Communicated by: Yoshikazu Nakamura
Translation ability of mitochondrial tRNAsSer with unusual secondary structures in an in vitro translation system of bovine mitochondria
Article first published online: 20 DEC 2001
Genes to Cells
Volume 6, Issue 12, pages 1019–1030, December 2001
How to Cite
Hanada, T., Suzuki, T., Yokogawa, T., Takemoto-Hori, C., Sprinzl, M. and Watanabe, K. (2001), Translation ability of mitochondrial tRNAsSer with unusual secondary structures in an in vitro translation system of bovine mitochondria. Genes to Cells, 6: 1019–1030. doi: 10.1046/j.1365-2443.2001.00491.x
- Issue published online: 20 DEC 2001
- Article first published online: 20 DEC 2001
- Received: 14 August 2001Accepted: 1 October 2001
Background Metazoan mitochondrial (mt) tRNAs are structurally quite different from the canonical cloverleaf secondary structure. The mammalian mt tRNASerGCU for AGY codons (Y = C or U) lacks the entire D arm, whereas tRNASerUGA for UCN codons (N = A, G, C or U) has an extended anti-codon stem. It has been a long-standing problem to prove experimentally how these tRNAsSer work in the mt translation system.
Results To solve the above-mentioned problem, we examined their translational abilities in an in vitro bovine mitochondrial translation system using transcripts of altered tRNASer analogues derived from bovine mitochondria. Both tRNASer analogues had almost the same ability to form ternary complexes with mt EF-Tu and GTP. The D-arm-lacking tRNASer GCU analogue had considerably lower translational activity than the tRNASerUGA analogue and produced mostly short oligopeptides, up to a tetramer. In addition, tRNASerGCU analogue was disfavoured by the ribosome when other tRNAs capable of decoding the cognate codon were available.
Conclusion Both mt tRNASerGCU and tRNASerUGA analogues with unusual secondary structure were found to be capable of translation on the ribosome. However, the tRNASerGCU analogue has some molecular disadvantage on the ribosome, which probably derives from the lack of a D arm.
All known tRNAs from bacteria, archaea and eukaryotic cytoplasm, as well as those from chloroplasts, and from the mitochondria of yeasts, fungi and plants, possess a canonical cloverleaf secondary structure. However, metazoan mitochondrial (mt) tRNAs are structurally quite different. In the most extreme cases, all the tRNAsSerGCU (AGY [Y = C or U] codon-specific tRNAsSer with the anti-codon GCU) lack the D arm (de Bruijn et al. 1980), whereas the isoacceptor tRNAsSerUGA (UCN [N = A, G, C or U] codon-specific tRNAsSer with the anti-codon UGA) in mammals possess the bases necessary for canonical tertiary interactions, but have a slightly variant secondary structure with an extended anti-codon stem (Yokogawa et al. 1991) (see Fig. 1).
For these bizarre tRNAs to function in the translation process, the mutual distance and orientation between the anti-codon and the CCA terminus must remain constant (Steinberg et al. 1994; Watanabe et al. 1994; Hayashi et al. 1998). We and other groups have shown, by means of chemical and enzymatic probing, NMR analysis and computer modelling that tRNASer GCU (de Bruijn & Klug 1983; Hayashi et al. 1997b) as well as tRNASerUGA (Watanabe et al. 1994; Hayashi et al. 1998) can be folded into an L form-like tertiary structure, which conforms to the above constraints. However, Steinberg et al. (1994) have proposed a somewhat different ‘boomerang’ model for tRNASerGCU arising from its slightly smaller dimensions than those of canonical tRNAs (Wolstenholme 1992). Thus, the precise nature of the tertiary structures of these atypical tRNAs remain to be elucidated.
The biological functions of mt tRNAsSer have been studied in only a limited number of processes, including aminoacylation (Ueda et al. 1985; Yokogawa et al. 1989; Ueda et al. 1992) and ternary complex formation with bacterial EF-Tu and GTP (Gebhardt-Singh & Sprinzl 1986). The decoding ability of these mt tRNAs with unusual secondary structures has not thus far been examined, presumably due to the nonavailability of an in vitro translation system for mitochondria and to the difficulty of preparing these tRNAs in sufficient amounts for experimental purposes. Of course, no one has been able to get a functional bovine mitochondria cell-free protein synthesis system programmed with natural mRNA. Recently, however, we succeeded in constructing an in vitro bovine mitochondrial translation system and were able to prepare adequate quantities of tRNA transcripts derived from bovine mt tRNAsSer (Kumazawa et al. 1991; Takemoto et al. 1995; Hayashi et al. 1997a). Consequently, as structural models of bovine mt tRNAsSer, we have used modified tRNA transcripts with an altered anti-codon (GAA) to decode poly(U), and with an altered acceptor stem (G3–U70) for the alanylation by Escherichia coli Ala-tRNA synthetase (the modified transcripts were named tRNASerGCU/GAA and tRNASerUGA/GAA, respectively).
In this paper, we demonstrate that while the two mt tRNASer analogues can function in translation, the efficiency of tRNASerGCU/GAA lacking the D arm is much lower than that of tRNASerUGA/GAA. This may be related to the fact that in mammalian mitochondrial genomes the frequency of occurrence of AGY codons is much lower than that of UCN codons.
