In extracto cryo-EM reveals eEF2 as a major hibernation factor on 60S and 80S particles
- RNA Therapeutics Institute, UMass Chan Medical School, United States
- Howard Hughes Medical Institute, UMass Chan Medical School, United States
Figures
Figure 1 with 2 supplements
In extracto cryogenic electron microscopy (cryo-EM) workflow for sample preparation and data collection.
(A) Sample and grid preparation from primate semi-permeabilized cells. Examples of an MCF-7 cell culture (left) and a grid with cell lysate (right) are shown. (B) Representative lysate micrograph (left) and two-dimensional template matching (2DTM) processing yielding positions and orientations of 60S subunits (right). (C) Examples of averaged maps (prior to classification) from typical single-particle picking pipelines (cyan) and from 2DTM processing (middle and gray). The close-up view highlights densities of 28S rRNA nucleotides (gray) and ribosomal protein residues (cyan). Also see Figure 2—figure supplement 1.
Figure 1—figure supplement 1
Examples of micrographs showing cellular context in BSC-1 lysates.
Scale bar for all micrographs is 50 nm.
Figure 1—figure supplement 2
Data-processing workflow.
Micrographs were binned using in-house Python script and then contrast transfer function (CTF) parameters were estimated from 2× binned micrographs using cisTEM, followed by two-dimensional template matching (2DTM) on binned data. To prepare the templates, we initially used both e2pdb2mrc.py and cisTEM’s simulate program to generate templates (see Methods). Subsequently, we switched to exclusively using the simulate program since it performed slightly better. Our goal in determining and optimizing search parameters for 2DTM was a balance between particle search precision and computational cost (see Methods). In practice, users must weigh the benefit of finer sampling against the substantial increase in runtime, particularly for large datasets. In some cases, the defocus search was turned off, and it was sufficient to use in-plane and out-of-plane angular steps of approximately 4.5° and 3.5°, respectively, and binned images with ~2 Å/pixel, to still obtain enough matches to obtain high-resolution reconstructions. To determine when defocus search was needed, micrographs were selected based on their fit resolution obtained during CTF estimation. From each dataset, we selected 10 micrographs representing the highest and lowest fit resolutions. Template matching was then performed using identical parameters, once with defocus search enabled and once with it disabled. The number of detected particles for each micrograph under both conditions was compared. When a significant difference was observed, most commonly for icy micrographs with low fit resolution, we enabled defocus search for that group of images. We found that these images/datasets appeared to have a higher background compared to in vitro reconstituted samples but less than in situ samples. The template matching results from these micrographs were subsequently combined with results from groups processed with defocus search disabled. After 2DTM, .star files were generated using cisTEM. The pixel size and directory to un-binned images were changed in this .star file, now called edited.star. A dummy project with a small group of micrographs was made and after performing 2DTM and generating a dummy .star file, the dummy .star file was substituted with edited .star. Then, we generated the refinement package, corresponding particle stacks, and .par files. Un-binning refers to generating particle stacks from the original micrographs at the original pixel size. Particle stacks and .par files were exported from cisTEM, and to speed up processing, binned image stacks (e.g. 2×, 4×, or 8×) were prepared using resample .exe, which is part of the Frealign v9.11 distribution. Initial 3D classification was carried out in Frealign v9.11. For each resulting class, per-class particle stacks and .star files were generated, and subsequent rounds of focused or global 3D classification were performed (see Methods). Software packages used at each step are indicated by color coding.
Figure 2 with 3 supplements
Ribosome and 60S particle distributions in control MCF-7 and starved MCF-7 cells.
(A–H) Eight cryogenic electron microscopy (cryo-EM) maps correspond to elongating ribosomes (A–E): codon sampling with eEF1A and A/T tRNA (A), non-rotated with A/A, P/P, E/E tRNAs and putative extended eEF1A (B), non-rotated pre-translocation with A/A, P/P, E/E tRNAs (C), rotated pre-translocation with hybrid-state A/P and P/E tRNAs (D), and post-translocation ribosome with eEF2, P, and E tRNAs (E); hibernating rotated ribosome with eEF2, SERBP1, and P/E tRNA (F); 60S subunits with eEF2 and E-tRNA (G); and vacant 60S subunits (H). (I–P) Eight cryo-EM maps corresponding to nutrient-deprived MCF-7 cell lysates comprise elongation states (I–L) similar to those in panels A–D; hibernating head-swiveled state with eEF2, CCDC124, pe/E tRNA, and putative La-related protein 1 (LARP1) (M), hibernating rotated ribosome with eEF2, SERBP1, and P/E tRNA (N); 60S with eEF2 and E-tRNA (O); 60S with eIF6 (P). (Q) Particle distributions among functional ribosome states in normal and nutrient-deprived MCF-7 cells. (R) Putative extended eEF1A conformation; the close-up box shows density with the extended eEF1A model fitted (green; PDB ID:8B6Z; Gemmer et al., 2023).
