Genes Genet. Syst. (2000) 75, p. 327–333 Tnr8, a foldback from rice

Chaoyang Cheng, Suguru Tsuchimoto, Hisako Ohtsubo, and Eiichi Ohtsubo* Institute of Molecular and Cellular Biosciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan

(Received 19 December 2000, accepted 4 Junuary 2001)

An insertion sequence 418 bp in length was found in one member of rice retroposon p-SINE1 in Oryza glaberrima. This sequence had long terminal inverted repeats (TIRs) and is flanked by direct repeats of a 9-bp sequence at the target site, indicative that the insertion sequence is a rice transposable element, which we named Tnr8. Interestingly, each TIR sequence consisted of a unique 9-bp terminal sequence and six tandem repeats of a sequence about 30 bp in length, like the foldback transposable element first identified in Drosophila. A homology search of databases and analysis by PCR revealed that a large number of Tnr8 members with sequence variations were present in the rice genome. Some of these members were not present at given loci in several rice species with the AA ge- nome. These findings suggest that the Tnr8 family members transposed long ago, but some appear to have mobilized after rice strains with the AA genome diverged. The Tnr8 members are thought to be involved in rearrangements of the rice genome.

in sea urchin (Hoffman-Liebermann et al., 1985), TFB1 in INTRODUCTION Chiromomus thummi (Hankeln and Schmidt, 1990), and Numerous kinds of transposable elements have been SoFT in Solanaceae (Rebatchouk and Narita, described in plants, and other organisms (for 1997). review, see Berg and Howe, 1989). Based on their trans- In this report, we show that a member of the rice position mechanisms, they are divided into two classes. retroposon p-SINE1 family (Umeda et al., 1991) in O. One is the DNA-type transposable element which moves glaberrima contains an FB-like element, named Tnr8, via a DNA form using a cut-and-paste mechanism and is which has TIRs that contain tandem repeats of a sequence characterized by the presence of terminal inverted repeats about 30 bp in length. This element is distinct from other (TIRs). The other is the retroelement, which moves via rice transposable elements. No sequence similarity was an RNA intermediate using a mechanism involving re- found between Tnr8 and the foldback transposable ele- verse transcription of the intermediate and integration of ments identified in other organisms. Tnr8 appears to cDNA into another location in the genome. All of these constitute a large family in the rice genome, whose mem- elements, however, generate a duplication of the target bers have a variation(s) in their sequences. We discuss site sequence of several bp long upon their transposition. the possibility that the Tnr8 members are involved in re- The Drosophila FB (foldback) element is a transposable arrangements of the rice genome. element with long TIRs, which contain tandem repeats of a conserved sequence (Potter et al., 1980). These TIRs MATERIALS AND METHODS vary in the length from several hundred bp to several kb due to a variation in the number of the tandem repeats materials. Rice strains with the AA genome, (Levis et al., 1982; Potter, 1982a; 1982b; Trutt et al., 1981). Oryza sativa (Nipponbare and IR36), O. glaberrima About 10% of the FB element members carry a well-con- (GMS1), O. longistaminata (W1451), O. meridionalis served 4-kb internal sequence between TIRs, whereas the (W1625), and O. glumaepatula (W1192) were obtained other members have no or a very short internal sequence from the Genetic Strains Research Center of National (Brierley and Potter, 1985). The FB element is supposed Institute of , Japan. Total genomic were to be a DNA-type transposable element (Potter, 1982a). isolated from these strains as described previously FB-like transposable elements with TIRs consisting of tan- (Ohtsubo et al., 1991). dem repeats have been identified in other organisms: TU Computer analysis. Primary nucleotide sequences Edited by Takashi Endo were analyzed with the program HarrPlot 1.2.2 and the * Corresponding author. E-mail: [email protected] GENETYX-Mac 10.1 system. The computer-assisted 328 C. CHENG et al.

