1. Structural Biology and Molecular Biophysics

Structures of translationally inactive mammalian ribosomes

  1. Alan Brown  Is a corresponding author
  2. Matthew R Baird
  3. Matthew CJ Yip
  4. Jason Murray
  5. Sichen Shao  Is a corresponding author
  1. Harvard Medical School, United States
  2. MRC Laboratory of Molecular Biology, United Kingdom
https://doi.org/10.7554/eLife.40486
Share this article
Cite this article
  1. Alan Brown
  2. Matthew R Baird
  3. Matthew CJ Yip
  4. Jason Murray
  5. Sichen Shao
(2018)
Structures of translationally inactive mammalian ribosomes
eLife 7:e40486.
https://doi.org/10.7554/eLife.40486
  2. Download BibTeX
  3. Download .RIS

Abstract

The cellular levels and activities of ribosomes directly regulate gene expression during numerous physiological processes. The mechanisms that globally repress translation are incompletely understood. Here, we use electron cryomicroscopy to analyze inactive ribosomes isolated from mammalian reticulocytes, the penultimate stage of red blood cell differentiation. We identify two types of ribosomes that are translationally repressed by protein interactions. The first comprises ribosomes sequestered with elongation factor 2 (eEF2) by SERPINE mRNA binding protein 1 (SERBP1) occupying the ribosomal mRNA entrance channel. The second type are translationally repressed by a novel ribosome-binding protein, interferon-related developmental regulator 2 (IFRD2), which spans the P and E sites and inserts a C-terminal helix into the mRNA exit channel to preclude translation. IFRD2 binds ribosomes with a tRNA occupying a noncanonical binding site, the ‘Z site’, on the ribosome. These structures provide functional insights into how ribosomal interactions may suppress translation to regulate gene expression.

https://doi.org/10.7554/eLife.40486.001

Introduction

Translation is an important point of regulation for gene expression. The overall levels and activities of ribosomes are implicated in cellular differentiation, developmental disorders, and cancers (Buszczak et al., 2014; Narla and Ebert, 2010). Red blood cell differentiation and function is especially sensitive to ribosome concentrations and activity. Humans produce over two million red blood cells every second (Dzierzak and Philipsen, 2013), with hemoglobin comprising 98% of total protein content of each cell. To achieve this, translation is optimized for hemoglobin production during red blood cell differentiation (Mills et al., 2016; Smith and McNamara, 1971). This translational program occurs predominantly in enucleated cells, precluding transcriptional control and relying solely on preexisting ribosomes and translational factors before the ribosomes are degraded upon terminal differentiation (Rowley, 1965).

Underscoring the importance of ribosomal activity in blood cell differentiation, mutations in ribosomal proteins, ribosome-assembly factors, and the ribosome-degradation machinery manifest as blood disorders (Draptchinskaia et al., 1999; Ebert et al., 2008) and are often found in cancers of blood cells (De Keersmaecker et al., 2013; Ljungström et al., 2016). A recent study demonstrated that mutations that cause Diamond-Blackfan anemia alter global ribosome levels, which dictates the mRNAs that are translated (Khajuria et al., 2018). Similarly, mutations in the ubiquitination machinery required to eliminate ribosomes during terminal red blood cell differentiation results in anemia in mice (Nguyen et al., 2017).

Although controlling absolute ribosome levels is an effective mechanism to regulate gene expression, assembling and degrading ribosomes is slow and requires significant energy consumption (Kressler et al., 2010; Warner, 1999). For faster responses, two mechanisms are known to acutely and reversibly repress global translation in eukaryotes. The first is phosphorylation of translation initiation factor eIF2α, which dampens protein synthesis by preventing the formation of productive translation initiation complexes (Wek et al., 2006). In mammals, four distinct kinases phosphorylate eIF2α in response to different cellular stresses, including nutrient deprivation, endoplasmic reticulum (ER) stress, viral infections, and heme deprivation (Chen, 2014). The second mechanism is translational repression by the yeast Stm1 protein or its mammalian homolog, SERPINE mRNA-binding protein 1 (SERBP1). Stm1 directly binds and sequesters 80S ribosomes (Balagopal and Parker, 2011; Van Dyke et al., 2006). This function of Stm1 is thought to preserve ribosomes during nutrient deprivation (Van Dyke et al., 2013) and may influence the degradation of some mRNAs (Balagopal and Parker, 2011).

Despite emerging evidence of the importance of global translation levels on short- and long-term gene regulation, other mechanisms for modulating ribosome activities remain poorly characterized. Here, we use electron cryomicroscopy (cryo-EM) to analyze translationally inactive ribosomes isolated from mammalian reticulocytes, the penultimate stage of red blood cell differentiation (Dzierzak and Philipsen, 2013). In addition to observing ribosomes silenced by SERBP1, we identify interferon-related developmental regulator 2 (IFRD2) as a novel factor capable of translationally inactivating ribosomes. IFRD2 directly binds 80S ribosomes, precluding binding of mRNA and P-site tRNA. We additionally observe deacylated tRNA occupying a noncanonical binding site on the mammalian ribosome that is exploited by viral internal ribosome entry site (IRES)-mediated translation. We observe tRNA in this site in the presence and absence of either IFRD2 and P-site tRNA. Together, these findings identify new ribosomal interactions that may modulate global translation activity during erythropoiesis and in other differentiating cells.

Results

Identification of translationally inactive ribosomes in reticulocyte lysate

We previously used a cell-free translation system derived from rabbit reticulocyte lysate to isolate various ribosomal complexes for cryo-EM analysis (Brown et al., 2015b; Shao et al., 2016). Despite biochemical enrichment for specific complexes, the target structures generally derived from 10% to 20% of the particles within each cryo-EM dataset. The remaining particles are various translation intermediates and inactive ribosomes that copurify with the target complex. To characterize these reticulocyte ribosomal populations, we reanalyzed a combined cryo-EM dataset of 971,191 ribosomal particles purified from in vitro translation reactions containing a dominant-negative release factor (DN-eRF1) added to enrich for translational termination complexes (Brown et al., 2015b). These datasets were imaged under identical conditions, with the only difference being the identity of the stop codon in the mRNA used to program the in vitro translation reactions. As all stop codons adopt the same structural fold (Brown et al., 2015b), these samples can be considered biochemically and structurally analogous.

Three-dimensional classification of the combined dataset of ribosomal particles identified nine structurally distinct classes of 80S ribosomes (Figure 1; Table 1). These included the target termination complex and known intermediates of translation. Notably, we did not observe any empty 80S ribosomes lacking a bound tRNA or protein factor. Four classes represented non-translating ribosomes based on the absence of canonical tRNAs. These structures fall into three types of translationally inactive 80S ribosomes (1) bound to eEF2 and SERBP1, (2) bound to IFRD2, and (3) containing a tRNA in a noncanonical binding site.

Figure 1 with 1 supplement see all
Download asset Open asset
Identification of translationally inactive ribosomal complexes in a cryo-EM dataset.

In silico classification of a cryo-EM dataset of ribosomal particles reveals known intermediates of translation and translationally inactive 80S ribosomes. The overall resolution and number of particles that make up the reconstruction are shown for each complex.

https://doi.org/10.7554/eLife.40486.002

SERBP1 traps eEF2 on different conformations of inactive ribosomes

The two most abundant classes of non-translating ribosomes are distinct conformations (rotated and unrotated) of 80S ribosomes bound to the elongation factor eEF2 (Figure 2—figure supplement 1A). eEF2 normally catalyzes the translocation of mRNA and peptidyl-tRNA during translation, although its association with inactive ribosomes is well established (Liu and Qian, 2016). Closer inspection reveals that all maps of non-translating ribosomes with eEF2 in our dataset also contain SERBP1 in the mRNA entrance channel (Figure 2A,B). It is possible that eEF2 requires SERBP1 to form stable interactions with inactive ribosomes. Our map resolution (3.4 Å) allows the reassignment of SERBP1 residues from a 5.4 Å-resolution cryo-EM structure of eEF2-SERBP1 on the human ribosome (Anger et al., 2013) to produce a model that is more consistent with the structure of Stm1 on the yeast ribosome (Ben-Shem et al., 2011) and cross-linking data (Liu et al., 2015).

Figure 2 with 2 supplements see all
Download asset Open asset
SERBP1 sequesters eEF2 on inactive ribosomes.

Comparison of (A) SERBP1 (SERPINE1 mRNA-binding protein 1) and eEF2 (eukaryotic elongation factor 2) bound to the rabbit ribosome with (B) Stm1 bound to the yeast ribosome (PDB 4V88). (C) Path of SERBP1 and (D) Stm1 through the mRNA entrance channel. (E) Interaction between SERBP1 and eEF2.

https://doi.org/10.7554/eLife.40486.004

SERBP1 and Stm1 interact with the A and P sites of the ribosomal mRNA channel via conserved residues (Figure 2—figure supplement 2), including a DRHS motif (residues 194–197 in SERBP1) that forms a 310-helix around nucleotide C1701 of 18S rRNA (Figure 2C,D; Figure 2—figure supplement 1B). The density of SERBP1 is less clear following residue 201 in the unrotated ribosome (Figure 2—figure supplement 1C), potentially the result of averaging splice variants that differ at this site (Figure 2—figure supplement 2). The C-termini of SERBP1 and Stm1 emerge from the mRNA channel to interact with eS10, eS12 and eS31, although the density is only sufficiently resolved to model a short helix of SERBP1 (Figure 2—figure supplement 1D). Unlike Stm1, we observe no interaction between the N terminus of SERBP1 and the ribosomal large subunit.

