Translation termination depends on the sequential ribosomal entry of eRF1 and eRF3

Abstract

Translation termination requires eRF1 and eRF3 for polypeptide- and tRNA-release on stop codons. Additionally, Dbp5/DDX19 and Rli1/ABCE1 are required; however, their function in this process is currently unknown. Using a combination of in vivo and in vitro experiments, we show that they regulate a stepwise assembly of the termination complex. Rli1 and eRF3-GDP associate with the ribosome first. Subsequently, Dbp5-ATP delivers eRF1 to the stop codon and in this way prevents a premature access of eRF3. Dbp5 dissociates upon placing eRF1 through ATP-hydrolysis. This in turn enables eRF1 to contact eRF3, as the binding of Dbp5 and eRF3 to eRF1 is mutually exclusive. Defects in the Dbp5-guided eRF1 delivery lead to premature contact and premature dissociation of eRF1 and eRF3 from the ribosome and to subsequent stop codon readthrough. Thus, the stepwise Dbp5-controlled termination complex assembly is essential for regular translation termination events. Our data furthermore suggest a possible role of Dbp5/DDX19 in alternative translation termination events, such as during stress response or in developmental processes, which classifies the helicase as a potential drug target for nonsense suppression therapy to treat cancer and neurodegenerative diseases.

Figure 1.

Dbp5, but not Rli1, requires eRF1 to associate with the ribosome at a stop codon. (A) The interaction of Rli1 with the ribosome is not dependent on eRF1. Western blot analysis of Rli1 co-IPs in wild-type and sup45-2 cells shifted to 37°C for 1 h reveal co-precipitation of the small ribosomal protein uS3 ( = Rps3) and the large ribosomal protein uL29 ( = Rpl35). Detection of Zwf1 served as non-binding control. (B) The interaction of Dbp5 with ribosomal subunits is decreased in sup45-2. Western blot analysis shows co-precipitation of uS3 and uL29 with Dbp5 IPs in wild-type and sup45-2 cells, shifted to 37°C for 1 h. (C) The interaction of mutant eRF1 with the ribosome is decreased. Western blot analysis of the co-precipitation of mutant eRF1 (sup45-2) with uL23 (top) or with uS3 (bottom) is shown. Aco1 served as a negative control. (D) Quantification of at least three independent experiments, one of which is shown in panels (A–C), which determine the amount of the co-precipitated proteins, measured with the Fusion SL detection system. (E) Ribosome profiles of wild-type and the translation elongation defective strain tef2-9 reflects the translational run-off in wild-type and a translational arrest in the elongation mutant. Wild-type and tef2-9 cells were shifted to 37°C for 1 h before the lysates were analysed in linear sucrose-density gradients without cycloheximide. (F) Rli1, but not Dbp5, is bound to ribosomes during translation elongation. Western blot analysis of the fractions, representing the total cellular amount of the indicated proteins, reveals their ratio in the 80S, polysomal or non-ribosomal, 40S and 60S fractions. Asc1 as a ribosome binding protein served as a positive control. (G) Quantification of four different western blot analyses shown in panel (F); *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 2.

Rli1 supports the recruitment of Dbp5 and eRF1 to the ribosome. (A) Dbp5 interacts RNA-independently with Rli1. Immunoprecipitation of TAP-Dbp5 in the presence of RNase A shows co-precipitation of Rli1-HA in western blot analysis. Detection of eRF1 served as positive and of Por1 as negative control. (B) The interaction between Rli1 and Dbp5 is decreased in sup45-2, shifted to 37°C for 1 h. Western blot analyses of Rli1-IPs reveal less co-precipitation of Dbp5, but no reduction of the ribosomal protein uS3 in sup45-2 compared to wild-type. Cdc28 served as negative control. (C) Quantification of four different experiments shown in panel (B). (D) Inhibition of translation elongation leads to a reduced interaction between Rli1 and Dbp5. Western blot analyses of co-IPs of Dbp5 with Rli1 upon treatment with 0.5 mg/ml cycloheximide (CHX) for 30 min are shown. Hem15 was detected as non-binding control. (E) Quantification of three different experiments shown in panel (D). (F) Overexpression of RLI1 partially rescues the growth defects of sup45-2, while the wild-type growth is not influenced. Serial dilutions of the indicated strains are shown upon growth on selective plates for 3 days at 35°C. (G) Overexpression of RLI1 suppresses the binding defect of eRF1 to the ribosome in sup45-2 cells. Co-IPs of eRF1 with uS3-GFP are shown in the indicated strains with or without high copy (HC) RLI1. (H) Quantification of four different experiments shown in panel (G); *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 3.

