Accuracy mechanism of eukaryotic ribosome translocation

Translation of the genetic code into proteins is realized through repetitions of synchronous translocation of messenger RNA (mRNA) and transfer RNAs (tRNA) through the ribosome. In eukaryotes translocation is ensured by elongation factor 2 (eEF2), which catalyses the process and actively contributes to its accuracy1. Although numerous studies point to critical roles for both the conserved eukaryotic posttranslational modification diphthamide in eEF2 and tRNA modifications in supporting the accuracy of translocation, detailed molecular mechanisms describing their specific functions are poorly understood. Here we report a high-resolution X-ray structure of the eukaryotic 80S ribosome in a translocation-intermediate state containing mRNA, naturally modified eEF2 and tRNAs. The crystal structure reveals a network of stabilization of codon–anticodon interactions involving diphthamide1 and the hypermodified nucleoside wybutosine at position 37 of phenylalanine tRNA, which is also known to enhance translation accuracy2. The model demonstrates how the decoding centre releases a codon–anticodon duplex, allowing its movement on the ribosome, and emphasizes the function of eEF2 as a ‘pawl’ defining the directionality of translocation3. This model suggests how eukaryote-specific elements of the 80S ribosome, eEF2 and tRNAs undergo large-scale molecular reorganizations to ensure maintenance of the mRNA reading frame during the complex process of translocation.

Translation of the genetic code into proteins is realized through repetitions of synchronous translocation of messenger RNA (mRNA) and transfer RNAs (tRNA) through the ribosome. In eukaryotes translocation is ensured by elongation factor 2 (eEF2), which catalyses the process and actively contributes to its accuracy 1 . Although numerous studies point to critical roles for both the conserved eukaryotic posttranslational modification diphthamide in eEF2 and tRNA modifications in supporting the accuracy of translocation, detailed molecular mechanisms describing their specific functions are poorly understood. Here we report a high-resolution X-ray structure of the eukaryotic 80S ribosome in a translocation-intermediate state containing mRNA, naturally modified eEF2 and tRNAs. The crystal structure reveals a network of stabilization of codon-anticodon interactions involving diphthamide 1 and the hypermodified nucleoside wybutosine at position 37 of phenylalanine tRNA, which is also known to enhance translation accuracy 2 . The model demonstrates how the decoding centre releases a codon-anticodon duplex, allowing its movement on the ribosome, and emphasizes the function of eEF2 as a 'pawl' defining the directionality of translocation 3 . This model suggests how eukaryote-specific elements of the 80S ribosome, eEF2 and tRNAs undergo large-scale molecular reorganizations to ensure maintenance of the mRNA reading frame during the complex process of translocation.
In eukaryotes, the complex process of translocation is ensured by eEF2, a GTPase that is indispensable for maintaining the correct mRNA reading frame. Many genetic and biochemical studies point to a critical role of the unique eEF2 post-translational modification diphthamide, which is located in domain IV and is conserved among eukaryotes and archaea. Organisms lacking diphthamide have reduced protein synthesis rates and increased occurrence of (−1) frameshifting 1,4 . Diphthamide is a target of several virulent toxins that inactivate eEF2 by ADP ribosylation and cause lethal effects 5 .
The current structural knowledge about dynamics of the eukaryotic elongation cycle has been provided by low-to-intermediate resolution cryo-electron microscopy (cryo-EM) reconstructions [6][7][8][9][10] . Although several late steps of translocation have been described, the level of detail achieved in these studies is insufficient to suggest a precise mechanism explaining fidelity of translocation. At the same time, understanding of the principles of the fidelity has become particularly crucial during the current viral pandemic, because programmed mRNA frameshifting is at the heart of the SARS-CoV-2 replication cycle 11 .
Here we present the crystal structure of eukaryotic 80S ribosomes from Saccharomyces cerevisiae trapped in intermediate translocation state. The X-ray crystal structure provides a detailed mechanism of tRNA translocation from A-to P-sites and highlights the specific role of eEF2 in the movement of the tRNA-mRNA module during the process. The crystal structure presented here uncovers the precise role of diphthamide and wybutosine, a heavily modified nucleoside at position 37 of eukaryotic phenylalanine tRNA, in stabilization of the codon-anticodon interactions during translocation and demonstrates how eukaryote-specific elements of the 80S ribosome, eEF2 and tRNA rearrange to ensure maintenance of the mRNA reading frame.