At the A site, SERBP1 residues 198 – 201 interact with domain IV of eEF2 (Figure 2E, Figure 2—figure supplement 1E). This small interface is unlikely to physically anchor eEF2 to the ribosome, and the presence of GDP in the nucleotide-binding pocket of eEF2 (Figure 2—figure supplement 1F) despite differences in the switch-loop conformations (Figure 2—figure supplement 1G) indicates that SERBP1 does not impair GTP hydrolysis by eEF2. The exact mechanism for how SERBP1 is able to trap eEF2 on the ribosome is unclear, although there are similarities with the eEF2-inhibitor sordarin (Pellegrino et al., 2018). Like SERBP1, sordarin traps eEF2 on the yeast ribosome without impairing GTP hydrolysis or inducing large-scale structural changes in eEF2 (Pellegrino et al., 2018). Instead, sordarin binds in a pocket between domains III, IV, and V, and increases interdomain contacts to subtly constrain the domains in conformations that prevent release. SERBP1 may also constrain domain rearrangements necessary for release through its interaction with eEF2 domain IV. Thus, SERBP1 simultaneously removes a ribosome and an abundant translational GTPase from active translation.

IFRD2 binds the ribosomal core and is incompatible with translation

The second most abundant subset of translationally inactive 80S ribosomes contain IFRD2 in the intersubunit space of a nonrotated ribosome, where it occludes tRNA binding to the P and E sites (Figure 3A) and mRNA binding to the mRNA channel. We identified IFRD2 by comparing the predominantly α-helical density with the secondary structure profiles of approximately 150 candidates identified by mass spectrometry of the cryo-EM sample (Supplementary file 1). IFRD2 matches the number and length of the helices present in the map and has a sequence that fits regions of the map with well-defined density (Figure 3—figure supplement 1A,B). Supporting this assignment, both IFRD2 and its paralog IFRD1 co-immunoprecipitate with endogenously tagged ribosomes from mouse embryonic stem cells (Simsek et al., 2017), although IFRD1 could be excluded here based on side-chain density (Figure 3—figure supplement 1B). IFRD proteins are generally described as transcriptional regulators implicated in cellular differentiation and various human diseases (Vietor et al., 2002), although their naming as ‘interferon-related’ is apparently due to a mistaken sequence similarity with mouse interferon-β (Tirone and Shooter, 1989). Our structure suggests that IFRD proteins regulate translation instead of, or in addition to, transcription.

Figure 3 with 3 supplements see all
Download asset Open asset
IFRD2 binding to ribosomes is incompatible with translation.

(A) IFRD2 (interferon-related developmental regulator 2) occupies the intersubunit space in a subset of ribosomes containing Z-site tRNA. (B) Model of IFRD2 colored by domain. (C) IFRD2 interacts with many of the same ribosomal elements as P-site tRNA. (D) The C-terminal α-helix of IFRD2 occupies the mRNA exit channel and follows the path taken by mRNA through the ribosome. (E) 200 nM ribosomes isolated from rabbit reticulocyte lysate were treated without or with 50 μg/mL RNase A or 10 mM EDTA for 5 min and subjected to native size fractionation on 10 – 50% sucrose gradients. Eleven fractions taken from the top were collected and subjected to immunoblotting for IFRD2. Absorbance readings at 254 nm are shown below.

https://doi.org/10.7554/eLife.40486.007
Table 1
Structurally distinct classes identified within a cryo-EM dataset of 971,191 ribosomal particles.
https://doi.org/10.7554/eLife.40486.011
Ribosomal
complex		Resolution (Å)Particles% of particlesEMDB
accession code
80S classes (unrotated)
80S • P5.921010.29234
80S • P • E3.823,9452.59235
80S • P • E • eRF1 • ABCE13.683,6828.6-
80S • Z4.210,7831.19236
80S • Z • P3.662,5606.49237
80S • Z • IFRD23.474,0317.69239
80S • eEF2 • SERBP1
(head swivel)		3.3146,80115.19240
80S classes (rotated)
80S • A/P • P/E3.833,6643.59241
80S • eEF2 • SERBP13.4133,48013.79242
Table 2
Model statistics.
https://doi.org/10.7554/eLife.40486.013
80S • Z • P80S • Z • IFRD280S • eEF2 • SERBP1
(head swivel)		80S • eEF2 • SERBP1
(rotated)
Model composition
 Protein residues11,54711,90112,76012,591
 RNA bases (+modified bases)5686 (40)5603 (26)5534 (36)5528 (40)
 Ligands (Zn2+/Mg2+)8/2978/3118/2978/298
Refinement
 Resolution (Å)3.63.43.33.4
 Map sharpening B factor (Å2)−101.9−95.5−97.1−95.1
 Average B factor (Å2)75.651.884.186.1
 Correlation coefficient volume mask0.844
0.865		0.825
0.853		0.847
0.862		0.847
0.862
Rms deviations
 Bond lengths (Å)0.0100.0090.0070.008
 Bond angles (°)1.21.11.11.0
Validation (proteins)
 MolProbity score1.791.641.681.71
 Clashscore, all atoms5.85.05.65.7
 Rotamer outliers (%)1.30.80.90.9
 EMRinger score2.873.242.862.97
Ramachandran plot
 Favored (%)94.094.394.594.0
 Outliers (%)0.10.10.10.1
 Validation (RNA)
 Probably wrong sugar puckers (%)1.61.51.61.4
 Bad backbone conformations (%)23.623.622.822.0
 PDB6MTB6MTC6MTD6MTE

The core of IFRD2 (residues 71 – 342), which follows an unstructured N-terminus absent in our reconstruction, adopts an Armadillo-type fold formed by six HEAT repeats (Figure 3B). Each HEAT repeat contains two α-helices linked by a short loop that together form a curved α-solenoid. A small globular domain of IFRD2 (residues 343 – 400) binds the concave surface of the α-solenoid and contacts each repeat. These core domains occupy the same space as a P-site tRNA and interact with many of the same ribosomal elements (Figure 3C). For example, an α-helix of IFRD2 (residues 375–384) contacts H69 of 28S rRNA at the same place as the D stem of a P-site tRNA, and IFRD2 interactions with uL5 and uL13 are reminiscent of those made by the P-site tRNA acceptor arm (Figure 3—figure supplement 1C). IFRD2 also interacts with the ribosomal small subunit through a C-terminal tail containing an α-helix that protrudes into the mRNA exit channel (Figure 3D, Figure 3—figure supplement 1D). Its path through the channel follows that of mRNA: it starts at the P site, extends through the E site, and ends by contacting ribosomal proteins uS11 and eS26 at the channel exit. The α-helix forms multiple electrostatic interactions with the rRNA phosphate backbone that lines the channel via highly conserved residues (Figure 3—figure supplement 2), suggesting that other IFRD proteins bind the ribosome with a similar insertion.

These interactions indicate that IFRD2 engages non-translating ribosomes with vacant mRNA channels and precludes further translation. Although a previous large-scale mass spectrometry study suggested that the ribosomal interaction of IFRD1 and IFRD2 are sensitive to RNase treatment (Simsek et al., 2017), treating reticulocyte lysate or ribosomes with RNase collapses mRNA-dependent polysomes but does not disrupt the stoichiometric association of endogenous IFRD2 with 80S ribosomes (Figure 3E; Figure 3—figure supplement 3A). This is consistent with our structure showing that IFRD2 binds to core ribosomal features independently of mRNA and tRNAs.

The observation that almost 8% of particles in the dataset contain IFRD2 suggests that it is highly abundant in reticulocytes. Indeed, IFRD2 appears to be ~8 fold less abundant on human HEK293T ribosomes compared to reticulocyte ribosomes (Figure 3—figure supplement 3B). Overexpressed Flag-tagged human IFRD2 in HEK293T cells also associates and copurifies with 80S ribosomes independent of RNase treatment (Figure 3—figure supplement 3C–E), further supporting the specific association of IFRD2 with ribosomes. We collected a small cryo-EM dataset of affinity purified IFRD2-containing ribosomes from HEK293T cells, which confirmed that IFRD2 binds human ribosomes identically as those observed in the reticulocyte sample (Figure 3—figure supplement 3F).

A new tRNA-binding site on the mammalian ribosome

All ribosomes bound to IFRD2 in the reticulocyte dataset also contain a tRNA in an extreme position on the ribosome past the E site that we call the ‘Z site’ (Figure 3A). Although we could not distinguish any IFRD2-containing reticulocyte ribosomes lacking Z-site tRNA, we do not observe Z-site tRNA in the reconstruction of IFRD2 on the human ribosome (Figure 3—figure supplement 3F), indicating that Z-site tRNA is not required for IFRD2 binding. In the combined dataset of reticulocyte ribosomes, Z-site tRNA also binds 80S ribosomes alone and with peptidyl-tRNA (Figure 1) but would clash with an E-site tRNA (Figure 4A).

Figure 4
Download asset Open asset
Z-site tRNA.

