Title: Wobble modification defect in tRNA disturbs codon-anticodon interaction in a mitochondrial disease
Abstract: Article3 September 2001free access Wobble modification defect in tRNA disturbs codon–anticodon interaction in a mitochondrial disease Takehiro Yasukawa Takehiro Yasukawa Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan Search for more papers by this author Tsutomu Suzuki Tsutomu Suzuki Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan Departmchent of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Bioscience Building 3S09, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8562 Japan Search for more papers by this author Norie Ishii Norie Ishii Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan Search for more papers by this author Shigeo Ohta Shigeo Ohta Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan Search for more papers by this author Kimitsuna Watanabe Corresponding Author Kimitsuna Watanabe Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan Departmchent of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Bioscience Building 3S09, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8562 Japan Search for more papers by this author Takehiro Yasukawa Takehiro Yasukawa Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan Search for more papers by this author Tsutomu Suzuki Tsutomu Suzuki Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan Departmchent of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Bioscience Building 3S09, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8562 Japan Search for more papers by this author Norie Ishii Norie Ishii Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan Search for more papers by this author Shigeo Ohta Shigeo Ohta Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan Search for more papers by this author Kimitsuna Watanabe Corresponding Author Kimitsuna Watanabe Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan Departmchent of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Bioscience Building 3S09, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8562 Japan Search for more papers by this author Author Information Takehiro Yasukawa1,2, Tsutomu Suzuki1,3, Norie Ishii2, Shigeo Ohta2 and Kimitsuna Watanabe 1,3 1Department of Chemistry and Biotechnology, Graduate School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan 2Department of Biochemistry and Cell Biology, Institute of Gerontology, Nippon Medical School, 1-396, Kosugi-cho, Nakahara-ku, Kawasaki, Kanagawa, 211-8533 Japan 3Departmchent of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Bioscience Building 3S09, 5-1-5 Kashiwanoha, Kashiwa, Chiba, 277-8562 Japan *Corresponding author. E-mail: [email protected] The EMBO Journal (2001)20:4794-4802https://doi.org/10.1093/emboj/20.17.4794 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info We previously showed that in mitochondrial tRNALys with an A8344G mutation responsible for myoclonus epilepsy associated with ragged-red fibers (MERRF), a subgroup of mitochondrial encephalomyopathic diseases, the normally modified wobble base (a 2-thiouridine derivative) remains unmodified. Since wobble base modifications are essential for translational efficiency and accuracy, we used mitochondrial components to estimate the translational activity in vitro of purified tRNALys carrying the mutation and found no mistranslation of non-cognate codons by the mutant tRNA, but almost complete loss of translational activity for cognate codons. This defective translation was not explained by a decline in aminoacylation or lowered affinity toward elongation factor Tu. However, when direct interaction of the codon with the mutant tRNALys defective anticodon was examined by ribosomal binding analysis, the wild-type but not the mutant tRNALys bound to an mRNA–ribosome complex. We therefore concluded that the anticodon base modification defect, which is forced by the pathogenic point mutation, disturbs codon–anticodon pairing in the mutant tRNALys, leading to a severe reduction in mitochondrial translation that eventually could result in the onset of MERRF. Introduction Mutations in tRNA genes encoded on mitochondrial DNA (mtDNA) are associated with a wide spectrum of human pathologies caused by mitochondrial disorders (Schon et al., 1997). A point mutation in the tRNALys gene at nucleotide position 8344, which is responsible for myoclonus epilepsy associated with ragged-red fibers (MERRF) (Fukuhara et al., 1980)—a major subgroup of the mitochondrial encephalomyopathies—was the first reported disease-related tRNA gene mutation (Shoffner et al., 1990). On the other hand, point mutations at nucleotide positions 