Designing mitochondrial serine tRNA transcripts with unusual secondary structures for a translation assay
Given the difficulty of isolating mt tRNAsSer from bovine liver mitochondria in amounts sufficient for biochemical study, we designed transcripts in which some residues were altered in such a way that would allow us to utilize the conventional in vitro translation assay with hot TCA precipitation and to obtain the transcripts in a high yield. First, the anti-codon sequences of both tRNASerGCU and tRNASerUGA were altered to GAA, corresponding to the UUU codon. This was done because: (i) it was considered desirable that the two tRNAs should have uniform codon/anti-codon interaction, and (ii) we have to use synthetic polynucleotides with consecutive codons as the mRNA for detecting appreciable amounts of translation products in the in vitro translation assay, because nobody has yet succeeded in using any natural mRNA in the assay. Therefore, it is most convenient to use poly(U) as the mRNA. Second, the G3–U70 base pair and A73 (in the standard numbering; Sprinzl et al. 1989) were introduced into the transcripts so that instead of being charged with Ser, they could be charged with Ala by E. coli alanyl–tRNA synthetase (Hou & Schimmel 1988; Shi et al. 1990), because poly or oligo(Ala)—but not poly or oligo(Ser)—can be precipitated by hot TCA in the procedure for detecting synthesized peptides. Third, all the base-pairs in the acceptor stem of both tRNAs except for G3–U70 were replaced by G–C pairs to enhance the transcription efficiency with T7 RNA polymerase, because it was previously found that: (i) even wild-type tRNASerGCU transcripts possessing the G1–C72 base pair had much lower yields than E. coli tRNA variants in transcription with T7 RNA polymerase (Ueda et al. 1992) and (ii) the amount of tRNASerGCU expressed in E. coli BL21(DE3) cells when the mt tRNASerGCU gene was transcribed by endogenously expressed T7 RNA polymerase increased exponentially with the number of G-C pairs in the acceptor stem (Hayashi et al. 1997a). The secondary structures of these modified tRNA transcripts are shown in Fig. 1 in comparison with those of their wild-type counterparts.
Aminoacylation activity of transcripts
tRNASerGCU/GAA and tRNASerUGA/GAA, prepared as previously described (Milligan & Uhlenbeck 1989) with a yield of 1.7 mg tRNASer transcripts per 10 mL reaction mixture, were alanylated with E. coli alanyl-tRNA synthetase (Hou & Schimmel 1988). The extents of alanylation of tRNASerGCU/GAA and tRNASerUGA/GAA were ≈10% and ≈50%, respectively. Since the identity determinant for E. coli alanyl-tRNA synthetase is preferentially located at G3–U70, the low aminoacylation activity of these tRNA analogues was not due to the alterations introduced at various sites (Fig. 1). To increase the alanylation extent of the analogues for the translation assay, nonalanylated tRNASer analogues were removed from Ala-tRNASer preparations by a hydrazine-coupled column (see Experimental procedures), which resulted in Ala-tRNASer analogues with a purity of more than 80% in both cases.
Ternary complex formation activities of transcripts with EF-Tu and GTP
The first step in examining the translational activity of the tRNA analogues was to check their ability to form ternary complexes with EF-Tu and GTP by means of a deacylation-protection assay (Pingoud et al. 1977). Since the tertiary structure of mt EF-Tu has been reported to be very similar to that of bacterial EF-Tu (Andersen et al. 2000), mt EF-Tu probably binds aminoacyl-tRNA in a similar manner to bacterial EF-Tu, which recognizes the 3′-terminal aminoacyl residue, the acceptor stem, and the T stem (Nissen et al. 1995). As shown in Fig. 2, both free Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA were deacylated very quickly, with the same half-life of 21 min, which was almost doubled when the tRNA analogues were mixed with mt EF-Tu·GTP.
The plots in Fig. 2 clearly demonstrate that the short half-lives of Ala-tRNASer GCU/GAA and Ala-tRNASer UGA/GAA were due to the amino acid residue attached to the tRNA moiety and not to the alterations in the acceptor stem. The tRNASer analogues could be charged with Ser by mt seryl-tRNA synthetase because the tRNA identity determinant for seryl-tRNA synthetase is known to be located in the T loop in the two serine tRNAs, with the T loop–D loop interaction being additionally required in the case of tRNASerUGA (Ueda et al. 1992; Shimada et al. 2001), both of which features were retained in the analogues. As shown in Fig. 2a, the free Ser-tRNASer GCU/GAA and wild-type Ser-tRNASerGCU had very similar half-lives (51–52 min), which were much longer than those of the ternary complex with Ala-tRNASerGCU/GAA (40 min). When these transcripts were complexed with EF-Tu·GTP, their half-lives were prolonged more than 10-fold. Similar results were obtained for Ser-tRNASerUGA/GAA and wild-type Ser-tRNASerUGA (Fig. 2b).
To obtain quantitative data on the ternary complex formation activities of Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA, gel-shift assays were performed, in which [14C] Ala-tRNASer analogues were incubated with increasing amounts of mt EF-Tu·GDPNP and the resulting mixture was electrophoresed on polyacrylamide gel. As shown in Fig. 3, upon formation of the complex, the band corresponding to the ternary complex appeared in the lower mobility region. The radioactivity of the band was quantified and the amount of free EF-Tu·GDPNP was calculated to obtain the apparent dissociation constant (Kd) by means of a Scatchard plot. The apparent Kd's for Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA were 14.2 and 7.2 µm, respectively.
