Research Article

Crystal Structure of the Eukaryotic 60S Ribosomal Subunit in Complex with Initiation Factor 6

See allHide authors and affiliations

Science  18 Nov 2011:
Vol. 334, Issue 6058, pp. 941-948
DOI: 10.1126/science.1211204

Abstract

Protein synthesis in all organisms is catalyzed by ribosomes. In comparison to their prokaryotic counterparts, eukaryotic ribosomes are considerably larger and are subject to more complex regulation. The large ribosomal subunit (60S) catalyzes peptide bond formation and contains the nascent polypeptide exit tunnel. We present the structure of the 60S ribosomal subunit from Tetrahymena thermophila in complex with eukaryotic initiation factor 6 (eIF6), cocrystallized with the antibiotic cycloheximide (a eukaryotic-specific inhibitor of protein synthesis), at a resolution of 3.5 angstroms. The structure illustrates the complex functional architecture of the eukaryotic 60S subunit, which comprises an intricate network of interactions between eukaryotic-specific ribosomal protein features and RNA expansion segments. It reveals the roles of eukaryotic ribosomal protein elements in the stabilization of the active site and the extent of eukaryotic-specific differences in other functional regions of the subunit. Furthermore, it elucidates the molecular basis of the interaction with eIF6 and provides a structural framework for further studies of ribosome-associated diseases and the role of the 60S subunit in the initiation of protein synthesis.

In all domains of life, protein synthesis is catalyzed by ribosomes. These giant ribonucleoprotein particles consist of two subunits with distinct functions. Information contained in mRNA is decoded by the small ribosomal subunit, whereas peptide bond formation is mediated by the RNA component of the large ribosomal subunit (1). Ribosome-associated factors bind to the large ribosomal subunit and interact with nascent polypeptides emerging from the ribosomal tunnel (2). In addition, the large subunit is a target for antibiotics that interfere with the peptidyl transferase reaction and with the progression of the nascent polypeptide chain through the tunnel (3). Over the past decade, our knowledge of prokaryotic translation has advanced with the publication of the crystal structures of the small (30S) and large (50S) ribosomal subunits as well as the complete (70S) prokaryotic ribosome (1, 46). In recent years, structures of prokaryotic ribosomes in complex with protein factors involved in translation have provided us with molecular snapshots of various stages of protein synthesis at atomic resolution (1, 79). However, our molecular understanding of protein synthesis in eukaryotes remains incomplete.

Eukaryotic ribosomes are considerably larger than their bacterial counterparts. Both eukaryotic ribosomal subunits contain numerous RNA expansion segments, which coevolved with many additional eukaryotic-specific ribosomal protein elements. As a consequence, the eukaryotic 60S subunit in yeast or T. thermophila has a total molecular weight of about 2 million daltons, whereas that of the 50S subunit in Escherichia coli is 1.3 million daltons.

The increased level of structural complexity of eukaryotic ribosomes reflects functional differences between prokaryotes and eukaryotes. First, ribosome biogenesis in eukaryotic cells is elaborate. It takes place in different cellular compartments and involves about 200 trans-acting proteins in the processing and modification of ribosomal RNA (rRNA), the import of ribosomal proteins into the nucleus, and the export of preribosomal subunits into the cytoplasm (10). Second, the regulation of protein synthesis is much more complex in eukaryotes and at the level of initiation mostly involves the 40S ribosomal subunit.

Recently, cryoelectron microscopy (cryo-EM) reconstructions of yeast and wheat germ 80S ribosomes at 5.5 to 6.1 Å resolution and a partial interpretation of crystallographic data from the 80S yeast ribosome at 4.15 Å resolution have provided us with the positions and topology of eukaryotic RNA expansion segments and several proteins for which homologous structures are known (1113). However, in these studies it was not possible to correctly assign and build many eukaryotic proteins. The recent crystal structure of the T. thermophila small ribosomal subunit (40S) in complex with eIF1 represents the first complete atomic model of the eukaryotic 40S ribosomal subunit (14), whereas a corresponding structure of the large ribosomal subunit (60S) has so far remained elusive.