(A) Position of Z-site tRNA on the ribosome in the presence of a P-site tRNA. Docked E-site tRNA (shown in cartoon representation) is incompatible with a Z-site tRNA. (B) Interactions between the Z-site tRNA (CCA nucleotides shown in stick representation) and the 60S ribosomal subunit. (C) Interaction between the anticodon stem-loop (anticodon nucleotides shown in stick representation) of Z-site tRNA and eS25. (D) Cryo-EM map of CrPV IRES (EMD-2599), with the L1.1 and stem-loop V (SLV) domains, which make similar interactions as Z-site tRNA, indicated, bound to the Kluyveromyces lactis ribosome. LSU – large ribosomal subunit; SSU – small ribosomal subunit; IRES – internal ribosome entry site; SLV – stem-loop V.

https://doi.org/10.7554/eLife.40486.012

Like the universal tRNA-binding positions, Z-site tRNA bridges the ribosomal subunits. Three distinct interactions are observed between the tRNA and the large subunit. First, the backbone of the 3’ CCA binds a lysine-rich stretch of eL42 (residues 27 – 30) (Figure 4B). Second, the acceptor arm binds the backbone of helix 68 of 28S rRNA (nucleotides 3730 – 3732), and third, the elbow of the Z-site tRNA interacts with the L1 stalk (Figure 4B). At the head of the 40S subunit, the anticodon stem-loop of the Z-site tRNA binds a positively charged site on the non-essential ribosomal protein, eS25 (Figure 4C). Although eS25 has not been shown to bind tRNAs previously, it is a common binding site for internal ribosome entry site (IRES) sequences utilized by many viruses, including human immunodeficiency virus (HIV) and the hepatitis C virus (Quade et al., 2015; Yamamoto et al., 2015), to initiate translation of viral RNA on host ribosomes. The Z-site tRNA most closely resembles the ribosome-binding conformation of the cricket paralysis virus (CPrV) IRES (Fernández et al., 2014), which also bridges eS25 and the ribosomal L1 stalk (Figure 4D). Thus, viruses appear to exploit the Z site for IRES-mediated translation.

Discussion

Our structural analysis of reticulocyte ribosomes in a large cryo-EM dataset has revealed new insights into ribosome inactivation in mammals. We identify new similarities between yeast Stm1 and mammalian SERBP1, reveal the interactions that sequester eEF2 on the ribosome, and redefine the IFRD family as ribosome-silencing proteins.

As proposed for Stm1, IFRD2 may repress translation during cellular stress. Previous studies indicate that IFRD2 expression is induced in specific circumstances, such as tetradecanoyl phorbol acetate treatment and serum replenishment after starvation (Varnum et al., 1989). In addition, IFRD proteins appear to escape translational repression by stress-induced eIF2α phosphorylation via an upstream open-reading frame (uORF) (Andreev et al., 2015; Zhao et al., 2010). IFRD proteins may also provide a general mechanism to reduce translational output, which may explain the pleiotropic effects of IFRD ablation on the differentiation of various cell types (Vietor et al., 2002).

Our results suggest that IFRD2 is especially abundant on ribosomes in reticulocytes, the immediate precursors to red blood cells. Reticulocytes also express enhanced levels of other factors that can suppress protein synthesis, including the heme-dependent eIF2α kinase HRI (Crosby et al., 1994). Considering the importance of ribosomal levels (Khajuria et al., 2018) and global translational regulation (Grevet et al., 2018) on red blood cell differentiation and function, IFRD2 may offer an additional layer of translational control to dictate blood cell fate.

While SERBP1 and IFRD2 may act redundantly to inactivate mammalian ribosomes, our structures suggest they target different subsets of ribosomes and may induce different fates. In yeast, Stm1-inactivated ribosomes are recycled after starvation by Dom34-Hbs1 (Pelota-Hbs1l in mammals) and Rli1 (ABCE1 in mammals) (van den Elzen et al., 2014). Whether IFRD2-silenced ribosomes are similarly rescued, and the physiological conditions and timescales of these processes, remain to be determined. The regulation of these two mechanisms may impact their role on global translation activity during different types of cellular differentiation and stress.

We additionally observe Z-site tRNA on all reticulocyte ribosomes containing IFRD2. Although we cannot rule out the possibility that a tRNA in the Z site results from an artefact of isolating these samples, the stable association of Z-site tRNA and the similarity with IRES interactions supports the idea that the Z site is a preferred binding site for certain RNA structures, potentially including cellular mRNAs. In yeast, knocking out eS25, a key interacting partner of Z-site tRNA, decreases global protein synthesis and slows growth (Ferreira-Cerca et al., 2005; Landry et al., 2009). A recent study also suggests that eS25 is expressed at substoichiometric levels relative to total ribosomes in mammals, and that ribosomes containing eS25 may preferentially translate subsets of mRNAs (Shi et al., 2017).

The Z-site may represent a late-stage intermediate of tRNA ejection downstream of the E site. This hypothesis is supported by the observation that Z-site tRNA can bind simultaneously with a P-site tRNA but not a canonical E-site tRNA, and that Z-site tRNA interacts with the L1 stalk, which interacts with all known tRNA-ejection intermediates in bacteria (Agrawal et al., 2000) (Zhang et al., 2018). However, the interaction of Z-site tRNA with the eukaryotic-specific proteins eL42 and eS25 suggests that the Z site is distinct from bacterial tRNA-ejection intermediates.

A Z-site tRNA ejection intermediate would be transient during active translation, occurring only before the next translocation event. Therefore, the presence of a stably bound Z-site tRNA could act as signature of a translationally incompetent ribosome regardless of whether it engages the Z site from the E site or from a pool of deacylated tRNAs in the cytosol. In this scenario, tRNA binding to the Z site is more likely when deacylated tRNA levels rise, for example by amino acid depletion, or when translational factors are limiting. Intriguingly, deacylated tRNAs, a yeast-specific tRNA ejection factor, and stalled ribosomes are all implicated in regulating the integrated stress response by the GCN2 kinase (Ishimura et al., 2016; Ramirez et al., 1991; Visweswaraiah et al., 2012).

Our unexpected observation of IFRD2 and a new tRNA position on the mammalian ribosome highlights the ability of mining cryo-EM datasets to reveal new biological interactions from heterogeneous samples. This strategy is directly applicable to identifying and visualizing binding partners of other biological assemblies.

Materials and methods

Key resources table
Reagent type
(species) or resource		DesignationSource or referenceIdentifiersAdditional
information
Cell Line
(H. sapiens)		HEK293TAmerican Type
Culture Collection		Cat# CRL-3216
RRID:CVCL_0063
Recombinant
DNA reagent		pcDNA3.1-3xFLAG-TEV-IFRD2This paperN/AMammalian expression vector expressing Flag-tagged IFRD2 behind a CMV promoter
AntibodyIFRD2InvitrogenCat# PA5-48833
RRID:AB_2634289		IB: 1:1000
AntibodyFLAG M2Sigma-AldrichCat# F3165
RRID:AB_259529		IB: 1:5000
AntibodyuL2AbcamCat# ab169538
RRID:AB_2714187		IB: 1:10000
AntibodyeS24AbcamCat# ab196652
RRID:AB)2714188		IB: 1:3000
AntibodyHRP anti-mouseJackson ImmunoResearchCat# 115-035-003
RRID:AB_10015289		IB: 1:5000
AntibodyHRP anti-rabbitJackson ImmunoResearchCat# 111-035-003
RRID:AB_2313567		IB: 1:5000
AntibodyFLAG M2
agarose resin		Sigma-AldrichCat# A2220
RRID:AB_10063035
Peptide,
recombinant protein		3X FLAG peptideSigma-AldrichCat# F4799
Peptide,
recombinant protein		RNase ASigma-AldrichCat# R6513
Chemical
compound, drug		Dulbecco's Modified Eagle Medium (DMEM)Gibco/Thermo FisherCat# 10569
Chemical
compound, drug		HyClone Fetal
Bovine Serum		GE Healthcare Life SciencesCat# SH30910.03
Chemical
compound, drug		Trypsin-EDTAGibco/Thermo FisherCat# 25200
Chemical
compound, drug		OptiMEMGibco/Thermo FisherCat# 31985
Chemical
compound, drug		TransIT 293MirusCat# MIR2705
Chemical
compound, drug		EDTAGrowCellsCat# MRGF-1202
Chemical
compound, drug		SuperSignal
West Pico		Thermo FisherCat# 34080
Biological
sample (O. cuniculus)		Rabbit
reticulocyte lysate		Green HectaresN/A
Software,
algorithm		RELION-2.0 or
RELION-2.1		Kimanius et al., 2016RRID:SCR_016274https://www2.mrc-lmb.cam.ac.uk/relion
Software,
algorithm		MotionCor2Zheng et al., 2017RRID:SCR_016499http://msg.ucsf.edu/em/software/motioncor2.html
Software,
algorithm		GCTFZhang, 2016RRID:SCR_016500https://www.mrc-lmb.cam.ac.uk/kzhang/
Software,
algorithm		UCSF ChimeraPettersen et al., 2004RRID:SCR_004097https://www.cgl.ucsf.edu/chimera/
Software,
algorithm		Coot v.0.8.9Brown et al., 2015aRRID:SCR_014222https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/
Software,
algorithm		I-TASSERZhang, 2008RRID:SCR_014627https://zhanglab.ccmb.med.umich.edu/I-TASSER/
Software,
algorithm		PHENIXAdams et al., 2010RRID:SCR_014224https://www.phenix-online.org/
Software,
algorithm		MolProbity v.4.3.1Chen et al., 2010RRID:SCR_014226http://molprobity.biochem.duke.edu/
Software,
algorithm		EMRingerBarad et al., 2015N/Ahttp://emringer.com/
Software,
algorithm		SBGridMorin et al., 2013RRID:SCR_003511https://sbgrid.org/
Software,
algorithm		PyMOLDeLano, 2002RRID:SCR_000305http://www.pymol.org
Software,
algorithm		SCIPIONde la Rosa-Trevín et al., 2016N/Ahttp://scipion.i2pc.es/
Software,
algorithm		MonoResVilas et al., 2018N/A

Plasmids and antibodies

Request a detailed protocol

The ORF of IFRD2 was subcloned into a pcDNA3.1-based vector after a CMV promoter and an N-terminal 3X Flag tag and TEV cleavage site. IFRD2 antibody was obtained from Invitrogen (PA5-48833). Anti-FLAG M2 monoclonal antibody (F3165), resin (A2220) and 3X FLAG peptide (F4799) were obtained from Sigma-Aldrich.

