| Nucleic Acids Research | Pages |
The presence of modified nucleotides is required for cloverleaf folding of a human mitochondrial tRNA
Introduction
Materials and Methods
Materials
Purification and sequencing of human mt-tRNALys
Cloning and in vitro transcription
Structural mapping in solution
Results And Discussion
Sequencing and characterization of modified nucleotides in human mt-tRNALys
Comparative solution mapping of native human mt-tRNALys and of its unmodified in vitro transcript
Solution structure of human mt-tRNALys in vitro transcript
Importance of epigenetic modifications in mt-tRNALys
The potential key role of a single methyl group in the folding process
Acknowledgements
References
The presence of modified nucleotides is required for cloverleaf folding of a human mitochondrial tRNA
ABSTRACT
INTRODUCTION
The development of in vitro transcription systems (1,2) has led to an explosion of biochemical results concerning structure-function relationships of many different RNAs (for example 3,4). The success of this approach is intrinsically related to the expectation that transcripts have correct folding, i.e. that allow expression of functional properties, an assumption widely verified even in the case of RNAs that contain post-transcriptional modifications in their native versions.
For tRNAs where modifications are most abundant, the in vitro approach has been extremely successful and allowed, for example, identification of sets of nucleotides responsible for aminoacylation identities (5-7). To our knowledge all tRNA molecules so far transcribed fold well. Indeed, in all cases where the structure of the transcripts has been investigated, either by thermal melting (for example 2,8,9), chemical and enzymatic probing in solution (8,9), NMR (10-12) or X-ray crystallography (13), it was shown that in the presence of magnesium they are similar to those of their fully modified counterparts. Thus, post-transcriptional modifications in natural tRNAsare believed not to be involved in directing cloverleaf folding of tRNAs, but in stabilizing the structures (reviewed in 14). In some cases, however, the absence of a modified base led to loss of biological activity of tRNA transcripts, but such effects were accounted for by the absence of an important identity signal and not by incorrect folding of the RNA (for example 15,16).
In this paper we present an example of a tRNA for which the in vitro transcript deprived of modified nucleotides does not fold into a cloverleaf, whereas the native tRNA does. Our example concerns human mitochondrial (mt) tRNALys, which deserves interest since point mutations within its gene have been found to be correlated with severe human pathologies (see for example 17,18). Sequencing of the tRNA as well as phylogenetic comparisons, in conjunction with comparative structural analysis of a series of variant in vitro transcripts, point to m1A9 as a potential key element for correct cloverleaf folding. These data are to our knowledge the first experimental arguments in favor of involvement of modified nucleotides in RNA folding. They demonstrate that the contribution of post-transcriptional modifications to RNA structure is not restricted to conformational stabilization but can be enlarged to more dramatic folding processes.
MATERIALS AND METHODS
Materials
Dimethylsulfate (DMS) was from Aldrich Chimie (St Quentin-Fallavier), imidazole (buffer grade), lead(II) acetate and aniline from Merck (Darmstadt), diethylpyrocarbonate (DEPC) and hydrazine from Sigma (St Louis) and kethoxal from US Biochemical (Cleveland). Radioactive [[gamma]-32P]ATP (3000 Ci/mmol), [[alpha]-32P]ATP (3000 Ci/mmol) and T4 polynucleotide kinase were from Amersham (Les Ulis). Nuclease S1 and RNases T1 and V1 were from Pharmacia (Paris), RNase T2 from Sigma andbacterial alkaline phosphatase from Appligène (Strasbourg). Restriction enzymes were from New England Biolabs (Beverly), avian myeloblastosis reverse transcriptase from Life Sciences (St Petersburg) and T4 DNA ligase from Boehringer-Mannheim. Phage T7 RNA polymerase was prepared according to Becker et al. (19). Streptavidin paramagnetic particles were from Promega (Madison) and Sephadex-G25 medium from Pharmacia. Oligonucleotides were synthesized on an Applied Biosystems 381A DNA synthesizer with phosphoramidite chemistry and purified by HPLC on a Nucleosil 120-5-C18 column (Bischoff Chromatography, Zymark-France, Paris). A 5[prime]-biotin-labeled oligonucleotide was from Eurogentec (Searing).
Purification and sequencing of human mt-tRNALys
Mitochondria were isolated from human placenta according to Gasnier et al. (20) and processed into mitoplasts by treatment with digitonin (21).Total mt-tRNAs were obtained by subsequent phenol extraction (22).Human mt-tRNALys was further purified by hybridization to a 5[prime]-biotin-labeled 30mer, complementary to its 3[prime]-end (23,24). Hybridization was at 70°C for 5 min in 6× SSC and cooling to room temperature was over a period of 40 min. Recovery oftRNA from the beads was by heating for 5 min at 80°C in 2 mM EDTA. About 2.5 µg pure tRNA,as judged by a RNase T1 ladder of a 32P-labeled aliquot, were obtained from 500 µg total mt-tRNA. Sequencing of the tRNA was according to Stanley and Vassilenko (25) and modified bases were identified on TLC plates by comparison with standards (26).