3243 and 3271 in the tRNALeu(UUR) (R = A and G) gene were found in patients with another major subgroup—mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS) (Goto et al., 1990, 1991; Kobayashi et al., 1990). By establishing cybrid cell clones in which mutant mtDNA derived from patients was transferred intercellularly into human mtDNA-lacking (ρ0) cells (King and Attardi, 1989; Hayashi et al., 1991), the above three mutations (i.e. those at nucleotide positions 8344, 3243 and 3271) were revealed to be directly involved in mitochondrial dysfunction. In particular, the A to G transition at nucleotide position 8344 in the tRNALys gene was demonstrated to cause severe impairment of mitochondrial protein synthesis and respiratory deficiency in cultured myoblasts from MERRF patients (Boulet et al., 1992) and in ρ0 transformants (Chomyn et al., 1991; Yoneda et al., 1994). Peptide analysis carried out to try and explain the molecular mechanism suggested the occurrence of premature translation termination resulting from a decrease in aminoacyl-tRNALys (Enriquez et al., 1995). On the other hand, in tissue biopsies from MERRF patients harboring the 8344 mutation, neither the relative abundance nor the aminoacylation of the mutated tRNALys was found to be affected (Börner et al., 2000), which raises the question of whether the amount of aminoacyl-tRNALys significantly affects the molecular pathogenesis of MERRF. Recently, we found that a modification normally present at the anticodon wobble nucleotide is absent in mitochondrial tRNALys with the MERRF mutation [tRNALys(A8344G)] (Yasukawa et al., 2000a), while the other modified nucleotides remain normal in the mutant tRNALys (Figure 1). In addition, modification at the wobble position is also defective in tRNALeu(UUR) with either the MELAS 3243 [tRNALeu(UUR)(A3243G)] or 3271 [tRNALeu(UUR)(U3271C)] mutation (Yasukawa et al., 2000b). The wild-type tRNAs have novel uridine derivatives at the wobble position, whose chemical structures have been identified recently (T.Suzuki, T.Suzuki, T.Wada, K.Saigo and K.Watanabe, in preparation). Both tRNAs have a taurinomethyl group at position 5 of the uridine, while tRNALys additionally has a thio-group at position 2 [tRNALyssU*UU and tRNALeu(UUR)UAA, respectively]. Uridine modification at the wobble position has been studied in terms of its apparent importance in codon–anticodon interaction (Yokoyama and Nishimura, 1995; Takai et al., 1996; Hagervall et al., 1998; Krüger et al., 1998). Ashraf et al. (1999) have also shown that modification of the wobble uridine, particularly thiolation at position 2, is essential for tRNALyssU*UU species having 2-thiouridine derivatives at the wobble position (5-methylaminomethyl-2-thiouridine (mnm5s2U) in Escherichia coli tRNALys and 5-methoxycarbonylmethyl-2-thiouridine in human cytoplasmic tRNALys3, respectively) to interact with codons on ribosomes. Figure 1.Cloverleaf structures of human mitochondrial tRNALys from wild-type cells (left) and from MERRF patient-derived cells with the A8344G mutation (right). sU* indicates the modified uridine, and U on a round black background, unmodified uridine. G on a square black background represents the point mutation. The other modified nucleosides were determined previously: 1-methyladenosine (m1A), 2-methylguanosine (m2G), pseudouridine (Ψ) and N6-threonino carbonyladenosine (t6A) (Helm et al., 1998; Yasukawa et al., 2000a). Download figure Download PowerPoint These findings prompted us to speculate that the MERRF-mutant tRNALys intrinsically loses its translational activity due to an impairment of codon–anticodon interaction, eventually leading to mitochondrial dysfunction. To verify our speculation, we estimated the translational ability of tRNA in vitro using mitochondrial translation components and were able to obtain clear experimental evidence that the mutant tRNALys with the unmodified anticodon UUU (tRNALysUUU) is actually unable to translate its cognate codons AAR. Furthermore, taking advantage of ribosomal binding analysis, we show that the mutant tRNALys does not bind to AAA- programmed ribosomes. Our findings lead us to conclude that the wobble modification defect is primarily responsible for dispossessing the mutant tRNALys of its cognate codon binding affinity, forcing the mutant tRNALysUUU to become translationally inactive, which subsequently results in mitochondrial dysfunction. This is the first evidence that a post-transcriptional modification deficiency causes a human disease. Results Mitochondrial translation and oxygen consumption in MERRF-mutant cybrid cells The cybrid cell lines used were constructed previously by the intercellular