Not eRF1 and eRF3 enter the ribosome together, but eRF1 and Dbp5. (A) The binding of Dbp5 with the ribosome-binding defective protein sup45-2 is increased, while its ribosome association is decreased. Western blot analyses of co-IPs with mutated eRF1 (sup45-2) and Dbp5 or uL29 are shown. Detection of Hem15 served as a non-binding control. (B) The interaction between eRF3 and eRF1 is decreased in the sup45-2 strain as shown in western blots of the eRF1 co-IP with eRF3. Por1 served as negative control. (C) Quantification of three different experiments shown in panels (A) and (B). (D) The interaction of eRF1 and eRF3 is decreased in a DBP5 mutant. Western blot analysis of eRF3 co-IPs with eRF1 in wild-type and the rat8-2 stain is shown. (E) Quantification of three different experiments shown in (D); *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4.

Nup159 recycles Dbp5-ATP also for translation termination. (A) Scheme of the reporter plasmids used in the dual reporter β-galactosidase luciferase assay. The lacZ gene, expressing β-galactosidase and the luc gene, expressing luciferase is either separated by the stop codon UAG or in frame. In the upper case, luciferase will only be expressed in case the stop codon is readthrough. The in-frame reporter serves as control to monitor basal expression levels and relate it to the stop codon containing construct. (B) Mutants of NUP159 show increased readthrough of the stop codon. The average readthrough activity of at least three independent experiments is shown after shift of all indicated strains to 37°C for 30 min. (C) High copy DBP5 rescues the increased stop codon readthrough of rat7ΔN. All strains were shifted to 37°C for 30 min. (D) The interaction of Dbp5 and eRF1 is disturbed in the recycling defective mutant rat7ΔN. Western blot analysis of a co-IP with Dbp5 and eRF1 is shown. Aco1 served as a negative control. (E) Quantification of three different experiments shown in panel (D). (F) The interaction of eRF1 and eRF3 and the ribosome is diminished in rat7ΔN cells. Western blot analysis of a co-IP with eRF1-GFP and eRF3 or the ribosome bound protein Asc1 is shown. (G) Quantification of three different experiments shown in panel (F). (H) Dbp5 and eRF1 directly interact in the presence of a non-hydrolysable ATP-analogue. An in vitro binding study with recombinant proteins in which GST-tagged Dbp5 or eRF3 were used in pull-down experiments in the presence of His-eRF1 and if indicated 1 mM AMP–PNP is shown in western blot analysis. GST alone served as a non-binding control; *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 5.

Dbp5 and eRF3 interaction with eRF1 is mutually exclusive. (A) The interaction site of eRF1 with Dbp5 and eRF3 overlap. Western blot analyses of pull-downs experiments with GST–Dbp5, GST–eRF3 or GST and His-eRF1 or His-eRF1∆25 lacking the last C-terminal amino acid residues are shown. All binding buffers contained 1 mM AMP–PNP. (B) A preformed complex of eRF1 and Dbp5 cannot be disrupted by eRF3. Western blot analysis of a competition assay of the indicated recombinantly expressed proteins is shown. Increasing amounts of eRF3 were added to the preformed complex of GST–Dbp5 and His-eRF1. Rli1 served as a positive control for eRF3 binding. All binding buffers for Figure 5 contained 1 mM AMP–PNP.