The 60S ribosomal subunit is subject to several regulatory processes during initiation. Binding of eukaryotic initiation factor 6 (eIF6) to the large ribosomal subunit inhibits subunit joining and thus prevents translation initiation (15). The crystal structures of isolated eIF6 and its archaeal homolog aIF6 in isolation have highlighted its pentameric shape (16). The mechanism by which eIF6 prevents 80S complex formation is most likely steric hindrance; however, the exact interaction between eIF6 (or aIF6) and the large ribosomal subunit is currently unclear, as chemical probing and low-resolution EM data have provided conflicting evidence (17, 18). In addition to its function as an anti-association factor, eIF6 has also been implicated in 60S maturation (19, 20).

Crystallization and structure determination. We present the crystal structure of the T. thermophila 60S subunit in complex with eIF6 (Fig. 1). Cocrystallization with the antibiotic cycloheximide yielded crystals diffracting to 3.5 Å resolution and permitted visualization of its binding site. Bulky and ordered side chains are clearly visible, and the registry for all well-ordered regions of the structure is unambiguous. We have also included in the model some flexible solvent-exposed loops of proteins for which the sequence register is less reliable because it was established by extrapolation from the last identifiable residue. Detailed crystallization and structure determination procedures are provided in table S1 and (21). Table S2 lists all proteins included in this structure, their UNIPROT codes, and homologs from yeast, bacteria, and archaea. Throughout this article the human nomenclature is used for all ribosomal proteins; where the yeast nomenclature differs, the name of the yeast homolog is included in parentheses (Fig. 1, C and D, and Fig. 2) (22). A PYMOL script that displays two Protein Data Bank (PDB) files corresponding to one complete 60S subunit and labels all proteins according to the human, yeast, and E. coli nomenclatures is available (21).

Fig. 1

Architecture of the 60S ribosomal subunit. (A and B) Views of the solvent-exposed (A) and 40S binding side of the 60S subunit (B) with color-coded RNA expansion segments. (C and D) Views of the solvent-exposed (C) and 40S binding side of the 60S subunit (D) with color-coded ribosomal proteins shown as ribbons.

Fig. 2

Evolutionary representation of ribosomal proteins of the 60S subunit. 60S large ribosomal subunit proteins are colored according to conservation. Protein cores found in all kingdoms are depicted in light blue, proteins with archaeal homologs are in gold, and proteins or protein extensions unique to eukaryotes are in red. Positions of N and C termini are indicated. Zinc ions are shown as green spheres.

Structure of the 26S and 5.8S rRNAs. The T. thermophila 60S ribosomal subunit contains three rRNA molecules (5S, 5.8S, and 26S rRNA). As a result of differences in pre-rRNA processing in eukaryotes, the 5.8S rRNA occupies the same region as the 5′ end of the bacterial 23S rRNA (figs. S1 and S2) (23).

The core structure of the 5.8S and 26S rRNA is highly homologous to the 23S rRNA seen in archaea and bacteria or the 5.8S and 25S rRNA in yeast. However, despite the similarities in overall topology, the expansion segments (ESs) cannot be readily superimposed because of differences in their sequences (fig. S3) (13). With the exception of ES31 and ES41, RNA expansion segments do not occupy the ribosomal subunit interface (Fig. 1, A and B). Two clusters of RNA expansion segments serve as binding platforms for eukaryotic-specific proteins or eukaryotic-specific extensions of conserved ribosomal proteins.