Cloning and in vitro transcription
A synthetic gene corresponding to the T7 RNA polymerase promoter directly connected to the downstream human mt-tRNALys sequence (27) and terminating at a BstN1 site was constructed from 10 overlapping and complementary oligonucleotides and cloned into pTFMA (derived from pUC 18; 8). Subclones derived from this gene were constructed using appropriate synthetic oligodeoxynucleotides as PCR primers and cloned into pTFMA. Amplification of the plasmids was with transformed Escherichia coli TG2 cells.
In vitro transcriptions of 0.1 mg/ml BstN1 linearized plasmid were performed for 3 h at 37°C in 40 mM Tris-HCl, pH 8.1 (at 37°C), 22 mM MgCl2, 5 mM dithiothreitol, 0.01% Triton-X100, 1 mM spermidine, 4 mM each nucleoside triphosphate and T7 RNA polymerase. After phenol/chloroform extraction transcripts were purified on denaturing 12% polyacrylamide gels, electroeluted, ethanol precipitated and desalted on Sephadex-G25. Sequencing of the extremities revealed systematic incorporation of a non-encoded A at the 5[prime]-end of the transcripts. Thus tRNAs are longer by 1 nt than expected. At their 3[prime]-ends transcripts terminate as expected with a CCAOH sequence. The 5[prime]-extension, present to the same extent (>95%) in all transcripts studied here, most likely originates from the unfavorable (see for example 28) 5[prime]-CACUG...-3[prime] starting sequence encountered by T7 RNA polymerase.A 1 nt 5[prime]-extension exists naturally in all tRNAHis species (29) and longer extensions have been introduced on purpose in tRNAAsp transcripts (30), which are efficient substrates of synthetases and fold properly into cloverleaves. The extra nucleotide in mt-tRNALys does not contribute to structural events occurring in the central domain of the tRNAs investigated here and has no effect on the conclusions drawn from this work. However, this nucleotide might be a reason for the non-aminoacylation properties of these transcripts by human mt lysyl-tRNA synthetase, as opposed to the active native tRNALys (to be published).
Structural mapping in solution
Experimental prerequisites. To ensure the proper conformation of transcripts during structural analysis, solution probing was performed at 25°C at pH 7.5 in the presence of 10 mM MgCl2 and 300 mM KCl. Optimal denaturation/renaturation conditions have been defined taking into account the following observations. Differential optical melting curves showed that in the presence of 10 mM MgCl2 total denaturation of the transcripts is achieved by heating to 85°C. However, after a denaturation/renaturation procedure performed under such conditions (2 min at 85°C, 10 mM MgCl2, slow cooling) strong degradation of the transcripts occurs, as verified on 5[prime]-32P-labeled molecules electrophoresed on denaturing polyacrylamide gels (not shown). To overcome this problem, alternative denaturation/renaturation procedures have been tested and their quality judged by comparative structure probing of the transcripts with RNase T2. The final protocol leading to the same stable conformation but low degradation level of tRNA consists of denaturation of the transcript at 60°C in the absence of Mg2+, slow cooling and subsequent addition of Mg2+ to a final concentration of 10 mM before addition of the structural probes.
End-labeling and renaturation of RNAs. Labeling of tRNAs at their 5[prime]-end was done with [[gamma]-32P]ATP and T4 polynucleotide kinase (31). Labeling at the 3[prime]-end resulted from [[alpha]-32P]ATP exchange in the presence of (ATP,CTP):tRNA nucleotidyltransferase (G.Keith, personal communication). Labeled tRNAs were purified from excess nucleotides by electrophoresis on 12% polyacrylamide gelsfollowed by passive elution for 2 h in 0.5 M NH4OAc, pH 6.0, 0.1mM EDTA, 0.1% SDS, 10mM Mg(OAc)2. After ethanol precipitation in the presence of 0.1 mg/ml pAp as carrier RNAs were redissolved in H2O, desalted on Sephadex-G25 spin columns and stored at -20°C.Transcripts were submitted either to statistical chemical modifications followed by specific chemical cleavage of the modified positions or to statistical cleavage by chemicals or enzymes. Location of cleavage sites within RNA was determined by electrophoretic separation of fragments on denaturing (7 M urea) polyacrylamide gels (12 or 15%). Alkaline ladders were obtained by incubation in 50mM NaHCO3 at pH 9.0 at 90°C for 10 min and guanine laddersby denaturing RNase T1 digestion (32). Signal analysis and quantification were on a FUJIX Bio-Imaging Analyzer BAS 2000 with Work Station Software (v.1.1) for volume integration of specific cleavage sites.