Table 1. Primers used for PCR

Primera Sequence (5' to 3')

dTnr8-P ＧＡＧＴＡＡＡＴＴＴＣＡＣＡＡＡＣＴＡＣＡ G23-46-3F ＡＴＴＧＧＣＣＴＧＡＡＴＧＡＴＴＧＧＧＴ Try1 ＧＡＧＡＡＣＣＡＡＴＧＡＡＡＡＧＡＧＴＧ AQ365837-2F ＧＡＣＧＡＡＣＧＧＧＡＴＣＡＧＣＧＴＧＴ AQ365837-2R ＡＴＣＴＴＴＴＴＡＣＣＴＴＧＧＡＣＣＡＡＴＴＡＡＣ AP000391-F ＡＴＧＴＧＴＡＧＧＧＴＧＡＣＴＧＧＡＴＧ AP000391-R ＣＡＡＣＴＣＧＧＴＧＧＧＣＣＡＣＧＴＴＧ AC051633-F ＣＴＧＣＡＧＡＴＣＧＴＣＡＡＴＡＡＣＧＧ AC051633-R ＡＴＧＴＧＣＴＡＡＧＣＣＡＡＴＴＴＣＧＧ AC027133-F ＣＣＡＣＡＴＧＣＡＴＧＧＣＧＣＡＧＴＡＧ AC027133-R ＴＡＴＧＴＧＣＡＣＡＣＧＣＴＴＡＧＣＡＴ

a Primer dTnr8-P was used to isolate the fragment with Tnr8-21, and others were used to amplify the fragments with each of Tnr8-1 - Tnr8-5.

nucleotide sequence searches for Tnr8 homologs in the DDBJ/Genbank/EMBL databases were done using the programs BLAST (Altschul et al., 1990) and Smith- Waterman search (Smith and Waterman, 1981). Mul- tiple sequences were aligned using the program CLUSTAL W (version 1.7) with some manual modifications.

Polymerase chain reaction (PCR), cloning, and DNA sequencing. The primers used are listed in Table 1. The PCR was carried out in a reaction mixture (20 µl) con- taining 100 to 200 ng total rice DNA and Ex Taq DNA polymerase (Takara). Thirty repeats of thermal cycling, consisting of denaturation at 94°C for 30 sec, annealing at 60–65°C for 30 sec and extension at 72°C for 1 min, were Fig. 1. (A) The structure of Tnr8. Tnr8 has TIRs, each consist- done. The PCR products were analyzed by electrophore- ing of a terminal 9-bp sequence (TSL or TSR; solid short arrows) sis in a 1.8 % agarose gel, and directly sequenced by using and tandem repeats (1L -6L or 1R - 6R; open arrows), in which units in the left and right TIR are numbered with a BigDye Terminator kit (PE Biosystem) and an ABI377 L and R in the order from each end of Tnr8, respectively. The sequencing apparatus. Some PCR products were cloned two TIRs flank a 40-bp internal sequence (a thin line). The un- into pGEM-T Easy Vector (Promega) and sequenced as derlined 9-bp sequences in the regions flanking Tnr8 represent described above. the target site sequence duplicated upon insertion. (B) The nucleotide sequence of Tnr8. The terminal 9-bp sequence and the tandem repeats are indicated. The 40-bp internal region is Accessions. Nucleotide sequence data appear in the shown in boldface. DDBJ/EMBL/GenBank International Nucleotide Se- quence Databases under the accession numbers AB052632 - AB052634. indicates that the insertion sequence in p-SINE1-r32 is a transposable element, and we therefore named it Tnr8 (Transposable element of rice No. 8). Tnr8 had long ter- RESULTS ―― ― ― minal inverted repeats (TIRs), 189 bp in length, which An insertion sequence (named Tnr8) in p-SINE1- flank an internal 40-bp sequence. Each TIR consisted of r32. A PCR-amplified fragment containing a p-SINE1 a 9-bp unique terminal sequence and six tandem repeats member (p-SINE1-r32) at a locus in the O. glaberrima ge- of a sequence about 30 bp in length (Fig. 1). These struc- nome was found to be larger than that from O. sativa. tural features in Tnr8 are distinct from those in other rice Nucleotide sequencing revealed that the fragment con- transposable elements, but are reminiscent of foldback tained an insertion of a 418-bp sequence which was not transposable elements such as FB of Drosophila, TU of sea present in p-SINE1-r32 at the same locus in the O. sativa urchin, and SoFT of Solanaceae plants, all of which have genome (Fig. 1). This insertion sequence was flanked by TIRs with tandem repeats. Tnr8, however, did not show direct repeats of a 9-bp sequence in p-SINE1-r32, which any sequence homology with the foldback transposable has been duplicated by the insertion event (Fig. 1). This elements. A foldback transposable element from rice 329

