Abstract
The P element in Drosophila melanogaster preferentially transposes into nearby sites. The local insertions display a preferential orientation toward the starting element. We investigated the mechanism of the P-element local transposition by isolating and characterizing local insertions in the male germline. We designed a genetic screen employing a marker gene that is carried in the P element and is dose sensitive. This dose effect allows isolation of flies containing newly transposed P elements in the presence of the starting element. A rapid molecular screen with PCR was used to identify 45 local insertions located within an ~40-kb genomic region on both sides of the starting element. Our system permits the isolation of the cluster progeny derived from a single insertion event, but none was isolated. The data suggest that local transposition occurs in the meiotic cell cycle. Nearly all of the local insertions were located within the promoter regions of the genes that were active in the male germline cells, suggesting that local insertions target predominantly active promoters. Our analysis shows that local transposition of the P element is highly regulated, displaying a cell-type specificity and a target specificity.
THE P-transposable element in Drosophila melanogaster is one of the best-characterized eukaryotic transposons (Engels 1989; Rio 1990). Full-length P elements are 2907 bp in length and encode an 87-kD transposase protein that is a site-specific endonuclease (Beall and Rio 1997). P-element transposition requires several sequences at both ends of the element, in addition to the 31-bp terminal inverted repeats (O'Hare and Rubin 1983; Kaufmanet al. 1989; Mullinset al. 1989). P elements transpose by a cut-and-paste mechanism and generate an 8-bp target site duplication upon insertion (O'Hare and Rubin 1983; Engelset al. 1990; Glooret al. 1991). Following excision, the P element induces a double-strand break at the donor site, which is usually repaired by gene conversion. Although P elements insert into genomic sequences nonrandomly, the target selectivity is independent of a specific sequence motif (O'Hare and Rubin 1983; Liaoet al. 2000). Nonrandom insertion of P elements is also reflected in the rate of P-element-induced mutations at different genes. The analysis of a collection of 3900 P insertion lines associated with recognizable lethal, semilethal, sterile, or visible phenotypes showed that 70% of these P elements were inserted preferentially into ~400 hotspot loci (Spradling et al. 1995, 1999). P elements also have a tendency to insert into the 5′ end of genes (Spradlinget al. 1995).
New insertions of the P element were frequently found near the existing P elements, exhibiting a structure with two closely located P elements (Hawleyet al. 1988; Roihaet al. 1988; Eggleston 1990). These insertions resulted from the P-element local transposition activity by which an element preferentially transposes into nearby sites (Toweret al. 1993; Zhang and Spradling 1993). These studies showed that the frequency of P-element local insertions into a minichromosome was increased more than 100-fold when the starting P element was located on the minichromosome. In the absence of selection, the local insertions account for as much as 20–40% of the total insertions throughout the genome (Zhang and Spradling 1993). Several studies showed that the local insertions were preferentially oriented in a reverse direction toward the starting element (Hawleyet al. 1988; Roihaet al. 1988; Eggleston 1990; Daniels and Chovnick 1993; Toweret al. 1993; Zhang and Spradling 1993; Delattreet al. 1995). In conjunction with the high frequency, this orientational preference of the local insertions has led to a hypothesis that the excised P element is associated with proteins that tether it to the donor site in a fixed orientation via protein-protein interaction, resulting in the preferential orientation and frequent insertions into nearby sites (Zhang and Spradling 1993).
Genetically engineered P elements have been extensively used in single P-element insertional mutagenesis to study the Drosophila genes (Cooleyet al. 1988; Spradling et al. 1995, 1999). Some researchers have also proposed using frequent local transposition as a mutagenesis method to study gene functions (Toweret al. 1993; Zhang and Spradling 1993). Because simply examining the expression of the marker gene carried in the P element could no longer identify a new insertion near the starting element, the isolation of local insertions was often complicated. To identify local insertions, several studies used the changes of P-carrying marker expression as an indicator of local transposition (Toweret al. 1993; Zhang and Spradling 1994; Zhang and Stankiewicz 1998). This was made possible by a starting element expressing the rosy+ marker at a low level due to its location in heterochromatin (position-effect variegation). Following activation by transposase, the production of a new insertion was indicated by elevated marker expression in the adult eyes. Local insertions were also isolated with changes of the phenotypes associated with the starting elements (Hawleyet al. 1988; Roihaet al. 1988; Daniels and Chovnick 1993: Toweret al. 1993; Delattreet al. 1995). In addition, an attempt was made to physically identify local insertions by using the molecular methods (Zhang and Spradling 1993).
The P element often transposes premeiotically in the developing germline, giving rise to a cluster of progeny derived from a single insertion event (Spradling 1988). P-element transposition in the meiotic cell cycle was also observed when the P element transposed locally (Daniels and Chovnick 1993). It was thought that association of the excised P element with the donor site leads to a local transposition event, while dissociation of the excised element from the donor site frees the transposon and leads to an insertion elsewhere in the genome (Zhang and Spradling 1993). In this report we devised experiments to address whether local transposition occurs prior to the meiotic cell cycle by examining the cluster events among the chromosomes carrying local insertions. By using the dose effect of a marker gene expression, a simple genetic screen was employed to isolate chromosomes carrying local insertions. These insertions were identified by using the polymerase chain reaction (PCR) with pairs of primers derived from the 31-bp inverted terminal repeats of the P element and from the genomic sequences. The data show that all 45 characterized insertions were produced from independent local transposition events. These insertions are located within the promoter regions of the genes that were transcriptionally active in the male germline cells, suggesting that local insertions target predominantly active promoters.