Nuclease mapping. Digestions with the various nucleases (S1, T1, T2 and V1) were for 10 min at 25°C in 20 µl 40 mM Tris-HCl, pH 7.5, 10 mM MgCl2 and 40 mM NaCl. For digestion with nuclease S1, 1 mM ZnCl2 was added. The reaction mixtures contained 3 µg total tRNA, 3[prime]-or 5[prime]-end-labeled transcripts ([sim]50 000 Cerenkov counts) and 10-3 or 2.5 × 10-3 U RNase T1, 6.6 × 10-2 or 0.2 U RNase T2, 3.5 × 10-2 or 10-1 U RNase V1 and 50 or 150 U nuclease S1. Reactions were stopped by immediate cooling on ice, addition of 20 µl 0.6 M NaOAc, 4 mM EDTA, 0.1 µg/µl total tRNA, phenol extraction and precipitation with 400 µl 2% LiClO4 in acetone (33).Mapping with chemical probes. Mapping with lead and imidazole was essentially as described above. A freshly prepared0.6 mMPb(OAc)2 solution in H2O was used. Incubation was for 10-40 min at a final concentration of 0.15 mM Pb(OAc)2. Imidazole as a probe (33) was added from a 4.4 M solution at pH 7.0 (25°C). After 18 h incubation at final concentrations of 1.5 and 2 M imidazole, reactions were stopped by acetone/LiClO4 precipitation (33).Modification of the N3 atom in C by DMS and the N7 position in A by DEPC were accordingto Peattie and Gilbert (32). Concentrations of chemicals and incubation times were optimized for mt-tRNALys transcripts. Reaction mixtures of 200 µl containing the transcripts ([sim]50 000 Cerenkov counts and 10 µg non-labeled tRNA) were supplemented with 10 µl diluted DMS solution (10-fold dilution in 100% ethanol) or with 10 µl pure DEPC. Three conditions were tested. Under native conditions DMS reactions were for 10 min at 25°C in 50 mM Na cacodylate, pH 7.5, 10 mM MgCl2, 300 mM KCl for N3 modification of C residues. Semi-denaturing conditions were as native conditions but the buffer was 50 mM Na cacodylate, pH 7.5, 1 mM EDTA. Denaturing conditions were as before but incubation was for 3 min at 55°C. DEPC modifications were as with DMS but incubation time was 40 min for the native and semi-denaturing conditions.
Figure 1. Sequence of human mt-tRNALys and comparison with mammalian mt-tRNALys species. (a) Hypothetical cloverleaf fold of the gene sequence of human mt-tRNALys. Note that the CCA 3[prime]-end is not encoded (27). (b) Sequence of human mt-tRNALys extracted from total placental tRNA as determined by the Stanley-Vassilenkov method (25). Six modified nucleotides are revealed. Their nomenclature is according to Sprinzl et al. (39). Numbering of residues is according to that of canonical cytosolic tRNAs (39); note the particular numbering in the D and T loop regions due to their atypical length, as is often the case in mitochondrial tRNAs. (c) Conserved residues in mammalian mt-tRNALys species.The consensus sequence is derived from 15 mitochondrial genome sequences and three mt-tRNALys species (39). Positions A9 (m1A), G10 (m2G), U34 and A37 are modified in all three as yet sequenced mammalian mt-tRNALys and are indicated by thick circles. Occasionally modified bases are indicated by thin circles. Dots correspond to non-conserved nucleotides and Y and R to semi-conserved pyrimidines and purines respectively. Note the variable length of the D and T loops of these tRNAs. The primer extension method (34,35) was used to detect methylation at the N1 position in A and the N3 position in C with DMS and modification at the N1 and N2 sites in G with kethoxal. Modifications were on 0.2 µg transcript in 50 mM Na cacodylate, pH 7.0, 5 mM MgCl2, 300 mM KCl for native conditions and in 50 mM Na cacodylate, pH 7.0, 1 mM EDTA for semi-denaturing conditions, in a final volume of 100 µl. For DMS modification reactions were for 10 min at 25°C in the presence of 0.5-2.5 µg DMS. For kethoxal modification incubations were for 10 min at 25°C in the presence of 1-10 µl freshly prepared solution of 20 mg/ml probe in 20% ethanol.Stabilization of adducts was at 0°C for 10 min through addition of Na borate (10 mM final). The modified transcripts were ethanol precipitated in the presence of carrier tRNA andanalyzed as described (34) with the modifications of Ehresmann et al. (35). A 5[prime]-labeled primer oligonucleotide ([sim]100 000 Cerenkov counts) complementary to residues A76-U62was annealed to modified or unmodified RNA by incubation at 60°C for 5 min and slow cooling to room temperature. Elongation was in 50 mM Tris-HCl, pH 8.0, 6 mM MgCl2, 40 mM KCl in the presence of 250 µM each dATP, dCTP, dGTP and dTTP and 1 U reverse transcriptase for 30 min at 42°C (final volume 20 µl). Sequencing was in parallel by the dideoxy termination method (36).Reactions were stopped by adding an equal volume of 0.5% SDS with 10 mM EDTA. Template RNA was hydrolyzed for 1 h at 60°C in the presence of KOH (0.25 M final). The mixtures were then acidified with acetic acid, ethanol precipitated and analyzed on denaturing 12 and 15% polyacrylamide gels.
RESULTS AND DISCUSSION
Sequencing and characterization of modified nucleotides in human mt-tRNALys
Pure human mt-tRNALys has been extracted from total placental mt-tRNA using a biotin-labeled 30mer DNA probe coupled to streptavidin-coated magnetic beads and complementary to the 5[prime]-half of the tRNA (23). The sequence of this tRNA (Fig. 1) has been established using 5[prime]-post-labeling techniques (25) in combination with thin layer chromatography (TLC). This enlarges the number of sequenced human mt-tRNAs to three, since only tRNASer(AGY) (37) and tRNAPro (38) have been investigated at the RNA level so far. The RNA sequence is identical to the genomic sequence (Fig. 1a), showing that no editing events occur during the maturation steps of tRNALys in human mitochondria. Six modified nucleotides are present, namely m1A9, m2G10, [Psi]27, [Psi]28 and hypermodified nucleotides at positions U34 and A37. The methylation at position A9 is specific to mt tRNAs, whereas the m2G at position 10 is commonly found in many canonical tRNAs (39). Pseudouridines at positions 27 and 28 are equally often found in mitochondrial and canonical tRNAs. Adenine 37 is modified to t6A and partially hypermodifed, probably to ms2t6A. The nature of the modified base at position 34 could not be identified by bi-dimensional TLC(in particular its migration properties deviate from those of cmnm5U, the modification reported for position 34 in rat mt-tRNALys; 40) and remains unknown.