Tnr8-homologous members in the rice genome. than rice. Fig. 2 shows schematic structures of 35 mem- A homology search in databases using Tnr8 as the query bers that we characterized, in which the Tnr8-1 is the first sequence revealed that many Tnr8 homologous members one found in p-SINE1-r32 of O. glaberrima. Although the were present in the O. sativa genome. No Tnr8 homolo- Tnr8 members showed high variations both in size and in gous members were, however, found in plant species other sequence, these members were essentially similar in struc-

Fig. 2. Structures of Tnr8-homolog members. Tnr8-1 is the first element found in O. glaberrima. Tnr8-21 is a member isolated as a PCR-amplified fragment from O. sativa using a primer hybridizing to two terminal regions of the TIRs in Tnr8. The other Tnr8 mem- bers were identified by a homology search of databases. The terminal 9-bp sequences (solid arrowheads), tandem repeats (open ar- rows), and conserved internal region (short thick lines) are shown. Broken lines indicate the absence of Tnr8 sequences. Deleted regions of length more than 4 bp are shown by blank spaces. Non-Tnr8 chromosomal sequences are shown by thin solid lines. Sizes of Tnr8 members are shown in bp. Note that structures of 14 members (Tnr8-22 - Tnr8-35) are not fully drawn, because sequence data are not available. Sizes of the duplicated target site sequences are shown in bp. Some Tnr8 members appear to be truncated in either of the two end regions, and thus the duplicated target site sequences in these members could not be assigned. Homology with the consen- sus sequence of Tnr8 is shown in percentage. The accession numbers to the sequences containing Tnr8 members are given. Note that Tnr8-27 and Tnr8-21 have an insertion of Tnr1 within the left tandem repeats or of RIRE10 in the internal region, respectively. Other structural features are explained in the legend of Fig. 1. 330 C. CHENG et al.