Mass spectrometry

Request a detailed protocol

Three samples for mass-spectrometry analysis were prepared to identify the proteins present in the samples used for cryo-EM (Supplementary file 1). The first was prepared exactly as for cryo-EM analysis (Brown et al., 2015b). Briefly, we performed 2 mL in vitro translation reactions of a model substrate encoding an N-terminal 3X Flag tag, the autonomously folding villin headpiece domain, and the cytosolic portion of Sec61β in the presence of 0.5 μM dominant negative eRF1 in which the catalytic GGQ motif is mutated to AAQ to trap ribosomes at stop codons. Immediately after the translation reaction, we affinity purified the stalled ribosome-nascent protein complexes via the N-terminal 3X Flag tag on the nascent protein using anti-Flag M2 agarose beads (Sigma). The reactions were incubated with M2 affinity resin at 4°C for 1 hr, followed by three sequential 6 mL washes with: (1) 50 mM Hepes pH 7.4, 100 mM KOAc, 5 mM Mg(OAc)2, 0.1% Triton X-100, 1 mM DTT, (2) 50 mM Hepes pH 7.4, 250 mM KOAc, 5 mM Mg(OAc)2, 0.5% Triton X-100, 1 mM DTT, and (3) 50 mM Hepes pH 7.4, 100 mM KOAc, 5 mM Mg(OAc)2, 1 mM DTT. Two sequential elutions were performed with 0.1 mg/mL 3X Flag peptide in the third wash buffer for 25 min each at room temperature. This elution was directly analyzed as ‘sample A’ (Supplementary file 1).

To reduce the identification of ribosomal protein peptides, we evenly divided the elutions into two halves. We maintained one half in physiological salt concentrations and adjusted the other half to a final concentration of 750 mM KOAc, 15 mM Mg(OAc)2. We centrifuged both samples at 100,000 rpm for 30 min in a TLA120.2 rotor to pellet ribosomes and strip off peripherally associated proteins under the high-salt conditions. The supernatants from both spins were precipitated with trichloroacetic acid (TCA) and analyzed by SDS-PAGE and Coomassie staining. The entire gel lanes were excised and submitted for mass spectrometry analysis as ‘sample B’ (physiological salt wash) and ‘sample C’ (high-salt wash). The proteins identified by mass-spectrometry are given in Supplementary file 1.

Cell, lysate, and ribosomal treatments

Request a detailed protocol

HEK293T cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) with high glucose and 10% fetal bovine serum and verified to be mycoplasma free. Transfections of Flag-tagged IFRD2 were performed when cells were 60 – 70% confluency using TransIT 293 (MIRUS) according to manufacturer instructions. 16 – 18 hr after transfection, we lysed cells in 50 mM Hepes pH 7.4, 100 mM KOAc, 5 mM Mg(OAc)2, 0.5% Triton X-100, 1 mM DTT, and protease inhibitor cocktail. Lysates were clarified by spinning at 20,000xg at 4°C for 10 min before being subjected to biochemical treatments.

Where applicable, rabbit reticulocyte lysate (Green Hectares; Figure 3—figure supplement 3A), 200 nM ribosomes (Figure 3E) isolated from reticulocytes or HEK293T cell lysate (Figure 3—figure supplement 3C) were treated with 50 μg/mL RNase A at 25°C for 5 min before sucrose gradients. RNase A efficiency was confirmed after sucrose gradients by measuring absorbance at 254 nm to confirm the collapse of polysomes into monosome fractions (Figure 3E). EDTA was included with ribosomes isolated from reticulocytes (Figure 3E) or ribosomes pelleted after Flag-tagged IFRD2 affinity purification (Figure 3—figure supplement 3E) at a final concentration 10 mM. All size fractionation were performed using 200 μL 10–50% sucrose gradients prepared in 50 mM Hepes pH 7.4, 100 mM KOAc, 5 mM Mg(OAc)2 without or with 10 mM EDTA.

Samples were layered on top of the gradients and spun at 55,000 rpm using a TLS-55 rotor in an OptimaMax ultracentrifuge (Beckman Coulter) for 30 min at 4°C using the slowest acceleration and deceleration settings. Eleven 20 μL fractions were manually collected from the top for immunoblotting analysis.

Affinity purification and cryo-EM analysis of human IFRD2-80S ribosome complexes

Request a detailed protocol

Four 10 cm plates of HEK293T cells were transfected as described above and passaged 1:4 the day after transfection. Two days later, the cells were lysed as described above and incubated with M2 anti-Flag agarose resin for 1 hr at 4°C. The resin was washed with 6 mL of lysis buffer and subjected to two sequential elutions using one column volume of 0.1 mg/mL 3X Flag peptide incubated at room temperature for 25 min. A portion of the elution was directly frozen onto cryo-EM grids, and the leftover elution was pelleted at 100,000 rpm for 40 min at 4°C in a TLA120.2 rotor to isolate ribosomes, which were subjected to biochemical analysis (Figure 3—figure supplement 3E).

For cryo-EM analysis, 3 μL of the pooled elution containing ~80 nM ribosomes were directly frozen onto glow-discharged Quantifoil Cu R2/2 grids with a thin (~50 Å) continuous layer of carbon using a Vitrobot Mark III at 4°C and 100% humidity with 30 s wait time and 3 s blot time. A dataset of 877 micrographs was collected using an automated data collection pipeline in SerialEM on a Talos Arctica operated at 200 kV. Images were acquired at a nominal magnification of 36,000x (corresponding to a pixel size of 1.169 Å) with a Gatan K2 direct electron detector in super-resolution mode and defocus values ranging from −1.3 to −3 μm. Each movie was acquired at 5 frames/s over a total exposure time of 8 s and a dose rate of 5.4 electrons/pixel/s.

Data processing

Request a detailed protocol

To take advantage of recent developments in cryo-EM image processing, we reprocessed datasets of ribosomes isolated in the presence of a dominant-negative eukaryotic release factor (DN-eRF1) (Brown et al., 2015b). All processing steps were performed within RELION-2.0 or RELION-2.1 (Kimanius et al., 2016). We used MotionCor2 (Zheng et al., 2017) to correct for global and local (5 × 5 patches) beam-induced motion and to dose weight the individual frames. The motion-corrected sums without dose weighting were used for CTF estimation with GCTF (Zhang, 2016). All motion- and CTF-corrected micrographs and their Fourier transforms were inspected manually to remove those that displayed astigmatism or ice contamination.

To generate reference templates for auto-picking, 2000 ribosomal particles were picked manually and subjected to 2D classification. The best five classes (out of 10) were low-pass filtered to 20 Å and used for auto-picking in RELION (Scheres, 2015). For autopicking, a mask diameter of 360 Å and a minimum interparticle distance of 200 Å was used. The autopicked particles were extracted with a box size of 400 pixels and sorted by a Z-score generated by subtracting the reference image from the extracted particles. The sorted particles were inspected manually and non-ribosomal particles with a low Z score discarded. Retained particles were subjected to reference-free 2D classification and the best-resolved classes selected. In total, 971,191 particles were retained after 2D classification.

All selected particles were subjected to an initial round of refinement using RELION’s 3D autorefine in which a 30 Å low-pass filtered map of the rabbit 80S ribosome with tRNAs but no factors (EMD-4129) (Shao et al., 2016) was used as a reference. Following refinement, a single round of three-dimensional classification without alignment was performed. This strategy was very effective at separating rotated and unrotated 80S ribosomes from 60S ribosomal subunits.

To isolate particles corresponding to different classes within the set of 80S particles we used multiple rounds of focused classification with signal subtraction (FCwSS) (Bai et al., 2015) centered on the A, P, E, and Z sites of the ribosome. These sites were chosen as they displayed nebulous density after the initial round of refinement, consistent with mixed occupancy. The use of FCwSS in classifying the data represents the most significant difference from how the maps were processed originally (Brown et al., 2015b).

Once a defined class had been isolated, the particles were realigned, and the resultant maps postprocessed in RELION. For post-processing, solvent masks were generated using relion_mask_create. Typically, these binary masks were generated from the final map from refinement low-pass filtered to 15 Å at a threshold of 0.015 or 0.02. The initial binary mask was extended by 4 Å in all directions and a raised-cosine edge was added to create a soft mask. During post-processing, phase-randomization was used to correct for the convolution effects of the solvent mask. Overall resolution estimates were calculated from Fourier shell correlations at 0.143 between the two independently refined half-maps. Final reconstructions were sharpened using automatically estimated B-factors (Rosenthal and Henderson, 2003).

Processing of the human IFRD2-80S ribosomal complex was performed in RELION 2.1 as described above. After 2D classification, 22,807 particles were subject to an initial refinement using EMD-4129 low-pass filtered to 40 Å as a reference. We used the refined map, which already had clear IFRD2 density, as a reference for 3D classification to isolate 80S ribosomes, followed by a round of FCwSS centered on IFRD2 density. A final refinement of 5714 particles resulted in an 8.3 Å map without postprocessing.