Most interestingly, human mt-tRNALys shares four base modifications with the three as yet sequenced mammalian mt-tRNALys species, namely those of rat (40), hamster (41) and cow (42). These are m1A9, m2G10, the modification of U34 and a hypermodification at position 37 (Fig. 1c). Each of the pseudouridines detected in human tRNA occurs in at least one of the three formerly sequenced tRNALys.
Comparative solution mapping of native human mt-tRNALys and of its unmodified in vitro transcript
The secondary structures of native fully modified human mt-tRNALys and of the corresponding in vitro transcript have been compared using RNase T2 as the strategic structural probe (Fig. 2). Unexpectedly, whereas the resultof probing of native tRNA accounts well for cloverleaf folding, that of the in vitro transcript does not. For instance, RNase T2 cleaves within regions of the T domain which are double-stranded in a cloverleaf. The enlarged cleavage domain as compared with the situation in native tRNA indicates a larger T loop in the transcript. The anticodon region shows similar reactivities in both molecules, albeit slightly enhanced in the transcript. In contrast, the D loop is differently protected in the transcript. Altogether these data indicate a rearranged structure of the in vitro transcript as compared with native tRNA, a conclusion strengthened by additional comparative probing experiments with nuclease S1 and RNases T1 and V1 (not shown).
Figure 2. RNase T2 probing of mt-tRNALys (a) and of its in vitro transcript (b). Typical single-strand-specific cleavage products of 5[prime]-labeled RNAs are shown on autoradiograms of 15% polyacrylamide-8 M urea gels. G stands for the RNase T1 ladder, L for the alkaline ladder, Ct for control incubations without probe and (1) and (2) for 0.066 and 0.2 U RNase T2 respectively. Domains are emphasized by bars at the right-hand side of the gels; they are named according to their occurrence in the native tRNA.
Solution structure of human mt-tRNALys in vitro transcript
Complete structural probing. To further investigate the unusual structure of the transcript a detailed analysis withfour nucleases (S1, T1, V1 and again T2) and five chemical probes (Pb2+, imidazole, DMS, DEPC and kethoxal) was undertaken. Typical autoradiograms are shown in Figure 3 (data for lead and imidazole mapping are not shown).
Figure 3. Enzymatic and chemical mapping of human wild-type mt-tRNALys transcript. (a-c) Cleavage products of 5[prime]-end-labeled molecules after treatment with nucleases (a), DMS (b) and DEPC (c) displayed on autoradiograms of 15% denaturing polyacrylamide gels. G, L and Ct as in Figure 2. T1, T2, S1, V1, statistical hydrolysis under native conditions with the corresponding enzymes; N, native conditions; SD, semi-denaturing conditions; D, denaturing conditions (according to definitions in Materials and Methods). Numbering on the left side of the panels refers to G residues, that on the right side to mapped nucleotides. Note that, depending on the cleavage mechanism, fragments of the same length migrate differently (35). (d and e) Primer extension products obtained by reverse transcription of RNA treated with DMS (d) and kethoxal (e) under native conditions. Locations of modifications are revealed on autoradiograms of 12% denaturing polyacrylamide gels. Lanes A and C are sequencing ladders generated in the presence of ddTTP and ddGTP respectively. Conditions 1-3 refer to different concentrations of probe (0.5, 1.0 and 2.5% v/v for DMS and 0.04, 0.08 and 0.20% w/v for kethoxal). Numbering on the right side refers to mapped nucleotides. Probing with nucleases (Fig. 3a) reveals three types of nucleotides: those exclusively accessible to the single-strand-specific enzymes (nucleases S1, T1 and T2), namely U33-U40 and A52-C60, those exclusively accessible to RNase V1 (A6-A7 and C68-G70) and those accessible to both kinds of structure-specific enzymes, i.e. A13-A23 within the D arm and A26-A31 in the 5[prime]-part of the anticodon stem. Lead acetate (8,42) and imidazole (33) map global conformations of tRNA and are sensitive to single-stranded domains. Pb(OAc)2 is, furthermore, known to induce strong site-specific hydrolysis when bound to specific positions in RNA (43,44). Lead cleavage occurs in three main stretches of nucleotides, i.e. G10-U28, C31-G39 and A50-U62, with the strongest cuts occurring at nt 11, 12, 34, 35, 36, 57 and 60. Imidazole-induced cleavage is found at every nucleotide between C11 and U62, albeit with different intensities. Of note, no cuts are observed in the theoretical acceptor stem at either extremity of the molecule and for the three adjacent nucleotides on both sides (nt 5-9 and 63-71) of the stem. Reactivity of adenines to DEPC under native conditions is strong for residues 14, 37, 38, 54, 58 and 60a and medium for 46, 52, 55 and 59. Adenines 13, 42 and 43 are only weakly reactive, while A7, 8, 9, 26, 29, 30, 44, 49, 50 and 69 are not reactive at all. Significant alkylation by DMS under native conditions, as revealed by direct strand scission, is observed at the N3 position of C11, 25, 32, 56, 57, 60 and 61. Residues C15 and C60b are weakly reactive, whereas C31, 63 and 68 are not reactive. Reactivities of the N3 position of cytosines revealed by reverse transcription generally confirmed the preceding data, with some exceptions. Whereas C55 and C56 were found to be very reactive, C11, 15, 25, 31 and 32 are not reactive. The N1 position of A13, 14, 26, 37, 42, 43, 49, 54, 58 and 59 is accessible to DMS, whereas A2, 7, 8, 9, 23, 29, 39, 44, 46 and 50 are non-reactive. Positions N1 and N2 of G10 and G39 are protected against kethoxal, those of G24 and G51 weakly accessible and those of G45 and G53 very reactive. Some reproducible degradation at CpA and UpA phosphodiester linkages is noted.