In vitro translation assay using Ala-tRNASer GCU/GAA and Ala-tRNASer UGA/GAA
Poly(U)-directed poly(Ala) synthesis with Ala-tRNASer GCU/GAA and Ala-tRNASer UGA/GAA was carried out in an in vitro translation system consisting of purified bovine mt ribosomes (55S), mt EF-Tu, mt EF-G, poly(U), and the energy regenerating system. Since the amino acid incorporation reached a plateau after 10 min, we continued the incubation for 20 min to ensure an adequate reaction (data not shown). Figure 4 shows that in the case of Ala-tRNASerUGA/GAA, the incorporation of Ala into the hot TCA-insoluble material rose to more than 3.5 pmol as the analogue input was increased, although this was still only 1/10 the amount of Ala-tRNASer input. In the case of Ala-tRNASerGCU/GAA, however, the incorporation of Ala quickly reached a plateau at below 0.5 pmol, which was only 1/7 the amount incorporated with Ala-tRNASerUGA/GAA. A plausible explanation for this difference may be that, as described in the next section, tRNASerGCU/GAA cannot synthesize peptides long enough to be precipitated by 10% hot TCA. In both cases, the low levels of Ala incorporation would have arisen not from limited activity within the mt translation system itself, but from the low translational abilities of the mt Ala-tRNASer analogues (including their marked instability as noted above), because incorporation of Phe from E. coli Phe-tRNAPhe with poly(U) in this system reached 70% of the input Phe-tRNAPhe (data not shown).
Degree of polymerization of products
To estimate the lengths of the peptides synthesized by Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA in the above reactions, peptides released from peptidyl-tRNA analogues were analysed by a reversed-phase HPLC. The elution profiles are shown in Fig. 5. Each peak was assigned by comparison with the retention times of commercially available authentic tetra-, penta- and hexa-(Ala)s, which were, respectively, 5, 6.5 and 8 min (indicated by the numbered arrows in Fig. 5). In this system, mono-, di-, and tri-(Ala)s could not be separated from one another. The elution profiles show that tRNASerGCU/GAA hardly produced oligo(Ala)s longer than a tetramer, whereas tRNASerUGA/GAA could produce oligo(Ala)s longer than a decamer.
To determine the length of oligo(Ala)s that could be precipitated by 10% hot TCA, the products obtained with Ala-tRNASerUGA/GAA were precipitated by 10% hot TCA and the precipitates were analysed by the same HPLC system as above. Radioactivity could be detected from oligomers longer than a hexamer; and no radioactivity was detected from pentamer or shorter peptides (shown by the dotted line in Fig. 5). This result confirmed that the hot TCA-insoluble oligo(Ala)s were hexamers or longer. The relative ratio of peptides produced by the two tRNA analogues in Fig. 4 coincides well with that calculated from the sum of the peak areas comprising oligo(Ala) peaks longer than a hexamer in Fig. 5 (data not shown). These findings demonstrate that tRNASerGCU/GAA lacking the D arm is actually able to function on the ribosomes, albeit with very low efficiency.
Translational activity of tRNASer analogues in the presence of conventional tRNAPhe transcript
The low translational activity of Ala-tRNASerGCU/GAA toward poly(U) suggested the possibility that the mt ribosome cannot accept for translation several consecutive tRNAs lacking the D arm. To investigate this, in vitro translation with either Ala-tRNASerGCU/GAA or Ala-tRNASerUGA/GAA was carried out in the presence of an E. coli tRNAPheGAA transcript possessing the conventional cloverleaf structure using crude mt ribosomes (Fig. 6) instead of pure 55S ribosome (Fig. 4). The transcript of E. coli[14C] Phe-tRNAPheGAA was added to the reaction mixture at molar ratios of 0.05, 0.125 and 0.25 toward [3H] Ala-tRNASerGCU/GAA or [3H] Ala-tRNASerUGA/GAA (20 pmol each). Although we optimized the input amount of crude mt ribosomes, the amount of Ala incorporated in the absence of the tRNAPhe transcript was only about 1/3 that shown in Fig. 4, when purified mt ribosome (55S) was used. However, the relative ratio of Ala incorporation between Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA remained unchanged. This low level of Ala incorporation may have been caused by nuclease contamination and/or other inhibitory factors in the crude mt ribosome extract.
Addition of Phe-tRNA enhanced the incorporation of Ala in the translation of both tRNASer analogues (Fig. 6), indicating that the low translational activity of the tRNASer analogues was improved with the help of conventional tRNAPhe. Nevertheless, the incorporation of [3H] Ala from Ala-tRNASerGCU/GAA into peptides with the addition of Phe-tRNAPheGAA was still lower than that of Ala-tRNASerUGA/GAA. If the low level of Ala incorporation in the translation of tRNASerGCU was only due to the difficulty of arranging tRNASerGCU side by side on the ribosome, Ala would have been incorporated to the same level in the cases of both tRNASer analogues with the help of Phe-tRNAPheGAA. In fact, [14C] Phe incorporation from Phe-tRNAPheGAA with Ala-tRNASer GCU/GAA was slightly higher than that with Ala-tRNASerUGA/GAA. This means that tRNAPhe was utilized more frequently in the translation with tRNASerGCU/GAA than in that with tRNASerUGA/GAA, implying that the D-arm-lacking Ala-tRNASerGCU/GAA was less favoured by the ribosomes when other tRNAs capable of decoding the cognate codon were available.
