| Edited by Etsuko Matsuura. Masanobu Itoh: Corresponding author. E-mail: mitoh@kit.ac.jp |
Transposons, or transposable genetic elements (TEs) occupy a significant portion of eukaryotic genomes (Craig et al., 2002; Kaminker et al., 2002; Bergman et al., 2006), exerting a large influence on the host genome by their self-propagating lifestyle. The P element of Drosophila melanogaster is one of the most popular model systems for studies of the evolution of TEs. It is a causative factor for P-M hybrid dysgenesis (for reviews, see Kidwell, 1994; Ashburner et al., 2005) and in the P-M system, P strains have both P element transposition-inducing and -repressing capacity, while Q strains have only repressing ability, and M strains have neither. M strains without P elements are called “true M”, and those with P sequences in the genome are M’ strains (Bingham et al., 1982). In the progeny of M females and P males, P-M hybrid dysgenesis occurs, which include temperature-dependent sterility, and elevated rates of mutation, chromosome rearrangement, and recombination (Kidwell et al., 1977).
P elements can be classified, by their structure and function, into four types. First are the full-size P (FP) elements (2907 bp) that encode both a 87 kDa transposase and a 66 kDa repressor (for forming P cytotype), and thus are called autonomous (O’Hare and Rubin, 1983). Production of the proteins is regulated by alternative splicing (for a review, see Rio, 2002). Second are some elements that have internal deletions and encode the repressor, but not the transposase (type I repressor elements: Gloor et al., 1993). Third are some other elements that feature neither the transposase nor the repressor because of larger deletions, but are capable of repressing P transposition by their specific protein product (type II repressor elements). Last are other small elements that have no function in P element regulation. Most small P elements have been demonstrated to be derived from FP elements by various internal deletions (O’Hare and Rubin, 1983; Black et al., 1987; O’Hare et al., 1992). The KP element is one type II repressor element (Andrew and Gloor, 1995), probably because the KP element-encoded protein inhibits transposition by competing with the transposase for the terminal binding sites in vitro (Lee et al., 1998).
It is thought that P-M phenotype (P, Q, or M) depends on the genomic P elements content (size classes and their number) (Black et al., 1987; Daniels et al., 1987; Jackson et al., 1988; Gloor et al., 1993; Brookfield, 1996; Quesneville and Anxolabéhère, 1998), although position effects are assumed to play important roles (Laski et al., 1986; Robertson and Engels, 1989; Biemont et al., 1990; Misra et al., 1993; Rasmusson et al., 1993; Ronsseray et al., 1998). On the other hand, a survey of recent wild populations demonstrated that FP and KP elements are two of the major P elements (FP + KP predominance) irrespective of their phenotypic variation in the North and South Americas (Ruiz and Carareto, 2003; Itoh et al., 2007), Australia (Itoh et al., 1999, Itoh and Boussy, 2002; Ogura et al., 2007), Africa (Itoh and Boussy, 2002), and Asia (Itoh et al., 2001, 2004). Itoh et al. (1999) also found only a weak correlation of the FP/KP ratio with type II repression of each line, suggesting that some P elements are inactive. However, molecular basis of their inactivation is not fully elucidated.
In order to know how the genomic P elements are inactivated, we have focused on genomic P elements in M’ strains, in which all of P elements appear to be inactive. The reasons why M’ strains harboring FP and KP elements in the genomes have neither transposition-inducing nor -repressing ability were examined. Here we present isolation and molecular characterization of P elements in an M’ strain established from a wild-caught female. DNA sequencing and fine mapping of 20 elements showed that inactivation of KP elements is caused by neither mutations nor constitutional suppression by heterochromatinization. We also discuss the possible explanations for the increase of KP elements without the type II repression abilities in the wild populations.
OM isofemale lines are established from wild-caught females in Chichijima, Bonin Islands in 2000 (Itoh et al., 2004). Harwich was used as a standard P strain, which contains about 40 copies of P elements in the genome. Canton-S was used as a standard true M strain. The strains y w snw and C(1)DX, y f / scJ6 B were used for snw hypermutability test. Flies were maintained on standard food medium at 20°C.
