Edit · 2026 01 11 232245 Nature S41594 024 01454 9

Saving needs a token. Log in or register to get one. Edits are public and licensed CC-BY-SA-4.0.

# Capturing eukaryotic ribosome dynamics in situ at high resolution | Nature Structural & Molecular Biology

**URL:** https://www.nature.com/articles/s41594-024-01454-9

**Saved:** 2026-01-11T07:22:50.337Z

---

Article
Published: 09 January 2025
Capturing eukaryotic ribosome dynamics in situ at high resolution
Jing Cheng, Chunling Wu, Junxi Li, Qi Yang, Mingjie Zhao & Xinzheng Zhang

Nature Structural & Molecular Biology 
volume
 32, 
pages
698–708 (2025)Cite this article

4132 Accesses

12 Citations

62 Altmetric

Metrics
details

Abstract

Many protein complexes are highly dynamic in cells; thus, characterizing their conformational changes in cells is crucial for unraveling their functions. Here, using cryo-electron microscopy, 451,700 ribosome particles from Saccharomyces cerevisiae cell lamellae were obtained to solve the 60S region to 2.9-Å resolution by in situ single-particle analysis. Over 20 distinct conformations were identified by three-dimensional classification with resolutions typically higher than 4 Å. These conformations were used to reconstruct a complete elongation cycle of eukaryotic translation with elongation factors (eEFs). We found that compact eEF2 anchors to the partially rotated ribosome after subunit rolling and hypothesize that it stabilizes the local conformation for peptidyl transfer. Moreover, open-eEF3 binding to a fully rotated ribosome was observed, whose conformational change was coupled with head swiveling and body back-rotation of the 40S subunit.

You have full access to this article via your institution.

Download PDF
Similar content being viewed by others
Deciphering molecular details of the RAC–ribosome interaction by EPR spectroscopy
Article Open access
21 April 2021
mRNA reading frame maintenance during eukaryotic ribosome translocation
Article 29 November 2023
eIF5B and eIF1A reorient initiator tRNA to allow ribosomal subunit joining
Article 22 June 2022
Main

Studying the conformational dynamics of protein machines by cryo-electron microscopy (cryo-EM) is an important endeavor. In recent years, multiconformational structural studies have been achieved primarily by in vitro single-particle analysis (SPA), yielding high-resolution structures. However, simulating the cellular environment in vitro is challenging, resulting in issues such as not observing all macromolecular activities1. In situ dynamic structural studies of proteins in cells provide a full complement of protein functional states compared to SPA but the resolution of in situ protein structures in multiple states is usually limited.

Protein synthesis performed by ribosomes is a complex process consisting of four steps: initiation, elongation, termination and ribosome recycling. Incorporating each new amino acid into the nascent peptide chain during elongation involves multiple conformational rearrangements of the ribosome. Transfer RNA (tRNA) base pairs with the mRNA codon template and ribosomes synthesize polypeptides from amino acids carried by the ribosome-bound tRNAs. The peptidyl transferase center (PTC) in the large subunit (LSU) catalyzes peptide bond formation and transfers amino acids to the growing peptide chain. Eukaryotic elongation factor 1A (eEF1A), combined with guanosine triphosphate (GTP), facilitates aminoacyl-tRNA transport to the A/T-site, while eEF2 assists ribosome translocation along the mRNA to ensure fidelity. The tRNAs at the A-site and P-site translocate to the P-site and E-site, respectively, and the vacated A-site receives the next aminoacyl-tRNA for peptide extension. Deacyl-tRNA at the E-site is released from the ribosome and particular eukaryotes, such as Saccharomyces cerevisiae2, also require eEF3 for this process.

During ribosomal translation, ribosomes, tRNA and mRNA undergo large-scale conformational changes. Some intermediate states have been analyzed by X-ray crystallography and cryo-EM, as previously reviewed3, but the complexity of eukaryotic systems has limited this analysis; therefore, many in vitro studies have focused on bacteria4,5. In addition, recent studies based on cryo-electron tomography have elaborated on cellular translation processes in eukaryotes6,7 and prokaryotes8. However, these studies only achieved high-resolution structural analysis of an averaged ribosome in different states, with the resolution of individual states limited by the number of particles used, resulting in an average resolution of 6–7 Å, which is insufficient for atomic model building. In addition, these studies may have missed conformations with relatively small populations when analyzing protein dynamics.

In this report, we investigate the eukaryotic ribosome of S. cerevisiae from cryo-focused ion beam (cryo-FIB) milled cells using the in situ SPA (isSPA) method9,10 and determined over 20 distinct in situ ribosome structures at resolutions ranging from 3.2 to 4.8 Å. Notably, most structures were complexed with eEFs, with several of these states previously unresolved. We reconstructed the elongation cycle from these high-resolution structures to present a complete eukaryotic elongation cycle.

In situ structure determination of ribosome conformations

S. cerevisiae lamellae (Fig. 1a,b) were thinned by cryo-FIB milling to visualize various conformations of the eukaryotic ribosome during translation. With a total of 451,700 filtered ribosome particles detected by graphics-processing-unit-accelerated isSPA (GisSPA)11 from ~40 lamellae, we achieved an overall resolution of 2.9 Å for the 60S subunit (Fig. 1c). The high-resolution map revealed the well-resolved bases of ribosomal RNA (rRNA) and amino acid side chains of the ribosomal proteins (Fig. 1d). The local resolution of 80S varies noticeably between the 40S and 60S regions because of 40S subunit structural dynamics. Thus, individual conformations were identified. To identify the multiple states in translation and study structural heterogeneity, we performed focused three-dimensional (3D) classifications on the basis of well-refined 80S ribosome particles without alignment using masks to include regions with heterogeneity through cryoSPARC11 (Methods). The 3D classification analysis yielded 24 different conformations (Fig. 2 and Extended Data Figs. 1 and 2) differentiated by the 40S subunit, tRNAs, eEF2, eEF3 and the L1 stalk. The resolution of the classes discussed below is listed in Tables 1–3.

Fig. 1: In situ S. cerevisiae ribosome structure.

a, Overview of the cellular lamella. The magenta crosses indicate the data collection areas. Scale bar, 1 μm. b, Micrograph showing the cellular context of S. cerevisiae. Scale bar, 50 nm. c, Local resolution map of the averaged 80S ribosome. d, Well-resolved bases of 25S rRNA (left) and side chains of ribosomal protein uL13 (right).

Full size image
Fig. 2: Ribosomal conformations and the translation elongation cycle.

a, The elongation cycle is represented by 11 structures, distinguished by eEFs and tRNAs. These structures correspond to distinct phases: mRNA decoding (pink), peptidyl transfer (blue) and tRNA translocation (orange). The states of the 40S subunit rotation (orange) and rolling (red), along with the positions of tRNAs and eEFs, are highlighted using different color shadings. b, The frequencies of occurrence for different states were calculated by the classified particle numbers. States corresponding to each process are colored according to a. Four states unassigned in the elongation cycle are colored in white. c, The angles of rotation (orange) and subunit rolling (red) for the 40S body and swiveling (blue) angles of the 40S head in 11 main conformational states, relative to S2a. d, The densities of four additional structures could not be placed in the sequential cycle. Among these, S10 contains only a tRNA in the A-site. The tRNAs and eEFs were colored the same as in a.

