Probing the secondary structure of salmon SmaI SINE RNA

Gene 365 (2006) 67 – 73 www.elsevier.com/locate/gene

Probing the secondary structure of salmon SmaI SINE RNA Hiroko Kawagoe-Takaki a,1 , Nobukazu Nameki a,1,2 , Masaki Kajikawa a , Norihiro Okada a,b,⁎ a

Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, 4259-B-21 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan b Department of Evolutionary Biology and Biodiversity, National Institute for Basic Biology, 38 Nishigonaka, Myodaiji-cho, Okazaki, Aichi 444-8585, Japan Received 13 June 2005; received in revised form 11 August 2005; accepted 7 September 2005 Available online 13 December 2005

Abstract SmaI is a short interspersed element (SINE) of the salmon genome, and is derived from tRNALys. We probed the secondary structure of SmaI SINE RNA by enzymatic cleavage and found that the RNA structure comprises three separate domains. The 5′-terminal region (the 5′ domain) forms a tRNA-like cloverleaf structure, whereas the 3′-terminal region (the 3′ domain) forms an extended stem-loop. The loop region is thought to be recognized by the reverse transcriptase (RT) encoded by the long interspersed element (LINE). The two structural domains are linked by a single-stranded region (the linker domain). Our melting profile analyses indicated the presence of two structural domains having different thermal stabilities, thus supporting the domain composition described above. Based on these results, we discuss the structural generality and evolutionary advantage of the domain composition of SINE RNA. © 2005 Elsevier B.V. All rights reserved. Keywords: Transposable element; Retroposon; Retrotransposon; Retrotransposition; LINE; Enzymatic probing

1. Introduction Short interspersed elements (SINEs) and long interspersed elements (LINEs; also called non-long terminal repeat (nonLTR) retrotransposon) are transposable elements in eukaryotic genomes that mobilize through an RNA intermediate. These elements are first transcribed into RNA, which is then reversetranscribed into complementary DNA (cDNA) that is subsequently integrated at a new location within the host genome. This “copy and paste” mechanism is called retrotransposition, a process that expands the number of SINEs and LINEs such that they often occupy a considerable portion of a eukaryotic genome. For example, there are ∼ 1,500,000 copies of SINEs and Abbreviations: cDNA, DNA complementary to RNA; kbp, kilobase pair; LINE, long interspersed element; non-LTR, non-long terminal repeat; PAGE, polyacrylamide gel electrophoresis; SINE, short interspersed element; TPRT, target-primed reverse transcription. ⁎ Corresponding author. Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, 4259-B-21 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan. Tel.: +81 45 924 5742; fax: +81 45 924 5835. E-mail address: [email protected] (N. Okada). 1 These authors made equal contributions to this work. 2 Present address: Department of Biological and Chemical Engineering, Faculty of Engineering, Gunma University, Kiryu, Gunma 376-8515, Japan. 0378-1119/$ - see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2005.09.027

∼850,000 copies of LINEs in the human haploid genome, which comprise ∼ 13% and ∼ 21% of the genome, respectively (Lander et al., 2001). Furthermore, these elements also are thought to have a significant impact on the evolution and complexity of eukaryotic genomes (Moran and Gilbert, 2002; Kazazian, 2004). LINEs are approximately 4–7 kilobase pair (kbp) long and encode an endonuclease (EN) and a reverse transcriptase (RT), both of which are required for LINE retrotransposition (Xiong and Eickbush, 1988a; Xiong and Eickbush, 1988b; Martin et al., 1995; Feng et al., 1996; Moran et al., 1996). The LINE protein is thought to bind its own RNA and return to the nucleus. The LINE EN creates a nick in the DNA of the host genome, and the LINE RT initiates reverse transcription of its own RNA from the 3′ hydroxyl group generated by this nick (Luan et al., 1993; Cost et al., 2002). Thus, reverse transcription is coupled with the target site cleavage, a process called targetprimed reverse transcription (TPRT) (Luan et al., 1993). The newly synthesized LINE copy is integrated into the host genome by TPRT. In contrast, SINEs are relatively short (approximately 100–500 bp) and do not encode proteins required for their own retrotransposition. Typical SINEs are composed of two parts, a tRNA-related region and a tRNA-unrelated region, except for human SINE Alu and mouse SINE B1, which have a