Figure 4. Theoretical cloverleaf and experimental secondary structure models of thehuman mt-tRNALys in vitro transcript. (a) Cloverleaf structure with unexplained reactivities. Nucleotides leading to the extended amino acid acceptor stem in (b) are circled. (b) Experimental secondary structures in agreement with the probing data. Two alternative conformers, b1 and b2, in equilibrium account for the obtained mapping data. In (a) and (b) specifications and intensities of cuts or modifications are as indicated in the key. Intensities of cuts or modifications are proportional to the darkness of the symbols: open, stippled and filled for weak, medium and strong cuts respectively. In addition, gray lines highlight regions which are cut by both double-strand- and single-strand-specific nucleases and the black line indicates complete protection of A8 and A9 towards any probe. The anticodon triplet as defined in the cloverleaf structure is shaded; note its shifted location in b2. These observations lead to the proposal of an alternative folding which best fits the data. Figure 4b presents two versions of this new folding in equilibrium, no longer resembling a tRNA but rather an elongated hairpin with several bulges. Both of the alternative structures share the same `acceptor' domain but have two different structures in their `anticodon' parts. These unexpected foldings account in particular for protection of nt 8, 9, 10, 63, 64 and 65 in the acceptor domain (highlighted by circles), including them in an extended stem thanks to their complementary sequences. These non-reactivities equal those in the rest of the acceptor stem (nt 1-7 and 65-72) and are in striking contrast to those of the adjacent single-stranded residues on either side (i.e. nt 11, 12, 61 and 62). Nucleotides at the 5[prime]-end of the molecule are cleaved by double-strand-specific RNase V1, confirming the formation of an acceptor stem. The second major and common feature of the proposed structures is the presence of a large loop formed by residues 52-62 in the central domain of the molecule. This loop is larger than the expected T loop, since most nucleotides from the 5[prime]-part of the T stem are very reactive to the different probes. In particular, N7 A52 is reactive to DEPC, N1 and N2 of G53 to kethoxal and G53 to RNase T1. On the opposite strand of the T stem it is worth noting the reactivity of N3 C61 to DMS and cleavage after residues C61 and U62 by single-strand-specific imidazole and lead acetate. Opposite to this loop is another single-stranded domain formed by the 5[prime]-part of the former D stem (nt 11-13). The 3[prime]-part of this domain (nt 22-24) is base paired with parts of the previous T stem (nt 49-51) and the variable loop (nt 47, 48). As illustrated in Figure 4, all reactivities which showed a high level of discrepancy from the cloverleaf structure fit with the newly proposed structure. The two alternative foldings proposed for the lower part of the hairpin structure account for the conflicting reactivities found in this domain. One structure presents the canonical anticodon domain with the lysine-specific UUU triplet in a central position and is supported by numerous single-strand-specific reactivities in the loop and RNase V1 cuts at nt 26-28 and 40-42 in the stem, almost identical to what has been observed for cytosolic human tRNALys (45). In the second structure, in accounting for the reactivities of nt 27-29 and 39-40 towards single-strand-specific probes, the anticodon triplet is partially involved in base pairing. Native PAGE in the presence of magnesium failed to separate the two conformers (not shown), which likely reflects a very low activation barrier between them.