Rationale of using modified tRNA analogues for the functional assay
It has been a long-standing question of whether bizarre tRNAs found in animal mitochondria—such as metazoan tRNAsSer lacking the D arm (de Bruijn et al. 1980) and most nematode tRNAs without the T arm (Okimoto & Wolstenholme 1990)—can function in the translation process with a similar efficiency to canonical tRNAs possessing the cloverleaf secondary structure. In this study, we succeeded in partially resolving this question by using an in vitro bovine mitochondrial translation system and two artificial tRNAs derived from bovine mt tRNAsSer (tRNASer GCU/GAA and tRNASer UGA/GAA) (Fig. 1).
Our rationale for using these particular modified tRNASer analogues for the functional assay of mt tRNAsSer is as follows. By restricting the alterations to the acceptor stem and anti-codon, the tertiary structures could be preserved as in the native tRNAsSer, because in both tRNAs the interactions that take place to form the tertiary structure do not involve the acceptor stem and anti-codon but the T loop and extra loop, and additionally the truncated D loop in the case of tRNASerGCU (Fig. 1). It has been reported that changing any of the base pairs in the acceptor stem hardly reduces the aminoacylation efficiency of tRNASerGCU (Ueda et al. 1992; Hayashi et al. 1997a), and 1H-NMR measurement has also revealed no appreciable changes in the conformation of the acceptor stem resulting from replacement of the existing acceptor base pairs of tRNASerGCU by G-C pairs (Hayashi et al. 1997b). The fact that both Ser-tRNASerGCU/GAA and Ser-tRNASerUGA/GAA formed ternary complexes with similar affinities to those of the wild-type Ser-tRNASerGCU and Ser-tRNASerUGA (Fig. 2) is further solid evidence that the alterations made in the acceptor stem of the tRNAsSer had no appreciable influence upon their ternary complex formation ability. A rather intriguing finding was that the half-lives of the Ala-tRNASer analogues were extremely short, and even the half-lives of their ternary complexes were shorter than those of free Ser-tRNASer analogues. This was reflected in the large Kd values for the Ala-tRNASer analogues and EF-Tu·GDPNP (14.2 and 7.2 µm), which were three orders of magnitude higher than those reported for mt EF-Tu·GTP and E. coli Phe-tRNAPhe (18 ± 4 nm; Cai et al. 2000). In addition to the instability of Ala-tRNASer analogues themselves, this may also be related to the implication that the affinity between aminoacyl-tRNA and EF-Tu is influenced by the amino acid residue attached to the tRNA (Louie et al. 1984).
In addition to the changes made in the acceptor stem, we also replaced the original anti-codon sequences (GCU or UGA) by GAA so that the anti-codon loop sequence would decode the UUU codon in each case. Such replacements in the anti-codon have often been employed, for example, in using suppressor tRNAs to decode a termination codon (Murgola 1995), so they would be unlikely to greatly influence the original function of the tRNA.
Translational ability of Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA
In the in vitro mt translation system, incorporation of Phe from E. coli Phe-tRNAPhe into the hot TCA-insoluble material reached 70% of the input Phe-tRNAPhe. This is similar to the case for mt Phe-tRNAPhe (our unpublished results), suggesting there is little difference between the translation efficiency of E. coli tRNAPhe with the canonical secondary structure and that of mt tRNAPhe lacking the D loop–T loop interaction in regard to poly(U)-directed poly(Phe) synthesis using the bovine mt translation system.
In contrast, incorporation of Ala from Ala-tRNASer UGA/GAA in the same system was only 10% of the input Ala-tRNASerUGA/GAA, while that from Ala-tRNASerGCU/GAA was as low as 1.2% (Fig. 4). As already noted, these low incorporation levels can be ascribed to the extreme instability of the Ala-tRNASer analogues, and also of their ternary complexes with EF-Tu·GTP, rather than to the alterations made in the analogues per se. The instability of the Ala-tRNASer analogues can be explained by the relative stability of the aminoacyl ester bond against hydrolysis; since the methyl group, the side chain of Ala, is the second-lowest electron donor for the carbonyl group among the side chains of the 20 amino acids normally found in proteins, the aminoacyl ester bond would be very labile. Although the relative affinity of Ala-tRNASer GCU/GAA toward EF-Tu is half that of Ala-tRNASerUGA/GAA (Fig. 3), incorporation of Ala from Ala-tRNASer GCU/GAA was just 1/7 that from Ala-tRNASerUGA/GAA (Fig. 4), which can be attributed to the difference in their tRNA structure, particularly the lack of the D arm in tRNASerGCU/GAA.