68

H. Kawagoe-Takaki et al. / Gene 365 (2006) 67–73

7SL RNA-related region instead of the tRNA-related region (Weiner, 1980; Ullu and Tschudi, 1984; Okada, 1991a,b). Recently, a novel class of SINEs, which has a 5S rRNA-related region instead of the tRNA-related region, was also found in the zebrafish genome (Kapitonov and Jurka, 2003). The tRNArelated region of SINEs derives from a host tRNA, and includes an RNA polymerase III-dependent internal promoter. In the tRNA-unrelated region, the 3′ tails of some SINEs are almost identical to those of certain LINEs (Ohshima et al., 1996; Okada et al., 1997). Recent experiments provide evidence for the SINE amplification mechanism. We previously isolated a SINE and LINE from the eel genome, named UnaSINE1 and UnaL2, respectively, which have a common 3′ tail of ∼ 60 bp (Kajikawa et al., 2005). Using these elements, we showed that UnaL2 can retrotranspose in cultured HeLa cells and that the conserved 3′ tail of UnaL2 is required for its own retrotransposition (Kajikawa and Okada, 2002). We also showed that UnaL2 can recognize the conserved 3′ tail of UnaSINE1 in trans and mobilize a transcript that has the 3′ tail of UnaSINE1 (Kajikawa and Okada, 2002). These results indicate that SINEs are mobilized by LINEs through their conserved 3′ tails. Human LINE L1 can mobilize human SINE Alu via the poly A tail at the 3′ end, although L1 and Alu have no such common 3′ tails (Dewannieux et al., 2003). During amplification, the UnaL2 RT appears to bind to the conserved 3′ tail of UnaSINE1 (and UnaL2) RNA, recognizing a specific structure (Kajikawa and Okada, 2002; Baba et al., 2004). In addition, SINE RNAs bind to proteins other than the LINE protein in cells. For example, SRP9/14 proteins of the signal recognition particle bind to Alu RNA (Sarrowa et al., 1997), and this binding is proposed to increase the retrotransposition efficiency of Alu SINE (Boeke, 1997; Dewannieux et al., 2003). Thus, SINE RNA structure dictates protein binding specificity, thereby influencing SINE retrotransposition efficiency. Little is known, however, about SINE RNA structure. To address this issue, we used RNA of the salmon SINE, SmaI, the 3′ tail of which is quite similar to that of UnaL2 and UnaSINE1 (Kajikawa et al., 2005). We used enzymatic cleavage to probe the RNA secondary structure of SmaI. We also obtained additional structural information by measuring thermal melting profiles of several SINE RNAs. 2. Materials and methods

SmaI in pSmaT7 using QuikChange Site-directed Mutagenesis kit (Stratagene, USA), producing an NdeI site (CATACG to CATATG). The resulting plasmid was designated pSmaH and was used to prepare the 5′ domain RNA of SmaI. The 3′ domain of SmaI was amplified from the pSmaT7 DNA by PCR using primers Sma1Fsl3F (5′-CTAATACGACTCACTATAGTATGCACACATGACTGTAAG-3′) and SmaAseIR. The PCR product was cloned in the pUC18/SmaI vector. The resulting plasmid was designated pSmasl3 and was used to prepare the 3′ domain RNA of SmaI. Using primers M13M4 (5′GTTTTCCCAGTCACGAC-3′) and M13RV (5′-CAGGAAACAGCTATGAC-3′), we amplified the SmaI sequences (and the T7 promoter) by PCR from pSmaT7, pSmaH or pSmasl3 and then digested the PCR products with AseI (pSmaT7 and pSmasl3) or NdeI (pSmaH). SmaI RNAs were transcribed from the resulting DNA fragments using the AmpliScribe T7 Transcription kit (Epicentre Technologies, USA) according to the manufacturer's instructions. The transcribed RNA was purified by 7 M urea/10% polyacrylamide gel electrophoresis (PAGE). The full-length HpaI SINE sequence was amplified from masu salmon (Oncorhynchus masou) genomic DNA by PCR using primers HpaF-T7 (5′-CTAATACGACTCACTATAGGGGGCGGCAGGGTAG-3′) and HpaStuR (5′-TTTAGGCCTTTATTTAACTAGGCAAGTC-3′). The PCR product was cloned into the pUC18/SmaI vector, and the resulting plasmid containing the HpaI general consensus sequence (Kido et al., 1995) was designated as pHpaIT7Gcons. The HpaI sequence (and the T7 promoter) was amplified from pHpaIT7Gcons DNA by PCR using primers M13M4 and M13RV, and the PCR product was digested with StuI. The full-length HpaI RNA was prepared from the resulting fragment by T7 transcription as described above. The full-length UnaSINE2 sequence was amplified from the plasmid UnaSINE2-8 (Accession number AB179634; Kajikawa et al., 2005) DNA by PCR using primers UnaS2FT7 (5′-CTAATACGACTCACTATAGGGGGCGATATAGCTCAG-3′) and UnaS2DraR (5′-ACTTTAAATGGTAAATGGACTGCA-3′). The PCR product was cloned in the pUC18/SmaI vector, and the resulting plasmid was designated as pUnS2T7. The UnaSINE2 sequence (and the T7 promoter) was amplified from pUnS2T7 DNA by PCR using primers M13M4 and M13RV and was digested with DraI. The full-length UnaSINE2 RNA was prepared from the resulting DNA fragment by T7 transcription as described above.

