Reconstitution of mammalian mitochondrial translation system capable of correct initiation and long polypeptide synthesis from leaderless mRNA

Abstract

Mammalian mitochondria have their own dedicated protein synthesis system, which produces 13 essential subunits of the oxidative phosphorylation complexes. We have reconstituted an in vitro translation system from mammalian mitochondria, utilizing purified recombinant mitochondrial translation factors, 55S ribosomes from pig liver mitochondria, and a tRNA mixture from either Escherichia coli or yeast. The system is capable of translating leaderless mRNAs encoding model proteins (DHFR and nanoLuciferase) or some mtDNA-encoded proteins. We show that a leaderless mRNA, encoding nanoLuciferase, is faithfully initiated without the need for any auxiliary factors other than IF-2mt and IF-3mt. We found that the ribosome-dependent GTPase activities of both the translocase EF-G1mt and the recycling factor EF-G2mt are insensitive to fusidic acid (FA), the translation inhibitor that targets bacterial EF-G homologs, and consequently the system is resistant to FA. Moreover, we demonstrate that a polyproline sequence in the protein causes 55S mitochondrial ribosome stalling, yielding ribosome nascent chain complexes. Analyses of the effects of the Mg concentration on the polyproline-mediated ribosome stalling suggested the unique regulation of peptide elongation by the mitoribosome. This system will be useful for analyzing the mechanism of translation initiation, and the interactions between the nascent peptide chain and the mitochondrial ribosome.

Figure 1.

Components used in the reconstituted mammalian mitochondrial translation system. (A) Translation factors. Initiation factors (IF-2mt and IF-3mt), elongation factors (EF-Tumt, EF-Tsmt and EF-G1mt) and termination and recycling factors (RF-1Lmt, EF-G2mt and RRFmt), 2 μg each, were resolved by 12% SDS-PAGE and stained with Coomassie Brilliant Blue (CBB). (B) Ribosomes. 55S ribosomes were separated on 6–38% (w/v) sucrose gradients while monitoring the absorbance at 260 nm.

Figure 2.

Synthesis of mtDNA-encoded proteins by the reconstituted mammalian mitochondrial translation system. (A–C) Translation reactions (coupled transcription, aminoacylation, and translation reactions) were performed using the indicated template DNA and E. coli tRNA mix in the presence of [³⁵S] methionine. After the 120 min reaction, the samples were treated with RNase A and fractionated by 15% SDS-PAGE. The schematic of the mRNA produced from the template DNA during the reaction is shown below the panel. (A) Reactions with the standard template DNA; (B) reactions with the control template DNA, in which the initiation AUG codon in the standard template DNA is mutated to ACG and (C) reactions with the control template DNA, in which the λcI sequence in the standard template DNA is deleted. Full gel images are shown in Supplementary Figure S2. Another set of experiment corresponding to (C) is shown in Supplementary Figure S2C. White arrowheads indicate the relevant translation products, and asterisks denote unknown translation products. DHFR, E. coli dihydrofolate reductase; ND1–6 and ND4L, NADH dehydrogenase subunits; ATP6 and ATP8, ATP synthase subunits; CO1–3, cytochrome c oxidase subunits; CytB, cytochrome b. Molecular weights of DHFR and the mtDNA-encoded proteins, without the λcI sequence (kDa): DHFR(18), ATP6(24.8), ATP8(8), CO3(30), CO2(25.6), CO1(57), CytB(42.7), ND1(36), ND2(39), ND3(13), ND4L(11), ND4(52), ND5(67), ND6(18).

Figure 3.

Dependence of the reconstituted mammalian mitochondrial translation system on each component. (A) nLuc synthesis: The schematic of the mRNA used in the reaction (upper). Translation reactions were performed using the mRNA and [³⁵S]methionine-labeled yeast aminoacyl-tRNA mix, with each factor subtracted from the reaction mixture. After the 120 min reaction, either the samples were treated with RNase A and resolved by Tricine SDS-PAGE (lower left), or aliquots were subjected to the nLuc assay (lower right). The amount of nLuc synthesized in ‘Complete’ corresponds to approximately 0.026 nM. Error bars represent the standard deviation from three independent experiments. In the rightmost lane of the gel, translation product of 3XFLAG mRNA in (B) is also shown, to compare with 2–5 kDa non-specific products of nLuc mRNA (see also Supplementary Figure S4). (B) 3xFLAG peptide synthesis: The schematic of the mRNA used in the reaction (upper). Translation reactions were performed as in (A). After the 120 min translation reaction, the samples were treated with RNase A and resolved by Tricine SDS-PAGE (lower left). The yields of the synthesized 3xFLAG peptides were assessed by the incorporation of [³⁵S]methionine into the peptide, and the amounts of the synthesized peptides at 120 min were plotted (lower right). The amount of 3xFLAG peptide synthesized in ‘Complete’ corresponds to ∼0.039 nM. Error bars represent the standard deviation from three independent experiments.

Figure 4.

Fusidic acid (FA) resistance of the reconstituted mammalian mitochondrial translation system. (A) Effect of FA on ribosome-dependent GTPase activity. E. coli EF-G/EF-G1mt/EF-G2mt (final concentration 0.5 μM) were incubated with ribosomes (final concentration 0.2 μM) and [γ-³²P]GTP in the presence of the indicated amounts of FA, and the [³²P]Pi release was measured. Left, assay with E. coli 70S ribosomes; right, assay with mitochondrial 55S ribosomes. The amounts of Pi released at 0 mM FA correspond to 1,400 pmol (E. coli EF-G), 650 pmol (EF-G1mt), and 170 pmol (EF-G2mt) on 70S ribosomes, and 4,300 pmol (EF-G1mt) and 1,600 pmol (EF-G2mt) on 55S ribosomes. Note that the GTPase activity of E. coli EF-G is not detectable on 55S ribosomes. (B) Effect of FA on 3xFLAG peptide synthesis (upper). 3xFLAG peptides were synthesized as in Figure 3B, in the presence of the indicated amounts of FA. After the 120 min translation reaction, the samples were treated with RNase A and resolved by Tricine SDS-PAGE (upper left), and the amounts of the synthesized 3xFLAG peptides were plotted (upper right). Error bars represent the standard deviation from three independent experiments. Effect of FA on the nLuc synthesis with E. coli translation system (PURE frex ver. 2.0) was similarly analyzed for reference (lower). The standard deviation from three independent experiments do not exceed 0.3%.

Figure 5.

Polyproline-mediated ribosome stalling in the reconstituted mammalian mitochondrial translation system. (A) The schematic of the mRNA used in the reaction. A polyproline motif (Pro x4 or Pro x12) was inserted in front of the nLuc coding region. (B) Translation reactions were performed using the indicated mRNA and [³⁵S]methionine-labeled yeast aminoacyl-tRNA mix, with 7 mM Mg²⁺. After the translation reaction was performed for the indicated time period, aliquots were subjected to the nLuc assay (left). The nLuc activities at the 120 min reaction point were plotted (right). Error bars represent the standard deviation from three independent experiments. (C) Translation reactions were performed as in (B). After the 120 min reaction, the samples were treated with RNase A and subjected to Tricine SDS-PAGE. Black arrowhead, full-length translation products; white arrowhead, stall products. (D) Translation reactions were performed as in (B), but with the indicated Mg²⁺ concentrations. The nLuc activities at 120 min were plotted. Error bars represent the standard deviation from three independent experiments.