MATERIALS AND METHODS
Drosophila strains: Flies were grown on standard corn meal/agar media at 22°. A stock containing a P-element insertion, EP(3)3583, was obtained from The Berkeley Drosophila Genome Project at the University of California, Berkeley.
PCR to amplify Drosophila genomic DNA: Drosophila genomic DNA was isolated as described in Ashburner (1989). The polymerase chain reactions were carried out using the Gene-Amp system (PE Applied Biosystems) under standard conditions on a Robocycler (Stratagene, La Jolla, CA). Two types of PCR primers were employed, which were derived from the P element (Rubin and Spradling 1983) and the genomic sequences. The following primers were derived from the P element:
Pp31: 5′ CGACGGGACCACCTTATGTTATTTCATCATG 3′(the 31-bp inverted terminal repeat and with an outward orientation)
PE5′: 5′ AATTCGTCCGCACACAAC 3′ (112 bp internal to the 5′ end of the P element and with an outward orientation)
PE3′: 5′ TCGCACTTATTGCAAGCA 3′ (72 bp internal to 3′ end of the P element and with an outward orientation)
The following primers were derived from genomic sequences, which were based on a genomic scaffold sequence in 67B (GenBank accession no. AE003552; Adamset al. 2000):
Pg-L: 5′ GACTTGGCCTGTTCCTTG 3′
Pg-R: 5′ GAGCCAGAAGATGCGAGA 3′
Pg-A1: 5′ AGCGGATAATGGCGTGTA 3′
Pg-A2: 5′ GATGTACGGCAGTATCGG 3′
Pg-A3: 5′ TAGCTGCACATTTGCTTG 3′
Pg-A4: 5′ GCGCGTACGACAACAACT 3′
Pg-A5: 5′ GTGCCTGGAGCTATAGCC 3′
Pg-A6: 5′ GCTCCTTGGACTTGTCCT 3′
Pg-B1: 5′ TTGTCTCTCCGCTCTCCT 3′
Pg-B2: 5′ ACATTCGCATAGTGCTGG 3′
Pg-C1: 5′ ATGTTCGCACTTCTTGCA 3′
Sequencing: Two rounds of PCR reactions were used to clone the genomic sequences flanking an EP element. The first PCR was carried out with a genomic primer near the inserted EP element and the EP3′ primer or the EP5′ primer, which were specific to the 5′ end or the 3′ end of the element. The PCR product was loaded on a gel of 0.7% low-melting-temperature agarose (FMC Bioproducts, Rockland, ME) in TAE and the band containing the amplified DNA was cut out. The gel-purified DNA was used as a template for a second PCR reaction with the same primers used in the first PCR reaction. The DNA product was purified using a QIAquick PCR purification kit (QIAGEN, Chatsworth, CA) and was used in a sequencing reaction (PE Applied Biosystem) with either the PE5′ or PE3′ primer.
RESULTS
The design of our genetic screen to isolate local insertions of the P element was based on two assumptions. One was that the starting element used to produce local insertions was mostly retained in local transposition. A previous study isolated local insertions near a starting P element on a minichromosome (Dp1187) by using a genetic screen (Toweret al. 1993). The starting element was present in nearly all of the minichromosomes carrying the insertions (17/19), although it had structural changes in two cases. Local insertions were also identified by the method of Southern hybridization (Zhang and Spradling 1993). Since this method detected the transposon molecularly rather than genetically, in theory it allowed the recovery of all types of P-element insertions in the absence of any selection. The results from this study showed that the starting element was present on all chromosomes carrying the local insertions (34/34).
The other assumption was that the expression of a marker gene in the P element would allow us to identify flies carrying two closely located P elements, i.e., a retained starting element and a new local insertion. The mini-white gene is derived from the X-linked white gene and is carried in many P-element transformation vectors (Pirrotta 1988). Because regulatory elements necessary for high levels of the white expression in the adult eyes are absent from the mini-white gene, flies carrying a mini-white transgene mostly display pale-yellow to orange-red eyes, depending on chromosomal sites where the transgene is inserted. The amount of eye pigmentation in mini-white transformants increases when the number of mini-white copies increases in the genome (Leviset al. 1985; Pirrottaet al. 1985). We employed this dose effect—i.e., the amount of eye pigmentation is proportional to the number of transgenic mini-white copies—to isolate potential strains carrying local insertions.
The EP(3)3583 insertion in 67B. (A) A schematic drawing of the genomic region around the EP(3)3583 insertion. Two P-element-specific primers (PE5′ and PE3′) and two genomic primers (Pg-L and Pg-R) are shown. The arrows indicate the directions of the primers (5′ to 3′). (B) Amplification of genomic DNA by PCR. The genomic DNA samples were isolated from wild type (lane 1) and the EP(3)3583/EP(3)3583 flies (lanes 2–4). The PCR products were produced from three pairs of primers: Pg-L and Pg-R (lane 1 and 2), Pg-L and PE3′ (lane 3), and Pg-R and PE5′ (lane 4). Molecular size standards (100-bp ladder) are shown on the left.