2.1. RNA preparation 2.2. Enzymatic probing The full-length SmaI sequence was amplified from chum salmon (Oncorhynchus keta) genomic DNA by PCR using primers SmaT7F (5′-CTAATACGACTCACTATAGGTCCTTCTGTAGCTCAGT-3′) and SmaAseIR (5′-TTTTGTTATTAATATGCCATTTAGCAGACG-3′). The PCR product (T7 promoter and full-length SmaI) was subcloned into pUC18/ SmaI, and the resulting plasmid containing the SmaI consensus sequence (Kido et al., 1991) was designated pSmaT7. The pSmaT7 plasmid was used to generate full-length SmaI RNA. We altered the C to T at position 80 relative to the 5′ end of

The enzymatic reactions were carried out essentially as described by Felden et al. (1997). 5′ and 3′ end-labeled RNAs were prepared as described Felden et al. (1997) and purified by 7 M urea/10% PAGE. Prior to the enzymatic digestion, the labeled RNA was heated at 85 °C for 3 min and cooled slowly to 20 °C in the reaction buffer as below. Digestion with the various ribonucleases (V1, S1, and T1) was performed on both 5′- and 3′-labeled RNAs (10,000–20,000 cpm per reaction, depending on the experiments) supplemented with 1 μg

H. Kawagoe-Takaki et al. / Gene 365 (2006) 67–73

total cellular tRNA. The following amounts of nuclease were added: 1 × 10- 2–5 × 10− 2 units RNase T1, 1 × 10− 2–5 × 10− 2 units RNase V1, and 0.4–10 units nuclease S1. Enzymatic reactions were incubated for 5 min at 20 °C in 20 μl buffer containing 100 mM sodium cacodylate (pH 7.5) and 20 mM magnesium chloride. Nuclease S1 reactions were incubated for 2 min at 20 °C in 20 μl buffer containing 100 mM sodium acetate (pH 4.5) and 20 mM magnesium acetate. The reactions were terminated by extraction with phenol-chloroform and precipitated with ethanol. Cleavage sites were resolved by 7 M urea/10% PAGE on two gels, run for either 2.5 h or 5 h at 1200 V. All probing gels included (1) an alkaline hydrolysis ladder of the relevant RNA as a size maker, prepared by dissolving the dried pellet from the labeled RNAs in 50 mM sodium hydrogen carbonate (pH 9.0) and incubating it for 5 min at 85 °C, and (2) a G ladder, obtained by partial digestion of RNA with 0.1 units RNase T1 for 10 min at 37 °C in a buffer containing 50 mM sodium cacodylate (pH 7.5) and 1 mM EDTA. 2.3. Melting profiles Absorbance at 260 nm and melting temperatures were monitored in a spectrophotometer (DU-640; Beckman, USA)

69

equipped with a temperature regulator and a six-cell holder. Before the measurements, RNAs were heated to 85 °C for 3 min in the absence of both monovalent and divalent cations and then cooled quickly on ice. RNA absorbance was measured in a buffer containing 10 mM Tris–HCl (pH 7.5) and 5 mM MgCl2. The first derivatives of the melting curves were calculated using the application included with the spectrophotometer. The data smoothing was according to general least-squares smoothing and differentiation using the convolution (Savitzky-Golay) method (Gorry, 1990). 3. Results 3.1. Enzymatic probing of SmaI SINE RNA The 143-nt RNA transcript corresponding to the full-length SmaI SINE was probed with three RNases: nuclease S1, which cleaves single-stranded regions; RNase T1, which cleaves at single-stranded G residues; and RNase V1, which cleaves double-stranded or stacked nucleotides. The RNA samples were 5′ or 3′ end-labeled, and the cleavage fragments were separated by PAGE. The results of the enzymatic probing experiments are shown in Fig. 1. We superimposed the cleavage sites on the