transfer of mtDNA from a MERRF patient, or from fetal human fibroblasts as a control, to ρ0 HeLa cells (King and Attardi, 1989; Hayashi et al., 1991). A mutant cybrid clone (ME1-4) exclusively harboring mtDNA with the A8344G MERRF mutation consumed oxygen at a significantly lower rate (1.7 ± 0.2 fmol/min/cell) than the control cybrid (Ft2-11) with the wild-type mtDNA (5.3 ± 0.8 fmol/min/cell) when the oxygen consumption was measured as described by King and Attardi (1989). This finding is consistent with previous observations (Chomyn et al., 1991; Yoneda et al., 1994). Figure 2A shows the mitochondrial translation products labeled with [35S]methionine for 50 min in the presence of a cytoplasmic translation inhibitor, emetine, in which the polypeptides were identified as reported previously (Hayashi et al., 1993). The labeling efficiency of all the polypeptides was markedly reduced, indicating that the overall rate of protein synthesis in ME1-4 cells was very slow as compared with that in Ft2-11 cells. A faint but reproducible additional band was detected (indicated by an arrow in Figure 2A), which probably corresponds to the premature translation termination product of cytochrome c oxidase subunit I (COI) as has been reported (Chomyn et al., 1991; Boulet et al., 1992; Yoneda et al., 1994; Enriquez et al., 1995). The intensity of the abnormal band may depend on the nuclear background of the cybrid cells. Pulse labeling for 30 min and 2 h gave similar results (not shown). Western blot analysis (Figure 2B) revealed severely decreased steady-state levels of not only mitochondrially encoded cytochrome c oxidase subunits I and II (COI and COII) but also of nuclearly encoded subunit IV (COIV) in the mutant cells. The steady-state levels of internal control mitochondrial proteins, dihydrolipoamide succinyltransferase (DLST) (Nakano et al., 1993) and a translocase of the outer mitochondrial membrane (TOM 40) (Pfanner and Geissler, 2001), were comparable in the mutant and wild-type cybrids. Figure 2.(A) [35S]Methionine incorporation into the Ft2-11 wild-type cybrid (Wt) versus the ME1-4 MERRF-mutant cybrid (MERRF). CO, cytochrome c oxidase subunit; ND, NADH-ubiquinone oxidoreductase subunit; A, ATP synthase subunit. The arrow indicates the reported premature translation termination product of COI (Enriquez et al., 1995). (B) Western blots of cytochrome c oxidase subunits (COI, COII and COIV), dihydrolipoamide succinyltransferase (DLST) and translocase of the outer membrane 40 (TOM 40) in Ft2-11 and ME1-4 cybrids. Download figure Download PowerPoint Steady-state amounts and extent of aminoacylation of mutant and wild-type tRNALys in cybrid cells Northern blot analysis showed no reduction in the steady-state amount of the mutant tRNALys(A8344G) in the ME1-4 cybrid compared with that of the wild-type tRNALys in the Ft2-11 cybrid (Figure 3, lanes 1 and 4, and Figure 4 at time 0). The lysylation of the wild-type and mutant tRNAsLys was also examined by separating aminoacylated and uncharged tRNAs using acid urea polyacrylamide gel electrophoresis followed by northern hybridization (Varshney et al., 1991). Although the amount of uncharged tRNALys in the mutant cybrid was more than that in the control, the extent of lysylation in the mutant tRNALys appeared not to be markedly reduced (Figure 3), being 80% in the mutant cybrid and 93% in the control as determined by careful and repeated quantification. Figure 3.Quantitative analysis of aminoacyl-tRNALys in cells. The upper and lower bands correspond to aminoacyl and uncharged tRNALys, respectively. Lanes 2 and 3 depict RNA samples prepared under acidic conditions at 4°C from Ft2-11 and ME1-4 cybrids, respectively; lanes 1 and 4 show the Ft2-11 RNA and ME1-4 RNA, respectively, both treated with alkali. Download figure Download PowerPoint Figure 4.Time-dependent degradation of tRNA in the respective cybrid clones in the presence of a potential inhibitor of mitochondrial transcription. (A) Examples of northern hybridization for tRNALys and tRNAIle in the wild-type cybrid (left panels) and in the MERRF cybrid (right panels). (B) Time courses of quantitative analysis for the wild-type tRNALys (filled circles) and the mutant tRNALys (filled squares), normalized with tRNAIle in the respective cybrid clones. The average values at the starting time were defined arbitrarily as 100. Each set of data represents the average of at least three independent northern blotting experiments (SD bars are within the circles and squares). Download figure Download PowerPoint Stability of mutant tRNALys and its wild-type counterpart in cells To verify the finding of comparable steady-state amounts of tRNALys in