It is intriguing that the incorporation of Ala from Ala-tRNASerGCU/GAA soon reached a plateau (at 0.5 pmol) despite further increases in the amount of input Ala-tRNASerGCU/GAA, and that Ala-tRNASer GCU/GAA produced oligo(Ala)s up to a tetramer, whereas Ala-tRNASerUGA/GAA gave longer peptides. A possible explanation for these findings is the occurrence of one or more rate-limiting steps in the process of Ala-tRNASerGCU/GAA elongation due to the absence of the D arm; one plausible possibility lies in the binding of Ala-tRNASerGCU/GAA to the ribosomal A site (Rodnina et al. 1994). This effect may attenuate the translation process when tRNASerGCU is used by mt ribosomes in vivo. An alternative explanation could be that tRNASerGCU/GAAs bearing short oligo(Ala)s are apt to drop off the ribosomes for some unknown reason. These possibilities will need to be examined in future work.
Relationship between frequency of occurrence of serine codons in various animal mt genomes and the number of nucleotides in the D arm region of tRNASerGCU
A number of metazoan mitochondrial DNA sequences have so far been reported (Boore 1999; and references therein), which provide us with information on codon usage in metazoan mitochondria. Table 1 shows the occurrence frequencies of UCN, AGY and AGN codons, and of tandem repeats of serine codons, in mitochondria of various animal species, chosen mainly from species whose tRNAsSer sequences at the RNA level have been investigated (Nishioka et al. 1994; Matsuyama et al. 1998; Tomita et al. 1998, 1999; Kondow et al. 1999). In the universal genetic code, UCN and AGY codons are assigned to Ser. In most invertebrate mitochondria, AGR (R = A or G) codons also correspond to Ser, so that the whole AGN family box belongs to Ser codons, which are thought to be read by tRNASerGCU (in Drosophila mitochondria AGG is an unassigned codon, and in nematode mitochondria AGN codons are read by tRNASerUCU).
|Animal species||Total number of serine codons||UGN||AGY or AGN||UCN-UCN||UCN-AGY or UCN-AGN||AGY-UCN or AGN-UCN||AGY-AGY or AGN-AGN||Number of nucleotides in D-loop/stem region of tRNASerGCU|
The total number of Ser codons in the whole genome is around 340–390 in most invertebrates, but it decreases to under 300 in prochordates and vertebrates. While the usage of AGY or AGN codons is over 100 in invertebrates and prochordates, it is much lower—around 50—in vertebrates. The most noticeable feature is that tandem repeats of AGY codons do not exist in mammalian (human, bovine, rat, and mouse) mt genomes, and occur only once in the chicken and Xenopus mt genomes. These observations strongly suggest the presence of some suppressive factors in the vertebrate mt translation system that makes the translation of AGY codons unfavourable. We speculate that this is most likely to be the lack of the D arm in tRNASer GCU. As is also shown in the right-most column of Table 1, the length of the tRNASerGCU D loop/stem appears to have become shorter during the evolution of animal species. Four mammalian tRNAsSer GCU have only 3–5 nucleotides in the region corresponding to the D loop/stem of normal tRNAs, whereas chicken and Xenopus tRNAsSer GCU have 13 and 10 nucleotides, respectively, which is similar to the cases of invertebrate tRNAsSerGCU.
Only in C. elegans mitochondria are AGN codons utilized more than UCN. This exception may be explained by the following factors: (i) the anti-codon of tRNASerUCU is UCU, which means the tRNA would be able to decode AGN codons much efficiently than tRNASerGCU; (ii) nematode tRNASerUGA also lacks the D arm, as does tRNASerUCU; (iii) the total chain length of tRNASerUGA is shorter than that of tRNASerUCU (Okimoto & Wolstenholme 1990). All of these factors would be greatly advantageous to the translational proficiency of tRNASerUCU in comparison to that of tRNASerUGA.
Our finding that tRNAsSerGCU/GAA lacking the D arm have a low translational ability compared to tRNAsSerUGA/GAA may be associated with the relationship between the codon usage of serine and the absence of the D arm in tRNASerGCU, as indicated in Table 1. It has been shown that in many organisms there is a strong correlation between the most frequently used synonymous codon and the most prevalent isoacceptor tRNA (Ikemura 1981, 1982; Yamao et al. 1991). In human mitochondria, although tRNASerGCU and tRNASerUGA have almost the same abundance, the AGY codons are utilized less frequently than UCN (King & Attardi 1993), which could also be related to the low translational activity of tRNASerGCU/GAA compared with that of tRNASerUGA/GAA. These possibilities will require further clarification.