Source data

Full size image
Table 1 Cryo-EM data collection, refinement and validation statistics (part 1)
Full size table
Table 2 Cryo-EM data collection, refinement and validation statistics (part 2)
Full size table
Table 3 Cryo-EM data collection, refinement and validation statistics (part 3)
Full size table

We reconstructed an intact translation elongation cycle of S. cerevisiae according to the assignment of the three eEFs (eEF1A, eEF2 and eEF3) and the locations of tRNAs binding to the ribosome. The elongation cycle consists of 11 main conformational states (Fig. 2a) and the corresponding percentages of particles accounted for by each state are listed in Fig. 2b, similar to previous studies in situ6,7,8,12. Of the 11 conformations (Fig. 2a), three states, state 8 (S8; A/T, P and E), S1 (A/T and P) and S2a (A and P), represent the ribosome undergoing the mRNA decoding process, collectively accounting for 46.2% of particles. Three A and P states are newly resolved and present the stepwise native peptidyl transfer process (S2b–S2d), comprising 28% of particles. Four classical states, S4 (A/P and P/E), S5 (ap/P and pe/E), S6 (ap/P and pe*/E) and S7 (P and E) and a new intermediate state of S3, collectively accounting for 15.8% of particles, belonging to tRNA translocation were obtained. The rotation of the 40S body and swiveling of the 40S head in each state were measured (Methods and Fig. 2c). Three of these main conformations were further subdivided into nine subconformations according to the presence of eEF3 (Extended Data Fig. 2a,c,d), wherein a new S4a, featuring the binding of eEF3 to 80S, was captured. Furthermore, four additional states may not be part of the classical cycle and are listed separately in Fig. 2d. The results were confirmed through two additional rounds of 3D classification (Supplementary Table 1).

Compact eEF2 binds to the ribosome during peptidyl transfer

Classifications on particles in the A and P state by focusing on 40S subunit and eEF2 binding region yielded seven different conformations. Among these, S2a belonged to tRNA accommodation and three states (S2b–S2d) were associated with the peptidyl transfer process. During peptide bond formation, the nascent peptide chain is transferred from the P-site tRNA to the newly incorporated A-site tRNA in the A and P state. Although 3D classification did not focus on the PTC, the resulting three states (S2b–S2d) represent the stepwise process of peptidyl transfer. S2e (A and P) and S2f (A and P) were intentionally not included as main states of tRNA accommodation or peptidyl transfer process, considering the controversy surrounding the description of peptide chain transfer because of the relatively low resolution in PTC regions. The remaining S9 (A and P), consisting of 3.3% of ribosome particles with heterogeneities in the 60S and 40S subunits (Fig. 2d and Supplementary Fig. 1), probably does not belong to the reconstructed elongation cycle.

In the unrotated S2a, tRNAs transition from the A/T and P state to the A and P state with the dissociation of eEF1A, associated with classical tRNA accommodation5,13. However, the density of the aminoacyl-tRNA elbow is poorer than that of the anticodon stem loop and the CCA of the A-site tRNA is not observed in the PTC, indicating that the accommodation of A-site tRNA is incomplete. The nascent peptide chain is attached to the P-site tRNA. S2b exhibits a eukaryote-specific ‘subunit rolling’ of a ~5° tilt of the small subunit (SSU) body relative to unrotated states, occurring toward the P stalk around the previously defined axis14,15 located on the upper part of helix 44 (h44) of 18S RNA and orthogonal to the well-known intersubunit rotation axis. This rolling is earlier than the subunit rolling observed in the pretranslocation classical state, as reported previously14,15,16, and may facilitate the accommodation of aminoacyl-tRNA5,14,15. The CCA of A-site tRNA inserts into the PTC and the nascent peptide chain is still connected to P-site tRNA in this state (Fig. 3a), indicating the prepeptidyl transfer state upon accommodation of A-site tRNA. An intermediate state, S2c, during peptide bond formation was also captured (Fig. 3b,c). Peptide chain nucleophilic attack by the incoming free α-amino group onto the carbonyl carbon is known to break the acyl linkage between the carbonyl carbon and the 3′ hydroxyl of the P-site tRNA17 and forms the peptide bond between the carbonyl carbon and the amino nitrogen. In S2c, the peptide chain is not connected to A-site or P-site tRNA at a contour value of 2.6 root-mean-square deviation (r.m.s.d.) (Fig. 3b). Decreasing the value to 1.8 r.m.s.d. revealed that the peptide chain connects to the two tRNAs simultaneously (Fig. 3c), suggesting that this intermediate represents the tetrahedral intermediate of the nucleophilic attack. The CCA of P-site tRNA is observed to interact in the PTC with G2620, G2621 and A2820 and that of A-site tRNA interacts with G2922, U2923, U2924 and U2953 (refs. 17,18,19) (Fig. 3d). Additionally, the density of the CCA of A-site tRNA is poorer than that in S2b and the aminoacyl residues in S2c appear to shift slightly relative to S2b (Extended Data Fig. 3). In S2d, the peptide chain is well connected to A-site tRNA, representing the final peptidyl transfer state (Fig. 3e). The 40S subunit remains stationary and in the rolling state throughout peptidyl transfer process (S2b–S2d).

Fig. 3: Presentations of cryo-EM densities in PTC during peptidyl transfer process.

a, The positions of the nascent peptide chain, CCA of P-site tRNA (carbon atoms of model colored in red) and A-site tRNA (carbon atoms of model colored in blue) in PTC of S2b. b,c, View of the nucleophilic attack in S2c at a contour value of 2.6 r.m.s.d. (b) and 1.8 r.m.s.d. (c). d, Surroundings of CCA of A-site tRNA (left; model colored in blue) and P-site tRNA (right; model colored in orange) in S2c. Densities are colored in gray. e, The peptide chain disconnects from P-site tRNA (red) and is connected to A-site tRNA (blue) in S2d, indicating that the peptidyl transfer process is complete. The contour level is 2.6 r.m.s.d. in a,e. Densities in a–c,e are shown in mesh representation.

Full size image

Interestingly, in addition to transferring the nascent peptide chain in the PTC, we observed a clear occupation of eEF2 in S2c. When docking the five domains of eEF2 into the EM maps independently (Fig. 4a,b and Supplementary Fig. 2a), we found that eEF2 exhibits a compact arrangement, which resembles the conformation of EF-G in prokaryotes20. Domains I and V connect with the 60S subunit, exhibiting a similarity to the typical extended conformation21. Domain II interacts with the 40S body through eS24, h15 and h17, whereas domain III contacts uS12 of the 40S body. Additionally, loop 1 of domain IV is proximal to the shoulder of the SSU through h17, and loop 3 through both h17 and h16. These observations indicate that compact eEF2 immobilizes the tilted SSU on LSU in S2c, where nucleophilic attack may occur, potentially ensuring a stable environment and optimal performance for peptidyl transfer. Compact eEF2 was also captured in unrotated S2e (Fig. 2d and Extended Data Fig. 4a,b), supporting observations from previous studies in prokaryotes20. Compared to eEF2 in S2c, we observed that, in S2e, the interactions between domains II and IV of eEF2 and the 40S subunit are absent. eEF2 in S2b, S2d and S2f has a weak density, which may represent intermediate states of eEF2 undergoing conformation changes.

Fig. 4: Compact and extended eEF2 during elongation cycle.

a, Overview of the compact eEF2 bound to ribosomes in S2c. The cryo-EM densities of eEF2 are colored differently: domain I, green; domain II, blue; domain III, yellow; domain IV, purple; domain V, magenta. b, Two views of the environment surrounding of compact eEF2 (colored by domain) in the ribosome. Gray and blue densities are nucleic acids and proteins of the LSU, respectively. Orange and pink densities are nucleic acids and proteins of the SSU, respectively. The atomic models of the ribosome are shown as a dark-gray ribbon. Domain I of eEF2 interacts with uL6 and uL10 of the 60S subunit and domain V interacts with H43, H44, H95 and uL11 of the 60S subunit. c, Less extended eEF2 in S4. d, Two views of the environment surrounding of less extended eEF2 in the ribosome. The arrangement and color are the same as b, with the new densities of tRNA interacted with domain IV of eEF2 colored in red. e, Superimposition of the models of compact and less extended eEF2. Helices in the less extended eEF2 are displayed as cartoon tubes. The domains colored as in b,d, while less extended eEF2 is represented by a lighter color compared to compact eEF2. f, Overall movement of eEF2 and tRNAs on the ribosome from S2c to S4. The conformational change is shown from the compact (purple, in S2c) to the less extended conformation (pink, in S4) of eEF2. The A-site tRNA, P-site tRNA and mRNA in S2c are colored in blue, red and yellow, respectively. The translocated tRNAs and mRNA in S4 are represented by light colors. g, The conformational change from the less extended (pink, in S4) to the fully extended conformation (light green, in S6) of eEF2. The tRNAs and mRNA in S6 are represented by light colors compared to those in S4. The distances and angles presented reflect the movement of eEF2.