ture to one another. Some members, however, did not primer that hybridizes to both end regions of Tnr8, con- carry the internal sequence, and others were truncated in tained an insertion of a 564-bp sequence that begins with an end region (Fig. 2). All the members with TIRs, ex- TG and ends with CA (Fig. 2), indicative of a solo-LTR of a cept one (Tnr8-3), were flanked by direct repeats of a 9-bp gypsy-type (here named RIRE10) in the sequence (Fig. 2). The 9-bp sequences were rich in the A- internal region. T content (90% on average; see Fig. 4B). Six members (Tnr8-1 ~ Tnr8-6) had a typical structure, Tnr8 members at given loci among rice species with with TIRs consisting of a 9-bp unique terminal sequence the AA genome. Five Tnr8 members (Tnr8-1 ~ Tnr8-5) and six tandem repeats of a sequence about 30 bp in length were examined by PCR for the presence or absence at (Fig. 2). Although the sequences of the tandem repeats given loci among rice species with the AA genome, using a differed from one another slightly, the repeats at the cor- pair of primers hybridizing to the regions flanking each responding positions in the two TIRs showed little differ- member. The generated PCR fragments showed length ence in either length or homology (Fig. 3A). These repeat polymorphism (Fig. 4A). A PCR fragment from O. sativa sequences are rich in the A-T content (about 80%), and was shorter than that with Tnr8-1 from O. glaberrima due each repeat sequence appears to be divided into the con- to the absence of Tnr8-1 in the p-SINE1-r32 sequence, as served and variable regions (Fig. 3A). Short internal se- described in the first section of Results. The PCR frag- quences in these members are significantly homologous to ment from O. meridionalis was, however, shorter than one another (Fig. 3B). The consensus internal sequence that from O. sativa (Fig. 4A). DNA sequencing revealed appear to have no significant protein coding frames (data that both Tnr8-1 and p-SINE1-r32 sequences were absent not shown). in O. meridionalis (Fig. 4A). Similarly, Tnr8-2 was ab- Two members were found to carry an insertion of a sent in O. meridionalis, but was present in the other spe- transposable element. Tnr8-27 contained an insertion of cies; Tnr8-4 was absent in O. longistaminata, but was a member of rice transposable element Tnr1 (about 200 present in the other species (Fig. 4A, B). bp in length; Umeda et al., 1991) in the left TIR (Fig. The PCR fragment with Tnr8-2 from O. glaberrima was 2). Tnr8-21, a Tnr8 member obtained by PCR using a a little longer than those from the other rice species (Fig.

Fig. 3. (A) Comparison of tandem repeat sequences with a consensus sequence of Tnr8. The consensus sequence of Tnr8 was com- pared with another consensus sequence (shown at the top), which was derived from the alignment of all the tandem repeat sequences. Open and solid bars represent the conserved and variable regions, respectively. The sequences underlined are the left and right terminal sequences of Tnr8 (TSL and TSR). Tandem repeat sequences at the corresponding positions in the left and right TIRs are aligned. (B) Nucleotide sequences of the internal regions of Tnr8 members. The consensus sequence of Tnr8 is shown at the top. Only nucleotides different from the consensus sequence are shown. Dashes represent nucleotides identical to those in the consen- sus sequence, and slashes represent deleted nucleotides. A foldback transposable element from rice 331

Fig. 4. (A) Ethidium bromide-stained agarose gels showing the PCR-amplified fragments with or without a Tnr8 member from various rice species with the AA genome. PCR was performed using a pair of primers which hybridize to the flanking regions of each of the Tnr8 members located at five loci. The DNA templates used were total DNA from: O. sativa cv. IR36 (lane 1), O. sativa cv. Nipponbare (lane 2), O. glaberrima (lane 3), O. longistaminata (lane 4), O. glumaepatula (lane 5). and O. meridionalis (lane 6). + and – indicate the presence and absence of an element, respectively. Tnr8 with a duplication (dup) or deletion (del) is indicated. (B) Structures of the PCR fragments with or without Tnr8. Nucleotide sequences of the flanking regions include the target site sequence (underlined). Broken lines represent the absence of the Tnr8 sequence and a target site sequence. Other structural features are explained in the legend of Fig. 2. (C) An alignment of seven tandem repeat sequences in the right TIR of Tnr8-2 in O. glaberrima. Note that the sequences of the second and third repeats (enclosed in a box) are almost the same.