Fig. 1. Enzymatic probing of SmaI RNA. (A) Resolution of 5′-labeled RNA cleavage products by 7 M urea/10% polyacrylamide gel electrophoresis. Experimental conditions for nuclease mapping were as follows: 5 × 10− 2 units RNase V1, pH 7.5 (lane V1), 10 units nuclease S1, at pH 7.5 (lane S1n) or 2 units nuclease S1, at pH 4.5 (lane S1a), and 5 × 10− 2 units RNase T1, pH 7.5 (lane T1); RNase V1, T1 and nuclease S1 (pH 7.5) reactions were incubated at 20 °C for 5 min and nuclease S1 (pH 4.5) reaction was incubated at 20 °C for 2 min. (B) Resolution of 3′-labeled RNA cleavage products by 7 M urea/10% polyacrylamide gel electrophoresis. Experimental conditions for nuclease mapping were as follows: 1 × 10− 2 and 5 × 10− 2 units RNase V1, pH 7.5 (lane V1, 1 and 2), 2 and 10 units nuclease S1 at pH 7.5 (lane S1n, 1 and 2), 0.4 and 2 units nuclease S1 at pH 4.5 (lane S1a, lane 1 and 2), and 1 × 10− 2 and 5 × 10− 2 units RNase T1, pH 7.5 (lane T1, 1 and 2); reactions at pH 7.5 were incubated for 5 min at 20 °C and reactions at pH 4.5 were incubated for 2 min at 20 °C. (A, B) The gels ran for 2.5 h (left) or 5 h (right) at 1200 V. A range of concentrations for each enzyme was used to identify the optimal conditions for probing. Lane C, incubation control, in the absence of enzyme; lane G, G ladder; lane L, alkaline hydrolysis ladder. Positions from the 5′ terminus are numbered at G's based on the G ladder and the alkaline ladder. Regions that are cleaved by RNase V1 (white bars) or nuclease S1 (black bars) are indicated.

70

H. Kawagoe-Takaki et al. / Gene 365 (2006) 67–73

secondary structure models predicted by the mfold program (Mathews et al., 1999; Zuker et al., 1999). Consequently, we propose a secondary structure model of the SmaI SINE RNA in which the tRNA-like cloverleaf is most similar to canonical tRNA (Fig. 2). The structure comprises three separate domains: the cloverleaf structure at the 5′ terminus (nt 1–76), the linking single-stranded region (nt 77–87), and the extended stem-loop at the 3′ terminus (nt 88–143). In this study, we refer to the three domains as the 5′ domain, the linker domain, and the 3′ domain, respectively. Furthermore, we apply tRNA terminology to the corresponding structural segments of the 5′ tRNA-like domain (for example, the acceptor stem). In the 5′ domain, the single strand-specific nucleases cleaved the D-, T-, and anticodon-loops (Fig. 2). The enzymatic probing data are essentially consistent with those using canonical tRNA transcripts (Lockard and Kumar, 1981). The double strand-specific nuclease V1 cleaved the anticodon stem, the first two base-pairs in the acceptor stem, and the last base-pair in the D-stem. In canonical tRNAs, the anticodon stem is generally strongly cleaved by RNase V1, whereas the acceptor stem, T- and D-stems are only weakly cleaved by this nuclease. The most significant differences in the cleaved sites between the 5′ domain and tRNAs occur in the acceptor stem region. The 5′ domain of the SmaI SINE has an inner loop in the acceptor stem region (Fig. 2). Because its first two base-pairs are clearly cleaved, the inner loop appears to form a stable irregular structure, as seen in various RNA structures (Schroeder et al., 2004). The linker domain connecting the 5′ and 3′ domains was strongly cleaved by the single strand-specific nucleases (Fig. 2), indicating that this region is unstructured. The probing analysis showed that the 3′ domain consists of an extended

stem and loop with two internal loops and two bulged loops (Fig. 2). As expected from the sequence analysis (Kajikawa et al., 2005), this domain has a hairpin region with a GGAUA loop that is common to SmaI SINE, UnaSINE1 and UnaL2. This hairpin region is thought to be a recognition site for the LINE RT (Kajikawa et al., 2005). This is the first report to demonstrate the presence of the recognition stem and loop in the full-length SINE RNA. The solution structure of the isolated stem and loop in the 3′ conserved region of eel LINE UnaL2 has been reported (Baba et al., 2004), and the stem and loop sequence is identical to that of SmaI SINE (Kajikawa et al., 2005) (17 nt; Fig. 2, boxed region). In the 1GGAUA5 loop, the U4 base is bulged out, exposed to solvent, while G1 and A5 form a sheared-type G-A pair, stacking with G2 and A3, respectively (Baba et al., 2004). The terminal loop structure is consistent with the probing results in that the loop region was well-cleaved by the single strand-specific nucleases. On the other hand, the internal loop in the middle of the stem was hardly cleaved, implying that this region forms an irregular structure. 3.2. Melting profile analysis of SINE RNAs To obtain additional structural information, we measured the thermal melting profile of the full-length SmaI RNA. Firstderivative plots of the melting data are shown in Fig. 3A. The first-derivative plot of the full-length RNA has two distinct peaks, having melting temperature (Tm) values of 49 °C and 67 °C (Fig. 3A, thick line). This result indicates that the fulllength SINE RNA has two thermal stability transitions, implying the presence of two structural domains that differ in thermal stability. To examine whether these two domains correspond to the 5′ and 3′ domains, we constructed two corresponding RNA

Fig. 2. Secondary structure model for SmaI SINE RNA showing the enzymatic probing results. Cleavages induced by RNase T1, V1, and nuclease S1 are shown by gray, white, and black arrowheads, respectively. Cleavage intensities are proportional to the size of the arrowheads. Cleavages induced under conditions used for the G ladder in Fig. 1 are shown as circled G's. Thin lines show base pairs. The putative recognition region for LINE RT is boxed by a dashed line (see Section 3.1). Typical tRNA terminology is applied to the corresponding structural segments of the 5′ tRNA-like domain.