With standard P and M strains, two kinds of crosses, cross A (Canton-S females × OM males) and cross A* (OM females × Harwich males) were performed at 29°C (Kidwell et al., 1977; Engels and Preston, 1980). F1 females were individually dissected and the GD score was calculated as the percentage of undeveloped ovaries. At the same time as each test, Harwich males were crossed to Canton-S females as a control and resulted in 100% GD, but the reciprocal cross, with Canton-S males and Harwich females, gave less than 5% dysgenic ovaries.
OM5 males were mated to y w snw females. F1 males were mated to four of C(1)DX, y f / scJ6 B females in vials at 25°C. The male F2 progeny were scored for sn+ and sne mutations (Roiha et al., 1988).
Genomic DNA was extracted from ten adults by standard methods. To investigate the genomic composition of P elements, DdeI digested genomic DNA was probed with the internal probe, pDBs189, a DdeI-BsiWI 189 bp fragment of the P element (Itoh et al., 2004). This hybridization system can detect full-size P elements as a 2.2 kb band and KP elements as a 0.4 kb band. To estimate the number of genomic P elements, BamHI digested genomic DNA was probed with a 1.8 kb BamHI-XhoI fragment, containing the P element sequence (position 1–730) and about 1.0 kb of the flanking 17C genomic region of pπ25.1 (O’Hare and Rubin, 1983). Probes were labeled with digoxigenin using a DIG labeling kit (Roche).
The copy number of P elements was evaluated by scanning densitometry with radioautographs of Southern hybridization using image analyzing software (NIH Image v. 1.62).
Genome DNA of OM5 flies was used for making the genomic library. Genomic libraries were constructed by partial digestion of genomic DNA by BamHI and ligation into lambda phage vectors (lambda EMBL2 or lambda FIX2). The phage plaques were probed with the BamHI-XhoI 1.8 kb fragment as for Southern hybridization. The nucleotide sequence of P element insertion was determined using internal P element primers; KPC1, KPC351, KPC701, KPNC389, KPNC739, KPNC1089, FPC901, FPC1215, FPC1501, FPC1803, FPC2118, FPC2401, FPC2801, FPNC104, FPNC804, FPNC1104, FPNC1404, FPNC1704, FPNC2003, and FPNC2303, while for the flanking sequences: KPC701, KPC1051, KPNC39, KPNC389, FPC2801, and FPNC104 (Table 1). DNA sequencing was performed using an automated sequencer (ABI PRISM 310NT) and a DNA sequencing kit (Applied Bio System, BigDye Terminator v3.1 Cycle Sequencing Kit).
 View Details | Table 1 The PCR primers used |
BamHI digested OM5 genomic DNA was self-ligated into circular DNA by T4 ligase. Both ends of P elements and their flanking sequences were amplified by PCR using a pair of primers IVP5 and IVP3 (Table 1) not amplifying the KP element. From the information on both flanking sequences, the whole P element sequence in the inverse PCR products was amplified by another PCR with the OM5 genomic DNA as a template and was subcloned with the flanking sequences into the PCR2.1 plasmid vector with a TA Cloning kit (Invitrogen). The nucleotide sequence of the amplicon was determined using internal P element primers as described above. All sequences obtained in this study were deposited in the DDBJ with accession nos. AB331374–AB331393.
Salivary glands of OM5 third-instar larvae were dissected and fixed in 45% acetic acid. Polytene chromosomes were spread and probed with the BamHI-XhoI 1.8 kb probe as for Southern hybridization (Fig. 1C). Slides were soaked in hybridization buffer (5x SSC, 1% blocking reagent, 0.1% N-lauryl sarcosine, 0.02% SDS, 20ng/ml probe DNA) at 60°C for 16 hours. After washing by 2x SSC, 0.1% SDS and blocking with 1% blocking reagent, 100mM Tris-HCl, 150mM NaCl, hybridization signals were detected using anti-digoxigenin-AP antibodies and stained with NBT/BCIP. Chromosome preparations were observed under phase contrast microscopy (OLYMPUS BX50) and images were taken with a CCD camera.