Figure 6.

eRF3–GDP binds Rli1 prior to the entry of Dbp5 and eRF1. (A) eRF3 is present on ribosomes stalled in translation elongation. Western blot analysis of the collected fractions of the tef2-9 gradient shown in Figure 1E with antibodies against eRF3, eRF1 and Asc1 are displayed. (B) Quantification of four different western blot analyses shown in panel (A). (C) Ribosome binding of eRF3 in sup45-2 is increased. Western blot analysis of co-IPs of eRF3 and the positive control Asc1 with uS3 (top) and uL29 (bottom) are shown. (D) Quantification of four different IPs shown in (C). (E) Rli1 binds nucleotide free eRF3 directly and releases eRF3-GTP. Western blot analysis of in vitro pull-down experiments with Rli1 is shown. (F–I) The ribosomal association of eRF1, eRF3 and Dbp5 is decreased in nup159 (F and G) or dbp5 (H and I) mutants. Western blot analyses of the uS3-co-precipitated proteins in the indicated strains are shown. (G and I) Quantification of four (G) and three (I) different IPs shown in panels (F) and (H), respectively. (J) Defects in eRF1 delivery partially suppresses the growth defects of trp5Δ. Serial dilutions of the indicated strains are shown upon growth on full medium agar plates at the indicated temperatures; *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 7.

The eIF3 complex binds to the ribosome after the Dbp5-mediated delivery of eRF1. (A) The eRF3 release factor Hcr1 does not bind to Dbp5 and only weakly to the ribosome. Co-IPs with GFP-tagged Hcr1 and as a control eRF1–GFP with Dbp5, eRF3 and the ribosomal protein Asc1 are shown. The asterisk indicates Hcr1–GFP. (B) Defects in the eRF1 delivery result in the absence of the eIF3 subunit Prt1 at the ribosome. Co-IPs of GFP-tagged Prt1 with HA-tagged Rli1, eRF1, Dbp5, eRF3 and the ribosomal protein Asc1 are shown in wild-type and sup45-2 strains. The asterisk indicates that the sup45-2 lysate lanes were exposed four times as long as the wild-type lanes. (C) The binding of the eIF3 subunits Prt1 and Nip1 are independent of the nucleotide association of Rli1. The GFP-tagged eIF3 subunits were precipitated and the co-precipitated GST-tagged Rli1 and Asc1 are shown.

Figure 8.

Stepwise entry model for translation termination. (Top) Step 1: Nucleotide-free Rli1 associates with the ribosomes as soon as the A-site is free. It binds to eRF3–GDP. Step 2: Rli1 supports the entry of Dbp5–ATP bound eRF1. Gle1/IP₆ stimulated ATP-hydrolysis of Dbp5–ATP leads to the proper positioning of eRF1 on the stop codon. Dbp5–ADP dissociates and is recycled at the nuclear pore complex by Nup159. Step 3: Dissociation of Dbp5–ADP allows the controlled interaction of eRF1 with eRF3. This in turn triggers the GTP recruitment of eRF3. Subsequent GTP hydrolysis leads to conformational changes in eRF1 allowing adjustments in its positioning in the ribosomal peptidyl-transferase center. eRF3–GDP dissociates in a complex with Hcr1, which was delivered by eIF3. Step 4: eRF3–GDP dissociation allows change of position and strong binding of Rli1–ATP that locks eRF1 in the position necessary to mediate peptidyl-tRNA hydrolysis. Step 5: Upon peptide release, ATP-hydrolysis of Rli1–ATP recycles the ribosomal subunits, which is supported by eRF1. (Bottom) Situation in which Dbp5 cannot deliver eRF1 to the ribosome that consequently results in the stop codon readthrough. Step1: Rli1 associates and binds eRF3–GDP. Step 2: eRF1 is not protected by Dbp5 and contacts eRF3 before being properly positioned, leading to premature GTP-binding, hydrolysis and the subsequent release of eRF1 and eRF3–GDP from the ribosome before the polypeptide chain and the tRNA are released. Because eRF1 had contact to eRF3 before it was placed in the optimal position, it dissociates at the same time as eRF3. Step3: A near-cognate tRNA gets access to the A-site, the stop codon is suppressed and translation elongation continues until the next stop codon is reached.