Architecture of the translocation complex
We determined the structure of the S. cerevisiae 80S ribosome translocation complex trapped in intermediate state by X-ray crystallography at 3.2 Å resolution ( Fig. 1a, b, Extended Data Fig. 1, Extended Data Table 1). It consists of S. cerevisiae 80S ribosomes bound with native S. cerevisiae eEF2, the nonhydrolyzable GTP analogue GMPPCP, mRNA and two tRNAs, and was determined in the absence of antibiotics, which customarily used for stabilization, suggesting that the model represents a bona fide state of the 80S ribosome.
The crystal structure of the ribosome complex represents an intermediate translocation state that has not been described before with two tRNAs trapped in the chimeric hybrid ap/P and pe/E transitory positions. In this state the small subunit (SSU) head has swivelled 13.5° and the SSU body has undergone 9° anticlockwise rotation relative to the large subunit (LSU) (Fig. 1a). The anticodon stem-loop (ASL) of the A-site tRNA is captured half-translocated between the A-and P-sites of SSU (12.4 Å out of a fully translocated distance of 24.1 Å), and the tRNA acceptor end contacts the P-loop of the peptidyl-transferase centre of LSU forming an ap/P chimeric hybrid state (Fig. 1a, b, Extended Data Fig. 2a, b).
The overall conformation of eEF2 resembles that of eEF2 in the 80S translocation intermediate-post-translocation (TI-POST)-1 state of a medium-resolution cryo-EM reconstruction 6 that depicts a late-translocation event with domain IV of eEF2 protruding approximately 7 Å deeper into the A-site of SSU from the perspective of the SSU body (Extended Data Fig. 3). However-in contrast to the TI-POST-1 state-in our complex, domain IV of eEF2 is engaged in an extensive stabilization network with the codon-anticodon duplex and with the decoding centre area (Fig. 1b-d, Extended Data Fig. 1d). The intricate network of arginine-rich regions of domain II and III of eEF2 with helix 5 of 18S rRNA is formed owing to strong rotation of the SSU body (9.85°). These contacts are not present in the less rotated (4.26°) TI-POST-1 or unrotated TI-POST-3 states of 80S from the cryo-EM models 6 (Extended Data Fig. 4a, b). This comparison suggests that eEF2 uncouples from the SSU body at the later steps of translocation where it is associated only with LSU and remains stationary relative to the rest of the ribosome. Collectively, these structural data corroborate an existing hypothesis describing a translocase (eEF2 or EF-G) as a 'locking pawl' that decouples tRNA-mRNA from the SSU head and body during back-rotation 3 .
Comparison of the current early-intermediate state with late-translocation states described by cryo-EM 6 reveals that the uS12 protein remains attached to domain III of eEF2 during SSU back-ratcheting. Located on the SSU shoulder, uS12 is implicated in translocation 12 and codon-anticodon duplex stabilization in the decoding centre 13 . It is plausible that during translocation, uS12 transmits SSU body back-rotation by pulling domain III of eEF2 that in turn retransmits it to switch II in the G-domain, stimulating GTP hydrolysis (Extended Data Fig. 4c, d). We found additional stabilization of domain IV by the N terminus of the eukaryote-specific protein eS30, which also interacts extensively with conserved decoding protein uS12 (Extended Data Fig. 4c). It can be assumed that eS30 co-evolved with eEF2, whose domain IV has 65 additional amino acids compared with its bacterial counterpart EF-G, to provide supplementary stabilization as well as enhancing propagation of conformational changes at the decoding site 13,14 . Other interactions between eEF2, tRNAs and the 80S ribosome are presented in Extended Data Figs. 5, 6.