The starting element: The starting P element for this local transposition study was an EP element located in 67B on chromosome 3 (Rorth 1996; Rorthet al. 1998; Liaoet al. 2000). Homozygotes for this insertion were viable and fertile. The genomic location of this inserted transposon, EP(3)3583, was confirmed by PCR. The PCR primers derived from the genomic sequence around EP(3)3583 and the ends of this EP element produced the predicted products (Figure 1). When DNA from wild-type flies was used as a template, a predicted product of 926 bp was produced with two PCR primers specific for the genomic sequence Pg-L (at −716) and Pg-R (at +210) (Figure 1B, lane 1). Because EP(3)3583 was inserted between Pg-L and Pg-R, no product was produced from the DNA of flies homozygous for EP(3)3583 with this primer set (Figure 1B, lane 2). Other PCR reactions of the EP(3)3583/EP(3)3583 DNA gave rise to the expected products. For example, the primer pair of Pg-L and PE3′ (specific to the 3′ end of the P element) gave rise to a predicted product of ~800 bp (the sum of 72 bp from the P element and 716 bp of the genomic DNA; Figure 1B, lane 3). Similarly, the primer pair of Pg-R and PE5′ (specific to the 5′ end of the P element) produced a predicted product of ~300 bp (the sum of 112 bp of the P element and 210 bp of the genomic DNA; Figure 1B, lane 4). These results confirmed the physical location of the EP(3)3583 element.
Genetic crosses used to produce local insertions in the 67B region. (A) Genetic crosses. Local transposition from the starting element, EP(3)3583, were generated in the F0 males with the Δ2-3 transposase source. The new insertions were recovered in the F1 males with elevated eye pigmentation. An asterisk (*) indicates a mutagenized chromosome that potentially carries a local insertion. (B) Eye pigmentation. Flies carrying a copy of the starting element (F0) displayed light-orange eye color (left). (a), (b), and (c) show the F1 flies with elevated eye pigmentation (right).
Producing local insertions in the 67B region on chromosome 3: The EP element contains mini-white as a marker (Rorth 1996) and flies carrying a copy of the EP(3)3583 insertion in 67B showed light-orange eye color (Figure 2B). To determine if the EP(3)3583 element was able to transpose, flies carrying a copy of EP(3)3583 and a transposase source were generated. The element was frequently excised in these flies, recognizable by the different phenotypes in nearly every adult eye, i.e., patches of white cells on a background of cells with light-orange pigmentation (data not shown). This simple test demonstrated that the EP(3)3583 element was able to transpose in the presence of a transposase source. We then activated EP(3)3583 with a transposase source, Δ2-3 (99B) (Robertsonet al. 1988), in w/Y; EP(3)3583/Δ2-3, Sb males (F0, Figure 2A). To isolate local insertions, 563 F0 males were crossed singly to w/w; +/+ females in vials. A total of 1554 Sb+ male progeny (F1) expressing the mini-white gene at elevated levels (Figure 2B) were collected and crossed singly to w/w; +/+ females. Up to 10 F1 males were taken from a given F0 parent and crossed singly to the females.
Local insertions on both sides of the EP(3)3583 starting element. The genomic map in 67B (thick line) was generated on the basis of a genomic scaffold sequence, AE003552, with the starting element at the center. The map is divided into seven segments (region A through region G), separated by the starting element and five genomic primers with arrows indicating the 5′–3′ direction (Pg-A1 through Pg-A5). The transcription units (FlyBase 1999) are shown as dark lines with arrows indicating the directions of the transcriptions. The triangles above the genomic map represent the local insertions and show their genomic locations. The number in parentheses above a triangle indicates the number of the independent local insertions at the location. Pp31 is a primer complementary to the 31-bp terminal repeats of the P element in an outward direction. The genomic positions of the primers are −6443 (Pg-A1), −3033 (Pg-A2), +2946 (Pg-A3), +6316 (Pg-A4), +9420 (Pg-A5), and +7302 (Pg-A6).
The change of eye color from light orange in the F0 males to dark orange or light red in the F1 males might be caused by the mini-white dose effect. Some of these F1 males might carry two copies of the EP elements: a locally inserted element and the starting element. Among 1554 crosses between the F1 males and the w/w; +/+ females, a total of 1444 crosses produced progeny. The remaining produced no progeny, because the involved F1 males were sterile. Among the fertile crosses, 287 produced F2 progeny displaying obvious different eye colors, indicating that they carried at least two unlinked EP elements. These strains involving the nonlocal transposition events were not characterized further. The rest of the 1157 stocks were subject to the molecular analysis described below.