mutant and wild-type cybrid cells, we further examined the degradation rates of the tRNAsLys and, as a control, of tRNAIle in the presence of a potential mitochondrial transcription inhibitor, ethidium bromide (EtBr) (Hayashi et al., 1990). Both the wild-type tRNALys and tRNAIle in the Ft2-11 cells appeared to degrade slightly faster than their respective counterparts in the ME1-4 cells (Figure 4A), which might result from a difference in sensitivity to EtBr between the two cell lines. However, the amounts of remaining tRNALys normalized at each time point by that of the remaining tRNAIle in the respective cybrid cells were comparable (Figure 4), showing that the point mutation occurring in the TΨC loop of tRNALys affects neither the stability of the tRNA nor the steady-state level in vivo. Aminoacylation efficiency and affinity for elongation factor Tu of mutant tRNALys in vitro Aminoacylation and elongation factor Tu (EF-Tu)-dependent hydrolysis protection assays were performed in vitro to observe the aminoacylation efficiency and affinity of tRNAsLys toward EF-Tu, respectively. For the following in vitro experiments, both the wild-type and mutant tRNAsLys were purified homogeneously by a solid-phase probing technique (Wakita et al., 1994) from respective mass cultures of wild-type and mutant cells. The initial rates of tRNALys lysylation were determined using mitochondrial lysyl-tRNA synthetase (KRSmt) purified from bovine liver by a conventional purification method (see Materials and methods). As shown in Table I, the kinetic parameters for the lysylation of the wild-type and mutant tRNAsLys did not differ significantly. Table 1. Kinetic parameters in aminoacylation of wild-type and MERRF-mutant tRNALys Substrate Km (μM)a Vmax (μM/s) Vmax/Km (relative) Wild-type tRNALys 0.71 2.9 × 10−3 1 MERRF tRNALys(A8344G) 0.18 2.4 × 10−4 0.32 aThe apparent Km values are given since the KRSmt used was a partially purified fraction. For the hydrolysis protection assay, recombinant bovine mitochondrial EF-Tu (EF-Tumt) was used and preparative aminoacylation of native tRNALys with or without the A8344G mutation was carried out with sufficient amounts of KRSmt and [3H]lysine. The stability of the ternary complex (Lys-tRNALys·EF-Tumt·GTP) against non-enzymatic deacylation under an alkaline pH condition was estimated by measuring the remaining radioactivity of [3H]Lys-tRNALys in the complex (Pingoud et al., 1977). Although the half-life of the wild-type tRNALys was about double that of the mutant tRNALys in the presence of each amount of EF-Tumt used (Figure 5C), the plots in Figure 5A and B show that both tRNAsLys were recognized efficiently by EF-Tumt. Figure 5.Stability of the ternary complex [3H]Lys-tRNALys·EF-Tumt·GTP against non-enzymatic deacylation. Time courses of remaining radioactivity of (A) wild-type and (B) mutant [3H]Lys-tRNAsLys determined at each indicated time point in the presence of the following amounts of EF-Tumt: 0.5 pmol (open triangles), 5 pmol (open circles), 25 pmol (filled squares), 37.5 pmol (filled triangles) and 50 pmol (filled circles), and in the absence of EF-Tumt (open squares). Values at the starting time (0 min) were defined as 100. (C) Half-life values (T1/2) of the wild-type [3H]Lys-tRNALys (open circles) and mutant [3H]Lys-tRNALys (open squares) plotted against the ratio of EF-Tumt/[3H]Lys-tRNALys initially added to the reaction mixture. Download figure Download PowerPoint Translational activities of tRNAs estimated in vitro using bovine mitochondrial translation components We examined whether the MERRF mutant tRNALys (A8344G), as well as the MELAS mutants tRNALeu(UUR) (A3243G) and tRNALeu(UUR)(U3271C), all lacking the modification only at the wobble position, could function in the translation elongation process. The in vitro assay was performed basically according to the literature (Eberly et al., 1985; Takemoto et al., 1995) using bovine mitochondrial components. To compare the translational efficiencies, all the tRNAs used were purified from the respective cell lines and equal amounts of the 3H-labeled aminoacyl-tRNAs prepared were put into the reaction mixtures as described in Materials and methods. Since we wished to emphasize the codon reading efficiency of the wild-type and mutant tRNAs, instead of natural mRNAs we used synthetic RNAs containing 30 repeats of cognate (AAA and AAG) and non-cognate (AAC and AAU) codons for tRNALys, and of cognate (UUA and UUG) codons for tRNALeu(UUR), respectively. Figure 6A and B shows that the mutant tRNALysUUU(A8344G) could not translate AAA and AAG codons, whereas the translation reaction proceeded quite efficiently with the wild-type tRNALys. Neither