Model building and refinement

Request a detailed protocol

Our 3.6 Å map of 80S•eRF1•ABCE1 complex was used as the starting point for model building. The model of the rabbit 80S ribosome in an unrotated state (PDB accession code 5LZV) (Shao et al., 2016) was fitted into this map using the ‘fit in map’ feature of Chimera (Pettersen et al., 2004). All non-ribosomal elements were deleted, and chemical modifications added to ribosomal RNA and proteins uL4 and eL40 using the cryo-EM structure of the human ribosome as a guide (Natchiar et al., 2017). This model was propagated to all other maps as individual 60S and 40S subunits. For the rotated and head-swiveled classes, the individual proteins and rRNA in the head of the 40S subunit were fitted as rigid bodies following docking of the ribosomal subunits.

Protein factors and tRNAs were modeled in Coot v0.8.9 (Brown et al., 2015a). P- and E-site tRNAs were extracted from the rabbit 80S ribosome (PDB accession code 5LZV) (Shao et al., 2016). The Z-site tRNA was modeled using the same sequence as for the E-site tRNA, although the density comes from a mixture of different tRNAs. The model for the L1 stalk was built using the crystal structure of the L1-stalk fragment from Haloarcula marismortui (PDB accession code 5Ml7) as a template. The rabbit L1 protein (NCBI ID: XP_002714675) was modeled by docking a comparative model generated using I-TASSER (Zhang, 2008) and morphing it to fit the density using Phenix.real_space_refinement (Afonine et al., 2018). The model for IFRD2 (NCBI XP_002713258) was built de novo from polyalanine helices placed into the map. Sequences were assigned to these helices based on side chain density and the helices connected by manual model building in Coot. eEF2 was modeled using the model of Sus scrofa eEF2 (PDB accession code 3J7P) (Voorhees et al., 2014) as a template. Human numbering and sequence (Uniprot ID: P13639) was used as sequence information is not available for rabbit eEF2. SERBP1 isoform X2 (XP_002715981) was built de novo into the density.

All models were refined using Phenix.real_space_refinement v1.13_2998 (Afonine et al., 2018). Each round of global real-space refinement featured five macro-cycles with secondary structure, rotamer, Ramachandran, and Cβ-torsion restraints applied. Secondary structure restraints were determined directly from the model and recalculated for each round of refinement. For the rRNA and tRNAs present in the models, hydrogen-bonding and base-pair and stacking parallelity restraints were applied. Additional restraints were applied for the chemical modifications of the ribosome and the diphthamide modification of eEF2. These restraints were generated using phenix.readyset. The high-resolution limit was set during refinement to match the nominal resolution obtained by postprocessing in Relion.

The final models were validated using MolProbity v.4.3.1 (Chen et al., 2010) and EMRinger (Barad et al., 2015), with final statistics given in table S2. Over-fitting was monitored using cross-validation (Amunts et al., 2014) (Figure 1—figure supplement 1).

Software used in the project were installed and configured by SBGrid (Morin et al., 2013).

Figures

Request a detailed protocol

Figure panels were generated using PyMOL (DeLano, 2002) or Chimera (Pettersen et al., 2004). Maps colored by local resolution were generated with unsharpened density maps using MonoRes (Vilas et al., 2018).

Data availability

All cryo-EM maps and models have been deposited in EMDB under accession codes 9234, 9235, 9236, 9237, 9239, 9240, 9241 and 9242. All models have been deposited in PDB under accession codes 6MTB, 6MTC, 6MTD and 6MTE.

The following data sets were generated
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9234. Rabbit 80S ribosome with a P-site tRNA (unrotated state).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9234
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9235. Rabbit 80S ribosome with a P- and E-site tRNA.
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9235
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9236. Rabbit 80S ribosome with a Z-site tRNA (unrotated state).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9236
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9237. Rabbit 80S ribosome with P- and Z-site tRNAs (unrotated state).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9237
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9239. Rabbit 80S ribosome with Z-site tRNA and IFRD2 (unrotated state).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9239
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9240. Rabbit 80S ribosome with eEF2 and SERBP1 (unrotated state with 40S head swivel).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9240
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9241. Rabbit 80S ribosome with A/P and P/E tRNAs (rotated state).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9241
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) EMBL-EBI Protein Data Bank
    ID EMD-9242. Rabbit 80S ribosome with eEF2 and SERBP1 (rotated state).
    http://www.ebi.ac.uk/pdbe/entry/emdb/EMD-9242
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) RCSB Protein Data Bank
    ID 6MTB. Rabbit 80S ribosome with P- and Z-site tRNAs (unrotated state).
    http://www.rcsb.org/pdb/explore/explore.do?structureId=6MTB
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) RCSB Protein Data Bank
    ID 6MTC. Rabbit 80S ribosome with Z-site tRNA and IFRD2 (unrotated state).
    http://www.rcsb.org/pdb/explore/explore.do?structureId=6MTC
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) RCSB Protein Data Bank
    ID 6MTD. Rabbit 80S ribosome with eEF2 and SERBP1 (unrotated state with 40S head swivel).
    http://www.rcsb.org/pdb/explore/explore.do?structureId=6MTD
    1. Brown A
    2. Baird MR
    3. Yip MCJ
    4. Murray J
    5. Shao S
    (2018) RCSB Protein Data Bank
    ID 6MTE. Rabbit 80S ribosome with eEF2 and SERBP1 (rotated state).
    http://www.rcsb.org/pdb/explore/explore.do?structureId=6MTE

References

  19. Book
    1. DeLano WL
    (2002)
    The PyMOL Molecular Graphics System
    Springer Open.
    1. Kressler D
    2. Hurt E
    3. Baβler J
    (2010) Driving ribosome assembly
    Biochimica Et Biophysica Acta (BBA) - Molecular Cell Research 1803:673–683.
    https://doi.org/10.1016/j.bbamcr.2009.10.009
    1. Varnum BC
    2. Lim RW
    3. Herschman HR
    (1989)
    Characterization of TIS7, a gene induced in Swiss 3T3 cells by the tumor promoter tetradecanoyl phorbol acetate
    Oncogene 4:1263–1265.

Decision letter

  1. Nahum Sonenberg
    Reviewing Editor; McGill University, Canada
  2. James L Manley
    Senior Editor; Columbia University, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Structures of translationally inactivated mammalian ribosomes" for consideration by eLife. Your article has been reviewed by James Manley as the Senior Editor, a Reviewing Editor and three peer reviewers. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The authors report cryo-EM structures of mammalian 80S ribosomes bound with SERBP1•eEF2 or with IFRD2. Importantly, they document a tRNA-binding, which differs from the canonical A, P and E sites. The structures allow near-atomic interpretation of interactions between the ribosomes, SERBP1 and eEF2, shedding new light on the mechanism of SERBP1. The work describes several new findings that will be of interest to researchers in and beyond the ribosome field.

Essential revisions:

The reviewers believe that your paper contains novel and interesting data, but that the paper needs revisions to (1) include reference to biochemical/biophysical work in the field (please see in the reviews below) and (2) provide a clearer explanation/review of the biological relevance of these structures as "inactivated" complexes. As you can see reviewer #1, in particular, is concerned whether the "inactive" states are just an artifact of isolation and treatment or whether they represent true inactive states that might reflect different physiological conditions. Also, there is the possibility that they represent intermediates (i.e. perhaps during recycling)?

Reviewer #1:

This is an interesting manuscript that reports on the structure of complexes of "inactivated" ribosomes. As such, these may offer little new insight into the process of translation but may be very important in elements relating to the regulation of translation and availability of ribosomal subunits. A major concern for this paper is whether the structures determined are biologically relevant as these inactive complexes have not been shown to directly be related to any process within the cell that might have generated such inactive particles (i.e. either as soluble complexes or as stress granules). Secondly, in the partial purification of ribosomal complexes, the authors use high speed centrifugation which is known to cause inactivation of ribosomes (Lu et al., 1997 or Scheck and Landau, 1982). Thus, there is the concern as to whether the complexes observed represent native structures or ones altered by hydrostatic pressure. However, that said, it is impressive that the authors can resolve so many structures from a large dataset and offers promise for the ability to readily obtain almost any ribosomal structure.

1) Title – it is not clear that these ribosomes have been inactivated, but rather that they are in an inactive state as the A, P and E sites or mRNA channel are blocked. There is the possibility that some of these complexes (i.e. 80S•P•E•eRF1•ABCE1) may represent intermediates in the translation cycle (and therefore, should not be considered inactive).

2) The authors use affinity tags for purification, but it is not clear that this process has actually worked. How is this population different from one where no affinity purification is used?

3) It is anticipated that a comparison of the yield of different proteins as determined by mass spectrometry should be reflected in the types of particles that are examined. Is there any data that would allow one to relate the amount of protein (as reflected in Supplementary file 1 to the different ribosomal classes as identified in Table 1)?

4) Supplementary file 1 – Initiation factors: the authors should identify the subunits of eIF2 as α, β and γ so as to not confuse the reader with eIF2A, a separate initiation factor (MW 65,000). Secondly, is there any explanation as to why eIF1, 1A, 5 or 5B were not found?

5) IFRD2 – The authors have failed to indicate that rabbit reticulocytes are essentially "interferon treated cells" in that they have elevated levels of several normally interferon induced proteins (PKR, oligo(A) synthase and RNAse L). Thus, this complex is likely to not normally be present in most tissues in the absence of exposure to interferon.

6) For the Z site tRNA, is it possible that this unique conformation is influenced by eIF5A which binds in this vicinity (i.e. would this small protein have been seen?)?

7) Discussion section – "tRNA binding to the Z site is more likely during periods of elevated deacylated tRNA levels.". This is not likely to occur as a slight increase in deacylated tRNA leads to the phosphorylation of eIF2 and the shutoff of protein synthesis (thus sparing the use of aminoacyl-tRNAs).