Detecting local insertions with PCR: We used a simple screen based on PCR technology to detect local insertions in the 1157 stocks carrying potential local insertions. Five PCR primers were designed according to the genomic sequences around the starting EP(3)3583 element (Pg-A1 through A5, Figure 3). The Pp31 primer was derived from the 31-bp inverted terminal repeats of the P element. We scanned the genomic region around the starting element for local insertions by using two separate PCR reactions. First, to examine local insertions to the left of the starting element, we used a set of three primers (Pp31, Pg-A1, and Pg-A2) in a reaction. Second, to examine local insertions to the right of the starting element, we used a set of four primers (Pp31, Pg-A3, Pg-A4, and Pg-A5) in another reaction. The genomic primers were ~3 kb apart in both reactions. These designs allowed local insertions to be identified by a combination of Pp31 and a genomic primer (Pg). In addition, local insertions immediately adjacent to the starting element would be detected by the Pp31 primer, which amplifies a genomic sequence between two P elements (i.e., the retained starting element and a local insertion).
To further improve the efficiency of the PCR screen, DNA samples were extracted from pooled flies from different stocks. We combined flies with elevated eye pigmentation from 5 stocks and isolated their genomic DNA (five flies per stock, with a total of 232 DNA pools for 1157 stocks).
From the screen, PCR products were obtained from 29 pools of the genomic DNA. DNA samples were then extracted from individual stocks in which these DNA pools were isolated. The presence of a local insertion in a stock was determined by PCR with the same primers used for the DNA pools. We isolated a total of 32 stocks carrying independent local insertions. This was because 3 of the 29 DNA pools actually contained two independent local insertions per pool, which were revealed when DNA from the stocks was individually examined by PCR. These data showed that the frequency of the local insertions within this genomic region (~21 kb, from ~−9000 to ~+12,000, Figure 3) was ~2.7% (32/1157). In addition, all 32 stocks containing the local insertions displayed strong eye pigmentation, as shown in Figure 2B, (a) and (b).
Localizing the insertions within several genomic subregions: The genomic sequences around the starting element were divided into seven genomic regions, regions A through G, by the sites of the genomic primers used in the PCR screen (Pg-A1 through A5, Figure 3). To map the 32 local insertions identified in the screen, two series of experiments were carried out. First, the local insertions within regions C and D were mapped by using the Pp31 primer, which amplifies the templates between a local insertion and the starting element. This analysis identified 10 insertions. These insertions were further mapped to ~+200 (3 insertions), +400 (2 insertions), +600 (2 insertions), +700 (1 insertion), and >+2500 (2 insertions) by using several additional genomic primers in regions C and D (data not shown).
Mapping the local insertions in region F. Five insertions in the region (1–5) were analyzed. The genomic DNA samples were isolated from the adult flies containing the insertions. The PCR products were produced with two sets of primers separately. One set was Pp31 and the genomic primer Pg-A4. The other set was Pp31 and the genomic primer Pg-A6. Molecular size standards (100-bp ladder) are shown on the left.
Second, 22 insertions were mapped to regions F (17 insertions) and G (5 insertions) by using pairs of primers containing Pp31 and a genomic primer (Pg-A1–Pg-A5). Among 17 insertions mapped to region F, 15 were located within a small segment of ~100 bp at ~+6600 (Figure 3). Mapping data for five examples of these insertions are shown in Figure 4. The PCR products from these insertions with Pp31 and Pg-4 (+6316) were ~300 bp in size (Figure 4), indicating that the insertions were located at ~+6600. The Pg-A6 primer is located at +7302 (Figure 3) and is oriented in a direction opposite to that of the Pg-A4 primer (Figure 3). The PCR products from the insertions at ~+6600 with Pp31 and Pg-A6 were ~750 bp (Figure 4), which agrees with the sizes of the predicted products from these insertions. In addition to the insertions at +6600, two additional insertions in region F were mapped to the sites at ~+6400 (Figure 3). In region G, 5 insertions were mapped at ~+10,100 with Pp31 and Pg-5 (+9420).
To determine the precise physical locations of the local insertions, we selected 12 local insertions from region D (5 insertions), region F (4 insertions), and region G (3 insertions). The genomic sites of these insertions were determined by cloning and sequencing (see materials and methods for details). The results are shown in Table 1. For example, the genomic sites for 3 insertions in region F (nos. 7, 8, and 9) were very close, at positions +6561, +6577, and +6582, respectively. Another insertion (no. 6) was ~200 bp away at position +6361. These analyses confirmed the PCR mapping data obtained by estimating the sizes of the PCR products (Figures 3 and 4).
The genomic sequences around the local insertions were further examined for an 8-bp target site duplication, a characteristic of P-element insertional activity (O'Hare and Rubin 1983; Liaoet al. 2000). Five local insertions were randomly selected for this analysis. The genomic DNA segments flanking both ends of the insertions were cloned with either the PE3′ or PE5′ primer in combination with one of several genomic primers (data not shown). The sequencing data showed that all of these insertions were flanked by an 8-bp target site duplication (Table 1).