 View Details | Fig. 1 Southern hybridization analyses of genomic P elements in the M’ strain OM5. (A) DdeI digested genome DNA was probed with an internal P element probe, pDBs189. The signals expected from FP (2.2 kb) and KP (0.4 kb) elements are marked. (B) BamHI digested genome was hybridized with the BamHI-XhoI 1.8 kb fragment as a probe. The 1.8 kb signals expected from the 17C region are indicated. (C) Structures of full-size P and KP elements. Restriction fragments due to DdeI digestion are denoted by white bars. Probes for Southern hybridizations are denoted by solid bars. ORFs are indicated by gray boxes. B: BamHI, Bs: BsiWI, D: DdeI, X: XhoI. |
P-M phenotypes of each OM isofemale line were examined by GD test. None of the lines had any capacity for induction, but a variety was shown in repression (Table 2). OM5 had neither transposition-induction nor –repression ability; 0% of cross A (0/81) and 100% of cross A* (83/83) sterility, thus being phenotypically an M strain. Additional GD tests were repeatedly performed and showed that the phenotype of OM5 had not been changed until each construction of a genomic library or inverse PCR experiment (data not shown).
 View Details | Table 2 P-M phenotype of the OM isofemale lines |
Furthermore, the snw hypermutability test was carried out to evaluate the transposition-inducing ability of OM5 in detail. The mutation rate, 0.0038 (three sne in 790 F2 progeny), was markedly lower than that of the standard P strain, Harwich, but not as limited as the true M strain, Canton-S (Table 2).
Genomic P elements of OM5 were analyzed by Southern hybridization. Hybridization with the internal probe to DdeI digested genomic DNA showed 2.2 and 0.4 kb signals, suggesting the existence of FP and KP elements (Fig. 1A). KP elements are likely to constitute the majority of P elements in the OM5 genome. Similar results were obtained for other OM lines (data not shown, but partly in the Fig. 1 in Itoh et al. (2004)). On the other hand, the BamHI-XhoI probe for BamHI digestion allowed estimation of the number of genomic P elements, because the signal for the 17C genomic region could be used as a loading control in the blotting (Fig. 1B). Assuming the copy number in Harwich as 40, that in the OM5 genome was estimated at about 25 using scanning densitometry. These results imply that OM5 is an M’ strain.
To isolate genomic P elements of OM5, genomic libraries were constructed and surveyed. As a result, 62 clones were isolated as candidates. After subtracting overlap, 19 P insertions were identified; 15 KP and four other defective P elements (Table 3). Nucleotide substitution of A32T, which was reported in the first KP elements by Black et al. (1987), was found in all KP clones, except for KP6E and KP71B, whose nucleotide at position 32 was unknown. Among these we determined the sequence of the whole element and both flanking regions for eight KP elements, but only part of and one side of the flanking region for the remaining seven. Each of the four defective P elements was caused by a different internal deletion. The dP20B element (804 bp) lacks position 165–2267, dP50B (632 bp) lacks 229–2503, dP55F (2313 bp) lacks 1100–1693, and dP57B (2009 bp ) lacks 327–1224, all having A at position 32.
 View Details | Table 3 Characteristics of P elements isolated from the M’ strain OM5 |
We determined each insertion point from the flanking sequences (Table 3). Most of the P insertions occurred inside or near genes: five KP elements were located to untranslated exons and four to introns. Transcriptional directions of eleven elements were the same and those of eight were reverse for that of each nearest gene. We could not determine the direction for dP50B, because it unusually had the same nucleotide sequence (at least 59 nt) in reverse direction at both ends, in other words, forming extended inverted repeats.
For isolating FP elements in OM5, we further performed inverse PCR with a set of primers that work for FP, but not for KP element insertion. As a result, six PCR products obtained had the same sequence as CG32369 (NM168232). We carried out PCR with another set of primers, P3Lfor and P3Lrev, (Table 1) to amplify the candidate FP element. The PCR product contained a P element, with a 2907 nucleotide sequence completely the same as the canonical autonomous P element. This was called FP66A (Table 3).
Cytological distribution of P elements was analyzed by in situ hybridization with salivary gland polytene chromosomes of OM5 (Fig. 2). Signals in each preparation were not always identical, because OM5 is an isofemale line. We totally identified 20 signals: four in X, four in 2L, seven in 2R, two in 2R, and three in 3R. No signals were observed in heterochromatic regions (centromere, telomere and the fourth chromosome), except for 60F in the 2R telomeric region. Signals were also detected in the 17C region of the X chromosome because of the probe containing a 17C genomic sequence.