Reviewer #2:

Brown and co-authors describe cryo-EM structures of mammalian 80S ribosomes bound with SERBP1•eEF2 (at a higher resolution than the previously published 80S•SERBP1•eEF2 structural work) or with IFRD2. In addition, a tRNA-binding site is described, which differs from the canonical A, P and E sites. These results were obtained by re-evaluation of a previously published cryo-EM data set. The structures allow near-atomic interpretation of interactions between the ribosomes, SERBP1 and eEF2, bringing insights into the mechanism of SERBP1, however the authors' description of the mechanism is not clear (see below). The 80S•IFRD2 structure is an exciting finding, which identifies a new protein's role in translation regulation. The new tRNA position near the E site echoes a recent work on the bacterial ribosome. The authors propose that this position is on the pathway of deacyl-tRNA dissociation from the E site. Overall this work describes several new findings that will be of interest to researchers in and beyond the ribosome field.

The following changes/clarifications are suggested:

1) The mechanistic implications of SERBP1•eEF2 complex remain unclear. Based on the finding of SERBP1 bound together with eEF2, the authors propose that SERBP1 depletes cellular eEF2. First, it would be helpful to expand the discussion and mention that yeast homologue Stm1 binds yeast ribosomes without eEF2. Do the mechanisms of translation regulation by Stm1 and SERBP1 differ in that the latter depletes not only the 80S but also eEF2? Alternatively, the authors may have missed 80S•SERBP1 particles in their data set. Such particles may even be more abundant than 80S•SERBP1•eEF2 but also more dynamic, thus difficult to average and classify. Second, the mechanistic description of how SERBP1 prevents eEF2 dissociation remains vague. Comparison with other available eEF2 structures (with sordarin or other drugs and/or GDPNP/GDP nucleotides) might bring insights into how the conformational dynamics of eEF2 and/or the ribosome are affected by SERBP1.

2) IFRD2 is proposed to regulate translation by binding to non-translating 80S. But could IFRD2 also bind vacant 40S and prevent translation initiation?

3) Describe the rotational state of the 80S bound with P-tRNA and Z-tRNA (likely non-rotated, but it would help to state this in the paper). Does IFRD2 induce a different 80S conformation?

Reviewer #3:

This manuscript describes the results of a refined analysis of a data set of nearly 1 million ribosome particles obtained from an in vitro rabbit reticulocyte translation reaction in which a dominant negative release factor had been added to enrich for translation termination complexes (Brown et al., 2015b). Evidently, individual complexes only represent a small portion (5-25%) of the ribosomes isolated by the procedures used and the authors have mined this reaction for complexes other than those that have already been published to define 9 new structures (described in summary in Figure 1). There is little information in the present manuscript that describes how the complexes examined here were specifically obtained from the reticulocyte lysate, but it would appear that they work under conditions in which only 80S particles form and that the isolated 80S peak is filled with a diverse array of distinct ribosome complexes. In the Brown et al., 2015 paper it appears to describe a FLAG-tag pull down of the nascent peptide, where focus was given to the eRF1-trapped sub-population of complexes. One of the 9 complexes defined here appears to be the same structure as that which was published in Brown et al., 2015b (the eRF1-ABCE1 POST complex), indicating a duplication or positive control of sorts. While I might be able to understand how the other P-site tRNA-containing complexes were isolated by a nascent polypeptide-directed pull-down approach, it is unclear how the complexes lacking P-site tRNA were obtained (i.e. the focus of the present investigations). I would strongly suggest further clarification on these points for the sake of reproducibility.

The two translationally inactive complexes discussed in detail in Figure 2, Figure 3 and Figure 4 are certainly interesting in regards to the mechanisms by which cells may "store" excess ribosomes and the quality of the data appear sufficient enough to support the claims that the authors make, which are largely qualitative descriptions of the SERBP1 and IFRD2 binding sites and the global positioning of the deacylated tRNA within the E site. In this regard, I think the paper is perfectly suitable for publication.

What appears to need further clarification and elaboration is the discussion surrounding the "Z" site for tRNA binding. Alternative positions for E-site tRNA have been described since the early 80s in the early fluorescence studies of eukaryotic ribosomes (Robertson and Wintermeyer, 1987; Rodnina and Wintermeyer, 1992). Movements of deacylated tRNA within the E site have also been specifically described in recent single-molecule fluorescence studies (Ferguson et al., 2015 and references therein). Although the authors state that their "Z" site position may be similar or related to the position (Eout) described by (Zhang et al., 2017), I don't see how they make this connection. I also don't understand the authors reference to the "known role [of the L1 stalk] in escorting deacylated tRNA from the ribosome", citing two bodies of work in bacteria that make no such claim or data to support such a claim. In this regard, is it really wise to give it a name other than E' or something of the sort? I fear that re-naming it the "Z" site will significantly complicate its description in future publications. The new position the authors see is clearly an "E site" position, yes?

Aside from these major points, the paper certainly seems publishable in its present format from my perspective.

https://doi.org/10.7554/eLife.40486.042

Author response

Reviewer #1:

This is an interesting manuscript that reports on the structure of complexes of "inactivated" ribosomes. As such, these may offer little new insight into the process of translation but may be very important in elements relating to the regulation of translation and availability of ribosomal subunits. A major concern for this paper is whether the structures determined are biologically relevant as these inactive complexes have not been shown to directly be related to any process within the cell that might have generated such inactive particles (i.e. either as soluble complexes or as stress granules).

Although SERBP1/Stm1 has been implicated in stress responses (Van Dyke et al., 2006)(Lee et al., 2014), inactive ribosomes may also occur outside periods of stress. Both SERBP1 and IFRD2 are expressed in most tissues at moderate levels (https://www.proteinatlas.org/ENSG00000214706-IFRD2/tissue and https://www.proteinatlas.org/ENSG00000142864-SERBP1/tissue) and are detected to be associated with ribosomes in unstressed cells (Simsek et al., 2017). This suggests that a percentage of ribosomes exist in inactivate states even in cells not undergoing stress. As we mention in the paper, one possibility is that ribosomes are inactivated as part of translational reprogramming that occurs during differentiation. As our ribosomes were purified from reticulocytes, the penultimate stage of erythrocyte differentiation, this may explain the presence of inactive ribosomes in our sample. Consistent with this, we observe higher IFRD2 levels in reticulocytes than in HEK293T cells (Figure 3—figure supplement 3B).

Secondly, in the partial purification of ribosomal complexes, the authors use high speed centrifugation which is known to cause inactivation of ribosomes (Lu et al., 1997 or Scheck and Landau, 1982). Thus, there is the concern as to whether the complexes observed represent native structures or ones altered by hydrostatic pressure. However, that said, it is impressive that the authors can resolve so many structures from a large dataset and offers promise for the ability to readily obtain almost any ribosomal structure.

Though we cannot completely rule out the possibility that high-speed centrifugation may inactivate some reticulocyte ribosomes, we believe there is strong evidence that the structures of inactive ribosomes we analyze represent physiological complexes and are not the result of technical artifacts:

First, the ability to detect known intermediates of translation (Figure 1) demonstrates that our purification scheme is capable to retaining ribosomes in near-native states. In addition, using the same centrifugation conditions as in this manuscript, we can isolate ribosomal pellets that translate just as efficiently in fractionated cell-free translation systems as complete reticulocyte lysate (Shao and Hegde, 2011), suggesting that these centrifugation conditions do not significantly inactivate ribosomes. Similar centrifugation conditions are used to isolate ribosomes to generate other prokaryotic and eukaryotic cell-free translation systems (Spedding, 1990; Shimizu et al., 2001), including the translation system used in Lu et al., 1997. We think this is consistent with the studies by Lu et al., 1997 and Scheck and Landau, 1982, which primarily analyze ribosome activity after applying high-pressure treatments in closed pressure vessels, rather than by centrifugation.

Second, the interaction between IFRD2 and ribosomes is also detected in a mass spectrometry study that did not use high-speed centrifugation (Simsek et al., 2017), as well as in immunoprecipitations of Flag-tagged IFRD2 from 293T cell lysate that was not subject to centrifugation (Figure 3—figure supplement 3D). We also detect endogenous IFRD2 quantitatively associated with ribosomes in reticulocyte lysate before (Figure 3—figure supplement 3A) and after (Figure 3E) highspeed centrifugation. Finally, recombinant IFRD2 only weakly associates with ribosomes after centrifugation when incubated with lysate (Author response image 1), unlike IFRD2 expressed in cells, which clearly interacts with ribosomes (Author response image 1, Figure 3—figure supplement 3B-E). Similarly, the association of SERBP1/Stm1 with ribosomes has also been observed by different methods in the literature (Zinoviev et al., 2015). These observations all suggest that these complexes assemble in physiological conditions and are not induced by centrifugation.

Author response image 1
Download asset Open asset
Assembly of IFRD2 onto ribosomes.

Recombinant IFRD2 purified from E. coli was incubated with cytosol isolated from HEK293T cells for 30 minutes at 32°C and subject to native size fractionation by centrifugation on a 10-50% sucrose gradient. Eleven fractions were collected from the top of the gradient and analyzed by SDS-PAGE and immunoblotting for the recombinant IFRD2 (top blot; quantified by blue curve). This revealed that only minor co-association of IFRD2 in ribosomal (80S) fractions. In contrast, transiently transfecting IFRD2 into HEK293T cells, followed by cytosol isolation and size fractionation (bottom plot; quantified by orange curve) reveals a distinct peak of IFRD2 co-association in ribosomal fractions (red arrow).

https://doi.org/10.7554/eLife.40486.017

1) Title – it is not clear that these ribosomes have been inactivated, but rather that they are in an inactive state as the A, P and E sites or mRNA channel are blocked. There is the possibility that some of these complexes (i.e. 80S•P•E•eRF1•ABCE1) may represent intermediates in the translation cycle (and therefore, should not be considered inactive).