Figure 5. Design (a) and structural probing (b) of selected variants of the in vitro transcript of human mt-tRNALys. (a) Three mutants (Kiw, Kew and Krw) have been constructed. Mutations introduced (hatched circles) are presented in the elongated hairpin structure found in the wild-type transcript (see Fig. 4b). For simplicity only the central part of the molecules are shown. These mutations lead to disruption of 1 or 2 bp. (b) Comparative enzymatic (upper) and chemical (bottom) structural mapping of the wild-type mt-tRNALys transcript (Kwt) and its three variants (Kiw, Kew and Krw). Cleavage products of 5[prime]-end-labeled molecules obtained after treatment with four structural probes at 25°C as revealed on autoradiagrams of 15% polyacrylamide denaturing gels. (Upper) Lanes 1 and 2 correspond to two concentrations of enzymatic probes (see Materials and Methods) with the lowest in 1 and the highest in 2. Chemical probing was under native (N), semi-denaturing (SD) or denaturing (D) conditions. G, L and Ct as in Figure 2. For additional experimental details see Materials and Methods. Note the large changes in reactivity observed within wild-type and variants, especially for G53 and G45 (upper left ), the T loop nucleotides (upper right ), C11 and C15 (bottom left ) and A8 and A9 (bottom right ). One implication of the structural probing of the transcript of human mt-tRNALys is that its elongated hairpin fold cannot adopt the canonical tRNA L-shaped tertiary structure. In agreement with this view, an unusual structure was suggested by transient electric birefringence measurements on this tRNA transcribed in vitro in a form presenting extended acceptor and anticodon stems. An unusual angle of 140° (90° for canonical tRNA) between the two elongated stems of the engineered molecule was measured (46).
Importance of epigenetic modifications in mt-tRNALys
The unusual structure of the mt-tRNALys transcript as opposed to cloverleaf folding of the native fully modified tRNA demonstrates the mandatory contribution of post-transcriptional modifications in folding of this RNA. This is, to our knowledge, the first experimental evidence for such a role of modified bases. In addition to their contribution to RNA function (47,48), modified nucleotides are already known to influence structural properties of RNA. However, these influences were restricted to stabilization of active conformations, as demonstrated by comparing native tRNAs with their in vitro transcribed counterparts (14), and to thermophilic organisms, where modified bases contribute to the increased thermostability (49). A role of modified nucleotides in RNA folding was proposed by Steinberg and Cedergren (50), who highlighted a correlation between the presence of N2-dimethylguanosine and possible alternative mt-tRNA conformers. This theoretical consideration, focused on the inability of the modified nucleotide to base pair, is conceptually related to our experimental data.
In a more general perspective, post-transcriptional modifications might potentially be able to direct folding of other RNAs and thus enlarge their potential fields of action. We suspect thatthis important new function could be uncovered in the special case of human mt-tRNALys, because of its peculiar sequence, allowing an unusually large base complementarity between the 5[prime]- and 3[prime]-ends (10 instead of 7 bp). Mitochondrial tRNAs are especially rich in A and U residues (60% in human; see the sequence compilation in 39) and thus present a large `folding space' (51). In other words, they have the potential to fold into a larger number of alternative foldings than cytosolic tRNAs. Modified bases in mt-tRNAs thus provide additional information, helping to overcome this problem by stabilizing the desired biologically important secondary structure and restricting the folding space. In human mt-tRNALys it is probably the 5-fold copy of a UUAR sequence which extends the folding space and thus enables alternative foldings to occur.
The potential key role of a single methyl group in the folding process
Sequence comparison of the three formerly sequenced mammalian mt-tRNAs and the herein sequenced human species highlights the presence of four conserved modified residues, located at the same positions, namely at nt 9, 10, 34 and 37 (see Fig. 1c). The two modifications in the anticodon loop will likely govern correct folding of the anticodon arm (52) and favor the structure with a 7 nt anticodon loop instead of that with a 6 nt loop (see Fig. 4b). Similarly, m1A9 and m2G10 are good candidates to interfere with the architecture of the tRNA core. The presence of a methyl group at a Watson-Crick position of A9 appears especially interesting, since it potentially impairs base pairing of this nucleotide. Interestingly, A9 is present in a single-stranded domain in the native tRNA, whereas it is base paired in the in vitro transcribed tRNA and contributes to formation of its extended amino acid accepting stem. In this structure, base pairs A8-U65, A9-U64 and G10-C63 (circled in Fig. 4b) occur at the expense of base pairs U65-A49, U64-A50 and C63-G51, characteristic of a cloverleaf fold. Consequently it becomes tempting to hypothesize that residue A9 and its post-transcriptional modification are key factors in the folding process of tRNALys.
To test this hypothesis variants which would weaken the extended base pairing pattern in the transcript, and in addition, mimic the potential effect of methylation of a Watson-Crick position were designed. Three mutants (Kiw, Kew and Krw) that would lead to disruption of the Watson-Crick interactions in base pair A9-U64 of the extended RNA helix (Fig. 5a) were constructed. This pair is at a median location in the new 3 nt box revealed in the hairpin fold and thus its disruption should have the largest destabilizing effect.