Full size image

We also found clearly resolved densities belonging to eS25 (aa 4–21) during the peptidyl transfer process in S2b–S2d (Extended Data Fig. 5c), which was not previously modeled. We used AlphaFold2 to predict the eS25 structure and aa 4–21 were predicted to form an α-helix, aligning well with our cryo-EM densities. This helix interacts with A-site tRNA, extends over P-site tRNA (Extended Data Fig. 5c) and may stabilize aminoacyl-tRNA during peptidyl transfer. However, the helical densities in unrotated S2a and S2e exhibit poor continuity (Extended Data Fig. 5a,b) and are absent in other states beyond the peptidyl transfer process.

Conformational changes of eEF2 during translocation

After peptidyl transfer, peptidyl-tRNA begins translocation from the A-site to P-site. S3–S8, with the peptide chain connecting with A-site tRNA, represent intact eukaryotic ribosomal translocation and are well sorted according to the positions of P-site and E-site tRNA (Fig. 2a). eEF2 adopts two different conformations during translocation, the less extended and fully extended states. In S3, the SSU body undergoes a back-rolling of ~1° and a large relative rotation of ~8° with respect to unrotated states. A weaker density of eEF2 compared to that in S2d and a flexible density of A-site tRNA (Supplementary Fig. 1) indicate that S3 represents an intermediate state at the beginning of translocation. Following another small SSU body rotation (Fig. 2c), a less extended conformation of eEF2 is observed in S4 (Fig. 4c,d and supplementary Fig. 2b), where domain IV does not overlap with the A-site on the 40S subunit, as previously resolved in situ8,22. The surroundings of less extended eEF2 are shown in Fig. 4d. Domain IV is primarily maintained by the presence of peptidyl-tRNA, with additional support from h34 and eS31. Domain III maintains its interactions with the SSU body through uS12, whereas domain I establishes interactions with 40S subunit through eS24, eS6, h8 and h14, facilitated by the rotation of the 40S subunit. Simultaneously, h5, near domain II, may stabilize eEF2 during its conformational change. As tRNAs move along the mRNA, the SSU head swivels ~13.3° and the SSU body undergoes an ~3.6° back-rotation in S5 (Fig. 2c). eEF2 is found to extend toward the A-site and reaches the fully extended state21. The SSU body continues with another small back-rotation of ~2° while the head swivels to ~17.2°, leading to tRNAs reaching S6 with a fully extended eEF2 (ref. 16). The interactions between domains I and II of eEF2 with SSU are mostly lost with the back-rotation of the SSU body (Extended Data Fig. 4c,d). A back-rolling of ~2.1° was observed between S5 and S6, which may, together with the back-rotation, facilitate tRNA–mRNA decoupling between the SSU body and head16,21. In S7 (ref. 7), the SSU body returns to the unrotated and unrolling state while the head swivels to 0°. Simultaneously, peptidyl-tRNA reaches the P-site and deacyl-tRNA occupies the E-site, indicating that the translocation process is nearing completion. eEF2 maintains a fully extended conformation until dissociation in S8. Another elongation cycle begins when eEF2 separates and eEF1A•tRNA combines with the next mRNA codon.

We compared the domain arrangement of compact eEF2 in S2c and less extended eEF2 in S4 by superimposing their models (Fig. 4e), revealing that domains I–III maintain a relatively fixed arrangement among each other while domains IV and V form another rigid body, undergoing a rotation of ~40° around an axis at their junction. This rotation causes domain IV of compact eEF2 to extend about 48 Å, mostly toward the A-site of the 40S subunit, whereas domain V shifts ~23 Å to be primarily closer to 40S subunit (Fig. 4e). The spatial changes of eEF2 resulting from conformational changes are illustrated using the 60S subunit as a reference to analyze the shifts and rotations of eEF2 domains relative to the ribosome. From compact to less extended, eEF2 domains I–III rotated ~30° while domains IV and V rotated ~60°, yielding a movement of ~80 Å for domain IV (Fig. 4f). From less extended to fully extended, eEF2 domains had a monolithic rotation of ~15°, leading to domain IV of S6 being extended primarily by ~20 Å toward the A-site (Fig. 4g).

Conformational changes of eEF2 from the compact state in S2c to the fully extended state in S6 do not affect the interactions of domain V with H95 of 25S rRNA and H43 and H44 of the P stalk (Extended Data Fig. 6). By aligning domain V of fully extended eEF2 with compact eEF2 on the ribosome in S2c, we observed steric hindrance caused by the P stalk. This results in a clash of domain I with the 60S subunit (Extended Data Fig. 7a), hindering the formation of extended eEF2. When the ribosome transitions to S4, the P stalk shifts ~7 Å toward tRNA (Extended Data Fig. 7b) and the 40S body rotates by ~6° to allow eEF2 to adopt the extended conformation. We found a stronger density of uL10 of the P stalk near domain I when focusing on the density of eEF2 in S3 (Supplementary Fig. 3).

Conformational rearrangement of eEF3 throughout elongation

In addition to the two conserved and shared elongation factors, eEF1A and eEF2, found in eukaryotes, there is a requirement for another adenosine triphosphatase (ATPase) known as eEF3 in the fungal translation system23,24,25. This ATPase contains two ATP-binding motifs and spans the head of the 40S to 60S subunits. Focusing on the overall cryo-EM maps of S4, S8 and S1, eEF3 essentially exhibits weak density, which may suggest a relatively low occupancy of eEF3. Therefore, a dedicated round of 3D classification focusing on the eEF3 region illustrated the 3D variability of eEF3 in these three states. S4 is divided into two substates, S4a (with stable eEF3 binding) and S4b (without eEF3 binding), as illustrated in Extended Data Fig. 2a. Three substates are identified in S8 (Extended Data Fig. 2c): stably bound eEF3 (S8a), flexibly bound eEF3 (S8b) and unbound eEF3 (S8c). Four substates are identified in S1 (Extended Data Fig. 2d): stable binding to the 80S (S1a), a less flexible binding state (S1b), a more flexible binding state (S1c) and a dissociated state (S1d), where eEF3 may dissociate from the 80S.

Two conformations of eEF3 were identified from eEF3-binding states: a stable closed-eEF3 conformation (Fig. 5a and Supplementary Fig. 2c) when bound to the unrotated ribosome in S1a and S8a (Extended Data Fig. 2c,d), which closely resembles the state obtained from SPA24, and a stable open-eEF3 conformation (Fig. 5b and Extended Data Fig. 2a) when bound to the fully rotated ribosome of S4a, which was not observed previously. The local resolution of open-eEF3 is lower than closed-eEF3 (Supplementary Figs. 2c,d and 4) because of inherent flexibility. When docking the eEF3 models into the EM densities, two ATP-binding domains (ABC1 and ABC2) of open-eEF3 (Fig. 5b) exhibit notable gaps between the HEAT domain and the ribosome compared to those of closed-eEF3 (Fig. 5a). The analysis of chromodomain (CD) is absent in our structures given that the density of CD is insufficient for model fitting. An overarching perspective shows that the closed-eEF3 exhibits a pronounced rotation relative to open-eEF3 around the junction of the ABC2 domain and the ribosome, primarily attributed to the rotation of the 40S subunit (Fig. 5c). When the HEAT domain and four-helix bundle (4HB) domain are superimposed, ABC1 and ABC2 domains, which are in complex with ATP or adenosine diphosphate (ADP), have a rotation of ~90° on the axis of the HEAT domain (Fig. 5d).