As shown in Figure 1 we only consider ribosomal complexes containing SERBP1 and IFRD2 to be inactive. To more accurately describe these complexes, we have changed the title to “Structures of translationally inactive mammalian ribosomes”.

2) The authors use affinity tags for purification, but it is not clear that this process has actually worked. How is this population different from one where no affinity purification is used?

We address this comment in detail in response to reviewer #3.

3) It is anticipated that a comparison of the yield of different proteins as determined by mass spectrometry should be reflected in the types of particles that are examined. Is there any data that would allow one to relate the amount of protein (as reflected in Supplementary file 1 to the different ribosomal classes as identified in Table 1)?

Our mass spectrometry analysis was run for protein detection of SDS-PAGE gel lanes and not quantification (e.g. via multiplexing of different samples for comparison). As proteins respond differently to trypsin digestion, some peptides ionize better than others, and detection efficiencies for ions with different m/z values are unequal, the peptides detected may not fully correlate with the protein composition of the sample. We do detect many more proteins by mass spectrometry (Supplementary file 1) than we identified by cryo-EM analysis (Table 1). This may be due to numerous factors, including relative protein abundance, mRNA-binding proteins that do not directly contact the ribosome, or the inability to visualize flexible ribosome-binding proteins.

4) Supplementary file 1 – Initiation factors: the authors should identify the subunits of eIF2 as α, β and γ so as to not confuse the reader with eIF2A, a separate initiation factor (MW 65,000). Secondly, is there any explanation as to why eIF1, 1A, 5 or 5B were not found?

We have edited Supplementary file 1 to include the names of the eIF2 subunits alongside their Uniprot IDs. We note that eIF5B did appear in our mass spectrometry analysis (G1TRL5_RABIT, IF2P_HUMAN). The absence of other initiation factors may be due to their absence from the sample, presence at concentrations below the detection limit, or technical limitations of the mass spectrometry experiment, particularly given the small sizes of eIF1 and 1A.

5) IFRD2 – The authors have failed to indicate that rabbit reticulocytes are essentially "interferon treated cells" in that they have elevated levels of several normally interferon induced proteins (PKR, oligo(A) synthase and RNAse L). Thus, this complex is likely to not normally be present in most tissues in the absence of exposure to interferon.

As we mention in the manuscript, the name “interferon-related developmental regulator” appears to be a misnomer that originated from a mistaken sequence alignment with mouse interferon-b (Tirone et al., 1989). We do not know of a direct connection between IFRD2 expression or function and interferon signaling, though we cannot exclude the possibility that IFRD2, like established interferon-induced proteins, may be upregulated in certain situations to inhibit translation. In addition, IFRD2 is expressed in many tissues (https://www.proteinatlas.org/ENSG00000214706-IFRD2/tissue), and we can also detect low levels of endogenous IFRD2 in HEK293T cells (Figure 3—figure supplement 3B). We therefore think that IFRD2 is a general mechanism of translationally repressing ribosomes.

6) For the Z site tRNA, is it possible that this unique conformation is influenced by eIF5A which binds in this vicinity (i.e. would this small protein have been seen?)?

eIF5A was our first thought when we saw density in this area prior to classification. However, following extensive classification, we see no evidence of eIF5A bound to any ribosomes. eIF5A has been observed in another cryo-EM study (Schmidt et al., 2016)

7) Discussion section – "tRNA binding to the Z site is more likely during periods of elevated deacylated tRNA levels.". This is not likely to occur as a slight increase in deacylated tRNA leads to the phosphorylation of eIF2 and the shutoff of protein synthesis (thus sparing the use of aminoacyl-tRNAs).

Work in yeast has shown that phosphorylation of eIF2α is induced within 15 min of amino acid starvation (Zaborske et al., 2009). This means elevated concentrations of deacylated tRNA can occur even if only transiently before eIF2 phosphorylation dampens translation. We have rephrased the text to make this clearer.

Reviewer #2:

1) The mechanistic implications of SERBP1•eEF2 complex remain unclear. Based on the finding of SERBP1 bound together with eEF2, the authors propose that SERBP1 depletes cellular eEF2. First, it would be helpful to expand the discussion and mention that yeast homologue Stm1 binds yeast ribosomes without eEF2. Do the mechanisms of translation regulation by Stm1 and SERBP1 differ in that the latter depletes not only the 80S but also eEF2?

The thinking that Stm1 binds the yeast ribosome in the absence of eEF2 is influenced by the absence of eEF2 in structures of the yeast 80S in complex with Stm1 (Ben-Shem et al., 2011). However, work from Nono Takeuchi-Tomita’s lab provides evidence that Stm1 stabilizes eEF2 on the yeast ribosome (Hayashi et al., 2017). It is therefore possible that Stm1 and SERBP1 function similarly within cells to deplete eEF2, but yeast eEF2 was lost during purification or did not crystallize.

Alternatively, the authors may have missed 80S•SERBP1 particles in their data set. Such particles may even be more abundant than 80S•SERBP1•eEF2 but also more dynamic, thus difficult to average and classify. Second, the mechanistic description of how SERBP1 prevents eEF2 dissociation remains vague. Comparison with other available eEF2 structures (with sordarin or other drugs and/or GDPNP/GDP nucleotides) might bring insights into how the conformational dynamics of eEF2 and/or the ribosome are affected by SERBP1.

We agree with the reviewer that SERBP1 probably binds ribosomes in the absence of eEF2 and cannot discount the presence of 80S•SERBP1 in our dataset. However, as eEF2 is one of the most highly expressed proteins in cells, ribosome•SERBP1 complexes may quickly recruit eEF2.

The exact mechanism for how SERBP1 prevents eEF2 from dissociating from the ribosome is difficult to infer from our two structural snapshots. However, it is clear that SERBP1 does not prevent GTP hydrolysis or induce large displacements in orientations of the domains of eEF2. Rather, its mechanism of action may be similar to sordarin, an antibiotic that also does not prevent GTP hydrolysis or induce large structural changes yet still traps eEF2 on the ribosome. Sordarin binds in a pocket between domains III, IV and V and in doing so increases interdomain contacts and subtly constrain the domains in conformations that prevent release (Pellegrino et al., 2018). As SERBP1 interacts only with domain IV of eEF2, it is likely that SERBP1 constrains the interdomain movements necessary for release by locking domain IV in a fixed position in the A site. We have altered the text to include a more detailed proposal for how we expect SERBP1 to prevent eEF2 dissociation.

2) IFRD2 is proposed to regulate translation by binding to non-translating 80S. But could IFRD2 also bind vacant 40S and prevent translation initiation?

IFRD2 interacts with both the small and large subunits, suggesting that the interaction is most stable with 80S ribosomes. Additionally, our immunoprecipitation experiments with FLAG-tagged IFRD2 did not isolate any 40S•IFRD2 complexes (as judged from analysis of our micrographs). Despite these observations, we cannot preclude the possibility that IFRD2 binds vacant 40S subunits before recruiting 60S subunits and in some cases a tRNA bound at the Z site. Regardless of the assembly pathway, IFRD2 would function to repress translation.

3) Describe the rotational state of the 80S bound with P-tRNA and Z-tRNA (likely non-rotated, but it would help to state this in the paper). Does IFRD2 induce a different 80S conformation?

The ribosome is in a non-rotated state in the presence of IFRD2 and/or Z-site tRNA. We have added this information to the manuscript.

Reviewer #3:

[…] There is little information in the present manuscript that describes how the complexes examined here were specifically obtained from the reticulocyte lysate, but it would appear that they work under conditions in which only 80S particles form and that the isolated 80S peak is filled with a diverse array of distinct ribosome complexes. In the Brown et al., 2015 paper it appears to describe a FLAG-tag pull down of the nascent peptide, where focus was given to the eRF1-trapped sub-population of complexes. One of the 9 complexes defined here appears to be the same structure as that which was published in Brown et al., 2015b (the eRF1-ABCE1 POST complex), indicating a duplication or positive control of sorts. While I might be able to understand how the other P-site tRNA-containing complexes were isolated by a nascent polypeptide-directed pull-down approach, it is unclear how the complexes lacking P-site tRNA were obtained (i.e. the focus of the present investigations). I would strongly suggest further clarification on these points for the sake of reproducibility.

The original aim of the purification strategy was to isolate 80S•P•E•eRF1•ABCE1 complexes using a FLAG-tagged nascent chain and a catalytically inactive mutant of eRF1 in cell-free translational reactions. The complexes lacking peptidyl-tRNA are likely ribosomes that either co-associate with the target ribosome-nascent-chain complexes or bind non-specifically to the anti-FLAG beads during purification. This apparent co-association and/or non-specific binding also occurs in our experiments with FLAG-tagged IFRD2, where 75% of immunoprecipitated ribosomes observed by cryo-EM lack IFRD2. The phenomenon of achieving enrichment rather than homogeneity is widely reported with many target cryo-EM structures coming from about 10% of the total data, for example the ribosomal complexes described in Voorhees and Hegde, 2015; Shen et al., 2015; Braunger et al., 2018.

The two translationally inactive complexes discussed in detail in Figure 2, Figure 3 and Figure 4 are certainly interesting in regards to the mechanisms by which cells may "store" excess ribosomes and the quality of the data appear sufficient enough to support the claims that the authors make, which are largely qualitative descriptions of the SERBP1 and IFRD2 binding sites and the global positioning of the deacylated tRNA within the E site. In this regard, I think the paper is perfectly suitable for publication.