The sites of the insertions showed interesting features about the local transposition. None of these sites was located within the genomic sequences inside the transcription units of the local genes (Figure 3 and Table 1). Instead, they were distributed within several base pairs to several hundred base pairs upstream of the transcription starting sites of four genes, including Hsp27, Hsp23, Hsp26, and Hsp22 (Table 1). These data suggested that the local transposition of EP(3)3583 targeted the promoter regions that are near the starting element. In addition, the data showed an intriguing correlation between the local insertions and the transcriptional activity of the targeted promoters. Three of the targeted genes, Hsp27, Hsp23, and Hsp26, were expressed in the male germline cells where the local transposition took place, as indicated by the expressed sequence tags (ESTs) in a testis-cDNA library (Table 2). Another targeted gene, Hsp22, was also expressed in the male germline, as shown in an immunostaining study (Michaudet al. 1997). In contrast, several other genes that had no ESTs from the male germline cells received no local insertions. These included Hsp67Ba, Hsp20, Hsp67Bb, and Hsp67Bc with the transcription starting sites at +4548, +8546, +11,058, and +11,670, respectively. All of these genes were well inside the range of the PCR detection, since the primers in the PCR screen were spaced ~3 kb apart (Figure 3). The observation that the local insertions produced in the male germline (Figure 1) were located within the promoter regions that were active in the same tissue suggested that the local transposition preferentially targeted the active promoters.
Local insertions into transcriptionally active genes: To test the hypothesis that active promoters were the preferential targets of local insertions, we further divided the local insertions into active and inactive promoters. The expression of the Hsp67Ba gene with the transcription starting site at +4548 was not detected in the male germline (Table 2). Thus, this gene was not predicted by the hypothesis to be a frequent target of the local transposition. The initial PCR screen did not find any insertions into Hsp67Ba (Figure 3). Although this result supported our hypothesis, it could also be explained if the primer set of Pg-A3 and Pp31 was less efficient in detecting the insertions within region E where Hsp67Ba resides. To confirm the observation that local insertions within the Hsp67Ba gene were not present in our screen, the following PCR experiment was carried out. By employing Pg-A6 and Pp31 in PCR, we screened the genomic DNA pools for local insertions into the genomic region containing Hsp26 and Hsp67Ba. All of the 17 insertions into Hsp26 (Figure 3), which had been isolated from the initial screen, were readily identified by the Pg-A6 and Pp31 pair. For example, 5 insertions isolated from the initial screen produced ~300-bp PCR products with Pp31 and Pg-4 (Figure 4). These insertions were also identified in the screen using Pp31 and Pg-6, which produced ~750-bp products (Figure 4). However, this screen found no new insertions into Hsp67Ba. These results provided strong evidence that the local transposition preferentially inserted the P element into the active promoters.
The positions of the local insertions and their adjacent genes
Our examination of local insertions was extended into three additional genes expressed in the male germline. These included the CG4080 gene, the eIF-4E gene, and the CG4452 gene. The CG4080 gene was expressed in the male germline, but no insertions were isolated within it in the initial screen (Figure 3 and Table 2). This failure of recovering local insertions in CG4080 could be the result of our initial experimental design, which did not cover the CG4080 promoter located outside the initial PCR range (Figure 3). To determine if CG4080 was a target of local transposition, PCR with three primers, Pg-B1, Pg-B2, and Pp31, was used to screen the genomic DNA pools for local insertions in this gene (Figure 5). Pg-B2 was a primer located at −13,347, orientating toward the CG4080 promoter (its transcription starting site at −14,085). The experiment isolated six local insertions in the CG4080's promoter region (Figure 5). Using PCR with the Pg-B2 and Pp31 pair mapped the genomic locations of these insertions to ~−14,000. The precise locations for two of these insertions were determined by cloning and sequencing as described in materials and methods. One was located at −14,055, while the other was only 2 bp away at −14,057 (Table 1, region I). These results revealed that CG4080, a transcriptionally active gene in the male germline, was a frequent target of the local insertions.
Relationship between gene expression in the male germline and the local insertions
The PCR screen with Pg-B1, Pg-B2, and Pp31 also recovered two insertions at ~ −17,000, which was determined by using the Pg-B1 (at −16,706) and Pp31 pair (Figure 5). The insertion site for one of these insertions was cloned and sequenced. This site, at −16,993 (Table 1, region H), was 283 bp upstream of the transcription starting site of eIF-4E (at −17,275, Table 1), which was expressed in the male germline (Table 2).
Another PCR screen with the Pg-C1 (+15,852) and Pp31 primer pair recovered five insertions at ~+17,100 (Figure 5). By cloning and sequencing, two of these insertion sites were located at the same position, +17,075 (Table 1, region J). These insertions were near the transcription starting sites of the CG4452 and Klp67A genes at +16,574 and +17,051, respectively (Table 1). The CG4452 gene was expressed in the male germline (Table 2).
The absence of the cluster events in the local transposition: In the spermatogenesis of D. melanogaster, a germline stem cell undergoes four rounds of mitotic divisions, producing 16 primary spermatocytes in a cyst (Lindsley and Tokuyasu 1980; Fuller 1993). The meiotic divisions generate 64 haploid round spermatids, which develop into mature spermatozoa. P element often transposes premeiotically in the developing germline cells, giving rise to a cluster of progeny derived from a single insertion event (Spradling 1988). Of 45 local insertions isolated in our screen (Figures 3 and 5), 31% (14/45) were collected from vials that produced a single F1 male with elevated eye pigmentation. The remainder, 69% (31/45), were collected from 31 individual transposition vials and examined as F1 males along with one or more siblings (ranging from two to nine) produced from a given F0 male. Although our screen allowed the recovery of the cluster progeny, no clusters of the local insertions were found throughout the entire screen. Instead, each one of the 45 local insertions was derived from an independent insertion event.