Diphthamide
The current crystal structure of 80S ribosome translocation intermediate reveals direct involvement of diphthamide in stabilization of a codonanticodon duplex in transition from the A-to P-site. eEF2 interacts with the mRNA codon exclusively via diphthamide, which protrudes into a cleft formed by mRNA, ap/P tRNA and rRNA (Fig. 1c, d). Assisted by His583 and Asp696 of eEF2, diphthamide contacts the codon-anticodon duplex minor groove. A similar pattern of interactions has been extensively described between the bacterial decoding centre mould and the codon-anticodon duplex in a classical unrotated state 13,14 (Fig. 1d). The closest resemblance is observed at the second base pair (BP2, Fig. 1d), where diphthamide together with Asp696 mimics stabilization of decoding nucleotides G577 and A1755 in 18S rRNA (bacterial G530 and A1492). Fixation of the first codon-anticodon pair is divided between diphthamide with His583 and His694 contacting the anticodon ribose, and the wybutosine modification of ap/P tRNA Phe of nucleotide 37 stacking on codon position +4 (BP1, Figs. 1d, 2a). The third codon-anticodon pair (BP3, Fig. 1d) is anchored by diphthamide only through its trimethylammonio moiety stabilizing the codon position. These findings are corroborated by numerous studies describing increased −1 frameshifting slippage when translation is performed with eEF2 lacking diphthamide 1,15 .
The mRNA backbone of the UUC codon paired to ap/P tRNA Phe is located in close proximity to the sugar-phosphate backbone of decoding nucleotides A1755-A1756 and G1757, whose backbone is bulged out because of partial 'flipped-in' and 'flipped-out' positions of their nucleoside moieties (Fig. 2b). This close positioning of mRNA and A1756-G1757 backbones is realized by ribose-phosphate zipper interactions 16 . The diphthamide trimethylammonio moiety additionally stabilizes negatively charged backbones by interaction with the phosphate group between A1756 and G1757 of 18S rRNA (BP3, Fig. 1d).
The difference in structure of helix 44 of SSU rRNA (h44), which forms a part of the decoding centre between bacteria and both eukaryotes and archaea might explain emergence of the diphthamide modification during evolution (Extended Data Fig. 7). The changes of decoding centre bulge, and flanking nucleotides could lead to increased flexibility of the eukaryotic decoding centre, which would require additional stabilization that was achieved by development of diphthamide modification of eEF2. In addition, the unique trimethylammonio moiety could have been evolutionally refined to reduce repulsion between juxtaposing negatively charged backbones of mRNA and the h44 decoding-centre loop at the early stages of translocation.

Unlocking of the decoding centre
In the early-intermediate translocation state, when diphthamide takes over stabilization of the codon-anticodon duplex, decoding nucleotides rearrange to initiate resetting of the decoding centre. Nucleotide A1755 of 18S rRNA adopts a flipped-in position, whereas A1756 is in a more flipped-out and somewhat flexible conformation. There is a noticeable  rearrangement of the pivotal intersubunit bridge B2a that is partially composed of A2256 of LSU helix 69 (H69) and decoding nucleotides of h44 (A1755-A1756) of SSU (Fig. 2c). In the present structure, adenosine 2256 of H69 protrudes towards sugar phosphate backbone between A1755-A1756, and contacts A1756 phosphate and A1755 ribose moieties. This connection allows H69 to transmit its movement directly to decoding adenines. During selection of tRNA on the ribosome, the decoding centre serves as a mould imposing restraints on codon-anticodon nucleotides via defined interactions 13,14 (Fig. 1d, right). To translocate mRNA by one codon, the decoding centre has to unlock from the bound codon-anticodon duplex. The present crystal structure with eEF2 suggests that destabilization of the decoding centre is initiated by anticlockwise rotation of SSU, which leads to rearrangement of the B2a bridge. The latterrearrangement induces displacement of H69 towards decoding h44, resulting in a new contact of A2256 of H69 with the decoding nucleotides A1755-A1756. Further rearrangements of the H69 tip lead to its synchronized movement with the decoding adenosines and triggers partial unlocking of the decoding centre from mRNA and tRNA (Fig. 2c). The trimethylammonio moiety of diphthamide contributes to these changes of A1755-A1756 and prevents decoding adenines from re-establishing their contacts with the codon-anticodon duplex (Fig. 1c, d).
Diphthamide-induced unlocking catalyses translocation by reducing the energy required for movement of the codon-anticodon pair. This function is supported by biochemical studies, which have demonstrated decreased protein synthesis rates in organisms lacking diphthamide 4,15 . Such an interpretation is also consistent with results of pre-steady-state kinetics studies on the bacterial translocation system 17 and may explain the inhibitory effects of the antibiotics paromomycin and viomycin that interfere with the resetting of the decoding nucleotides by binding to the decoding centre region of h44 and reduce the rate of translocation by about 160-and more than 10,000-fold, respectively [18][19][20] .