What appears to need further clarification and elaboration is the discussion surrounding the "Z" site for tRNA binding. Alternative positions for E-site tRNA have been described since the early 80s in the early fluorescence studies of eukaryotic ribosomes (Robertson and Wintermeyer, 1987; Rodnina and Wintermeyer, 1992). Movements of deacylated tRNA within the E site have also been specifically described in recent single-molecule fluorescence studies (Ferguson et al., 2015 and references therein). Although the authors state that their "Z" site position may be similar or related to the position (Eout) described (Zhang et al., 2017), I don't see how they make this connection. I also don't understand the authors reference to the "known role [of the L1 stalk] in escorting deacylated tRNA from the ribosome", citing two bodies of work in bacteria that make no such claim or data to support such a claim. In this regard, is it really wise to give it a name other than E' or something of the sort? I fear that re-naming it the "Z" site will significantly complicate its description in future publications. The new position the authors see is clearly an "E site" position, yes?

Aside from these major points, the paper certainly seems publishable in its present format from my perspective.

We decided to call the site as the “Z site” for a number of reasons. Firstly, the naming of E-site intermediates in bacteria is confusing and we did not want to add to this confusion. Published terms include Eout (Zhang et al., 2017), E2 (Agrawal et al., 2000), E´ (Robertson et al., 1986), and F site (Wower et al., 2000). We felt a different letter would help distinguish the site from reported intermediates in bacterial ribosomes. Secondly, using E´ (or similar) may give the impression that this tRNA position is definitely an intermediate of ejection from the E site. Although this is a plausible explanation, we cannot ignore the possibility that a tRNA can only bind to the Z site when the normal passage of tRNAs through the ribosome is stopped. Thirdly, “Z” as the last letter of the alphabet evokes the extreme position on the ribosome adopted by the tRNA.

The comparison with Eout (Zhang et al., 2017) was based on the way in which the anti-codon stem-loop of the tRNAs in both complexes interact with the extremity of the small subunit. However, given that the Z-site is specific to eukaryotes, we have removed this comparison. We have also removed the references to the role of the L1 in escorting deacylated tRNAs, except to say that the L1 stalk interacts with all known tRNA ejection intermediates.

https://doi.org/10.7554/eLife.40486.043

Article and author information

Author details

  1. Alan Brown

    Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, United States
    Contribution
    Conceptualization, Supervision, Investigation, Writing—original draft, Writing—review and editing
    For correspondence
    alan_brown@hms.harvard.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0021-0476

  2. Matthew R Baird

    Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, United States
    Contribution
    Investigation, Writing—review and editing
    Competing interests
    No competing interests declared

  3. Matthew CJ Yip

    Department of Cell Biology, Harvard Medical School, Boston, United States
    Contribution
    Investigation, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2505-9987

  4. Jason Murray

    MRC Laboratory of Molecular Biology, Cambridge, United Kingdom
    Contribution
    Investigation, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8679-8281

  5. Sichen Shao

    Department of Cell Biology, Harvard Medical School, Boston, United States
    Contribution
    Conceptualization, Supervision, Investigation, Writing—original draft, Writing—review and editing
    For correspondence
    sichen_shao@hms.harvard.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-2679-5537

Funding

Harvard Medical School

  • Alan Brown
  • Matthew R Baird
  • Matthew CJ Yip
  • Sichen Shao

International Retinal Research Foundation

  • Alan Brown

E. Matilda Ziegler Foundation for the Blind

  • Alan Brown

Eunice Kennedy Shriver National Institute of Child Health and Human Development

  • Jason Murray

Charles H. Hood Foundation

  • Sichen Shao

Richard and Susan Smith Family Foundation

  • Sichen Shao

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

The authors thank V Ramakrishnan and RS Hegde, M Skehel and the MRC-LMB mass-spectrometry facility, Harvard Research Computing and SBGrid for computing support, E Fischer and SHarrison for access to their GPU workstations, the cryo-EM facility at the University of Massachusetts Worcester for help with cryo-EM data collection, L Hollingsworth and T Rapoport for comments, the Shao lab for helpful discussions, and NVIDIA Corporation for the donation of a Quadro P6000 GPU. This work was supported by Harvard Medical School, the International Retinal Research Foundation, the E Matilda Ziegler Foundation, the Richard and Susan Smith Family Foundation, and the Charles H Hood Foundation.

Senior Editor

  1. James L Manley, Columbia University, United States

Reviewing Editor

  1. Nahum Sonenberg, McGill University, Canada

Publication history

  1. Received: July 26, 2018
  2. Accepted: October 12, 2018
  3. Accepted Manuscript published: October 24, 2018 (version 1)
  4. Version of Record published: November 9, 2018 (version 2)

Copyright

© 2018, Brown et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 6,687
    Page views
  • 971
    Downloads
  • 41
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.

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)

  1. Alan Brown
  2. Matthew R Baird
  3. Matthew CJ Yip
  4. Jason Murray
  5. Sichen Shao
(2018)
Structures of translationally inactive mammalian ribosomes
eLife 7:e40486.
https://doi.org/10.7554/eLife.40486

Categories and tags

Further reading

    1. Structural Biology and Molecular Biophysics

    The prolactin receptor scaffolds Janus kinase 2 via co-structure formation with phosphoinositide-4,5-bisphosphate

    Raul Araya-Secchi, Katrine Bugge ... Birthe B Kragelund
    Research Article

    Class 1 cytokine receptors transmit signals through the membrane by a single transmembrane helix to an intrinsically disordered cytoplasmic domain that lacks kinase activity. While specific binding to phosphoinositides has been reported for the prolactin receptor (PRLR), the role of lipids in PRLR signalling is unclear. Using an integrative approach combining NMR spectroscopy, cellular signalling experiments, computational modelling and simulation, we demonstrate co-structure formation of the disordered intracellular domain of the human PRLR, the membrane constituent phosphoinositide-4,5-bisphosphate (PI(4,5)P2) and the FERM-SH2 domain of the Janus kinase 2 (JAK2). We find that the complex leads to accumulation of PI(4,5)P2 at the transmembrane helix interface and that mutation of residues identified to interact specifically with PI(4,5)P2 negatively affects PRLR-mediated activation of signal transducer and activator of transcription 5 (STAT5). Facilitated by co-structure formation, the membrane-proximal disordered region arranges into an extended structure. We suggest that the co-structure formed between PRLR, JAK2 and PI(4,5)P2 locks the juxtamembrane disordered domain of the PRLR in an extended structure, enabling signal relay from the extracellular to the intracellular domain upon ligand binding. We find that the co-structure exists in different states which we speculate could be relevant for turning signalling on and off. Similar co-structures may be relevant for other non-receptor tyrosine kinases and their receptors.

    1. Biochemistry and Chemical Biology
    2. Structural Biology and Molecular Biophysics

    The structural basis of the multi-step allosteric activation of Aurora B kinase

    Dario Segura-Peña, Oda Hovet ... Nikolina Sekulic
    Research Article

    Aurora B, together with IN-box, the C-terminal part of INCENP, forms an enzymatic complex that ensures faithful cell division. The [Aurora B/IN-box] complex is activated by autophosphorylation in the Aurora B activation loop and in IN-box, but it is not clear how these phosphorylations activate the enzyme. We used a combination of experimental and computational studies to investigate the effects of phosphorylation on the molecular dynamics and structure of [Aurora B/IN-box]. In addition, we generated partially phosphorylated intermediates to analyze the contribution of each phosphorylation independently. We found that the dynamics of Aurora and IN-box are interconnected, and IN-box plays both positive and negative regulatory roles depending on the phosphorylation status of the enzyme complex. Phosphorylation in the activation loop of Aurora B occurs intramolecularly and prepares the enzyme complex for activation, but two phosphorylated sites are synergistically responsible for full enzyme activity.

    1. Neuroscience
    2. Structural Biology and Molecular Biophysics

    Cannabidiol sensitizes TRPV2 channels to activation by 2-APB

    Aaron Gochman, Xiao-Feng Tan ... Andres Jara-Oseguera
    Research Article

    The cation-permeable TRPV2 channel is important for cardiac and immune cell function. Cannabidiol (CBD), a non-psychoactive cannabinoid of clinical relevance, is one of the few molecules known to activate TRPV2. Using the patch-clamp technique, we discover that CBD can sensitize current responses of the rat TRPV2 channel to the synthetic agonist 2-aminoethoxydiphenyl borate (2-APB) by over two orders of magnitude, without sensitizing channels to activation by moderate (40°C) heat. Using cryo-EM, we uncover a new small-molecule binding site in the pore domain of rTRPV2 in addition to a nearby CBD site that had already been reported. The TRPV1 and TRPV3 channels are also activated by 2-APB and CBD and share multiple conserved features with TRPV2, but we find that strong sensitization by CBD is only observed in TRPV3, while sensitization for TRPV1 is much weaker. Mutations at non-conserved positions between rTRPV2 and rTRPV1 in either the pore domain or the CBD sites failed to confer strong sensitization by CBD in mutant rTRPV1 channels. Together, our results indicate that CBD-dependent sensitization of rTRPV2 channels engages multiple channel regions, and that the difference in sensitization strength between rTRPV2 and rTRPV1 channels does not originate from amino acid sequence differences at the CBD binding site or the pore domain. The remarkably robust effect of CBD on TRPV2 and TRPV3 channels offers a promising new tool to both understand and overcome one of the major roadblocks in the study of these channels – their resilience to activation.