4A). DNA sequencing revealed that Tnr8-2 in O. including the terminal region of the left TIR and the abut- glaberrima had seven repeats in the left TIR, unlike the ting flanking sequence (Fig. 4B). Tnr8-2 with six repeats in O. sativa (Fig. 4B). The 2R We failed to amplify fragments with Tnr8-1, Tnr8-2 or and 3R repeats of Tnr8-2 in O. glaberrima had almost the Tnr8-5 from O. longistaminata, and a fragment with Tnr8- same sequence (Fig. 4C), indicating that the 2R repeat has 4 from O. meridionalis, probably because the binding sites been duplicated in Tnr8-2. The PCR fragments with of primers used for PCR have been lost due to mutations Tnr8-3 from O. longistaminata and O. meridionalis were in these rice species. about 30 bp longer than the fragment from O. sativa (Fig. 4A). DNA sequencing revealed that a deletion which was DISCUSSION found in the 6R repeat of Tnr8-3 in O. sativa was absent in O. meridionalis (Fig. 4B). The PCR fragments with We have identified Tnr8 as a novel rice transposable Tnr8-5 from O. sativa cv. IR36, O. glaberrima and O. element with long TIRs consisting of a unique terminal meridionalis were smaller than those from O. sativa cv. sequence and tandem repeats. There are transposable el- Nipponbare and O. glumaepatula (Fig. 4A). Tnr8-5 in O. ements carrying TIRs with tandem repeats in Drosophila, sativa IR36 was found to have deletion of a 40-bp sequence sea urchin, Chiromomus thummi, and Solanaceae plants. 332 C. CHENG et al.