Preparation of tRNA
Mitochondrial tRNASerGCU and tRNASerUGA analogues were prepared by in vitro transcription using T7 RNA polymerase as previously described (Milligan & Uhlenbeck 1989): the anti-codon was altered to GAA, the third base pair in the acceptor stem was replaced by G3-U70, the discriminator base of tRNASerGCU was altered to A73, and all the base pairs in the acceptor stem, except for G3-U70 were replaced by G-C (see Fig. 1). A 10-fold excess of 5′-GMP over 5′-GTP was added to the transcription reaction mixture to obtain the 5′-monophosphate termini of the tRNAs. Transcripts were purified by denaturing polyacrylamide gel electrophoresis and stored in deionized water. Alanylation was carried out in 30 mm Tris/HCl (pH 7.8), 5 mm MgCl2, 10 mm KCl, 1 mm dithiothreitol, 2 mm ATP, 27 µm[3H]l-alanine (2.07 TBq/mmol) or 21 µm[14C]l-alanine (5.62 GBq/mmol), 1 A260 unit of the relevant tRNASer analogue, and 2 µg recombinant E. coli Ala-tRNA synthetase (Hou & Schimmel 1988). Serylation was performed in 100 mm Hepes/KOH (pH 7.8), 15 mm MgCl2, 60 mm KCl, 5 mm dithiothreitol, 2 mm ATP, 1 mm spermine, 33 µm[14C]l-serine (5.59 GBq/mmol), 0.4 A260 unit of the relevant tRNASer, and 6 µg recombinant bovine mt Ser-tRNA synthetase (Yokogawa et al. 2000). The aminoacylated-tRNA was extracted by acidic phenol after adjusting the pH of the reaction mixture to 4.5 by the addition of acetic acid. The aqueous phase was concentrated by ethanol precipitation and the recovered RNA fraction was dissolved in 20 mm sodium acetate (pH 5.5).
Purification of alanyl-tRNASer analogues
Cyanogen bromide-activated Sepharose 4B coupled with hydrazine was prepared as previously described (Cuatrecasas 1970). Following the alanylation reaction, uncharged tRNA was oxidized at the 3′-terminal adenylate residue with a periodic acid in a mixture containing 0.1 m sodium acetate (pH 5.0), 10 mm MgCl2, and 10 mm NaIO4. After ethanol precipitation, the material was loaded on to the above-mentioned column (1 A260 unit tRNA/mL resin), recovered (the material was not adsorbed by the column), and concentrated by ethanol precipitation. The sample was dissolved in 20 mm sodium acetate (pH 5.5).
Preparation of ribosomes and elongation factors
Mitochondria were prepared from fresh bovine liver according to the literature (O'Brien & Denslow 1996). Crude mt ribosomes were prepared as reported (Eberly et al. 1985) and stored in 55S buffer: 20 mm Tris/HCl (pH 7.6), 20 mm MgCl2, 80 mm KCl, and 6 mm 2-mercaptoethanol. 55S mt ribosome was purified by sucrose density gradient (SDG) centrifugation in the 55S buffer containing sucrose. Eighty A260 units of crude mt ribosome dissolved in 0.6 mL 55S buffer without sucrose were layered on to a 6–38% (w/v) sucrose gradient, followed by centrifugation at 17 500 r.p.m. for 18 h using a SW28 rotor (Beckman). Fractions containing 55S mt ribosome were collected by ultracentrifugation at 35 000 r.p.m. over a period of 24 h using a 70Ti rotor (Beckman). The purified 55S mt ribosome was stored in the 55S buffer. Recombinant mt EF-Tu was expressed from a plasmid kindly provided by Dr L.L. Spremulli (Woriax et al. 1995). The enzyme carrying a His-tag at the C-terminus was purified on a Hi-Trap chelating column (Amersham Pharmacia Biotech) and used for the translation assay without removing the His-tag. Mitochondrial EF-G was partially purified as described (Chung & Spremulli 1990) more than 100-fold over the crude mitochondrial extract by two successive column chromatography steps.
The deacylation-protection assay was basically performed according to Pingoud et al. (1977). The reaction mixture contained 75 mm Tris/HCl (pH 7.4), 75 mm NH4Cl, 15 mm MgCl2, 7.5 mm dithiothreitol, 0.1 mm GTP, 60 µg/mL bovine serum albumin, 2.25 mm phosphoenolpuruvate, 2.3 units/mL pyruvate kinase, 1.2 µm mt EF-Tu, and 0.2 µm[14C] aminoacyl-tRNA. The reaction mixture was pre-incubated without [14C] aminoacyl-tRNA at 30 °C for 5 min and [14C] aminoacyl-tRNA was then added. The deacylation reaction was carried out at 30 °C.
Gel-shift analysis of the ternary complex
Mitochondrial EF-Tu·GDPNP complex was kindly provided by N. Hino (Department of Integrated Biosciences, The University of Tokyo). The gel-mobility shift assay was carried out as follows. Reaction mixtures (15 µL) were prepared containing 10 mm Tris/HCl (pH 7.5), 2 mm magnesium acetate, 13 mm ammonium acetate, 2 mm GDPNP, 3.3 µm[14C] Ala-tRNASer, and mt EF-Tu·GDPNP at various concentrations (2–8 µm). Following incubation at 37 °C for 12 min, 1.5 µL of a solution containing 50% glycerol and 0.05% bromophenol blue was added, and the mixture (16.5 µL) was loaded on to 5% polyacrylamide gel (0.5 mm thickness) containing 50 mm Tris/HCl (pH 6.8), 10 mm magnesium acetate, 65 mm ammonium acetate, 1 mm EDTA, 1 mm dithiothreitol and 10 µm GDPNP. Electrophoresis was carried out at 50 mA for 1 h at 4 °C, after which the gel was dried on a Whatmann 3MM CHR filter. The dried gel was exposed to an imaging plate, followed by analysis using a Fuji Photo Film BAS1000 bioimaging analyser. The apparent dissociation constant for the ternary complex was calculated by means of a Scatchard plot.