Figure 6. Potential contribution of m1A9 to the folding of native human mt-tRNALys. The figure highlights the hypothetical role of the modification at A9 in the core of the tRNA in hindering Watson-Crick interaction with residue U64. Such an interaction occurring in the transcript cannot be formed in the native molecule due to the methyl group on position N1 of A9. Dashed lines correspond to Watson-Crick interactions in the elongated hairpin structure presented in Figure 4b. The solution structure of these three variants was investigated under the same experimental conditions as those used to establish the hairpin fold of the wild-type mt-tRNALys transcript. Whereas their behaviors resemble each other very closely, it is strikingly different from that of the wild-type and is indicative of an important reorganization of folding of the variants as compared with the wild-type transcript (Kwt). Figure 5b presents close-up views of autoradiograms with significant cleavage products of the variants in comparison with those of Kwt that support this conclusion. So, the RNase T1 cleavage pattern detects drastic changes in reactivities of G53 and G45 in the two types of RNAs (upper left panel in Fig. 5b). RNase T2 cuts are restricted to a smaller domain within the variants (upper right panel in Fig. 5b). Chemical probing with DMS and DEPC highlights large changes in reactivities, especially for C11, C15 and A9 (two bottom panels in Fig. 5b). For instance, A9, which is paired with U64 in the hairpin model (Figs 4 and 5a), is not reactive in Kwt and is strongly modified in the structural variants. Along the same lines, direct chemical probing of the N3 of C with DMS reveals negligible reactivity of C11 in the variants, whereas this position is clearly reactive in the wild-type, both under native and semi-denaturing conditions. In addition, both the lead and imidazole cleavage patterns (not shown) indicate large changes in reactivity at ntC11, U12, 50-53 and 61-65 between Kwt and the variants. These residues are reactive in the wild-type and protected in the variants. Alternatively, A8 and A9 are totally protected against these two probes in Kwt and become strongly reactive within the variants. In summary, and as hypothesized, the structural variants fold in a cloverleaf as does native tRNA. In conclusion, the present data supply strong support for a central role of residue A9 and its post-transcriptional modification in the establishment of the structure of native human mt-tRNALys. As illustrated in Figure 6, m1A9 would prohibit the Watson-Crick interaction A9-U64 present in our model and consequently hinder formation of the extended amino acid accepting stem. This would in turn allow formation of canonical T and D stems and thus of a cloverleaf. Thus the post-transcriptional modification of nt 9 may have a major role in folding of the tRNALys structure. A direct in vitro demonstration of this role relies on studies of chimeric molecules with individual modified residues. A correlation with in vivo events opens numerous interesting questions about the chronology of transcription, processing and modification of the tRNA. Similar questions have already been discussed for `hyperprocessing' of tRNAMeti (53).
ACKNOWLEDGEMENTS
We are indebted to Dr C.Marsac for numerous exchanges and discussions as well as for her kind and constant support. We thank Dr G.Keith for his helpful comments and for 3[prime]-end-labeling of the transcripts. This work was supported by grants from the Centre National de la Recherche Scientifique (CNRS), Université Louis Pasteur (Strasbourg) and Association Française contre les Myopathies (AFM). M.H. was supported by a fellowship from the Fonds der Chemischen Industrie and a TMR grant from the EC. H.B. was supported by an AFM grant.
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 24 Mar 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Messmer, J. Putz, T. Suzuki, T. Suzuki, C. Sauter, M. Sissler, and F. Catherine Tertiary network in mammalian mitochondrial tRNAAsp revealed by solution probing and phylogeny Nucleic Acids Res., November 1, 2009; 37(20): 6881 - 6895. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Messmer, A. Gaudry, M. Sissler, and C. Florentz Pathology-related mutation A7526G (A9G) helps in the understanding of the 3D structural core of human mitochondrial tRNAAsp RNA, August 1, 2009; 15(8): 1462 - 1468. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Rorbach, A. A. Yusoff, H. Tuppen, D. P. Abg-Kamaludin, Z. M.A. Chrzanowska-Lightowlers, R. W. Taylor, D. M. Turnbull, R. McFarland, and R. N. Lightowlers Overexpression of human mitochondrial valyl tRNA synthetase can partially restore levels of cognate mt-tRNAVal carrying the pathogenic C25U mutation Nucleic Acids Res., May 1, 2008; 36(9): 3065 - 3074. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Bonnefond, C. Florentz, R. Giege, and J. Rudinger-Thirion Decreased aminoacylation in pathology-related mutants of mitochondrial tRNATyr is associated with structural perturbations in tRNA architecture RNA, April 1, 2008; 14(4): 641 - 648. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hengesbach, M. Meusburger, F. Lyko, and M. Helm Use of DNAzymes for site-specific analysis of ribonucleotide modifications RNA, January 1, 2008; 14(1): 180 - 187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Ozanick, J. M. Bujnicki, D. S. Sem, and J. T. Anderson Conserved amino acids in each subunit of the heteroligomeric tRNA m1A58 Mtase from Saccharomyces cerevisiae contribute to tRNA binding Nucleic Acids Res., November 29, 2007; 35(20): 6808 - 6819. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Oliva, A. Tramontano, and L. Cavallo Mg2+ binding and archaeosine modification stabilize the G15 C48 Levitt base pair in tRNAs RNA, September 1, 2007; 13(9): 1427 - 1436. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Maniura-Weber, M. Helm, K. Engemann, S. Eckertz, M. Mollers, M. Schauen, A. Hayrapetyan, J.-C. von Kleist-Retzow, R. N. Lightowlers, L. A. Bindoff, et al. Molecular dysfunction associated with the human mitochondrial 3302A>G mutation in the MTTL1 (mt-tRNALeu(UUR)) gene Nucleic Acids Res., December 2, 2006; 34(22): 6404 - 6415. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Bailey, G. Robinson, A. Pinner, L. Chamlee, E. Ulasova, M. Pompilius, G. P. Page, D. Chhieng, N. Jhala, A. Landar, et al. S-adenosylmethionine prevents chronic alcohol-induced mitochondrial dysfunction in the rat liver Am J Physiol Gastrointest Liver Physiol, November 1, 2006; 291(5): G857 - G867. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Helm Post-transcriptional nucleotide modification and alternative folding of RNA Nucleic Acids Res., February 1, 2006; 34(2): 721 - 733. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. C. KURSCHAT, J. MULLER, R. WOMBACHER, and M. HELM Optimizing splinted ligation of highly structured small RNAs RNA, December 1, 2005; 11(12): 1909 - 1914. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Mollers, K. Maniura-Weber, E. Kiseljakovic, M. Bust, A. Hayrapetyan, M. Jaksch, M. Helm, R. J. Wiesner, and J.-C. von Kleist-Retzow A new mechanism for mtDNA pathogenesis: impairment of post-transcriptional maturation leads to severe depletion of mitochondrial tRNASer(UCN) caused by T7512C and G7497A point mutations Nucleic Acids Res., September 30, 2005; 33(17): 5647 - 5658. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Hao, M.-W. Zhao, Z.-X. Hao, Y.-N. Yao, and E.-D. Wang A T-stem slip in human mitochondrial tRNALeu(CUN) governs its charging capacity Nucleic Acids Res., June 22, 2005; 33(11): 3606 - 3613. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Sakurai, T. Ohtsuki, and K. Watanabe Modification at position 9 with 1-methyladenosine is crucial for structure and function of nematode mitochondrial tRNAs lacking the entire T-arm Nucleic Acids Res., March 21, 2005; 33(5): 1653 - 1661. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Levinger, M. Morl, and C. Florentz Mitochondrial tRNA 3' end metabolism and human disease Nucleic Acids Res., October 11, 2004; 32(18): 5430 - 5441. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Armengaud, J. Urbonavicius, B. Fernandez, G. Chaussinand, J. M. Bujnicki, and H. Grosjean N2-Methylation of Guanosine at Position 10 in tRNA Is Catalyzed by a THUMP Domain-containing, S-Adenosylmethionine-dependent Methyltransferase, Conserved in Archaea and Eukaryota J. Biol. Chem., August 27, 2004; 279(35): 37142 - 37152. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Uusimaa, S. Finnila, A. M. Remes, H. Rantala, L. Vainionpaa, I. E. Hassinen, and K. Majamaa Molecular Epidemiology of Childhood Mitochondrial Encephalomyopathies in a Finnish Population: Sequence Analysis of Entire mtDNA of 17 Children Reveals Heteroplasmic Mutations in tRNAArg, tRNAGlu, and tRNALeu(UUR) Genes Pediatrics, August 1, 2004; 114(2): 443 - 450. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. SISSLER, M. HELM, M. FRUGIER, R. GIEGE, and C. FLORENTZ Aminoacylation properties of pathology-related human mitochondrial tRNALys variants RNA, May 1, 2004; 10(5): 841 - 853. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Ando, T. Tanaka, and Y. Kikuchi Substrate Shape Specificity of E. coli RNase P Ribozyme Is Dependent on the Concentration of Magnesium Ion J. Biochem., April 1, 2003; 133(4): 445 - 451. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. M. Wittenhagen, M. D. Roy, and S. O. Kelley The pathogenic U3271C human mitochondrial tRNALeu(UUR) mutation disrupts a fragile anticodon stem Nucleic Acids Res., January 15, 2003; 31(2): 596 - 601. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ambrogelly, I. Ahel, C. Polycarpo, S. Bunjun-Srihari, B. Krett, C. Jacquin-Becker, B. Ruan, C. Kohrer, C. Stathopoulos, U. L. RajBhandary, et al. Methanocaldococcus jannaschii Prolyl-tRNA Synthetase Charges tRNAPro with Cysteine J. Biol. Chem., September 13, 2002; 277(38): 34749 - 34754. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. N. Mikhailov, J. Rozenski, E. V. Efimtseva, R. Busson, A. Van Aerschot, and P. Herdewijn Chemical incorporation of 1-methyladenosine into oligonucleotides Nucleic Acids Res., March 1, 2002; 30(5): 1124 - 1131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. McCulloch, B. L. Seidel-Rogol, and G. S. Shadel A Human Mitochondrial Transcription Factor Is Related to RNA Adenine Methyltransferases and Binds S-Adenosylmethionine Mol. Cell. Biol., February 15, 2002; 22(4): 1116 - 1125. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.-X. Guan, J. A. Enriquez, N. Fischel-Ghodsian, R. S. Puranam, C. P. Lin, M. A. Maw, and G. Attardi The Deafness-Associated Mitochondrial DNA Mutation at Position 7445, Which Affects tRNASer(UCN) Precursor Processing, Has Long-Range Effects on NADH Dehydrogenase Subunit ND6 Gene Expression Mol. Cell. Biol., October 1, 1998; 18(10): 5868 - 5879. [Abstract] [Full Text]
|
|
|
|
|
|
E. Tolkunova, H. Park, J. Xia, M. P. King, and E. Davidson The Human Lysyl-tRNA Synthetase Gene Encodes Both the Cytoplasmic and Mitochondrial Enzymes by Means of an Unusual Alternative Splicing of the Primary Transcript J. Biol. Chem., November 3, 2000; 275(45): 35063 - 35069. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||