Fig. 5: eEF3 binds to and dissociates from the ribosome in different conformations.

a, Bottom: the overall densities of closed-eEF3 in complex with the unrotated ribosome (S1a). Detailed zoomed-in view highlighting eEF3 domains: HEAT, pink; 4HB, green; ABC1, orange; ABC2, blue. Top: close-up view of model located in closed-eEF3 region. The eEF3 domains are colored consistently with their densities, while the model of the ribosome is colored in gray. b, Presentation of open-eEF3 in complex with the fully rotated ribosome (S4a). The domains of eEF3 are colored darker compared to the closed-eEF3 in a, with helical models displayed as a cartoon tube. The model of rRNAs are colored in red. c, Comparison of closed-eEF3 and open-eEF3 under superimposition of LSU models of S1a and S4a. d, Swinging between the ABC1 and ABC2 domains of the two eEF3 conformations under superimposition of HEAT and 4HB domains. The colors are the same as in a. e, Binding states of eEF3 vary with the ribosome states. Stable and flexible binding eEF3 states are illustrated in orange and light orange, respectively. The intermediate color between orange and light orange in the second pattern represents the occurrence of closed-eEF3 in S6. The states of tRNAs and SSU are indicated below the corresponding ribosomes. The tRNA colored by light yellow in the fourth pattern represents the release of E-site tRNA in S1a. The purple outline represents eEF2 in the elongation cycle.

Full size image

Throughout the elongation cycle, open-eEF3 initially binds to the ribosome when the 40S subunit reaches the fully rotated state of S4a (Fig. 5e and Extended Data Fig. 2a). As the 40S subunit undergoes body back-rotation and large head swiveling, the stability of eEF3 binding to the ribosome weakens, leading to flexible eEF3 density in S5 and subsequently transitioning to a closed conformation in S6 (Extended Data Figs. 2b and 8). However, this quality of eEF3 density in S6 is lower than that of the closed-eEF3 observed in S1a. The flexible density of eEF3 in S5 indicates a conformational change of eEF3 coupled with head swiveling and back-rotation of ribosomal SSU (Fig. 5e). Accompanying the head swiveling of the 40S subunit from S4 to S6, the E-site tRNA is transferred ~14 Å and the L1 stalk opens by ~20 Å in the presence of the 40S body alignment (Extended Data Fig. 9a).

Following the dissociation of extended eEF2 and the binding of the eEF1A•tRNA after S7, strong densities are observed for eEF3 in the closed conformation in S8a and S1a. In S8a, the L1 stalk continues to open ~19 Å relative to S6, facilitating the release of the E-site tRNA from the ribosome. The L1 stalk adopts an open conformation in three substates identified from S8. After releasing the E-site tRNA in S1 (Fig. 2a), the L1 stalk of the 60S subunit gradually switches to the closed state. In the stable eEF3-binding state of S1a, the L1 stalk of 80S adopts the open state, whereas, in the dissociated state of S1d, L1 switches to the closed state. To verify whether the transition of L1 stalk from open to closed is sequential, we classified particles belonging to the A/T and P state (S1) into three classes according to the state of the L1 stalk: open state, intermediate state and closed state (Extended Data Fig. 9b) with the percentage of the stable eEF3-binding state calculated as 73%, 26% and 0%, respectively. From the results, we infer that the presence of E-site tRNA does not affect the binding state of eEF3 and the closure of the L1 stalk must occur after eEF3 dissociation.

Discussion

In this study, we resolved the dynamic structures of the S. cerevisiae ribosome in a native translation environment at near-atomic resolution. By tracking tRNAs and eEFs, we gained insight into the ribosomal elongation cycle. We captured the compact conformation of eEF2 during peptidyl transfer, occurring much earlier than previous observed extended eEF2 (refs. 20,26). Additionally, we found an open-eEF3 conformation binding to fully rotated ribosome, whose conformational change may facilitate the head swiveling of 40S subunit.

Subunit rolling is a common phenomenon of SSU in eukaryotic peptidyl transfer, facilitating aminoacyl-tRNA accommodation14,15. In our findings, rolling SSU was fixed by compact eEF2 to the LSU in S2c, where the peptide chain undergoes nucleophilic attack by the free α-amino. It is obvious that compact eEF2 binds to the ribosome at or before S2c. The absence of CCA of A-site tRNA in PTC of unrotated ribosome indicates uncompleted accommodation. Therefore, we hypothesize that the unrotated S2e, characterized by compact eEF2 binding and previously unassigned in the elongation cycle, corresponds to A-site tRNA accommodation, given its unrotated SSU and absence of CCA in the PTC. The helix of eS25 exhibits lower densities in S2a and S2e compared to S2b–S2d (Extended Data Fig. 5), suggesting that S2e is close to S2a. Additionally, subunit rolling reduces openings to the intersubunit space on the A-site side, hindering eEF2 binding in S2c.

We present a model (Extended Data Fig. 10) to illustrate the life cycle of compact eEF2 during elongation. After dissociation of eEF1A and eEF3, aminoacyl-tRNA approaches the A-site but CCA does not insert into the PTC. In S2a, unrotated ribosomes serve as substrates for compact eEF2 in the cytoplasm and facilitate ribosome recruitment. The compact eEF2 binds only to the 60S subunit in S2e, as it cannot reach the SSU body in the unrotated ribosome. Upon eEF2 binding, the ribosome undergoes a conformational change, tilting the 40S subunit by ~5° toward the P stalk, and attempts to establish an interaction with eEF2 that gives rise to a flexible density of eEF2 in S2b. When eEF2 stably interacts with both 60S and 40S subunits, the CCA is inserted into the PTC and transpeptidation proceeds in S2c. After peptidyl transfer, the SSU prepares for back-rolling and rotation, causing eEF2 to exhibit weak density again in S2d. Compact EF-G binding to the unrotated 70S ribosome before peptidyl transfer in Escherichia coli has been captured20, which supports our hypothesis. Recently, Hassan et al. showed that subunit rolling is common in prokaryotes and eukaryotes27, suggesting that compact eEF2 or EF-G in the ribosomal elongation cycle is ubiquitous.

Our results indicated that eEF2 has a pivotal role as a ‘molecular arm’ in the eukaryotic elongation cycle. During peptidyl transfer process, loops of domain IV of compact eEF2 function as ‘fingers’ within the contracted arm, skillfully capture the tilted SSU body and facilitate the seamless transfer of the nascent peptide. As the cycle progresses to translocation, the arm gradually extends to reach the A-site. Simultaneously, the fingers have a crucial role in stabilizing the mRNA reading frame, effectively ensuring the directionality of the translocation process21,28, which corresponds to S5–S7 in our results.

eEF2 is a GTP-binding protein29,30,31 and has crucial roles during the ribosomal elongation cycle. Majumdar et al.16 resolved an extended eEF2•GDP-binding ribosome in Giardia intestinalis, featuring a uniquely positioned ‘leaving phosphate’. This structure aligns with our S6 state, as confirmed by comparing tRNA positions to SSU body alignment, thus revealing that GTP hydrolysis occurs during the final steps of tRNA translocation16,28. In addition, we hypothesize that the large SSU rotation at the beginning of translocation (S3 and S4) induces the conformational change of eEF2 and this change does not require GTP hydrolysis by eEF2. The small back-rolling of ~1° from S2d to S3 weakens the interaction between domains II and IV of eEF2 and the SSU body, facilitating the subsequent extensive body rotation in S3. However, the interaction between domain III and uS12 of SSU is maintained, accompanying the SSU rotation and dragging the compact eEF2 in rotation around the junction of domain I and 60S subunit. Simultaneously, the closure of H95 in 25S rRNA removes steric barriers for extending eEF2. Moreover, uL10 of the P stalk likely has a key role in facilitating the movement of domain V (Supplementary Fig. 3).