Tnr8, however, shares no sequence homology with these which, however, are not present in the TIRs, but in their foldback transposable elements, and also differs in other internal regions. It should be noted here that tandem structural aspects. Tnr8 (360–471 bp) is smaller than repeat sequences in these elements are different in size Solanaceae plant SoFT members, SoFT1 and SoFT2 (697 and sequence from those in the TIRs of Tnr8. and 1043 bp, respectively; Rebatchouk and Narita, 1997), We found 25 Tnr8 members by a homology search in the and much smaller than Drosophila FB and sea urchin TU, 10.4-Mb sequence of the rice genome released by the Rice which are as large as several kb (Potter et al., 1980; Genome Research Program of Japan (on September 7, Hoffman-Liebermann, 1985). Tnr8 has a unique 9-bp 2000). In the 23 PAC clones identified to have a Tnr8 sequence at the terminal region in the TIRs, whereas TU member(s), only 2 clones contained two Tnr8 members, has a unique 8-bp sequence at the terminal region, in suggesting that members of the Tnr8 family are randomly which, however, the left 8-bp sequence is not homologous distributed in the rice genome. Given that the rice ge- to the right one. The sizes of tandem repeats in Tnr8 (29– nome size is 430 Mb (Arumuganathan and Earle, 1991), 32 bp) are, however, strikingly similar to those in FB (31 the copy number of Tnr8 can be estimated to be 1030 in bp) and SoFT (29–31 bp). We have also shown in this the rice genome. This number is higher than that of FB report that the tandem repeat sequence of Tnr8 is divided (about 30 copies ; Truett, 1981) and TU (about 400 copies; into conserved and variable regions, like that of FB and Hoffman-Liebermann, 1985). SoFT1 (Potter, 1982a; Rebatchouk and Narita, 1997). A Tnr8 member, Tnr8-2, in O. glaberrima had a dupli- The structure of Tnr8 suggests that its transposition is cation of the 2R repeat (see Fig. 4B, C), and another mem- DNA-mediated, as suggested for the FB elements (Potter, ber, Tnr8-3, in O. sativa had a deletion of the 6R repeat 1982a). The tandem repeat sequences significantly dif- (see Fig. 2 and Fig. 4B). There were variations in the fered in size and homology from one another, but those at number of tandem repeats in TIRs in other members (see corresponding positions in two TIRs showed little Fig. 2), as also observed in the FB element members (Le- difference. This indicates that the TIRs are recognized vis et al., 1982; Potter, 1982a). It has been argued that by transposase to form a transpososome with a tight ter- an unequal sister-chromatid exchange (Szostak and Wu, tiary structure (Fig. 1) that may play an important role in 1980; Hawley and Marcus, 1989) is responsible for the its transposition. Potter (1982a) has proposed a two- expansion and contraction of tandem repeat sequences. stage binding process of transposase, in which it first binds However, has recently been argued to be loosely but specifically to the tandem repeats, and then the predominant mechanism of regulating the expansion/ moves until it encounters the termini. Alternatively, two contraction of an rDNA array in Saccharomyces cerevisiae different kinds of transposases, which bind to the tandem (Gangloff et al., 1996; Kobayashi et al., 1998). A mecha- repeats and the terminal sequence, respectively, might be nism by which expansion/contraction of the sequence produced, as revealed in the maize En/Spm family trans- occurs during replication has been proposed, posable elements with short TIR sequences and subtermi- in which DNA replication fork blockage stimulates double nal sequences consisting of short repetitive sequences in strand breakage and gap repair between non-aligned sis- one or the other orientation (Gierl, 1996). The tandem ter chromatids (Kobayashi et al., 1998). The variation in repeats in Tnr8 and other foldback elements may serve to the number of tandem repeats in TIRs of Tnr8 may also increase the size of the target recognized by a transposase be due to gene conversion events during chromosome rep- like the subterminal repetitive sequences in En/ lication. Spm. Note here, however, that the Tnr8 members iden- The heterogeneity of the Tnr8 family members suggest tified are short and appear to have no significant protein- that they transposed long ago. Our observation that coding frames, suggesting that they are non-autonomous Tnr8-21 and Tnr8-27 contained insertions of other elements, in which a putative transposase-coding region(s) transposable elements, such as a retrotransposon and c has been deleted. One FB element, FBw , has been shown Tnr1, respectively, may support the above suggestion. to have a protein-coding region, but the actual function of However, our finding that several Tnr8 members were the encoded protein has not been established (Templeton, absent at given loci in some rice species with the AA 1989). genome indicates that some members appear to have mo- The length of the target-site sequence of Tnr8 is 9 bilized after rice strains with the AA genome diverged. bp, whereas that of TU and SoFT are 8 and 10 bp, The FB elements have been reported to be involved in respectively. It is known that a maize transposable ele- chromosomal rearrangements (for a review, see Bingham ment, MuDR (Hershberger et al., 1995), Arabidopsis and