H. Kawagoe-Takaki et al. / Gene 365 (2006) 67–73

71

RNA, comprise two different structural domains connected by a linker region. 4. Discussion

Fig. 3. First-derivative plots of UV melting curves of SINE RNAs. (A) Firstderivative plots of UV melting curves of the full-length SmaI RNA, the fragments corresponding to the 5′ and 3′ domains, and yeast tRNAPhe. (B) Firstderivative plots of UV melting curves of UnaSINE2 SINE RNA and Hpa I SINE RNA.

fragments and measured their melting profiles. The first-derivative plot of the 5′ domain fragment shows that the Tm value is similar to the second-transition Tm value of the full-length RNA (67 °C; Fig. 3A, dashed thick line). For reference, we also measured the melting profile of yeast tRNAPhe, and the firstderivative plot shows that the Tm value is 68 °C (Fig. 3A, thin line). On the other hand, the Tm value of the 3′ domain fragment (54 °C; Fig. 3A, dashed thin line) is lower than that of the 5′ domain fragment (67 °C), although it is about 5 °C higher than that of the first peak of the full-length RNA. These results indicate that the two putative domains of the full-length RNA essentially correspond to the 3′ and 5′ domains. However, the difference in Tm between the 3′ domain fragment and the first transition of the full-length RNA implies that the 3′ domain of the full-length RNA has less thermal stability than the 3′ domain alone. The 3′ domain may interact with the 5′ domain via the internal loop of the 3′ domain, and the interaction may negatively affect the thermal stability. This hypothesis remains to be tested in future experiments. Furthermore, we measured the thermal melting profiles of two other full-length SINE RNAs, UnaSINE2 SINE from eel and HpaI SINE from salmon. The first-derivative plots show that these RNAs have two distinct Tm peaks (Fig. 3B). The Tm values of UnaSINE2 are 52 °C and 68 °C, and those of HpaI are 56 °C and 63 °C. These results suggest that these SINE RNAs, like the SmaI SINE

The secondary structure of the SmaI SINE RNA has three separate domains: the cloverleaf structure (the 5′ domain), the linker domain, and the extended stem-loop (the 3′ domain). The 5′ and 3′ domains are independent structural domains that differ in thermal stability. Because the LINE RT is believed to bind to the 3′ domain to recognize the corresponding SINE RNA before reverse transcription, this structural separation is likely to be useful for RT recognition. The disruption of the 3′ domain structure is presumably required to initiate reverse transcription. We propose that during the reaction for the 3′ domain region, the 5′ domain maintains its structure owing to this structural separation, which prevents both interference from the remainder of the RNA and degradation. The probing experiments show that there is significant similarity between the secondary structure of the SmaI SINE 5′ domain and tRNAs. The D-stem-loop and the anticodon stemloop of SmaI SINE RNA are identical to those of canonical tRNAs, not only with regard to length but also the invariant nucleotides involved in tertiary interactions. Significant differences exist only in the T-stem-loop and acceptor stem regions. The acceptor stem of SmaI SINE RNA is shorter than that of canonical tRNAs by two base pairs, and the T-stem-loop is shorter by two base pairs in the stem and one nucleotide in the loop, although nucleotides conserved in canonical tRNAs occur in the loop. Such irregularity of the T-stem-loop is often observed in mitochondrial tRNAs (Watanabe and Osawa, 1995). According to the melting profiles, this SINE structure is similar in thermal stability to that of tRNAs. Taken together, it is likely that the SINE 5′ domain forms an L-shaped tertiary structure similar to tRNAs. Importantly, if the 5′ domain of a SINE is too similar to tRNA in sequence and structure, it will be processed by tRNA-processing enzymes, which modify precursor tRNAs during several processing steps to generate functional tRNA molecules (Deutscher, 1995). In particular, if the tRNA 3′endonuclease (RNase Z) were to cleave the 3′ end of the acceptor stem-like region of the SINE RNA, this cleavage would be lethal to SINEs. In SmaI RNA, however, the irregularity of the acceptor stem, which includes an internal loop, may help to prevent the RNA from being cleaved by the 3′-endonuclease, because this enzyme requires that the substrate has normal baseparing in the acceptor stem (Schiffer et al., 2001; de la SierraGallay et al., 2005). Besides, tRNA-modifying enzymes may recognize the SINE tRNA-like structure. tRNAs always have various modified nucleosides that are important for the specificity, fidelity, and/or efficiency of the tRNA functions (Björk, 1995). In fact, it was reported that SmaI SINE RNA can be modified by HeLa pseudouridylate synthetase(s), and the modified positions, which occur in the 5′ domain, are identical to those in human tRNA1Lys (Matsumoto et al., 1986). In particular, if SINE RNAs were to have some modified “bulky” bases in vivo, the modifications would directly disturb the reverse