The first cluster is composed of ES7 and ES39 and, to a lesser extent, ES9 and ES12 (fig. S4A). ES7 is the largest expansion segment of the 60S ribosomal subunit and can be divided into three parts (helices ES7A, ES7B, and ES7C). ES7A is evolutionarily variable in length and bridges the distance between ribosomal proteins RPL28 and RPL13 (Fig. 3, A and B). In most eukaryotes ES7A is stabilized by RPL28, whereas in budding yeast RPL28 is missing and ES7A is considerably shorter (11, 12). ES7B is positioned orthogonally with respect to ES7A, and the two helices thus define boundaries around the 5S rRNA and the juxtaposed ES9 and ES12 (Fig. 1A and Fig. 3A). ES7C protrudes from between ES7A and ES7B and is involved in several RNA-protein interactions (Fig. 1A and Fig. 3, A to D). ES39 is composed of two helices linked by an internal loop, which facilitates a sharp bend between the helices while also packing against the surface of the 60S ribosomal subunit. The opened loop structure also serves as a binding platform for the eukaryotic-specific protein RPL6 (Fig. 3, C and D).

Fig. 3

Architectural features of 60S ribosomal proteins. (A) Overview of ribosomal proteins in the vicinity of ES7 and ES39. Conserved proteins are shown in white; eukaryotic-specific proteins or protein extensions are color-coded. (B) Schematic representation of long-range and quaternary interactions of ribosomal proteins illustrated in (A). (C) Detailed view of ribosomal proteins interacting with ES39. (D) Schematic representation of the architecture of SH3 domain–containing proteins in the vicinity of ES39. SH3 domains are shown as large cylinders.

The second cluster is formed by ES5, ES19, ES31, ES20, and ES26, with outer boundaries of this region defined by helices H18 and H58 (fig. S4B). An internal loop between H18 and ES5 facilitates a sharp bend, which allows ES5 to be sandwiched between H18 and ES4 (the 5′ end of the 26S rRNA and the 3′ end of the 5.8S rRNA). ES19 is surrounded from one side by ES4 and ES3 of the 5.8S rRNA and from the other side by ES31, which contains a three-way junction, as well as ES20 and ES26, which together with ES31 create a binding platform for RPL27 (fig. S5, A and B).

Ribosomal proteins. The structure of the 60S ribosomal subunit contains 42 proteins of which 16 are present in all domains of life, 20 are shared between eukarya and archaea, and 6 are eukaryotic-specific (Fig. 2). Most ribosomal proteins contain eukaryotic-specific extensions, which are critical for establishing an intricate protein-RNA network (Fig. 3). Some of these extensions approach the conserved core regions of the ribosome, such as the active site and the exit tunnel (Fig. 4). The eukaryotic-specific proteins RPL22, RPL29, RPL36, and RPL28, as well as the eukaryotic/archaeal proteins RPL13, RPL34, and RPL38, are not homologous to other ribosomal proteins of known structure and hence expand the range of protein folds found in ribosomes (Fig. 2).

Fig. 4

Architecture of the peptidyl transferase center, the binding site of cycloheximide, and the exit tunnel of the 60S subunit. (A) Superposition of the T. thermophila 26S RNA active-site region (light blue) with T. thermophilus 23S rRNA (PDB code 2WDL, gray). Highlighted RNA elements are the active-site adenosine A2808 (bacterial A2451, red), the 8 o’clock helix (yellow), the P-loop (pink), and the T. thermophilus P-site tRNA (green). The N terminus of RPL29 is displayed as a purple sphere. (B) Architecture of ribosomal proteins in close vicinity to the active site. Eukaryotic-specific extensions of RPL4 (blue), RPL21 (yellow), RPL10 (green), and RPL3 (orange) are colored in a lighter shade. The locations of the peptidyl transferase center (PTC; red), RPL8 (L2) (pink), and RPL29 (purple) are indicated. (C and D) FobsFcalc difference Fourier map (green, contoured at 3.5σ) showing the binding site of cycloheximide at the tRNA E-site of the 60S subunit. Shown in (C) is a view from the inside of the subunit. The difference density reveals that the binding site of cycloheximide superimposes with A76 of the E-site tRNA (purple), based on superposition with the archaeal 50S subunit in complex with E-site tRNA mimic (38). The mutations in RPL27A (L28) (magenta) and RPL36A (L42) (marine) result in cycloheximide resistance. The rRNA is colored gray, with the proximal base C2754 (C2765 in yeast) in orange. Shown in (D) is a view toward the tRNA E-site [opposite direction relative to (C)]. The difference density reveals that the shape and the size of the density matches the molecular structure of cycloheximide shown in the inset. (E) Clipped view of the T. thermophila 60S ribosomal subunit showing the polypeptide exit tunnel with rRNA and proteins as gray and light blue surfaces, respectively. RPL4 (marine) and its eukaryotic-specific extensions at the surface and in the exit tunnel (red) are shown as ribbons. Superimposed elements include an aminoacylated P-site tRNA (green, PDB codes 2WDL and 2WDK) and the macrolide antibiotic erythromycin (yellow, PDB code 1YI2). The position of the P-site aminoacyl moiety of the tRNA is shown as a green sphere. (F) Conservation of RNA and protein elements around the exit tunnel. Eukaryotic-specific RNA and protein elements are color-coded; conserved regions are in gray.