the mutant nor the wild-type tRNALys showed translational activity for non-cognate codons under the conditions used (Figure 6C and D). In addition, the translational activities of the two MELAS-mutant tRNAsLeu(UUR) for the cognate codons also decreased to a greater or lesser extent (Figure 6E and F). This result clearly supports our hypothesis that the MERRF-mutant tRNALys loses its translational activity, at least in vitro. Figure 6.Summary of in vitro translational elongation activities. (A–D) Translational efficiencies of the wild-type tRNALyssU*UU (Wt) and mutant tRNALysUUU with the A8344G mutation (MERRF) toward codons AAA, AAG, AAC and AAU. (E and F) Translational efficiencies of the wild-type tRNALeu(UUR)UAA (Wt) and two mutant tRNAsLeu(UUR)UAA with the A3243G (MELAS 3243) or U3271C (MELAS 3271) mutation toward codons UUA and UUG. The radioactivity of the [3H]aminoacyl-tRNA initially added to the reaction mixture was defined arbitrarily as 100. Each set of data represents the average of three independent experiments, with bars showing the SD. Download figure Download PowerPoint Binding affinity of mutant tRNALysUUU for AAA-programmed ribosomal small subunits Although the above experiment showed that the mutant tRNALysUUU almost completely lost its translational activity in vitro, it was essential to prove that this translational deficiency arose from the wobble modification defect itself but was not a direct result of the A to G replacement in the TΨC loop (Figure 1). For this purpose, we considered it appropriate to apply the binding assay of AAA-programmed E.coli ribosomal small subunits. Since a crystallographic analysis has shown that only the tRNA anticodon stem–loop (ASL) region, but not the TΨC loop, is located on the small subunits when tRNA binds to the ribosomes (Ogle et al., 2001; Yusupov et al., 2001), the possibility of the mutant nucleotide directly affecting codon–anticodon pairing in this binding assay can be excluded. As is clearly shown in Figure 7, the wild-type tRNALyssU*UU bound efficiently to AAA-programmed small subunits, but the mutant tRNALysUUU did not. This finding suggests that the defective translational activity of the MERRF-mutant tRNALys is caused by an inability to form codon (AAA)–anticodon (unmodified UUU) base pairs on the ribosomes. Figure 7.Binding of the wild-type tRNALyssU*UU (filled circles) and the MERRF-mutant tRNALysUUU (filled squares) to AAA-programmed ribosomal small subunits. Three independent experiments were performed and the average values are plotted on the graph, with the bars indicating the SD. Download figure Download PowerPoint Discussion As has been reported previously (Chomyn et al., 1991; Yoneda et al., 1994; Enriquez et al., 1995), the overall rate of mitochondrial protein synthesis was shown to be severely impaired in MERRF-mutant cybrid cells (Figure 2A). As expected, the steady-state amounts of mitochondrially encoded COI and COII proteins were very low (Figure 2B). The level of nuclear-encoded COIV was also found to decrease markedly in the mutant cybrid, which presumably can be accounted for by enhanced intramitochondrial degradation previously observed in ρ0 cells in which mitochondrial subunits were absent, and as a consequence cytochrome c oxidase could not be built up (Nijtmans et al., 1995). This might also be the case with NADH-ubiquinone oxidoreductase, because seven of the subunits are produced from mtDNA. It is possible to speculate that if a very small number of mitochondria exclusively had the wild-type tRNALys, while the vast majority had only the mutant counterpart, the faint radiolabeled protein bands in Figure 2A would be derived from normal mitochondrial translation provided by a small number of mitochondria. Enriquez et al. (1995) found decreased availability of aminoacyl-tRNALys with the 8344 mutation, which they postulated most probably brought about the translational defect in their mutant osteosarcoma-derived cybrid clones. In contrast, we found that the stability and steady-state amount of the mutant tRNALys(A8344G) in the ME1-4 cybrid remained unchanged when compared with its wild-type counterpart in the Ft2-11 cybrid (Figure 4). Furthermore, the extent of aminoacylation of the mutant tRNALys was comparable to that of the wild-type in the respective cybrids (Figure 3). Consistent with our findings, Börner et al. (2000) reported that neither the relative abundance nor the aminoacylation of mutated tRNALys was reduced in tissue biopsies from MERRF patients. The reason for the discrepancy with the result reported by Enriquez et al. (1995) is unknown, but it may be related to the cell types used. Nevertheless, our findings