Hydrolysis of GTP (ATP for eEF3) occurs when the eEFs maintain ribosome translation, along with conformational changes depending on whether they are bound to GTP (ATP for eEF3) or GDP (ADP for eEF3). eEF3 contains two ATP-binding domains, thus carrying two ATP molecules24,25. We hypothesize that the open-eEF3 with two ATPs binds to the ribosome in the fully rotated ribosome of S4a. One ATP hydrolysis occurs during the conformational change of eEF3 from open (S4a) to closed (S6) states and may facilitate head swiveling and E-site tRNA translocation to exiting. In S4, the E-site tRNA binds to the L1 stalk and the translocation of tRNA toward the L1 stalk potentially drives the opening of the L1 stalk, facilitating the release of the E-site tRNA. We hypothesize that the opening of the L1 stalk and release of E-site tRNA occur without additional ATP hydrolysis by eEF3. A second ATP hydrolysis may contribute to dissociating eEF3 from the ribosome, where eEF3 binding is destabilized before dissociating in S8. If this occurs, the second ATP hydrolysis causes a conformational change of eEF3 that disrupts the interaction interface between eEF3 and the ribosome to cause eEF3 dissociation from the ribosome.

The particles for reconstruction were filtered on the basis of the presence of the SSU, which was not included in our template. Notably, the tRNAs and eEFs analyzed in the elongation cycle were also not included in the template. Additionally, the frequencies used for detection by GisSPA ranged from 1/400 to 1/8 Å−1. Consequently, the high-resolution signals above 1/8 Å−1 in the ribosome remained unaffected by template bias. Furthermore, following the final 3D classification, the densities of the LSU included in the template and the SSU beyond the template exhibited equivalent quality in each stable state. These comprehensive analyses collectively indicate that our results were hardly affected by model bias. In addition, the Z coordinates of ribosomal particles can be obtained through defocus searching but we did not perform an accuracy analysis of the Z coordinates in this study. The accuracy was likely influenced by factors such as the frequency of templates used. In this regard, Lucas et al.32 reported that the precision of two-dimensional template matching in the Z dimension, relative to template matching in binned 3D tomograms, was approximately 6 nm.

The single LSU and ribosomal conformations in the initiation and termination states were not observed even though we successfully identified conformations at a percentage of 1.5%. Thus, the absent states in our study present lower percentages than 1.5% in S. cerevisiae cells. We obtained 451,700 ribosomes after collecting data for 2 days. The large dataset increased the resolution of the ribosome at each state and facilitated the capture of an intermediate state with an estimated occupancy as low as ~1%. The use of GisSPA provides notable advantages in terms of high data acquisition throughput.

Methods
Cell culture

S. cerevisiae strains BY4741 were grown in yeast peptone dextrose medium at 30 °C to an optical density at 600 nm (OD600) of 0.6–0.8. The cells were collected and resuspended in the same medium and the concentration of the suspension was adjusted to an OD600 of 2.

Cryo-EM sample preparation and FIB milling

First, 3 µl of S. cerevisiae cells were applied to a glow-discharged copper Quantifoil 200 mesh R1.2/1.3 grid. After 30 s of preblotting, the grid was plunge-frozen in liquid ethane using a Leica EM GP (Leica Microsystems) with 6–8 s of backside blotting at 16 °C with the chamber at 70% humidity.

Cryo-lamellae were prepared by cryo-FIB milling using the Aquilos 2 FIB scanning EM instrument (Thermo Fisher Scientific). Before milling, samples were sputter-coated with metallic platinum for 10 s with a 15-mA beam current, then coated with organo-platinum for 6 s through the gas injection system and finally coated with metallic platinum for 10 s with the same settings. Lamellae were roughly milled to ~300 nm using a decreasing Ga+ ion beam from 0.3 nA to 50 pA with a milling angle of 8°. A 10-pA ion beam was used for the final polish to obtain ~150-nm thickness.

Cryo-EM data acquisition and image processing

Approximately 40 cryo-lamellae were collected by applying a single-particle data collection method33 using an FEI Titan Krios G3 (Thermo Fisher Scientific) equipped with a Gatan GIF K2 camera (Gatan). Before collection, an offset start angle of 8° was set to compensate for the milling angle, ensuring that the lamellae were perpendicular to the electron beam, as previously described11,34. Data collection was performed with SerialEM16 and the beam-image shift data were collected17 using a nominal magnification of 105,000 with a calibrated pixel size of 0.66 Å per pixel in super-resolution mode.

Two dose rates were used during data collection; 219 micrographs were irradiated with 35 e− and dose-fractionated to 40 frames, while the remainder were collected with a total dose of 50 e− and dose-fractionated to 50 frames. The energy slit width was set to 20 eV and the nominal defocus was set between −0.8 and −1.2 μm.

Structure determination

Motion correction was performed with MotionCor2 (ref. 35) and micrographs were dose-weighted and binned with a factor of 2. The contrast transfer function (CTF) parameters of micrographs were estimated by CTFfind4 (ref. 36) in RELION. Then, the micrograph file in ‘star’ format was converted to ‘lst’ format for the detection task. The 60S subunit structure of yeast (EMD-6615)37 was used as the initial 3D template in GisSPA11, which is relatively more homogeneous than the 40S subunit. The localization task was performed by generating 1,649 2D projections (ignoring the in-plane rotation) that uniformly covered the sphere model with a 5° angular step. We performed the initial detection task on the first 500 micrographs with overlapping parameter k as 3 and output score threshold as 6.8, resulting in 7,043 particles in total. A round of focused refinement by RELION38 on these particles yielded a structure of 60S at a resolution of 5.9 Å.

Considering the scale difference between the original template (EMD-6615) and the particles in our micrographs, which influenced the detection efficiency of GisSPA, we segmented the 60S subunit from the newly reconstructed map at 5.9-Å resolution and used it as the template for a new round of searching by GisSPA. To remove duplicated detections, potential particles within a translational distance of ~10.6 Å and Euler distance of 8° were merged as one particle. Only frequencies from 1/400 to 1/8 Å−1 were used in the calculations for both localization tasks. A total of 808,736 potential particles from the 3,702 micrographs were kept after removing repeats that met the new cutting threshold of 6.2. The particle file in ‘lst’ format was converted back to ‘star’ format for particle extraction in RELION. Particles were extracted in two bins with a box size of 200 after implementing a round of 3D classification using the 3D template in detection as the initial model and the lowpass filter set to 20 Å. Accordingly, ~56% of the particles including the 40S subunit density were selected and the remainder showing bias features were thrown out (Extended Data Fig. 1). Because the cryo-EM data on each grid were collected individually, we divided them into nine optical groups for two rounds of CTF parameter refinement, one round of beam tilt estimation and two rounds of orientational and translational refinements; the resolution was estimated as 3.0 Å in the 60S region. To further improve the resolution, a round of Bayesian polishing was carried out and another round of focused refinement was applied to the polished particles; the resolution in 60S subunit was improved to 2.9 Å. The distributions of the final selected particles in several representative micrographs are depicted in Supplementary Fig. 5.