In comparison to the prokaryotic 50S ribosomal subunits, the eukaryotic 60S subunit contains a vast network of additional protein-protein and protein-RNA interactions. These are particularly striking around ES7 and ES39, which form a central nexus on the back of the large ribosomal subunit (Fig. 1, A and C, and Fig. 3). The structure of the 60S subunit also reveals some additional architectural features not previously observed in bacteria or archaea. Long helices of eukaryotic-specific extensions of RPL4 and RPL7 extend above the surface of the subunit and form crane-like structures, bridging numerous eukaryotic-specific RNA expansion segments and proteins (Fig. 3, A and B). These features have not been observed in the 70S ribosome, where tertiary protein-RNA interactions are mostly mediated by positively charged extensions of ribosomal proteins that weave between the rRNA. Additionally, several proteins of the 60S subunit form a network of interactions by forming interprotein β sheets (Fig. 3B). In this regard, the eukaryotic-specific extension of RPL21 is used as a central mediator with two functions: First, it links the bridging helices by forming secondary structure elements with both RPL7 and RPL18A (L20). Second, together with eukaryotic-specific protein RPL29, it sandwiches ES12 and thus anchors it on the surface of the 60S subunit (Fig. 3, A and B).

RPL18A (L20) also has a second function in the stabilization of ES39 and its associated archaeal/eukaryotic ribosomal proteins, three of which [RPL6, RPL13A (L16), and RPL14] contain Src homology 3 (SH3) domains (Fig. 3, C and D). It contacts ES39 and also forms a cradle, in which RPL14 is oriented toward ES39 and a eukaryotic-specific extension of RPL13A (L16). Positioned on top of ES39 is RPL6, which acts as a cornerstone by contacting several RNA expansion segments (ES7A, ES7B, ES7C, and ES39) (Fig. 3, C and D). The interaction between RPL6 and ES39 is further stabilized by C-terminal helical extensions from RPL14 and RPL13A (L16) (Fig. 2 and Fig. 3, C and D). In the vicinity of ES39, three SH3 domains are used for both protein-protein and protein-RNA interactions (Fig. 3, C and D). Although other SH3 domain–containing proteins are scattered throughout the 60S subunit, these three proteins form an archaeal/eukaryotic-specific cluster.

Between the two RNA clusters, long terminal extensions of RPL13 and RPL36 mediate contacts among the tip of ES7A, ES9, and the base of ES5 (fig. S5C). Mutations of RPL36, RPL5, RPL11, and RPL35A (L33) have been associated with Diamond-Blackfan anemia, a congenital red blood cell aplasia, which can also be caused by mutations in 40S ribosomal proteins (24). During ribosome biogenesis, the 5S rRNA is assembled into the pre-60S subunit together with its flanking proteins RPL5 and RPL11 (fig. S5D). Therefore, mutations in these proteins most likely result in ribosome assembly defects (25, 26).

