A Method for the Rapid Identification of Mutated Genes Genetics, March 1, 2005; 169(3): 1779 - 1785. [Abstract] [Full Text] [PDF]

				C. Miskey, Z. Izsvak, R. H. Plasterk, and Z. Ivics The Frog Prince: a reconstructed transposon from Rana pipiens with high transpositional activity in vertebrate cells Nucleic Acids Res., December 1, 2003; 31(23): 6873 - 6881. [Abstract] [Full Text] [PDF]

				D. A. O'Brochta, N. Sethuraman, R. Wilson, R. H. Hice, A. C. Pinkerton, C. S. Levesque, D. K. Bideshi, N. Jasinskiene, C. J. Coates, A. A. James, et al. Gene vector and transposable element behavior in mosquitoes J. Exp. Biol., November 1, 2003; 206(21): 3823 - 3834. [Abstract] [Full Text] [PDF]

				E. R. Lozovsky, D. Nurminsky, E. A. Wimmer, and D. L. Hartl Unexpected Stability of mariner Transgenes in Drosophila Genetics, February 1, 2002; 160(2): 527 - 535. [Abstract] [Full Text] [PDF]