72

H. Kawagoe-Takaki et al. / Gene 365 (2006) 67–73

transcription process by LINE RT and prevent SINEs from retrotransposing. Thus, while the SINE internal promoter and the structural stability of the tRNA-related region have been evolutionarily retained for successful amplification of SINEs, sequence substitutions or structural alterations of the tRNArelated region of SINE RNAs seem to have occurred during evolution to escape recognition by tRNA-processing or tRNAmodifying enzymes (A. Weiner, personal communication). The need to limit recognition by tRNA-processing and tRNA-modifying enzymes suggests that the 5′ domain may ultimately deviate from the canonical tRNA tertiary structure. This scenario is supported by the secondary structure of BC1 RNA, a small brain-specific non-mRNA, which is transcribed from a locus of rodent SINE called ID elements (Rozhdestvensky et al., 2001). The tRNA-related region of BC1 RNA, derived from tRNAAla (75% sequence similarity), does not fold into the tRNA-like cloverleaf structure, but rather assumes an extended stem-loop structure. During evolution, the tRNA-related region of BC1 RNA has undergone a structural transition from a tRNA-like structure to a rod-like structure while maintaining the internal promoter for transcription and RNA stability. The SmaI SINE may still be in the process of structural transition from a tRNA-like structure to a different structure in order to escape tRNA processing or modification. By determining the structure of the tRNA-related region of SINE RNA, we might be able to estimate the time at which a SINE was newly born. The changes in the sequence and structure of the tRNA-like domain may be important for its function. Recently, it was reported that some SINE RNAs function via specific interactions with cellular proteins. Transcription of mRNA in cells is repressed in response to heat shock. During the heat shock response in mouse cells, the small noncoding RNA transcribed from mouse B2 SINE associates with RNA polymerase II and represses mRNA transcription (Allen et al., 2004; Espinoza et al., 2004). BC1 RNA (one of the ID elements), expressed exclusively in brain, exists as a ribonucleoprotein complex in cells and is specifically transported into dendritic processes of neurons where it may play a role in regulating translation at these sites (Tiedge et al., 1991; Cheng et al., 1996). In these processes, it is believed that the 5′ domains of SINE RNAs play key roles in protein binding, since only the 5′ domains vary in sequence and structure (i.e., such variations may determine the contingent of proteins to which the RNAs bind). The plasticity of the 5′ domain may be one of the important characteristics of SINEs that facilitate the acquisition of new cellular functions. In contrast, the 3′ terminal region plays essential roles in amplifying SINEs, the progeny of which are dependent on LINE replication system. Hence, if a LINE becomes extinct, the corresponding parasitic SINE can no longer amplify its own copy. However, the SINE may survive, provided that it exchanges its 3′ terminal region with that of any of the LINEs that are still active. In fact, SINEs that have the same 5′ terminal region but different 3′ terminal regions have been identified, and the 3′ terminal regions are similar in sequence to those of certain LINEs (Kido et al., 1995; Okada et al., 1997; Gilbert and Labuda, 1999). Recently, ten distinct families of

