Ribosome rescue and translation termination at non-standard stop codons by ICT1 in mammalian mitochondria

Abstract

Release factors (RFs) govern the termination phase of protein synthesis. Human mitochondria harbor four different members of the class 1 RF family: RF1Lmt/mtRF1a, RF1mt, C12orf65 and ICT1. The homolog of the essential ICT1 factor is widely distributed in bacteria and organelles and has the peculiar feature in human mitochondria to be part of the ribosome as a ribosomal protein of the large subunit. The factor has been suggested to rescue stalled ribosomes in a codon-independent manner. The mechanism of action of this factor was obscure and is addressed here. Using a homologous mitochondria system of purified components, we demonstrate that the integrated ICT1 has no rescue activity. Rather, purified ICT1 binds stoichiometrically to mitochondrial ribosomes in addition to the integrated copy and functions as a general rescue factor, i.e. it releases the polypeptide from the peptidyl tRNA from ribosomes stalled at the end or in the middle of an mRNA or even from non-programmed ribosomes. The data suggest that the unusual termination at a sense codon (AGA/G) of the oxidative-phosphorylation enzymes CO1 and ND6 is also performed by ICT1 challenging a previous model, according to which RF1Lmt/mtRF1a is responsible for the translation termination at non-standard stop codons. We also demonstrate by mutational analyses that the unique insertion sequence present in the N-terminal domain of ICT1 is essential for peptide release rather than for ribosome binding. The function of RF1mt, another member of the class1 RFs in mammalian mitochondria, was also examined and is discussed.

Figure 1. ICT1 shows codon-independent peptide-release activity on 55S mitoribosomes.

(A) ICT1 is integrated within the 39S large mitoribosomal subunit in a 1∶1 stoichiometry. The amounts of 55S-integrated ICT1 on 55S ribosomes were estimated by Western blotting, using the purified ICT1 protein as a standard (Std. ICT1; upper panel). 55S mitoribosomes (lower left panel) or mitoribosomal subunits (lower right panel) were separated on 15%–30% (w/v) sucrose gradients (containing 10 mM Tris-HCl [pH 7.4], 80 mM NH₄Cl, 8.2 mM MgSO₄, and 1 mM DTT, for 55S; 10 mM Tris-HCl [pH 7.4], 200 mM KCl, 2 mM MgCl₂, 2 mM GDP and 1 mM DTT, for subunits), and the fractions were analyzed by Western blotting with an ICT1 antibody. IB, immunoblotting. (B) Exogenous ICT1 exhibits codon-independent peptide-release activity on 55S mitoribosomes. The ac[³H]Phe release activities of ICT1, RF1Lmt and RF1mt on 55S ribosomes were tested in the presence of mRNAs and ac[³H]Phe-tRNA^Phe. Ribosomes were programed by mRNAs encoding MF-UAA (stop) or MFV (sense). The results were evaluated relative to the 100% value, when all of the ac^[3H]Phe-tRNA^Phe bound to ribosome was hydrolyzed; 100% value corresponds to 0.9 pmol of ac[³H]Phe. Results represent the average of at least three independent experiments. The error bars indicate standard deviation (SD). (C) ICT1 shows peptide release activity on 55S mitoribosomes in the absence of mRNA. Peptide-release assays on 55S ribosomes were performed as in (B), but in the absence of mRNA. 55S [+] and 55S [−] indicates the assays in the presence and absence of 55S mitoribosomes, respectively. Results represent the average of at least three independent experiments. The bars on the graph indicate SD.

Figure 2. ICT1 can function in the middle of the mRNA.

(A) ICT1 significantly functions with stop (UAA) and stall (AGA) mRNAs. Multi-round translation assays with E. coli 70S ribosomes were performed with the indicated peptide release factors, using mRNAs depicted in the upper. Each mRNAs differs at the 3′-terminus. Nonstop, no stop codon and no 3′UTR; stop, UAA stop codon followed by 3′UTR; sense, AGA sense codon followed by 3′UTR. The ribosomes stall at the AGA codon, since only phenylalanyl-tRNA synthetase and leucyl-tRNA synthetase are present as aminoacyl-tRNA synthetase sources. The amounts of synthesized peptide were evaluated with nonstop (upper right), stop (lower left), and stall (lower right). Also note that ICT1 functions more efficiently when the A-site is vacant (compare ICT1[WT] among nonstop, stop and sense). The results represent the average of at least three independent experiments. The bars on the graph indicate SD. For details, see the text. (B) ICT1 exhibits peptide-release activity on the polysome similar to that of puromycin. The polysome breakdown assay was performed as described previously . E. coli polysomes (2.0 A₂₆₀) were incubated with ribosome recycling factors (RRFmt, 15 µg; EF-G2mt, 30 µg) in the presence of the indicated peptide release factors (RF1Lmt or RF1mt, 60 µg; ICT1, 50 µg; puromycin, 10 µM), and subjected to sucrose gradient centrifugation.

Figure 3. Structural comparison of ICT1 with the classical class 1 release factors.