and those of Börner et al. (2000) both indicate that even without a reduction of aminoacylation, respiration declines in MERRF patients' cells. While the above analysis is salient, it was still considered important to elucidate the reason for the slightly reduced aminoacylation level in the mutant cybrid (Figure 3). The apparent kinetic parameters determined by a standard in vitro aminoacylation assay using bovine KRSmt showed that the lysine accepting activities (Vmax/Km values) of the wild-type and mutant tRNAsLys were similar, suggesting that the reduced aminoacylation level has no pathogenic significance (Table I). Because bovine KRSmt had the activity to aminoacylate human tRNALys as well as its bovine counterpart and the kinetic parameters in these aminoacylations were comparable (data to be published elsewhere), the region around the mutation point may not include an identity element(s) of the human, or bovine, tRNALys toward the corresponding KRSmt. For these reasons, the parameters obtained using bovine KRSmt would appear to be reasonable. Since bovine EF-Tumt bound efficiently to human aminoacyl-tRNALys, as would be expected from the fact that bovine and human EF-Tusmt are 95% identical in their sequences (Woriax et al., 1995), we used bovine EF-Tumt instead of its human counterpart for an EF-Tu-dependent hydrolysis protection analysis. Figure 5C shows that the half-life of Lys-tRNALys was shortened in the case of the mutant tRNALys, which implies that the binding affinity of EF-Tumt toward the mutant tRNA decreased to some extent. This finding is intriguing in seeking an explanation for the slight decrease in the ratio of aminoacyl-tRNALys to total tRNALys in mutant cybrid cells (Figure 3). The reduced affinity of the aminoacylated mutant tRNALys toward EF-Tumt presumably could be a candidate for explaining the decreased aminoacylation in the osteosarcoma background cybrid reported by Enriquez et al. (1995). However, in this study, the difference in affinity toward EF-Tumt between the wild-type and mutant tRNAsLys would not affect the translational efficiency in vivo since the ratio of aminoacylated tRNALys to total tRNALys was not greatly reduced in the mutant compared with that in the control cells (Figure 3). Additionally, in the in vitro translation reaction mixture without mRNA, the rates of deacylation in the wild-type and mutant tRNALys were observed to be almost identical (not shown). Since, for technical reasons, we used bovine enzymes in the two in vitro experiments above, the possibility that the results were due to heterogeneous enzymes cannot be excluded completely. Nevertheless, taking the in vivo and in vitro analyses together, it is reasonable to consider that the amount of aminoacylated tRNA is not the main factor underlying mitochondrial dysfunction in MERRF-mutant cells. We have already noted that the rate of mitochondrial protein synthesis was severely impaired in MERRF-mutant cybrid cells. At the same time, it has been proven that modification of a wobble uridine is indispensable for the tRNALyssU*UU species to bind poly(A) on ribosomal small subunits (Ashraf et al., 1999), and we recently found that the wobble modification is lacking in the MERRF-mutant tRNALys (Yasukawa et al., 2000a). Modifications of wobble bases are, in general, known to be very important for regulating the translational efficiency and accuracy of codon–anticodon base pairing. In addition, as shown in Figure 3, the absence of the wobble modification in the mutant tRNALys did not affect the aminoacylation, which is similar to the case in an in vivo study in which aminoacylation of E.coli tRNALys was shown to be unaffected by a specific lack of the wobble uridine modification (mnm5s2U) (Hagervall et al., 1998). There fore, we examined whether the mutant tRNALys is deprived of translational activity. For this purpose, we estimated the mutant tRNALysUUU activity in the translation elongation process in vitro using mitochondrial components and demonstrated that the mutant tRNALysUUU did indeed lose a significant amount of its essential function of translating lysine codons (Figure 6A and B). The magnitude of the decline in the translational activity of the mutant tRNA may have been overestimated to a certain extent, because unlike the actual circumstances in the mitochondrion, the mRNAs used contained repeats of the test codon. Furthermore, the in vitro conditions were heterogeneous. Nevertheless, we consider the in vitro assay appropriate to emphasize the defect of the mutant tRNA. Mistranslation of non-cognate codons was not observed in this system (Figure 6C and D). This finding is, unexpectedly, different from t