Local insertions in regions H, I, and J. These regions were divided by the genomic primers Pg-B1, Pg-B2, and Pg-C1 (arrows). The genomic map and the symbols are as described in Figure 3. The genomic positions of the primers are −16,704 (Pg-B1), −13,347 (Pg-B2), and +15,854 (Pg-C1).
Retaining of the starting element on the chromosomes carrying the local insertions. Eight chromosomes carrying the insertions in region F (1–8) were examined for the presence of the starting element. The PCR products were produced either with the Pp31 and Pg-L primer pair (a) or with the Pp31 and Pg-R primer pair (b). Molecular size standards (100-bp ladder) are shown on the left.
Retaining of the original element: As described above, one of the assumptions in our strategy to isolate the local insertions was that the starting element was retained following local transposition. The fate of the starting element on the chromosomes carrying the local insertions was determined by using PCR with two primer sets. One pair of the primers contained Pp31 and Pg-L, whereas the other pair contained Pp31 and Pg-R (Figures 1 and 3). An example of these examinations is shown in Figure 6. Out of eight chromosomes carrying the local insertions, seven retained the starting element, indicated by the production of an ~750-bp product with the Pp31 and Pg-L primer pair and an ~250-bp product with the Pp31 and Pg-R primer pair. The exceptional chromosome produced a ~250-bp product with Pp31 and Pg-R, but not with Pp31 and Pg-L (8a and 8b, Figure 6). Thus, this chromosome retained the 5′ end of the starting element, but had a rearrangement around the 3′ end. These analyses showed that nearly all of the chromosomes carrying the local insertions (43/45) retained the starting element. Two exceptional chromosomes retained the 5′ end of the starting element, although they carried the DNA rearrangement around the 3′ end as described above.
Preferential orientation of the local insertions: The orientation of the local insertions relative to the starting element was determined by using PCR with a genomic primer around a local insertion and either the PE5′ or PE3′ primer. A significant preference for the insertion orientation was observed. Among a total of 40 local insertions randomly chosen for this analysis, 28 were in the opposite orientation (head to head or tail to tail), while 12 were in tandem orientation (head to tail). However, the preference for the insertion orientation was present only among the insertions located to the 5′ side of the starting element. The vast majority of these insertions displayed a head-to-head orientation (24/32 or 75%). In contrast, 8 insertions 3′ to the starting element appeared to be randomly distributed in either orientation (50% vs. 50%).
Rearrangement associated with the local transposition: The screen with the primer set of Pp31 and Pg-6 identified not only the local insertions within the Hsp26 gene (Figure 4), but also a group of five chromosomes with irregular properties. The PCR products of these unusual chromosomes with Pp31 and Pg-6 were the same size as the regular insertions in Hsp26, which was ~750 bp (Figure 4). However, no PCR products were seen in the PCR experiments with Pp31 and Pg-4. The results suggested that these chromosomes carried local insertions within Hsp26, but the insertions were accompanied by genomic rearrangements. Furthermore, the PCR experiments with Pp31 and Pg-R produced no products from these chromosomes. These analyses suggested that the unusual insertions in Hsp26 were associated with a deletion between the insertions in Hsp26 and the starting element.
DISCUSSION
Efficient recovery of local insertions with simple genetic and molecular techniques: The genetic screen used to isolate the local insertions in this report was facilitated by the dose effect of the mini-white marker gene and the retaining of the starting element in local transposition. The molecular strategy used to detect the local insertions was based on the efficient polymerase chain reaction that was rapid and sensitive in the screens. The frequency of recovering the local insertions is ~4% (45/1147) within a genomic interval of ~40 kb (~−20,000–~+20,000). In contrast, a previous nonselective screen showed, by using Southern hybridization, that the local insertions into a minichromosome account for 20–40% of all transpositions throughout the genome (Zhang and Spradling 1993). At least three factors contribute to the difference between the observed frequency of these two screens.
First, the previous screen searched for local insertions in a genomic interval that is seven times larger than the current screen (~300 kb vs. ~40 kb). Second, it is possible that the elevated expression of the mini-white marker gene in some of the F1 males was not due to a new insertion. We noted that the flies carrying the local insertions all displayed highly elevated eye pigmentation as illustrated in Figure 2B, (a) and (b). Roughly one-third of the examined 1554 F1 males were flies with slightly increased eye pigmentation [Figure 2B (c)]. However, none of them was found to contain local insertions. Some of these F1 males were associated with genomic DNA rearrangements around the starting element, which were revealed by PCR experiments (data not shown). The increased mini-white expression in some of the F1 males could also be caused by DNA rearrangements internal to the starting element that were induced by the transposase activity. These DNA rearrangements, rather than new insertions, may give rise to the increased mini-white expression in a sizable fraction of the F1 males. Finally, the PCR method employed in our screen to identify the local insertions would not detect the local insertions within the starting element, since the Pp31 primer is located in the inverted terminal repeats of the P element. The local insertions transposed onto the starting element account for 44% of all the local insertions (15/34) in a study that detected the insertions by using Southern hybridization (Zhang and Spradling 1993). In addition, ~12% of the total local insertions (4/34) was located <200 bp outside the starting element. This current screen with PCR did not isolate these insertions (0/45). It remains possible that the PCR screen failed to detect the small products amplified between two P elements that were separated by <200 bp genomic DNA.