(A) Comparison of the domain-structures of Thermus thermophilus RF2, E. coli ArfB/YaeJ, and mouse ICT1. For Thermus thermophilus RF2, the catalytic domain 3 is indicated in blue, the stop codon recognition domain (consisting of domains 2 and 4) is shown in orange, and domain 1 is colored red. The switch loop is highlighted in green. For ArfB/YaeJ and ICT1, the N-terminal globular domain is shown in blue. (B) Comparison of 3D-structures of Thermus thermophilus RF2 (PDB code 2WH1), E. coli ArfB/YaeJ (PDB code 4DH9), and mouse ICT1 (PDB code 1J26). Each domain of RF2, ArfB and ICT1 is shown in the same colors as in (A). The structure of the C-terminal tail of ICT1 has not been determined, and is not shown. The positions of the GGQ-loop and the insertion sequence are indicated. The sketches are the mirror pictures, so that the switch loop of RF2 or the insertion sequence of ICT1 is depicted in front, and the E-, P- and A-site are from left to right on the ribosome. (C) Amino acid sequence alignment of various class 1 release factors. Domain 3 of E. coli RF1, E. coli RF2, human RF1Lmt and human RF1mt were compared with E. coli ArfB (full length) and human ICT1 (amino acid residue position of 61th–206th), using the BoxShade program. The GGQ motif, the unique insertion sequence within ArfB and ICT1, and the C-terminal extension are underlined. The basic residues in the insertion sequence of ICT1 are highlighted in red. (D) The positions of the basic residues in the insertion sequence of ICT1. For the ICT1[α2] mutant, K124, K126 and R129 were substituted with alanine simultaneously. The sketch is the mirror picture, so that it is accorded with Fig. 3B.

Figure 4. The peptide-release activity of ICT1 requires the insertion sequence in the N-terminal globular domain.

(A) Peptide-release assays on 55S ribosomes without mRNA were performed as in Figure 1C, with the indicated peptide release factors. ICT1[GSQ], GGQ motif is mutated to GSQ; ICT1[ΔC], C-terminal 14 amino acid residues are truncated. 55S [+] and 55S [-] indicates the assays in the presence and absence of 55S mitoribosomes, respectively. Results represent the average of at least three independent experiments. The bars on the graph indicate SD. (B) Multi-round translation assays with E. coli 70S ribosomes were performed with the indicated peptide release factors, using the mRNAs depicted on the left. The utilized mRNA has no stop codon and no 3′UTR at the 3′-terminus. The amounts of synthesized peptide were evaluated. Red closed squares, ICT1[WT]; red open squares, ICT1[GSQ]; green closed squares, ICT1[ΔC]; green open squares, ICT1[α2]. Results represent the average of at least three independent experiments. The bars on the graph indicate SD. For details, see the text.

Figure 5. The insertion sequence in the N-terminal globular domain of ICT1 does not affect ribosome binding.

(A) 55S mitoribosomes (upper panel) or E. coli 70S ribosomes (lower panel) were mixed with the indicated ICT1 proteins, and fractionated on 15%–30% (w/v) sucrose gradients. The fractions were analyzed by Western blotting with an ICT1 antibody. For the binding to 55S mitoribosomes, His-tagged ICT1 proteins were used. The arrows indicate the exogenous His-tagged ICT1 and endogenous ICT1. Intact His-tagged ICT1 proteins are hardly observed in the top fractions. Proteins indicated with asterisks are proteolytic products of His-tagged ICT1, rather than ICT1 that has been chased from 55S ribosomes. (B) The indicated amount of ICT1 proteins (open circles, ICT1[WT]; closed circles, ICT1[α2]) were incubated with E. coli 70S ribosomes (f.c. 0.25 µM). The ribosome-bound ICT1 was recovered by a filtering technique, and quantified by Western blotting against ICT1. The amount of ribosome-bound ICT1 per 70S ribosome was plotted against the amount of ICT1. Error bars represent SD of repeated measurements. (C) 55S ribosomes were mixed with the indicated ICT1 proteins and incubated in the presence of BS³ (f.c. 4 mM). The cross-linked products were analyzed by Western blotting with an ICT1 antibody. Note that the cross-linked proteins with the integral ICT1 (single asterisk) are different from those with the exogenous ICT1 (two asterisks, for example).

Figure 6. A model for the functions of ICT1 depicting the three scenarios involving the factor's activity.

Ribosomes might stall for many reasons such as non-stop mRNA (left, Stalled ribosome at the end of mRNA), the defect in the membrane insertion of the nascent chain, the non-standard stop codon (middle, Stalled ribosome in the middle of mRNA), and so on. Stalled ribosome might further lose the mRNA (right, Stalled ribosome without mRNA). Defective initiation complex might exist (middle panel, stalled ribosome in the middle of mRNA; right, stalled ribosome without mRNA). ICT1 is a versatile rescue factor, which can take care of all type of stalled ribosomes by exerting its peptide-release activity at ribosomal A-site. Ribosome-free ICT1 is not yet detected in mitochondria. Accordingly, the ribosome-integrated ICT1, which is not present at A-site, is potentially released from the ribosomes in response to the ribosomal stall, and function at the ribosomal A-site. Alternatively, It is also possible that ICT1 is overexpressed in response to a ribosome stalling in order to produce ribosome-free ICT1, which acts on the stalled ribosome independently of the integrated ICT1.