In vitro translation assay
The reaction was carried out as previously described with partial modification (Takemoto et al. 1995). The reaction mixture (40 µL) contained 50 mm Tris/HCl (pH 8.5), 7 mm MgCl2, 45 mm KCl, 0.5 mm spermine, 1 mm dithiothreitol, 2.5 mm phospoenolpyruvate, 2.5 units/mL pyruvate kinase, 0.5 mm GTP, 0.3–1 µm[3H] Ala-tRNASer analogue, 25 nm purified mt ribosome (55S) (for the in vitro mitochondrial translation assay using Ala-tRNASerGCU/GAA and Ala-tRNASerUGA/GAA) or 1.5 A260/mL crude mt ribosomes (to determine the degree of polymerization of the products and the translational activity of the in vitro translation of the tRNASer analogues in the presence of conventional tRNAPhe transcripts), 1.25 µm recombinant mt EF-Tu, 300 mg/mL mt EF-G, and 1 mg/mL poly(U). After incubation at 37 °C for 20 min, the reaction was terminated by adding 10% trichloroacetic acid (TCA) and the mixture was heated at 80 °C for 30 min The radioactive TCA-insoluble precipitate of polymerized alanine was trapped on an Advantec cellulose filter (A045B025A) and the radioactivity was measured using a liquid scintillation counter.
HPLC analysis of translation products
After the translation reaction mixture (80 µL, containing 0.5 µm[3H] Ala-tRNASer analogue) was incubated at 37 °C for 20 min, 0.5 N NaOH was added to increase the pH to 12 and the mixture was further incubated for 1 h at 37 °C. The reaction mixture was then neutralized with acetic acid and analysed by HPLC. The chromatography was performed on a TSK ODS-80Ts column (φ 4.6 mm × 75 mm) with a linear gradient of Solution A (0.1% aqueous TFA) and Solution B (0.1% TFA in acetonitrile) at a flow rate of 1 mL/min. Elution was accomplished in 40 min Fractions were collected at 30 ss intervals and the radioactivities of samples mixed with Ultima Gold (Packard) were measured using a liquid scintillation counter. To determine the length of oligo(Ala) that could be precipitated by 10% TCA, the following procedures were adopted. Oligo(Ala) produced by the in vitro translation system using Ala-tRNASerUGA/GAA was trapped on a cellulose filter (ADVANTEC A045B025A), which was then dissolved in 0.1% TFA and 70% 2-propanol. The sample was evaporated to dryness, suspended in water (3 mL), and filtered through a Millipore Ultrafree centrifugal filter (pore size = 0.45 µm). The volume of the filtrate was reduced to 100 µL by an evaporator and analysed by HPLC as described above.
We thank Takeo Suzuki, Nobukazu Shimada, Yoshihiro Shimizu and Narumi Hino for their kind gifts of materials and Dr Linda L. Spremulli for providing a plasmid of mt EF-Tu gene. We also thank the Radioisotope Center, The University of Tokyo, for allowing us to use its facilities. This work was supported by a Grant-in Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
- 2000) High resolution crystal structure of bovine mitochondrial EF-Tu in complex with GDP. J. Mol. Biol. 297, 421–436.DOI: 10.1006/jmbi.2000.3564 , , , (
- 1999) Animal mitochondrial genomes. Nucl. Acids Res. 27, 1767–1780. (
- 1983) A model for the tertiary structure of mammalian mitochondrial transfer RNAs lacking the entire ‘dihydrouridine’ loop and stem. EMBO J. 2, 1309–1321. & (
- 1980) A mammalian mitochondrial serine transfer RNA lacking the ‘dihydrouridine’ loop and stem. Nucl. Acids Res. 8, 5213–5222. , , , et al. (
- 2000) Interaction of mitochondrial elongation factor Tu with aminoacyl-tRNA and elongation factor Ts. J. Biol. Chem. 275, 20308–20314. , , , (
- 1990) Purification and characterization of elongation factor G from bovine liver mitochondria. J. Biol. Chem. 265, 21000–21004. & (
- 1970) Protein purification by affinity chromatography. J. Biol. Chem. 245, 3059–3065. (
- 1985) Bovine mitochondrial ribosomes elongation factor specificity. J. Biol. Chem. 260, 8721–8725. , , (
- 1986) Ser-tRNAs from bovine mitochondria form ternary complexes with bacterial elongation factor Tu and GTP. Nucl. Acids Res. 14, 7175–7188. & (
- 1997a) Expression of bovine mitochondrial tRNASerGCU derivatives in Escherichia coli. Nucl. Acids Res. 17, 3503–3507. , , (
- 1998) Higher-order structure and thermal instability of bovine mitochondrial tRNASerUGA investigated by proton NMR spectroscopy. J. Mol. Biol. 284, 57–69.DOI: 10.1006/jmbi.1998.2151 , , (
- 1997b) Assignment of imino proton signals of G-C base pairs and magnesium ion binding: An NMR study of bovine mitochondrial tRNASerGCU lacking the entire D arm. J. Biochem. (Tokyo) 121, 1115–1122. , , , , , (
- 1988) A simple structural feature is a major determinant of the identity of a transfer RNA. Nature 333, 140–145. & (