 View Details | Fig. 2 An example of in situ hybridization of salivary gland polytene chromosomes of OM5. Preparations of salivary gland chromosomes of third instar larvae were hybridized with the BamHI-XhoI 1.8 kb fragment as a probe. Hybridization signals are numbered according to their cytological map. The asterisk indicates the chromosomal region 17C detected by the genomic sequence in the probe. |
Since the hypothetical horizontal invasion in the middle of the 20th century (Daniels et al., 1990; Houck et al., 1991; Powell and Gleason, 1996), P elements have spread quickly and become almost ubiquitous in D. melanogaster (Ashburner et al., 2005). However, the relationship of genomic P elements and the P-M phenotype in wild populations is not fully clear, partly because previous profiling of genomic P elements was mainly based on Southern hybridization. We isolated and fine-mapped one FP, 15 KP and four of other internally deleted elements in the genome of our M’ strain. This is the first list of genomic P elements of an M’ strain.
The number of P elements in the OM5 genome was preliminary estimated at about 25 by Southern hybridization and at least 20 by cytological detection. These results are quite consistent, because some P insertions can be heterozygous in an isofemale line. We totally cloned 20 of P sequences from OM5, although hybridization of polytene chromosomes clearly showed some P sequences not yet isolated, for instance at locations 35, 42, and 82 in the chromosomal region (Fig. 2). We also failed to detect cytological signals for the P sequences isolated, for instance, for dP20B, KP22B, and KP61C. Consequently, the number of P elements was assumed to be between 25 and 30 in OM5.
O’Hare et al. (1992) scrutinized P elements in a strong P strain, π2, by cloning and DNA sequencing. Insertion positions were cytologically mapped. They found seven FP and two KP elements among the 26 elements cloned. Of the 17 internally deleted defective elements, they identified two copies of HP element (1112 bp by deletion 891–2685), which was assumed to have similar repressing function to KP elements. In OM5, we could find FP and KP elements, but neither any copy of HP nor other internally deleted elements found in π2. Considering that π2 has 30–50 genomic P elements, our results are comparably informative. Nitasaka and Yamazaki (1994) examined genomic P elements of P and Q strains. They isolated many FP, KP, and other defective elements from genomic libraries; four, two, and 21 in a strong P strain, five, one, and eleven in a weak P strain, and one, zero, and eight in a Q strain, respectively. Consequently, it is likely that the current wild M’ strains are represented by many copies of KP elements accompanied with a few copies of FP and other defective P elements.
Only one copy of FP element isolated from OM5, FP66A (2907 bp), has a perfectly canonical P nucleotide sequence (O’Hare and Rubin, 1983) and was located 627 bp upstream of CG32369, whose biological function is unknown. In addition to the results of GD testing, a very low level of P transposition-inducing ability was shown by the snw hypermutability test. The present results consistently showed little transposase and no repression, suggesting that all genomic P elements are inactive in OM5. However, we cannot completely exclude a possibility that the observed value of mutability was a result of a combination of some transposition-inducing ability of FP66A and a repressing ability of the genomic KP elements.