3D classification and refinement

Ribosome 3D classifications were accomplished using cryoSPARC39 with the imported refined particles from RELION-3.2. First, focused classification with a smooth shape mask covering the A-site, P-site and E-site tRNA positions was performed. This resulted in four major classes with different tRNA occupancies: A/T and P (150,024 particles), A and P (201,280 particles), A/T, P and E (37,868 particles) and ‘P’ and ‘E’ (62,528 particles). In the second round, we continued to classify particles belonging to the P and E class focused on the 40S region and this resulted in four tRNA translocation classes divided from the ‘P’ and ‘E’ class: A/P and P/E (S4), ap/P and pe/E (S5), ap/P and pe*/E (S6) and P and E (S7) (Extended Data Fig. 1).

We found densities with lower occupancy presented in the P stalk binding site in the A and P state; hence, a third round of classification focused on the P stalk binding site was carried out on particles in the A and P state, resulting in an empty binding site class, compact eEF2 stable binding class and compact eEF2 density scattered class. Each class showed flexibilities in the 40S subunit; therefore, a classification focused on the 40S region was performed on each class. Particles were further classified into the following states: unrotated 40S without compact eEF2 binding (S2a), rotated 40S with compact eEF2 flexible binding (S2b), rotated 40S with compact eEF2 stable binding (S2c), rotated 40S with flexible compact eEF2 density (S2d), unrotated 40S with stable compact eEF2 binding (S2e), an intermediate state (S2f) and an intermediate class (S3) positioned between the A and P and the A/P and P/E states. In addition, 14,925 particles (S9) belonging to the A and P state had a lower resolution of 4.8 Å than other classes within similar particles, showing severe flexibilities in the 40S subunit, and 7,588 particles only contained tRNA in the A-site (S10) with the 40S subunit partially rotated, similar to S2b and S2c.

To investigate the occupancies of eEF3 in S1, S4 and S8, three rounds classifications with a soft mask focus on eEF3 region were individually performed on the particles in three classes, resulting in stable eEF3 (S1a, S4a and S8a), flexible eEF3 (S1b, S1c and S8b) and no eEF3 (S1d, S4b and S8c). Classifications of particles in the A/T and P state focused on the L1 stalk were performed with a soft mask including regions of the L1 stalk.

The masks used above were generated by the ‘relion_mask_creation’ command implemented in RELION38 and the input maps were segmented from the overall reconstruction in Chimera40 and lowpass-filtered to 20 Å. The resolutions of all states were obtained through the final refinement focused on the 60S subunit. The local resolution was calculated using cryoSPARC39.

Determination of SSU movements

The movement analyses of the 40S subunit were conducted using Chimera40. The rotational angles, axes and matrices of the 40S body between any state and S2a were obtained by aligning the 60S subunit of all maps. A custom script was used to decompose the corresponding rotational angles along the axes of 40S body rotation and rolling. The rotational axis between S4 and S5 was chosen as the axis of 40S rotation, while the axis of subunit rolling was aligned with the upper part of h44 of 18S rRNA, identified in Chimera. Additionally, the angles of 40S head swiveling were calculated by Chimera through the alignment of the 40S body between each state and S2a.

Model building

The S. cerevisiae 60S subunit model was built on the basis of a previous crystal structure (Protein Data Bank (PDB) 4V88)41. It was split into separate rRNA and protein parts and fit into the cryo-EM density. The nonrotated 40S subunit was built on the basis of another structure (PDB 7OSM)21. For rotated 40S, yet another structure (PDB 7B7D)24 was used. In addition, the refinement of the PTC in the 25S rRNA model referred to a 2.4-Å crystal structure (PDB 8CVJ)4. AlphaFold2 (ref. 42) was used to complete the eS25 structure.

The eEF1A model was derived from a cryo-EM structure (PDB 5LZS)43. Using a previous structure (PDB 2P8W)30 as a template to build the extended conformation of eEF2, the compact conformation of eEF2 needed to rigid-body fit the five domains (domains I–V) divided as previously described44 into the cryo-EM map by Chimera40. Another model (PDB 1N0V)44 was used for the rearrangement of the eEF2 compact conformation. A previous eEF3 model (PDB 7B7D) was separated into five domains (HEAT, 4HB, ABC1, ABC2 and CD)2 that were fitted into the cryo-EM map separately as a rigid body in Chimera. The tRNA (PDB 7B7D) and the aminoacyl-tRNA model (PDB 8CVJ) were derived from different existing structures. The mRNA was built using another structure (PDB 6XHW)45 as a template. The nascent peptide was built de novo by constructing a poly(alanine) sequence.

The domains were fitted into each electron density map as rigid bodies in Chimera40 using the ‘fit in map’ function. Then, separate domains were manually adjusted using ‘real_space _refine’ and merged into a whole model in Coot46. The ribosome model was further refined by iterating in Coot and PHENIX47. Validation was performed using MolProbity7 (ref. 48).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The electron density maps were deposited to the EM Data Bank under accession codes EMD-38654 (S8a), EMD-38655 (S8), EMD-38656 (S1), EMD-38657 (S1a), EMD-38658 (S2a), EMD-38659 (S2b), EMD-38660 (S2c), EMD-38661 (S2d), EMD-38662 (S2e), EMD-38663 (S2f), EMD-38664 (S3), EMD-38665 (S4), EMD-38666 (S4a), EMD-38667 (S5), EMD-38668 (S6), EMD-38669 (S7), EMD-38670 (S9) and EMD-38671 (S10). The atomic models were deposited to the PDB under accession codes 8Z71 (S1a), 8XU8 (S2c), 8YLD (S4a) and 8YLR (S6). Source data are provided with this paper.

Code availability

The custom script used for the determination of SSU movements was uploaded to GitHub (https://github.com/zhao-xigua/Rotation).

References

Dunstone, M. A. & de Marco, A. Cryo-electron tomography: an ideal method to study membrane-associated proteins. Philos. Trans. R. Soc. B 372, 20160210 (2017).

Article
 
Google Scholar

Andersen, C. B. et al. Structure of eEF3 and the mechanism of transfer RNA release from the E-site. Nature 443, 663–668 (2006).

Article
 
CAS
 
PubMed
 
Google Scholar

Korostelev, A. A. The structural dynamics of translation. Annu. Rev. Biochem. 91, 245–267 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Syroegin, E. A., Aleksandrova, E. V. & Polikanov, Y. S. Insights into the ribosome function from the structures of non-arrested ribosome–nascent chain complexes. Nat. Chem. 15, 143–153 (2023).

Article
 
CAS
 
PubMed
 
Google Scholar

Loveland, A. B., Demo, G. & Korostelev, A. A. Cryo-EM of elongating ribosome with EF-Tu•GTP elucidates tRNA proofreading. Nature 584, 640–645 (2020).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Gemmer, M. et al. Visualization of translation and protein biogenesis at the ER membrane. Nature 614, 160–167 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Hoffmann, P. C. et al. Structures of the eukaryotic ribosome and its translational states in situ. Nat. Commun. 13, 7435 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Xue, L. et al. Visualizing translation dynamics at atomic detail inside a bacterial cell. Nature 610, 205–211 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Cheng, J., Li, B., Si, L. & Zhang, X. Determining structures in a native environment using single-particle cryoelectron microscopy images. Innovation 2, 100166 (2021).

CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Cheng, J. & Zhang, X. Optimizing weighting functions for cryo-electron microscopy. Biophys. Rep. 7, 152 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Cheng, J. et al. Determining protein structures in cellular lamella at pseudo-atomic resolution by GisSPA. Nat. Commun. 14, 1282 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Xing, H. et al. Translation dynamics in human cells visualized at high resolution reveal cancer drug action. Science 381, 70–75 (2023).

Article
 
CAS
 
PubMed
 
Google Scholar

Rodnina, M. V. Decoding and recoding of mRNA sequences by the ribosome. Annu. Rev. Biophys. 52, 161–182 (2023).