chimeric retropesdogenes were identified in the human genome (Buzdin et al., 2003). Intriguingly, in some families, the 3′ parts represented a 5′-truncated L1 mRNA, whereas their 5′ parts originated from different RNAs; such as 5S rRNA and U6 RNA. The chimera formation can be explained by the template switching mechanism during L1 reverse transcription (Buzdin et al., 2003). Thus, it is possible that the 3′ terminal region of SINEs is replaced by a new 3′ tail provided by certain LINE through RNA template switching via reverse transcription. If so, the structural separation of SINEs would be convenient for the exchange of the 3′ terminal region. The linker domain may be used as the point of exchange because it is unstructured; thus the LINE RT may be able to get near the linker domain during template switching. Our results suggest that SINE RNAs generally contain three separate domains: two structured domains and the unstructured linking domain. This characteristic domain composition seems to have allowed successful and continuous amplification of SINEs in eukaryotic genomes during evolution. Acknowledgments We thank Dr. Akitsugu Takasu at RIKEN Genomic Sciences Center for helpful suggestions with the melting profile analysis. We also gratefully acknowledge helpful discussions with Professor Gota Kawai at Chiba Institute of Technology. This work was supported by a Grant-in-Aid to N.O. from the Ministry of Education, Culture, Sports, Science and Technology of Japan. References Allen, T.A., Von Kaenel, S., Goodrich, J.A., Kugel, J.F., 2004. The SINEencoded mouse B2 RNA represses mRNA transcription in response to heat shock. Nat. Struct. Mol. Biol. 11, 816–821. Baba, S., Kajikawa, M., Okada, N., Kawai, G., 2004. Solution structure of an RNA stem-loop derived from the 3′ conserved region of eel LINE UnaL2. RNA 10, 1380–1387. Björk, G.R., 1995. Biosynthesis and function of modified nucleosides. In: Söll, D., RajBhandary, U.L. (Eds.), tRNA; Structure, Biosynthesis, and Function. ASM Press, Washinton, D.C., pp. 165–205. Boeke, J.D., 1997. LINEs and Alus—the polyA connection. Nat. Genet. 16, 6–7. Buzdin, A., et al., 2003. The human genome contains many types of chimeric retrogenes generated through in vivo RNA recombination. Nucleic Acids Res. 31, 4385–4390. Cheng, J.G., Tiedge, H., Brosius, J., 1996. Identification and characterization of BC1 RNP particles. DNA Cell Biol. 15, 549–559. Cost, G.J., Feng, Q., Jacquier, A., Boeke, J.D., 2002. Human L1 element targetprimed reverse transcription in vitro. EMBO J. 21, 5899–5910. de la Sierra-Gallay, I.L., Pellegrini, O., Condon, C., 2005. Structural basis for substrate binding, cleavage and allostery in the tRNA maturase RNase Z. Nature 433, 657–661. Deutscher, M.P., 1995. tRNA processing nucleases. In: Söll, D., RajBhandary, U.L. (Eds.), tRNA; Structure, Biosynthesis, and Function. ASM Press, Washinton, D.C., pp. 51–65. Dewannieux, M., Esnault, C., Heidmann, T., 2003. LINE-mediated retrotransposition of marked Alu sequences. Nat. Genet. 35, 41–48. Espinoza, C.A., Allen, T.A., Hieb, A.R., Kugel, J.F., Goodrich, J.A., 2004. B2 RNA binds directly to RNA polymerase II to repress transcript synthesis. Nat. Struct. Mol. Biol. 11, 822–829. Felden, B., Himeno, H., Muto, A., McCutcheon, J.P., Atkins, J.F., Gesteland, R. F., 1997. Probing the structure of the Escherichia coli 10Sa RNA (tmRNA). RNA 3, 89–103.

H. Kawagoe-Takaki et al. / Gene 365 (2006) 67–73 Feng, Q., Moran, J.V., Kazazian Jr., H.H., Boeke, J.D., 1996. Human L1 retrotransposon encodes a conserved endonuclease required for retrotransposition. Cell 87, 905–916. Gilbert, N., Labuda, D., 1999. CORE-SINEs: eukaryotic short interspersed retroposing elements with common sequence motifs. Proc. Natl. Acad. Sci. U. S. A. 96, 2869–2874. Gorry, P.A., 1990. General least-squares smoothing and differentiation by the convolution (Savitzky-Golay) method. Anal. Chem. 62, 570–573. Kajikawa, M., Okada, N., 2002. LINEs mobilize SINEs in the eel through a shared 3′ sequence. Cell 111, 433–444. Kajikawa, M., Ichiyanagi, K., Tanaka, N., Okada, N., 2005. Isolation and characterization of active LINE and SINEs from the eel. Mol. Biol. Evol. 22, 673–682. Kapitonov, V.V., Jurka, J., 2003. A novel class of SINE elements derived from 5S rRNA. Mol. Biol. Evol. 20, 694–702. Kazazian Jr., H.H., 2004. Mobile elements: drivers of genome evolution. Science 303, 1626–1632. Kido, Y., et al., 1991. Shaping and reshaping of salmonid genomes by amplification of tRNA-derived retroposons during evolution. Proc. Natl. Acad. Sci. U. S. A. 88, 2326–2330. Kido, Y., Saitoh, M., Murata, S., Okada, N., 1995. Evolution of the active sequences of the HpaI short interspersed elements. J. Mol. Evol. 41, 986–995. Lander, E.S., et al., 2001. Initial sequencing and analysis of the human genome. Nature 409, 860–921. Lockard, R.E., Kumar, A., 1981. Mapping tRNA structure in solution using double-strand-specific ribonuclease V1 from cobra venom. Nucleic Acids Res. 9, 5125–5140. Luan, D.D., Korman, M.H., Jakubczak, J.L., Eickbush, T.H., 1993. Reverse transcription of R2Bm RNA is primed by a nick at the chromosomal target site: a mechanism for non-LTR retrotransposition. Cell 72, 595–605. Martin, F., Maranon, C., Olivares, M., Alonso, C., Lopez, M.C., 1995. Characterization of a non-long terminal repeat retrotransposon cDNA (L1Tc) from Trypanosoma cruzi: homology of the first ORF with the ape family of DNA repair enzymes. J. Mol. Biol. 247, 49–59. Mathews, D.H., Sabina, J., Zuker, M., Turner, D.H., 1999. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure. J. Mol. Biol. 288, 911–940. Matsumoto, K., Murakami, K., Okada, N., 1986. Gene for lysine tRNA1 may be a progenitor of the highly repetitive and transcribable sequences present in the salmon genome. Proc. Natl. Acad. Sci. U. S. A. 83, 3156–3160. Moran, J.V., Gilbert, N., 2002. Mammalian LINE-1 retrotransposons and related elements. In: Craig, N.L., Craigie, R., Gellert, M., Lambowitz, A.M. (Eds.), Mobile DNA. ASM Press, Washington D.C., pp. 836–887.