The KP element was shown to be capable of repressing P transposition by competing with the transposase for DNA-binding sites, and this ability depends on the dose of the protein encoded by KP elements (Lee et al., 1998). Even a single copy of a transgenic KP element can repress P mobility (Rasmusson et al., 1993; Simmons et al., 2002). This finding notwithstanding, no repression was detected in OM5 that has at least 15 copies of KP elements. There are a number of possible explanations for their inactiveness in the M’ strain. First, mutations might impair the effects of KP. However, we found that the KP elements cloned have only a specific nucleotide substitution (T at position 32) and internal deletion (808–2560), but no other nucleotide changes. Therefore, impairment by mutations is not the case in OM5. Second, most elements might exist in centromeric or telomeric heterochromatin regions. TEs comprise 52% of the heterochromatic sequences in the Drosophila genome but only a few of them are active (Kaminker et al., 2002; Hoskins et al., 2002). However, no P element accumulated in such regions of OM5, with the one exception of KP60F (Fig. 2). Third, epigenetic gene silencing could result in suppression concerned with heterochromatin protein 1 (HP1) (Eissenberg et al., 1990). HP1 sites were recently examined in the Drosophila euchromatic region (Fanti et al., 2003), but there was no correspondence with the P insertion sites in OM5 (data not shown). In addition, our results indicated that many KP elements insert inside of probably indispensable genes for flies. For example, KP84E is located 3’ UTR of puckered, which has essential roles related to the JNK cascade (Martin-Blanco et al., 1998) and KP70C inserts in an intron of hsc70Cb, which encodes an important protein in stress responses and protein folding (Deak et al., 1997). It is therefore unlikely that these KP elements are thoroughly suppressed together with such genes. Since Black et al. (1987), there were many reports of characteristics of KP elements having type II repression, but not of inactive KP element. For instance, a naturally occurring KP element, KP(6), showing repression in the GD sterility and the snw mutability was reported by Rasmusson et al. (1993), who identified it by nucleotide sequence spanning the deletion and mapped it to the 13F region of the X chromosome, while no repression of KP(6) was found in a recent snw hypermutability test (Simmons et al., 2002). Although some changes might occur in the KP element between the two experiments, the inactive element was not exactly evaluated. We reported here the first KP elements fully molecularly characterized.
On the other hand, a small P element, named SP (517 bp by deletion of 187–2576), was shown to suppress P transposition by means of its antisense RNA (Rasmusson et al., 1993; Simmons et al., 1996). Such antisense RNAs need to be generated by readthrough from the 3’ outside of the elements. At least five elements, KP22B, dP57B, KP61C, KP70C, and KP84E, may be able to produce such RNAs, because they oppositely insert into the intron or 3’ UTR of any gene (Table 3). It is unlikely that antisense RNA of P elements play important roles for the P-M phenotype in OM5.
Taken together, we found no sufficient illustration for inactivity of FP, KP, and small P elements in OM5. One possibility is that only a few insertion sites may let P elements exhibit their potential. The Δ2–3(99B) element is one well known stable transposase source, in which a single copy of P transgene exists in the 99B chromosomal region (Robertson et al., 1988), but its transposase activity varies depending on its insertion site (Laski et al., 1986). Ronsseray et al. (1996) demonstrated that two tandem copies of complete P elements inserted in TAS (telomeric associated sequence) at the tip of the X chromosome, 1A region, are sufficient for repressing P mobility. In addition, an upstream-half deleted (1–871) defective P element inserted near TAS was found capable of forming a P cytotype (Marin et al., 2000). There was no P insertion near the region 1A in OM5. These results consistently support the assumption that a few insertions at special regions are responsible for the P-M phenotype and that there are no such functionally active elements in OM5 and other M’ strains. From this point of view, comparison of the present result with the genomic P elements content of other OM lines, especially of the Q strains would be interesting.
It has been thought that widely prevailing of KP element is a fruit of eliminating harmful effect of P element transposition (type II repression). However, our present results suggest that KP elements do not always repress P mobility, because most KP elements had the canonical nucleotide sequence, but no repression in OM5. The wild population of OM, to which OM5 belongs, is composed of only Q and M’ strains and is one of the most highly insulated populations in the Pacific Ocean (Itoh et al., 2004). In such a population, P transposition should be totally suppressed because of little transposase activity. Consequently, it is unlikely that KP elements can increase in copy number as a result of their repressing ability in the OM population.
For explanation of the world-wide predominance of KP elements, alternative possibilities need to be addressed (Itoh and Boussy, 2002). If KP elements can transpose higher efficiently than other P elements size classes, the transpositional advantage can make them predominant in the population. The nucleotide substitution at position 32, just adjacent to the 5’ inverted repeat, may influence activity of the transposase. Alternatively, KP elements may have a tendency to avoid incomplete homology-dependent DNA gap repair as Tam3 element in Antirrhinum majum (Yamashita et al., 1999). If this is the case, they can be highly stable among the P sequences and thus can increase the number gradually. Actually, present study found no small defective elements with T at position 32 suggesting derived from KP elements. In both cases, evolutionary success of KP elements is closely associated with their selfishness as TE, because accompanying by FP elements is necessary for KP elements to survive in a competition among the P elements in the host gene pool.
We thank N. Nanba for her help in stock keeping.
|
Index