Article
 
CAS
 
PubMed
 
Google Scholar

Budkevich, T. V. et al. Regulation of the mammalian elongation cycle by subunit rolling: a eukaryotic-specific ribosome rearrangement. Cell 158, 121–131 (2014).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Behrmann, E. et al. Structural snapshots of actively translating human ribosomes. Cell 161, 845–857 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Majumdar, S. et al. Insights into translocation mechanism and ribosome evolution from cryo-EM structures of translocation intermediates of Giardia intestinalis. Nucleic Acids Res. 51, 3436–3451 (2023).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Tirumalai, M. R., Rivas, M., Tran, Q. & Fox, G. E. The peptidyl transferase center: a window to the past. Microbiol. Mol. Biol. Rev. 85, e00104–e00121 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Martin Schmeing, T., Huang, K. S., Strobel, S. A. & Steitz, T. A. An induced-fit mechanism to promote peptide bond formation and exclude hydrolysis of peptidyl-tRNA. Nature 438, 520–524 (2005).

Article
 
PubMed
 
Google Scholar

Lapeyre, B. & Purushothaman, S. K. Spb1p-directed formation of Gm2922 in the ribosome catalytic center occurs at a late processing stage. Mol. Cell 16, 663–669 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Lin, J., Gagnon, M. G., Bulkley, D. & Steitz, T. A. Conformational changes of elongation factor G on the ribosome during tRNA translocation. Cell 160, 219–227 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Djumagulov, M. et al. Accuracy mechanism of eukaryotic ribosome translocation. Nature 600, 543–546 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Carbone, C. E. et al. Time-resolved cryo-EM visualizes ribosomal translocation with EF-G and GTP. Nat. Commun. 12, 7236 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Chakraburtty, K. & Kamath, A. Protein synthesis in yeast. Int. J. Biochem. 20, 581–590 (1988).

Article
 
CAS
 
PubMed
 
Google Scholar

Ranjan, N. et al. Yeast translation elongation factor eEF3 promotes late stages of tRNA translocation. EMBO J. 40, e106449 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Qin, S., Xie, A., Bonato, M. & McLaughlin, C. Sequence analysis of the translational elongation factor 3 from Saccharomyces cerevisiae. J. Biol. Chem. 265, 1903–1912 (1990).

Article
 
CAS
 
PubMed
 
Google Scholar

Stark, H., Rodnina, M. V., Wieden, H.-J., Van Heel, M. & Wintermeyer, W. Large-scale movement of elongation factor G and extensive conformational change of the ribosome during translocation. Cell 100, 301–309 (2000).

Article
 
CAS
 
PubMed
 
Google Scholar

Hassan, A. et al. Ratchet, swivel, tilt and roll: a complete description of subunit rotation in the ribosome. Nucleic Acids Res. 51, 919–934 (2023).

Article
 
CAS
 
PubMed
 
Google Scholar

Flis, J. et al. tRNA translocation by the eukaryotic 80S ribosome and the impact of GTP hydrolysis. Cell Rep. 25, 2676–2688 (2018).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Bourne, H. R., Sanders, D. A. & McCormick, F. The GTPase superfamily: conserved structure and molecular mechanism. Nature 349, 117–127 (1991).

Article
 
CAS
 
PubMed
 
Google Scholar

Taylor, D. J. et al. Structures of modified eEF2•80S ribosome complexes reveal the role of GTP hydrolysis in translocation. EMBO J. 26, 2421–2431 (2007).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Kaul, G., Pattan, G. & Rafeequi, T. Eukaryotic elongation factor‐2 (eEF2): its regulation and peptide chain elongation. Cell Biochem. Funct. 29, 227–234 (2011).

Article
 
CAS
 
PubMed
 
Google Scholar

Lucas, B. A. et al. Locating macromolecular assemblies in cells by 2D template matching with cisTEM. eLife 10, e68946 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Wu, C., Huang, X., Cheng, J., Zhu, D. & Zhang, X. High-quality, high-throughput cryo-electron microscopy data collection via beam tilt and astigmatism-free beam-image shift. J. Struct. Biol. 208, 107396 (2019).

Article
 
CAS
 
PubMed
 
Google Scholar

You, X. et al. In situ structure of the red algal phycobilisome–PSII–PSI–LHC megacomplex. Nature 616, 199–206 (2023).

Article
 
CAS
 
PubMed
 
Google Scholar

Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332 (2017).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Rohou, A. & Grigorieff, N. CTFFIND4: fast and accurate defocus estimation from electron micrographs. J. Struct. Biol. 192, 216–221 (2015).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Wu, S. et al. Diverse roles of assembly factors revealed by structures of late nuclear pre-60S ribosomes. Nature 534, 133–137 (2016).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife 7, e42166 (2018).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017).

Article
 
CAS
 
PubMed
 
Google Scholar

Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Ben-Shem, A. et al. The structure of the eukaryotic ribosome at 3.0 Å resolution. Science 334, 1524–1529 (2011).

Article
 
CAS
 
PubMed
 
Google Scholar

Bryant, P., Pozzati, G. & Elofsson, A. Improved prediction of protein–protein interactions using AlphaFold2. Nat. Commun. 13, 1265 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Shao, S. et al. Decoding mammalian ribosome–mRNA states by translational GTPase complexes. Cell 167, 1229–1240 (2016).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Jørgensen, R. et al. Two crystal structures demonstrate large conformational changes in the eukaryotic ribosomal translocase. Nat. Struct. Mol. Biol. 10, 379–385 (2003).

Article
 
Google Scholar

Svetlov, M. S. et al. Structure of Erm-modified 70S ribosome reveals the mechanism of macrolide resistance. Nat. Chem. Biol. 17, 412–420 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D 60, 2126–2132 (2004).

Article
 
PubMed
 
Google Scholar

Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D 66, 213–221 (2010).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Williams, C. J. et al. MolProbity: more and better reference data for improved all‐atom structure validation. Protein Sci. 27, 293–315 (2018).

Article
 
CAS
 
PubMed
 
Google Scholar

Download references

Acknowledgements

We thank L. F. Kong for cryo-EM data storage and backup. The project was funded by the National Natural Science Foundation of China (32150010, 32325027, 31930069 and 32241027), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB37040101), the National Key R&D Program of China (2021YFA1301501), the Key Research Program of Frontier Sciences at the Chinese Academy of Sciences (ZDBS-LY-SM003), the Basic Research Program Based on Major Scientific Infrastructures, Chinese Academy of Sciences (JZHKYPT-2021-05) and the Key Laboratory of Biomacromolecules, Chinese Academy of Sciences (ZGD-2023-05).

Author information
Author notes

These authors contributed equally: Jing Cheng, Chunling Wu, Junxi Li.

Authors and Affiliations

Key Laboratory of Biomacromolecules (CAS), National Laboratory of Biomacromolecules, CAS Center for Excellence in Biomacromolecules, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China

Jing Cheng, Chunling Wu, Junxi Li, Qi Yang, Mingjie Zhao & Xinzheng Zhang

University of Chinese Academy of Sciences, Beijing, China

Junxi Li, Qi Yang, Mingjie Zhao & Xinzheng Zhang

Contributions

X.Z. supervised the project. C.W. and J.L. prepared the S. cerevisiae cells, milled the cells to cellular lamellae using cryo-FIB and collected the cryo-EM data. J.C. performed the data processing. J.C., C.W. and J.L. conducted the data analysis. Q.Y. assisted with milling the lamellae. M.Z. wrote the script for calculating SSU movements. J.L. conducted the model building. X.Z., J.C., C.W. and J.L. wrote the initial draft. All authors contributed to the discussion of the results and revision of the paper.

Corresponding author

Correspondence to Xinzheng Zhang.

Ethics declarations
Competing interests

The authors declare no competing interests.