In analogy to the first RNA cluster, an extensive protein network exists in the vicinity of the second expansion segment area around ES31, ES20, and ES26 (fig. S4B). Anchored between helices H55 and H58, the archaeal/eukaryotic-specific protein RPL34 adopts an extended topology and harbors a zinc finger fold (figs. S4B, S5, and S6). Whereas its N terminus is deeply buried within the ribosomal core, its C-terminal helical half projects toward the surface of the 60S subunit, where it interacts with RNA expansion segments ES20 and ES31 as well as with the SH3 domain–containing, eukaryotic-specific ribosomal protein RPL27 (fig. S5, A and B). Positioned on top of this protein-RNA surface, RPL27 is involved in the interaction with two kink-turn binding proteins, RPL7A (L8, homologous to archaeal L7ae) and RPL30. Both RPL7A (L8) and RPL30 have additional functions apart from their ribosomal association. Previous studies have shown how yeast RPL30 recognizes its own pre-mRNA to prevent splicing (27, 28). Furthermore, archaeal L7ae is incorporated into the H/ACA ribonuclear protein family of pseudouridine synthase complexes, which use guide RNAs for targeted pseudo-uridylation (2730) (fig. S7). RPL38 and RPL22 are localized in the vicinity of RPL27 and the exit tunnel (Fig. 1). RPL38 has recently been shown to have a role in the selective recruitment of the Hox gene mRNAs to the 80S ribosomes in mice (31), whereas RPL22 can interact with noncoding RNAs of the Epstein-Barr virus (fig. S8) (21, 32).

With the crystal structures of the 40S and 60S ribosomal subunit from T. thermophila, we can now visualize the evolutionary expansion of the ribosome from bacteria to eukaryotes (fig. S9) (14). This is particularly evident for the coevolved areas around ES6 of the 40S subunit and the RNA-protein cluster in the vicinity of ES31 of the large subunit, where helical protein extensions emerge from the 60S subunit and play a role in intersubunit contacts (fig. S9) (13). Hence, long α-helical extensions emerge as mediators of long-range interactions within the 60S subunit as well as intersubunit contacts with the 40S subunit (Fig. 3 and fig. S9).

The peptidyl transferase center, the exit tunnel, and antibiotic binding sites. The 26S rRNA in the vicinity of the peptidyl transferase center is highly conserved throughout all kingdoms of life. Like the 23S rRNA in the 50S subunit, the 26S rRNA contains all regions essential for catalysis and substrate binding including the A-, P-, and E-sites, the 8 o’clock helix (helix 93), and the P-loop (5, 6). As a result, the rRNA structure of the active site can be readily superimposed with its bacterial counterpart from Thermus thermophilus (Fig. 4A). At the level of proteins, however, important differences can be observed with respect to their possible roles in the stabilization of the rRNA (Fig. 4, A and B). Eukaryotic-specific protein RPL29 extends from the surface of the 60S subunit to within 17 Å of the active-site adenosine A2808 (A2820 in yeast) (Fig. 4, A and B). Although genetic studies in yeast and mice have shown that RPL29 is not essential, deletion of this protein leads to reduced growth, presumably as a result of protein synthesis defects in both organisms (31, 33, 34). Interestingly, several other proteins (RPL4, RPL10, and RPL21) that are also located in close proximity to the active site are additionally interconnected on the surface of the 60S subunit by eukaryotic-specific extensions (Fig. 4B).