Figure 2—figure supplement 1
Maximum-likelihood three-dimensional (3D) classification of the dataset collected from BSC-1 cell extracts.
Classification was performed on 8× binned stacks with particle alignment disabled.
Figure 2—figure supplement 2
Maximum-likelihood three-dimensional (3D) classification of the nutrient-deprived MCF-7 extract.
Classification was performed on 8× and 4× binned stacks with particle alignment disabled.
Figure 2—figure supplement 3
Maximum-likelihood three-dimensional (3D) classification of the MCF-7 extract cryo-EM data.
Classification was performed on 8× and 4× binned stacks with particle alignment disabled.
Figure 3 with 5 supplements
In extracto cryogenic electron microscopy (cryo-EM) of rabbit reticulocyte lysates.
(A) NLuc mRNA construct used to monitor NanoLuciferase translation (top) and real-time translation kinetics (bottom). The graph represents the mean ± SEM from three independent experiments. (B–L) Cryo-EM maps of the major classes resulting from three-dimensional (3D) classification: (B) initiation complex with eIF5B; (C–F) elongating ribosomes: codon-sampling with eEF1A (C), post-translocation with eEF2 ap/P and pe/E tRNAs (D), pre-translocation hybrid state (E), and with GTPase density putatively assigned to DRG1/DRG2 next to A/A tRNA (F). (G–J) Hibernating ribosomes: 40S-rotated ribosome with eEF2, eIF5A, and SERBP1 (G); 40S-head-swiveled ribosome with eEF2, CCDC124, and LARP1 in the mRNA tunnel (H); 40S head-swiveled with eEF2, IFRD2, and LARP1 in the mRNA tunnel (I); 40S-head-swiveled with IFRD2 and LARP1 (J); 60S subunit with eEF2 domain IV open (K); 60S with eEF2 domain IV closed (L): the extent of domain IV movement in 60S-bound eEF2.
Figure 3—figure supplement 1
Ribosome classes feature EBP1, the Sec translocon, or unidentified density (UID) next to the polypeptide exit tunnel.
(A) Representative classes obtained by three-dimensional (3D) classifications with a mask covering the binding site of EBP1 yield 80S predominantly bound with EBP1 in MCF-7 and rabbit reticulocyte lysate (RRL), and additional classes consistent with the Sec translocon or unidentified (likely heterogeneous) densities in BSC-1 lysates. The bar graph shows particle percentages corresponding to different classes in BSC-1. (B) A close-up view of EBP1 in RRL 80S ribosomes with eEF2, eIF5A, and SERBP1. (C) A close-up view of membrane-bound ribosomes in BSC-1, with the structural models of Sec61, translocon-associated protein (TRAP), and signal sequence receptor subunit 3 (SSR3) shown (PDB ID 8BTK; Jaskolowski et al., 2023).
Figure 3—figure supplement 2
Maximum-likelihood three-dimensional (3D) classification of rabbit reticulocyte lysate (RRL) datasets with or without nanoluciferase-encoding mRNA.
Classification was performed on 8× binned stacks with particle alignment disabled.
Figure 3—figure supplement 3
RRL dataset cryo-EM map statistics and a reconstruction example.
(A) Fourier shell correlation (FSC) curves for rabbit reticulocyte lysate (RRL)-derived maps (masked), as a function of inverse resolution. (B) RRL 80S ribosomes with putative DRG1 (or DRG2), eIF5A in the E site, and tRNAs in the A and P sites.
Figure 3—figure supplement 4
60S subunit polypeptide tunnel is occupied with an uncharacterized protein resembling Rei1 (rabbit reticulocyte lysate [RRL]).
(A) Cryogenic electron microscopy (cryo-EM) map (mesh) and model of 60S with eEF2 open. Density corresponding to the tunnel protein is shown in gray. (B) Cryo-EM map (mesh) and model of 60S with eEF2 closed. (C) Superposition of the yeast 60S biogenesis factor Rei1 (PDB ID 5APN; Greber et al., 2016) with the tunnel density (gray). Rei1 is shown in red, and the rest of the yeast 60S model is in gray. Cryo-EM maps were softened with the B-factor of 50 Å2. 60S structure superpositions were performed in ChimeraX. (D) Superposition of the mammalian 60S biogenesis factor ZNF622 (PDB ID 9GMO; Akers et al., 2025) with the unidentified tunnel density (gray). ZNF622 is shown in purple, and the rest of the superposed 60S model is in gray.
Figure 3—figure supplement 5
The linker helix (aa 402–412) of IFRD2 adopts different conformations in the presence and in the absence of eEF2.
(A, B) Cryogenic electron microscopy (cryo-EM) maps (mesh) and models of 80S with LARP1 and IFRD2, in the presence of eEF2 (A) and in the absence of eEF2 (B). (C) Superposition of models from panels A and B. IFRD2 in the presence of eEF2 is shown in magenta, and in the absence of eEF2 is shown in purple. The steric clash between the IFRD2 linker helix and domain IV of eEF2 is indicated by a circle.