- 1981) Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes. J. Mol. Biol. 146, 1–21. (
- 1982) Correlation between the abundance of yeast transfer RNAs and the occurrence of the respective codons in protein genes. Differences in synonymous codon choice patterns of yeast and Escherichia coli with reference to the abundance of isoaccepting transfer RNAs. J. Mol. Biol. 158, 573–597. (
- 1993) Post-transcriptional regulation of the steady-state levels of mitochondrial tRNAs in HeLa cells. J. Biol. Chem. 268, 10228–10237. & (
- 1999) An extra tRNAGly (U*CU) found in ascidian mitochondria responsible for decoding non-universal codons AGA/AGG as glycine. Nucl. Acids Res. 27, 2554–2559.DOI: 10.1093/nar/27.12.2554 , , , , (
- 1991) Interaction of bovine mitochondrial phenylalanyl-tRNA with ribosomes and elongation factors from mitochondria and bacteria. Biochim. Biophys. Acta 1090, 167–172. , , , et al. (
- 1984) Relative affinities of all Escherichia coli aminoacyl-tRNAs for elongation factor Tu-GTP. J. Biol. Chem. 259, 5010–5016. , , , (
- 1998) A novel wobble rule found in starfish mitochondria. J. Biol. Chem. 273, 3363–3368.DOI: 10.1074/jbc.273.6.3363 , , , , (
- 1989) Synthesis of small RNAs using T7 RNA polymerase. Meth. Enzymol. 180, 51–62. & (
- 1995) Translational suppression. In: Trna: Structure, Biosynthesis, and Function (eds D.Söll & U.L.RajBhandary), pp. 491–509. Washington, DC : American Society for Microbiology. (
- 1994) Structural analysis of chicken mitochondrial tRNAsSer. Nucl. Acids Symp. Ser. 31, 187–188. , , , , (
- 1995) Crystal structure of the ternary complex of Phe-tRNAPhe, EF-Tu, and a GTP analog. Science 270, 1464–1472. , , , et al. (
- 1996) Bovine mitochondrial ribosomes. Meth. Enzymol. 264, 237–248. & (
- 1990) A set of tRNAs that lack either the TΨC arm or the dihydrouridine arm: towards a minimal tRNA adaptor. EMBO J. 9, 3405–3411. & (
- 1977) Ternary complex formation between elongation factor Tu, GTP and aminoacyl-tRNA: an equilibrium study. Eur. J. Biochem. 78, 403–409. , , , , (
- 1994) Transient conformational states of aminoacyl-tRNA during ribosome binding catalyzed by elongation factor Tu. Biochemistry 33, 12267–12275. , , (
- 1990) A nucleotide that enhances the charging of RNA minihelix sequence variants with alanine. Biochemistry 29, 3621–3626. , , , (
- 2001) Dual mode recognition of two isoacceptor tRNAs by mammalian mitochondrial seryl-tRNA synthetase. J. Biol. Chem., in press. , , (
- 1989) Compilation of tRNA sequences and sequences of tRNA genes. Nucl. Acids Res. 17(Suppl.), 1–172. , , , , (
- 1994) Fitting the structurally diverse animal mitochondrial tRNAsSer to common three-dimensional constraints. J. Mol. Biol. 236, 982–989. , , (
- 1995) The ability of bovine mitochondrial transfer tRNAMet to decode AUG and AUA codons. Biochimie 77, 104–108.DOI: 10.1016/0300-9084(96)88112-0 , , , et al. (
- 1999) Codon reading patterns in Drosophila melanogaster mitochondria based on their tRNA sequences: a unique wobble rule in animal mitochondria. Nucl. Acids Res. 27, 4291–4297. , , , , , (
- 1998) 7-Methylguanosine at the anticodon wobble position of squid mitochondria tRNASerGCU: molecular basis for assignment of AGA/AGG codons as serine in invertebrate mitochondria. Biochim. Biophys. Acta 1399, 78–82. , , (
- 1985) Large scale isolation and some properties of AGY-specific serine tRNA from bovine heart mitochondria. J. Biochem. (Tokyo) 98, 1275–1284. , , (
- 1992) The T-loop region of animal mitochondrial tRNASer (AGY) is a main recognition site for homologous seryl-tRNA synthetase. Nucl. Acids Res. 20, 2217–2222. , , , (
- 1994) Higher-order structure of bovine mitochondrial tRNASerUGA: chemical modification and computer modeling. Nucl. Acid Res. 22, 5378–5384. , , , et al. (
- 1992) Animal mitochondrial DNA: structure and evolution. Int. Rev. Cytol. 141, 173–216. (
- 1995) Cloning sequence analysis and expression of mammalian mitochondrial protein synthesis elongation factor Tu. Biochim. Biophys. Acta 1264, 347–356. , , (
- 1991) Levels of tRNAs in bacterial cells as affected by amino acid usage in proteins. Nucl. Acids Res. 19, 6119–6122. , , , , (
- 1989) Purification and characterization of two serine isoacceptor tRNAs from bovine mitochondria by using a hybridization assay method. Nucl. Acids Res. 17, 2623–2638. , , , (
- 2000) Characterization and tRNA recognition of mammalian mitochondrial seryl-tRNA synthetase. J. Biol. Chem. 275, 19913–19920. , , , et al. (
- 1991) A novel cloverleaf structure found in mammalian mitochondrial tRNASer (UCN). Nucl. Acids Res. 19, 6701–6105. , , , et al. (