Preferential targeting of local transposition and active promoters: The local transposition shares similar features with the interchromosomal transposition. The preference of P-element insertions for the 5′ end of genes was reported previously (Spradlinget al. 1995). A strong preference was also seen in our experiment of local transposition, which isolated the local insertions exclusively in the 5′ end of the genes (Figures 3 and 5). Similar to the nonrandom distribution of the interchromosomal insertions in the target genes (Spradlinget al. 1995), the local insertions are nonrandomly distributed in the local genes. Some local genes were targeted multiple times, whereas others were not targeted at all. For example, 17 independent insertions were located in Hsp26, but none was in Hsp67Ba (Figure 3).
The results from our initial screen displayed a strong correlation between the local insertion sites and the active promoters in the male germline where the local transposition took place. These promoters, including Hsp27, Hsp23, Hsp26, and Hsp22, were frequent targets of the local insertions. Further screens into regions H, I, and J revealed that the 5′ ends of three more active genes, CG4080, eIF-4e, and CG4452, were also targets of local transposition. Unlike the insertions in the Hsp genes, the insertions into these genes were located inside the transcription units as described in FlyBase (1999). However, these genes have multiple transcription starting sites that were used differentially in development. First, a database analysis showed that the testis-specific EST sequences for CG4080 (AT05655, AT18359, AT18379, and AT30244) are 45 bp shorter at the 5′ end than the EST sequences used to define the transcription starting site in the FlyBase (1999). Second, the testis-specific EST sequence for eIF-4e (AT17353) is >1000 bp shorter at the 5′ end than the EST sequences used to define the transcription starting site in the FlyBase (1999). These analyses suggest that the local insertions into CG4080 and eIF-4e were within the testis-specific promoter regions, as for the insertions in the Hsp genes.
Three insertions into region J were located 501 bp upstream of the transcription starting sites of the CG4452 gene, which is an active gene in the male germline (Table 2). However, since the genomic sequence between the transcription starting sites of CG4452 and the Klp67A gene is <500 bp, these insertions were actually located 24 bp inside the 5′ end of the Klp67A transcription unit. Although the testis-specific EST sequences for Klp67A were not found, this gene was also active in the male germline as shown in Northern analysis (Stewartet al. 1991). Thus, it is possible that the insertions in region J resulted from the preferential local insertion of P elements into the active Klp67A promoter.
Mechanism of the P-element local transposition: P elements are thought to transpose by a nonreplicative cut-and-paste mechanism (Engels 1989; Glooret al. 1991). Transposition starts by an excision of the element, resulting in a double-strand break. The break is usually repaired by gene conversion, which copies sequences from a template on either the sister chromatid or the homolog. Our observation that the starting element was retained on the chromosomes carrying the local insertions indicates that the local insertions were generated by integration of the excised P element into either the sister chromatid or the repaired chromatid that used the sister chromatid as the template.
In conventional experiments, the P element on one chromosome is activated and transposes onto a different chromosome. The interchromosomal transposition often takes place premeiotically in the developing germline, producing a cluster of progeny derived from a single insertion event. In P-element-mediated transformation, the clusters were found in >50% of the siblings when two or more siblings of a given G0 fly (developed from an injected embryo) were tested (Spradling 1988). Although no published data show how frequently clusters occur in interchromosomal transposition, these events unquestionably occur (our unpublished observation). A. Spradling (personal communication) has analyzed the target sequences of an entire set of 1866 P-element insertions that were generated by retaining two or more siblings. He showed that 28% of these insertions (526/1866) displayed the typical cluster characteristics; i.e., a sibling had an insertion at a site identical to that of one or more other siblings. Hence, if one takes a new line from such a collection, one has an ~28% chance that it is part of a cluster of two or more siblings. The data demonstrated that the P element frequently transposes prior to the completion of DNA replication at the 16-cell stage of primary spermatocytes, resulting in the interchromosomal clusters.
The absence of the cluster events in our experiment (0/45) suggests that local transposition of P elements occurs either mostly or entirely during the meiotic cell cycle, resulting in two heterozygous sister chromatids. We propose that the difference between an interchromosomal transposition and a local transposition results from differential timing when the transposition takes place. Premeiotic transposition could lead to dissociation of an excised element from the donor site and to insertion into a new site. This often produces the observed cluster of interchromosomal transposition. In contrast, transposition during the meiotic cell cycle may predominantly lead to an aborted process in which an excised element fails to dissociate from the donor site and inserts locally. In meiosis, there might be a reduction of factors required for an excised element to dissociate from the donor site. The occurrence of local transposition in the meiotic cell cycle has been previously demonstrated by recovering both sister chromatids of a chromosome after P-element activation in a meiotic mutant background (Daniels and Chovnick 1993). Our model for P-element local transposition and transposition to unlinked sites is reminiscent of the observation that mutually exclusive mechanisms are responsible for intramolecular and intermolecular transpositions of the IS911 transposon in bacteria (Polardet al. 1992). It is also possible that the accessibility of genomic target sites by the excised P element is limited in the meiotic cell cycle. Thus, the target preference of local transposition could be explained if the excised P element in the meiotic cell moves along the nearby chromosome region and predominantly inserts into the open chromatin containing an active promoter.