Wybutosine
In the current structure, we observe clear density for the tRNA Phe hyper-modification wybutosine in position 37 in an authentic binding state on the ribosome. Wybutosine consists of wyosine base with 4-methoxy-3-[(methoxycarbonyl)amino]−4-oxobutyl group (group R; Fig. 2a, Extended Data Fig. 2c). Wyosine cross-strand stacks with the first codon nucleoside paired to ap/P tRNA. Group R stretches towards the third nucleotide (G3) of the codon coupled to the anticodon of pe/E tRNA and forms two hydrogen bonds between its methylcarboxyl portion and G3, hence, upgrading the triplet codon-anticodon interaction with ap/ P tRNA to a quadruplet interaction (Fig. 2a). To our knowledge, this is the first time a tRNA modification has been seen to directly influence two adjoint codon-anticodon pairs by interactions with mRNA. Interactions between group R of wybutosine and the third position of the adjacent codon coupled to pe/E tRNA are possible because of a closer distance between ap/P tRNA and pe/E tRNA (Extended Data Fig. 2a). This situation is not achievable with tRNAs in classical A/A and P/P states and during the early stage of translocation where both tRNAs are positioned more than 10 Å apart, indicating that the main stabilization role of wybutosine modification is strongly manifested in the intermediate translocation states. This also implies that the (−1) ribosomal frameshifting events with hypomodified wybutosine tRNA are likely to occur during intermediate and late steps of translocation. These findings substantiate previous studies explaining why lack or alteration of wybutosine derivatives in tRNA Phe increases the (−1) programmed ribosomal frameshifting frequency and is associated with poor survival in patients with cancer 2,21 , and also demonstrates how the presence of wybutosine influences the efficiency of viral ribosomal frameshifting 22 .

Conclusion
The unique state of the translocating eukaryotic ribosome described here demonstrates eEF2 functioning as a pawl 3 during translocation (Figs. 2d, 3, Extended Data Fig. 4a, b). Comparison of the early translocation-intermediate complex described here with the late translocation-intermediate TI-POST-1 6 shows that eEF2 is sturdily anchored on LSU via domains I and V (Fig. 2e, Supplementary Video) with tRNAs retaining very similar positions in both states of the 80S ribosome, while SSU head and body move around them. The comparison suggests that domain IV of eEF2 uncouples tRNA-mRNA from the SSU body and head allowing these domains to return to their pretranslocation positions without pulling tRNA and mRNA back with them. Interestingly, an increase in the head swivel from around 13.5° (in the early intermediate) to about 18° (in TI-POST-1) is a result of the body back-rotation that the head cannot follow because it remains attached to tRNAs (Fig. 2e).

Article
This probably creates a strain in the SSU neck region that eventually forces the head to unbind from tRNAs and to rotate back (Supplementary Video). Thus, the last steps of translocation are achieved by back retracement of the head to a non-rotated state and rebinding of tRNA-mRNA (one codon further down) to the SSU P-site environment. The translocation process is enabled by Brownian intersubunit rotations of the ribosome 23,24 . The GTP-bound state of eEF2 on the ribosome stays in rigid conformation, as it is observed in current structure of early-translocation intermediate as well as in late-translocation intermediates 6 (Fig. 2d, e), thus serving as a pawl that ensures directionality of translocation process. The described model supports findings showing that the energy stored in the eEF2•GTP state is sufficient to promote translocation 25 and suggests that hydrolysis of GTP does not occur until the late steps of the process. It has been shown previously that domain IV of eEF2•GTP state can reach a codon-anticodon duplex of the P-site tRNA of the non-rotated ribosome 6 . However, experiments using fluorescence resonance energy transfer and other techniques 6,26 have reported that translocation induced by eEF2 with a non-hydrolyzable analogue of GTP is prone to reversion, demonstrating a critical role of GTP hydrolysis in promoting unidirectionality of translocation. GTP hydrolysis and inorganic phosphate release are most likely to occur during the late steps of translocation during SSU body back-ratcheting, when hydrolysis of GTP is stimulated by the movement of uS12-eEF2 domain III (Extended Data Fig. 4c, d). Conformational changes of eEF2 induced by GTP hydrolysis enable unbinding from the mRNA-tRNA module in a manner that prevents pulling the codon-anticodon duplex back to the A-site. Similarly, the bacterial homologue of eEF2 undergoes a large rotation in domain III before its dissociation from the ribosome 27 .

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-021-04131-9.

80S ribosome purification
Purification of the 80S ribosomes from the JD1370-∆Stm1 yeast strain 7 was carried out according to the previously described protocol 29 , with some modifications. The crude ribosomes obtained by precipitation with 8.5% PEG 20,000 were re-suspended in buffer M (30 mM Hepes-KOH, 10 mM MgCl 2 , 50 mM KCl, 8.5% mannitol, 0.5 mM EDTA-KOH, 2 mM DTT, pH 7.55) and MgCl 2 and KCl concentrations were slowly adjusted to 10 and 500 mM (10/500), respectively. The ribosomal suspension was then incubated on ice for 35 min with mild vortexing. The ribosomes were applied on the 6% sucrose cushion, which was prepared in the same dissociation conditions (10/500) and layered over the 10-30% sucrose gradient as in ref. 7 with 5 mM spermidine (5 mg of ribosomes per SW28 tube). After the overnight centrifugation selected fractions of the 80S ribosomes were collected and the ribosomes were precipitated by PEG 20,000.