73

Moran, J.V., Holmes, S.E., Naas, T.P., DeBerardinis, R.J., Boeke, J.D., Kazazian Jr., H.H., 1996. High frequency retrotransposition in cultured mammalian cells. Cell 87, 917–927. Ohshima, K., Hamada, M., Terai, Y., Okada, N., 1996. The 3′ ends of tRNAderived short interspersed repetitive elements are derived from the 3′ ends of long interspersed repetitive elements. Mol. Cell. Biol. 16, 3756–3764. Okada, N., 1991a. SINEs. Curr. Opin. Genet. Dev. 1, 498–504. Okada, N., 1991b. SINEs: short interspersed repeated elements of the eucaryotic genome. Trends Ecol. Evol. 6, 358–361. Okada, N., Hamada, M., Ogiwara, I., Ohshima, K., 1997. SINEs and LINEs share common 3′ sequences: a review. Gene 205, 229–243. Rozhdestvensky, T.S., Kopylov, A.M., Brosius, J., Huttenhofer, A., 2001. Neuronal BC1 RNA structure: evolutionary conversion of a tRNA(Ala) domain into an extended stem-loop structure. RNA 7, 722–730. Sarrowa, J., Chang, D.Y., Maraia, R.J., 1997. The decline in human Alu retroposition was accompanied by an asymmetric decrease in SRP9/14 binding to dimeric Alu RNA and increased expression of small cytoplasmic Alu RNA. Mol. Cell. Biol. 17, 1144–1151. Schiffer, S., Helm, M., Théobald-Dietrich, A., Giegé, R., Marchfelder, A., 2001. The plant tRNA 3′ processing enzyme has a broad substrate spectrum. Biochemistry 40, 8264–8272. Schroeder, R., Barta, A., Semrad, K., 2004. Strategies for RNA folding and assembly. Nat. Rev. Mol. Cell Biol. 5, 908–919. Tiedge, H., Fremeau Jr., R.T., Weinstock, P.H., Arancio, O., Brosius, J., 1991. Dendritic location of neural BC1 RNA. Proc. Natl. Acad. Sci. U. S. A. 88, 2093–2097. Ullu, E., Tschudi, C., 1984. Alu sequences are processed 7SL RNA genes. Nature 312, 171–172. Watanabe, K., Osawa, S., 1995. tRNA sequences and variations in the genetic code. In: Söll, D., RajBhandary, U.L. (Eds.), tRNA; Structure, Biosynthesis, and Function. ASM Press, Washinton, D.C., pp. 225–250. Weiner, A.M., 1980. An abundant cytoplasmic 7S RNA is complementary to the dominant interspersed middle repetitive DNA sequence family in the human genome. Cell 22, 209–218. Xiong, Y., Eickbush, T.H., 1988a. Functional expression of a sequence-specific endonuclease encoded by the retrotransposon R2Bm. Cell 55, 235–246. Xiong, Y., Eickbush, T.H., 1988b. The site-specific ribosomal DNA insertion element R1Bm belongs to a class of non-long-terminal-repeat retrotransposons. Mol. Cell. Biol. 8, 114–123. Zuker, M., Mathews, D.H., Turner, D.H., 1999. Algorithms and thermodynamics for RNA secondary structure prediction: A practical guide. In: Barciszewski, J., Clark, B.F.C. (Eds.), RNA Biochemistry and Biotechnology. Kluwer Academic Publishers, pp. 11–43.