Zachar, 1989). Rearrangements involving individual transposable elements, Tnat1 and Tnat2 (Noma and FBs occurred at a high frequency of about 1 per 1000 Ohtsubo, 2000), and rice transposable elements, Tnr2 and . The region flanked by two FB elements Tnr4 (Han et al. 2000), also generate the duplication of a can be deleted by homologous recombination, and genomic 9-bp sequence at the target site. These elements also regions (sometimes as large as several hundred kb) have long TIRs (70–220 bp), as well as tandem repeats, flanked by two FB elements can transpose to a new A foldback transposable element from rice 333 locus. Tnr8, like FB, is assumed to have also contributed trols of rDNA redundancy in Drosophila. Annu. Rev. Genet. to the genetic diversity of the Oryza genus by accelerating 23, 87–120 Hershberger, R. J., Benito, M.-I., Hardeman, K. J., Warren, C., the recombination process at various positions along the Chandler, V. L., and Walbot, V. (1995) Characterization of chromosomes. As described in this report, a Tnr8 mem- the major transcripts encoded by the regulatory MuDR trans- ber, Tnr8-3, with two TIRs was not flanked by direct re- posable element of maize. Genetics 140, 1087–1098. peats of the 9-bp target site sequence (see Figs. 2 and 4 B). Hoffman-Liebermann, B., Liebermann, D., Kedes, L. H., and This supports the above assumption, because genomic re- Cohen, S. N. (1985) Tu elements: a heterogeneous family of modularly structured eucaryotic transposons. Mol. Cell. arrangements mediated by a transposable element cause Biol. 5, 991–1001. one of the flanking sequences to change. Kobayashi, T., Heck, D. J., Nomura, M., and Horiuchi, T. (1998) Expansion and contraction of ribosomal DNA repeats in Sac- We thank Dr. N. Kurata for providing us rice strains. This charomyces cerevisiae: requirement of replication fork block- work was supported by a Grant-in-Aid for Scientific Research ing (Fob1) protein and the role of RNA polymerase I. Genes from the Ministry of Education, Science, Sports and Culture of Dev. 12, 3821–3830. Japan, and by a grant from the Ministry of Agriculture, Forestry Levis, R., Collins, M., and Rubin, G. M. (1982) FB elements are and Fisheries of Japan. the common basis for the instability of the WDZL and WC Droso- phila mutations. Cell 30, 551–565. Noma, K., and Ohtsubo, E. (2000) Tnat1 and Tnat2 from REFERENCES Arabidopsis thaliana: novel transposable elements with tan- dem repeat sequence. DNA Res. 7, 1–7. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, Ohtsubo, H., Umeda, M., and Ohtsubo, E. (1991) Organization of D. J. (1990) Basic local alignment search tool. J. Mol. Biol. DNA sequences highly repeated in tandem in rice 215, 403–410. genomes. Jpn. J. Genet. 66, 241–254. Arumuganathan, K., and Earle, E. D. (1991) Nuclear DNA con- Potter, S. S. (1982a) DNA sequence of a foldback transposable tent of some important plant species. Plant Mol. Biol. element in Drosophila. Nature 297, 201–204. Report. 9, 208–218. Potter, S. S. (1982b) DNA sequence analysis of a Drosophila Berg, D. E., and Howe, M. M. (1989) Mobile DNA, American Soci- foldback transposable element rearrangement. Mol. Gen. ety for Microbiology, Washington, DC. Genet. 188, 107–110. Bingham, P. M., and Zachar, Z. (1989) Retrotransposon and the Potter, S. S., Truett, M. A., Philips, M., and Maher, A. (1980) FB transposon from Drosophila melanogaster. In: Mobile Eucaryotic transposable genetic elements with inverted ter- DNA. (eds: Berg, D. E., and Howe, M. M.), pp. 485-502. Ameri- minal repeats. Cell 20, 639–647. can Society for Microbiology, Washington, DC. Rebatchouk, D., and Narita, J. O. (1997) Foldback transposable Brierley, H. L., and Potter, S. S. (1985) Distinct characteristics of elements in plants. Plant Mol. Biol. 34, 831–835. loop sequence of two Drosophila foldback elements. Nucl. Smith, T. F., and Waterman, M. (1981) Identification of common Acids Res. 13, 485–500. molecular subsequence. J. Mol. Biol. 147, 195–197. Gangloff, S., Zou, H., and Rothstein, R. (1996) Gene conversion Szostak, J. W., and Wu, R. (1980) Unequal crossing over in the plays the major role in controlling the stability of large tan- ribosomal DNA of Saccharomyces cerevisiae. Nature 284, 426– dem repeats in yeast. EMBO J. 15, 1715–1725. 430. Gierl, A. (1996) The En/Spm Transposable Element of Maize. Templeton, N. S., and Potter, S. S. (1989) Complete foldback trans- (1996) In: Transposable elements (eds.: Saedler, H., and Gierl, posable elements encode a novel protein found in Drosophila A.), pp. 145–160, Springer-Verlag Berlin Heidelberg. melanogaster. EMBO J. 8, 1887–1894. Han, C. C., Frank, M. J., Ohtsubo, H., and Ohtsubo, E. (2000) Truett, M. A., Jones, R. S., and Potter, S. S. (1981) Unusual struc- New transposable elements identified as insertions in rice ture of the FB family of transposable elements in Drosophila. transposon Tnr1. Genes Genet. Syst. 75, 69–77. Cell 24, 753–763. Hankeln, T., and Schmidt, E. R. (1990) New foldback transpos- Umeda, M., Ohtsubo, H., and Ohtsubo, E. (1991) Diversification able element TFB1 found in histone genes of the midge of the rice waxy gene by insertion of mobile DNA elements Chironomus thummi. J. Mol. Biol. 215, 477–482. into . Jpn. J. Genet. 66, 569–586. Hawley, R. S., and Marcus, C. H. (1989) Recombinational con-