Purification of native eEF2
The isolation procedure of native eEF2 was mainly based on the previously described protocol 30 , with changes in several steps. First, we used a fresh culture of yeast strain JD1370-∆Stm1 grown to an A 600 of 5-6 and cells were lysed in a microfluidizer. Second, instead of S-Sepharose, source-Q and uno-Q ion-exchange columns we used SP-Sepharose, Q-Sepharose and introduced a gel filtration with Sephadex-200 as the final purification step. The final sample was stored in pH 7.5 buffer consisting of 20 mM Tris-HCl, 5 mM MgCl 2 , 100 mM KCl, 10% glycerol and 1 mM DTT.
Purification and aminoacylation of tRNAs S. cerevisiae tRNA Phe ('chemical block') was aminoacylated according to the protocol as described 31 with minor modifications. After three rounds of phenol-chloroform-isoamyl alcohol extraction S. cerevisiae Phe-tRNA Phe was purified on a column DeltaPack, C4 300A, 5mkm, 3.9 × 150 mm HPLC column (Waters) using a ethanol gradient as described 32 . The final sample was stored in 20 mM NH 4 CH 3 CO 2 pH 5.0 at −80 °C. Escherichia coli tRNA fmet was prepared and then aminoacylated and formylated according to 33 . After phenol extraction, fMet-tRNA fmet was purified by hydrophobic chromatography using TSK-gel Phenyl-5PW column, and the final sample was stored in 20 mM NH 4 CH 3 CO 2 pH 5.0 at −80 °C.

Data collection and structure refinement
Data collection was carried out at 100 K at beamline PX1 -X06SA at the Swiss Light Source synchrotron at 1.0 Å wavelength using DA+ data acquisition and analysis software 34 . Data were integrated and scaled with the XDS program 35 . The search model was generated from the previously published structure of the yeast 80S ribosome 29 (PDB ID 4V88). The initial molecular replacement solution was refined in PHENIX by rigid-body refinement with the 40S and 60S subunits treated as rigid bodies. After initial refinement, density corresponding to the mRNA, tRNAs, eEF2 as well as ribosome rearrangements became clearly visible in the difference electron density map. The crystal structure of eEF2 (PDB ID 1N0U) was docked into the density and manually adjusted before refinement. Refinement was carried out in alternating cycles of automated refinements using PHENIX with manual refinement and model building in COOT resulting in a model with Ramachandran favoured, allowed and outliers of 83.4%, 14.4% and 1.1% respectively. A summary of refinement and data collection statistics is displayed in Extended Data Table 1. All figures were generated using PyMOL.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.

Data availability
Coordinates and structure factors have been deposited with the Protein Data Bank under accession code 7OSM. Previously published models that were used for analysis and comparison are also available in the Protein Data Bank with accession codes 4V88, 4V6F, 1N0U, 3J77, 3J78, 4V6F, 6GZ3 and 6GZ5. In this state, the U1191-C1637 interaction is broken, however, U1191 remains in contact with C35 of pe/E tRNA. A similar situation was described for the bacterial ribosome with the pe/E tRNA in a structure modelling spontaneous translocation without EF-G 37 . Compared to the intermediate eukaryotic state reported here, the bacterial ap/P tRNA coupled to mRNA is shifted towards the P-site indicating translational reading frame slippage in the factor-free system. This comparison shows the importance of role that translocase eEF2 (or EF-G) fulfils in the coordinated movement of tRNA-mRNA during translocation.

Corresponding author(s): Marat Yusupov, Gulnara Yusupova
Last updated by author(s): Oct 4, 2021 Reporting Summary Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy The atomic coordinates and structure factors for the determined crystal structure have been deposited in the Protein Data Bank with the accession code 7OSM (translocation intermediate complex). Previously published models that were used for analysis and comparison are also available from the Protein Data Bank with accession codes 4V88, 4V6F, 1N0U, 3J77, 3J78, 4V6F, 6GZ3, 6GZ5.

March 2021
Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative.

Sample size
The data set was complete and with a high redundancy as can be seen from the data statistics in Table 1 in accordance with generally accepted crystallographic methods.
Data exclusions No data were excluded from analysis. During data processing heterogeneous crystals which did not scale correctly with the main dataset were removed from further consideration.

Replication
All biochemical and crystallization procedures were successfully replicated more than 5 times in independent experiments.
Randomization No randomization was required for the reported experiments, as all variables could be controlled.

Blinding
Blinding is not applicable to crystallographic experiments. When needed during analysis of the models we compared our complex structure with previously published native structures.

Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.