The local insertions 5′ to the starting element showed a preferential direction (P < 0.01) in the head-to-head orientation compared with the head-to-tail orientation (75% vs. 25%). This orientational preference suggests that the excised P element is physically restricted to the donor site in local transposition. Similar preference was also seen in a previous study that recovered the local insertions in the absence of selection (Zhang and Spradling 1993). To explain the orientational preference in local transposition, it was proposed that proteins and the excised element would form a complex that is tightly associated with the donor site and maintains the excised element in a fixed orientation (Zhang and Spradling 1993). This complex would then lead the excised transposon to insert into a nearby site with the preferential orientation. The assembly of such a complex on the excised P element was implicated in in vitro assays showing that the P-element termini resisted exonucleolytic degradation (Beall and Rio 1997). These assays also showed that transposase made a staggered cut at the P-element termini, producing a 3′ end at the terminus of the P element and a 5′ end at 17 bp within the P-element 31-bp inverted terminal repeats. It is possible that one end of the P element is cleaved while the other end is still covalently linked to the donor site. The staggered cleavage activity could cause the P element with one cleaved terminus to be fixed in an orientation on the donor site, giving rise to a local insertion in a reverse direction toward the starting element.
Practice of local transposition in mutagenesis:Although local transposition occurs frequently, its success in mutagenesis has been limited. This was mainly because of technical difficulties in detecting local insertions in the presence of the starting element (Toweret al. 1993; Zhang and Spradling 1993, 1994; Zhang and Stankiewicz 1998). In this report, we combined a simple genetic screen employing the mini-white dose effect and a simple molecular screen with PCR detection. The data show that this combination can be used to efficiently isolate P-element insertions near an existing element. Thus, our system is a new addition to the genetic tool of isolating mutant alleles without the need of knowing the mutant phenotypes (Rong and Golic 2000, 2001).
The utility of this method to isolate local insertions can now be applied to most genes in the Drosophila genome. First, a collection of 2266 unselected EP insertions has been made available recently (Liaoet al. 2000). These insertions are distributed throughout the Drosophila genome and they are a good resource to be used as the starting elements for local mutagenesis. Although most of the local insertions so far characterized were mapped within a short distance from the starting element (~10–20 kb), local insertions were seen as far as 100 kb away from the starting elements (Toweret al. 1993; Zhang and Spradling 1993). Second, with the completion of sequencing the Drosophila genome (Adamset al. 2000), the genomic sequences needed to make the PCR primers have been readily available.
An obstacle to the use of local transposition is the target selectivity of the local insertions. As shown in this report, P elements inserted predominantly into the promoter regions that were transcriptionally active in the male germline. Local transposition in females may provide an alternative approach to mutagenize Drosophila genes. The local insertions produced in the female germline displayed a distribution pattern different from that of the male germline (Zhang and Spradling 1993). These data suggest that P-element local transposition to mutagenize a specific gene could be achieved more efficiently by using either males or females to carry out the local transposition, depending on the expression pattern of the target gene. A readily available resource for determining gene expression in the germline is the EST collection from the female germline (the GM library) and the male germline (the AT library; Rubinet al. 2000). However, further investigation of local transposition in females is needed to determine if active promoters in the female germline are preferentially targeted by local insertions as in the male germline.
The transpositional activity of the P element is accompanied by deficiencies around the starting element (Tsubota and Schedl 1986; Salzet al. 1987; Zhang and Spradling 1993; Prestonet al. 1996). We have also recovered five apparent deletions flanking the EP(3)3583 element with the Pp31 and Pg-A6 primer set. Mechanistically, these deficiencies were most likely produced from a local transposition followed by a deletion between the new insertion and the starting element (Zhang and Spradling 1993). This is because the deletions were mapped between the starting element and a small interval in region F where 17 local insertions were mapped (Figure 3). When activated by transposase, two closely located P elements induced deletion of the genomic sequence between the elements (Cooleyet al. 1990). It is also possible that the deletions were induced by P-element-induced male recombination around the starting element (Prestonet al. 1996). Regardless of how the deletions were produced, the local insertions described in this report are a resource for generating deficiencies around the EP(3)3583 element in 67B.
Acknowledgments
We thank the Berkeley Drosophila Genome Project at the University of California for the EP(3)3583 stock and the Biotechnology Center at the University of Connecticut for primer synthesis and sequencing. Our special thanks go to Allan Spradling for sharing with us the unpublished data of the interchromosomal clusters. This work was supported in part by the U.S. National Science Foundation Grant MCB-0077817 and a grant from the University of Connecticut Research Advisory Council.
Footnotes
-
Communicating editor: S. Henikoff
- Received September 14, 2001.
- Accepted December 10, 2001.
- Copyright © 2002 by the Genetics Society of America