Peer review
Peer review information

Nature Structural & Molecular Biology thanks Suparna Sanyal and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: Sara Osman, in collaboration with the Nature Structural & Molecular Biology team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data
Extended Data Fig. 1 Image processing workflow for the identification of the eleven main states.

The classification process can be divided into four steps: focused classification on the tRNA path region, focused classification on the eEF2 binding region, focused classification on the E-site tRNA binding region, and focused classification on the 40S subunit. The number of particles, corresponding state number, and resolutions are indicated under each structure.

Extended Data Fig. 2 Exhibitions of the coexistence of eEF3 with eEF2 or eEF1A in different states.

(a) Two sub-states classified from S4 by including eEF3 (S4a, 52%) or not (S4b, 48%). (b) The eEF3 shown in S5, S6, and S7. (c) Three sub-states divided from A/T, P and E state (S8) depending on the presence of eEF3: stable bound to ribosome (S8a), showing flexibility (S8b), and not bound (S8c). The percentages of particles accounted for by each state are listed below the corresponding maps. (d) Four sub-states divided from A/T and P state (S1) depending on the presence of eEF3: S1a (21%) is the state with stably bound eEF3, S1b (39%) has a little flexibility in eEF3, S1c (19%) shows severer flexibility in eEF3 compared to S1b, and S1d (20%) lost connection with eEF3. The 80S, eEF1A, eEF2, and eEF3 are colored in gray, green, purple, and orange, respectively. The tRNAs are colored consistent with Fig. 2.

Source data

Extended Data Fig. 3 Density presentation of the three intermediate states (S2b-d) during peptidyl transfer.

(a) The mesh mode presentation depicts the fitting of the atomic models of S2b onto three maps (S2b-d). The atomic models of the A-site tRNA (carbon atoms of model colored in blue) align well with the S2b and S2d maps; however, the fit is less accurate for A76 and the aminoacyl residue in S2c. (b) Comparison of the maps among three states. The maps of S2b, S2c and S2d are colored in blue, gray and purple, respectively. The densities of the CCA and aminoacyl residues of the A-site tRNA in S2b and S2d overlap significantly, indicating their similarity. In contrast, the density of the CCA in S2c is less well-resolved compared to the other two states, showing a noticeable shift (highlighted by a red circle) in the superimposed maps.

Extended Data Fig. 4 Surroundings of compact and fully extended eEF2 in S2e and S6.

(a) Overview of compact eEF2 bound to ribosomes (gray) in S2e. The densities of eEF2 are colored differently: domain I (green), domain II (blue), domain III (yellow), domain IV (purple), and domain V (magenta). (b) The environment surrounding compact eEF2 (colored by domain) in the ribosome. Gray and light blue densities represent nucleic acids and proteins of LSU, respectively. Orange and pink densities represent nucleic acids and proteins of SSU, respectively. The models of the ribosome are shown in dark gray ribbon. (c) Overview of fully extended eEF2 in S6. (d) The environment surrounding of fully extended eEF2 in the ribosome. The arrangement and color are the same as in (b) with the new densities of tRNA interacting with domain IV of eEF2 are colored in red.

Extended Data Fig. 5 Density comparison of eS25 helix in different ribosomal states.

Left panel: Carton diagram depicting A-site tRNA (blue), P-site tRNA (red) and eS25 (green) in S2a (a), S2e (b) and S2b (c). The yellow ribbon represents mRNA, and densities are shown in transparent gray. Right panel: A close-up view focusing on eS25. The densities of eS25 in S2a (a) and S2e (b) exhibit low quality. In contrast, the helix in S2b (c) shows high continuity with good side chain densities. This helix crosses the P-site tRNA and interacts with the A-site tRNA during peptidyl transfer for comparison.

Extended Data Fig. 6 The interactions of domain V of eEF2 and 25S rRNA in different ribosomal states.

All maps are represented by gray densities under transparency. The models of H95, H43, and H44 of 25S rRNA are highlighted by a cartoon ribbon in light blue, while domain V of eEF2 is colored in blue in compact manner (a), pink in less extended manner (b), and green in fully extended manner (c).

Extended Data Fig. 7 Superimposing models of compact and fully extended eEF2 onto the ribosome in S2c.

(a) Incompatibility of fully extended eEF2 with the 80S ribosome (light yellow) of S2c. The models of fully extended eEF2 were positioned by matching domain V (magenta) with compact eEF2 models. Domain I (green) crashes with the 60S subunit because of steric hindrance. Domains II, III, and IV are colored in blue, yellow, and purple, respectively. (b) Presentation of P stalk movement of 7 Å from S2c (gray) to S4 (red), with compact eEF2 colored in purple and fully extended eEF2 colored in pink.

Extended Data Fig. 8 Atomic models of eEF3 fitted with the corresponding region of EM maps of S6.

(a) The closed-eEF3 models (orange) docking into the EM map (gray) of S6 exhibit a good fit. (b) When replacing the models with the open-eEF3, a significant portion of the models extend beyond the boundaries of the EM map. The EM maps were low-pass filtered to 5 Å.

Extended Data Fig. 9 L1 stalk opens along with the E-site tRNA translocation.

(a) The E-site tRNA undergoes a 14.1 Å movement (left) towards the exit, transitioning from S4 (orange cartoon) to S6 (blue cartoon), as calculated from the translocation of G2467. Correspondingly, the L1 stalk exhibits an opening of ~20 Å (right), with the L1 stalk densities of S4 and S6 represented by orange and blue, respectively. (b) Top row: Transformation of L1 stalk from open to closed in “A/T, P” states obtained by classification on S1 with a focus on the L1 stalk. Bottom row: Corresponding cross-section of the inter-subunit space containing tRNAs. The L1 stalk is colored in pink, and eEF1A is colored in green, A- and P-site tRNAs are highlighted in blue and orange, respectively.

Extended Data Fig. 10 Hypothesis on kinetic of compact eEF2 during elongation cycle.

Top row: Cartoons depicting eEF2 conformational changes. The LSU, SSU, and eEF2 are shown in gray, yellow, and purple respectively. The nascent peptide chain (blue balls) is attached to the P-site tRNA (red), the aminoacyl residue (red balls) is associated with the A-site tRNA (blue). In the absence of accommodation of the A-site tRNA, the CCA, blurred in the PTC, is outlined by a dotted line and shaded in a lighter color. Stable and flexible binding eEF2 are highlighted in purple with a solid outline and light purple with a dotted outline, respectively. Bottom panels: Corresponding EM map.

Supplementary information
Supplementary Information
Reporting Summary
Peer Review File
Source data
Source Data Fig. 2

Statistical source data.

Source Data Extended Data Fig. 2

Particle number, percentage and resolution of states.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article
Cite this article

Cheng, J., Wu, C., Li, J. et al. Capturing eukaryotic ribosome dynamics in situ at high resolution. Nat Struct Mol Biol 32, 698–708 (2025). https://doi.org/10.1038/s41594-024-01454-9

Download citation

Received
04 August 2023

Accepted
19 November 2024

Published
09 January 2025

Version of record
09 January 2025

Issue date
April 2025

DOI
https://doi.org/10.1038/s41594-024-01454-9

Share this article

Anyone you share the following link with will be able to read this content:

Get shareable link

Provided by the Springer Nature SharedIt content-sharing initiative

Subjects
Cryoelectron microscopy
Ribosome
This article is cited by
Integrated structural biology of the native malarial translation machinery and its inhibition by an antimalarial drug
Leonie AntonWenjing ChengChi-Min Ho

Nature Structural & Molecular Biology (2025)

Decreased cytoplasmic crowding via inhibition of ribosome biogenesis can trigger Candida albicans filamentous growth
Antonio SerranoCharles PuernerRobert Alan Arkowitz

Nature Microbiology (2025)

Cancel

Preview · live