Figure 4 with 2 supplements
The occupancy of the mRNA tunnel in hibernating ribosomes from rabbit reticulocyte lysate (RRL).
(A) The 80S ribosome with eEF2, eIF5A, and SERBP1 (DC: decoding center). 40S body and head rotation degrees are shown relative to 80S with eEF1A•aa-tRNA and P-tRNA (PDB 5LZS). (B) Close-up view of cryogenic electron microscopy (cryo-EM) density for SERBP1 in the mRNA tunnel. (C) Density for SERBP1 interaction with eEF2. (D) The 80S ribosome with eEF2, IFRD2, and LARP1. (E) Close-up view of LARP1 density in the mRNA tunnel. (F–G) Views of the density for the LARP1 helix interacting with 40S ribosomal proteins and RNA. (H–J) Similar densities for LARP1 in CCDC124- and IFRD2-bound ribosomes.
Figure 4—figure supplement 1
N-terminal tail of uL14 in the presence of eEF2.
(A) Structure with eIF5A, eEF2, and SERBP1 (this work, rabbit reticulocyte lysate [RRL]). (B) Gallus gallus eEF2 bound to translocation-like ribosomes (PDB ID 8Q87 Nurullina et al., 2024).
Figure 4—figure supplement 2
Interaction of eEF2 with the isolated 60S subunit (rabbit reticulocyte lysate [RRL]).
(A) eEF2 in a compact conformation (relative to the second, more open conformation observed in this work). (B) Interaction with the sarcin-ricin loop (SRL) and uL14. The cryogenic electron microscopy (cryo-EM) map (mesh) was softened with a B-factor of 25 Å2. (C) Domain IV of eEF2 binds next to helix 69 of 28S rRNA. The map was softened with a B-factor of 30 Å2.
Figure 5
Interactions of eEF2 with the GTPase-activating center in hibernating ribosomes.
(A) Overall view of the 80S structure with eEF2, eIF5A, and SERBP1. (B) Cryogenic electron microscopy (cryo-EM) density and model of the eEF2 GTPase center at the sarcin-ricin loop (SRL). (C) GDP density in the GTPase center of the predominant hibernating ribosomes. (D, E) Distinct conformations of the N-terminal tail of uL14, highlighted in red, in ribosomes bound with eEF2•eIF5A•SERBP1 (D) and with eEF1A•GDP (PDB ID, 5LZS) (E).
Figure 6
Binding sites of hibernation factors overlap with ribosomal functional centers.
(A–D) Comparison of a translating ribosome (PDB 5LZS and mRNA from PDB 4V6F Jenner et al., 2010; panel A) with hibernating ribosomes identified in this work. (E–G) Superposition of translating and hibernating ribosomes illustrates that the key ribosomal functional centers are all shielded by hibernation factors (eEF1A, A/T-tRNA, P-tRNA, and E-tRNA are from PDB 5LZS and mRNA from PDB 4V6F). Panel E shows the view from panels B–D rotated by 36°.
Tables
Table 1
Comparison of in extracto cryogenic electron microscopy (cryo-EM) with traditional in vitro and in situ cryogenic electron tomography (cryo-ET) approaches to structural biology.
| Cryo-EM method Features | In vitro (traditional single particle) | In extracto (this work) | In situ (e.g. cryo-ET) |
|---|---|---|---|
| Interactions with cellular components | – * | +/– | + |
| Purified macromolecules | + | +/– (can be added) | – |
| Time-resolved cryo-EM | + | + | – |
| Fast sample and grid preparation, and data collection to enable near-atomic resolution | + (hours to an overnight session) | + (1–3 overnight sessions) | – (days/weeks) |
| Efficient particle identification (low # of false positives) | + | + | +/– |
-
*+ Means straightforward, – difficult/impossible, +/– possible.
Additional files
-
Supplementary file 1
Tables A-E.
(Table A) Particle picking strategies for the RRL+angiogenin dataset. (Table B) Ribosome states observed in lysates from different cell types. (Table C) Rotation of the 40S small ribosomal subunit head or body relative to non-rotated 80S ribosomes: with an accommodation-like eEF1A•aa-tRNA (PDB ID 5LZS; Shao et al., 2016) or with tRNAs in three classical sites (this work). Negative numbers report on the ‘underrotated’ subunits, i.e., rotated in the direction opposite to that measured from non-rotated (classical) to rotated (hybrid-state) pre-translocation ribosomes. (Table D) Cryo-EM data collection and model refinement statistics for 80S and 60S complexes from RRL. (Table E) Percentages of particles corresponding to 80S or 60S classes in RRL with NLuc mRNA and in RRL without NLuc mRNA.
- https://cdn.elifesciences.org/articles/110114/elife-110114-supp1-v1.docx
-
MDAR checklist
- https://cdn.elifesciences.org/articles/110114/elife-110114-mdarchecklist1-v1.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
In extracto cryo-EM reveals eEF2 as a major hibernation factor on 60S and 80S particles
eLife 15:RP110114.
https://doi.org/10.7554/eLife.110114.3