Prokaryotic and eukaryotic ribosomes are targets for numerous antibiotics. Cycloheximide and other glutarimide antibiotics are inhibitors of protein biosynthesis with high specificity for large eukaryotic ribosomal subunits (35, 36). This family of molecules is thought to inhibit translation elongation by interfering with deacetylated tRNA in the exit site of the ribosome, but the precise position of their binding site is currently limited by a lack of structural information (35, 37). Difference electron density maps calculated with our structure reveal that cycloheximide binds in a tight pocket on the 60S subunit (Fig. 4, C and D, and fig. S10) that was previously identified as the binding site for nucleotides C75 and A76 of E-site tRNA in the archaeal ribosome (38). Although the shape and size of the difference density agree very well with the bilobal structure of cycloheximide, we have not modeled it into the density because we cannot unambiguously assign the orientation of the molecule at 3.5 Å (Fig. 4, C and D). Our data are in agreement with the observation that point mutations in RPL36A (L42) and RPL27A (L28) in Chlamydomonas reinhardtii and budding yeast can result in resistance to cycloheximide (39, 40). Similarly, RNA footprinting analysis showed that a single cytosine (C2754; C2765 in yeast) is in the vicinity of the cycloheximide binding site (35). Therefore, despite their very different chemical structures, the mechanism of cycloheximide inhibition of eukaryotic protein synthesis may be related to the mechanism of mycalamide A and 13-deoxytedanolide, natural products with potent antitumor activities that sterically hinder tRNA binding to the exit site of the large subunit (41, 42).

Several antibiotics target the exit tunnel of prokaryotic ribosomes (43). The interior of the 60S ribosomal exit tunnel is mostly conserved, with some notable differences relative to the prokaryotic ribosomes. The site of constriction in the exit tunnel 20 Å below the peptidyl transferase center contains a highly conserved eukaryotic-specific insertion in the distal loop of RPL4, which interacts with helix 23 of the 26S rRNA (Fig. 4E and fig. S11). Hence, changes in the shape of the exit tunnel may result in reduced accessibility of the exit tunnel for larger macrolides. Interestingly, the recombinant insertion of six amino acids in the analogous region of bacterial L4, although too far away to directly contact the antibiotic, was shown to result in resistance to erythromycin and negatively affected SecM-mediated stalling of protein synthesis (44, 45). Another critical difference in the rRNA structure in this region of the ribosomal tunnel is the presence of G at position 2395 of the 26S rRNA (G2400 in yeast), which corresponds to A2058 in E. coli and G2099 in H. marismortui. The A2058G mutation in E. coli ribosomes prevents erythromycin binding, whereas the G2099A mutation in H. marismortui promotes it, implying that erythromycin binding depends on an A in this position (fig. S12) (46).

The region of the 60S ribosomal subunit at the tunnel exit is mostly conserved, with eukaryotic features restricted to the second tier of proteins surrounding the exit (Fig. 4F). Eukaryotic elements that could affect the binding of factors interacting with emerging nascent chains are the small ES24 of the 26S rRNA and the eukaryotic-specific protein RPL22 (Fig. 4F). Eukaryotic ES27 has been seen in EM reconstructions in two conformations, one of which positions it directly above the tunnel exit (11, 12). We observe weak electron density, insufficient to build an atomic model, corresponding to ES27 in the conformation above the tunnel in one of the four 60S molecules. In agreement with the observed flexibility, it has been shown that ES27 swings away from the tunnel exit area toward the L1 protuberance of the 60S subunit when the eukaryotic ribosome interacts with the protein-conducting channel (47). In general, the surface surrounding the tunnel exit is very flat, with no bulky eukaryotic features added in this region; this may be attributable to the requirement of the ribosome to approach the flat surface of the membrane during synthesis of membrane proteins (fig. S13).

RPL40—a ribosomal ubiquitin fusion protein. RPL40 is a zinc finger protein expressed as a ubiquitin fusion (figs. S6 and S14) (48). The structure of the 60S subunit reveals that RPL40 is positioned in close proximity to the sarcin-ricin loop and the elongation factor binding site (Fig. 5, A and B). The position of the unprocessed, N-terminal ubiquitin domain would sterically prevent elongation factor binding to the large ribosomal subunit. This is analogous to the role of the N-terminal ubiquitin domain of RPS27A (S31), which extends into the A site of the decoding center on the small ribosomal subunit and would prevent tRNA binding (14). It was previously shown that the cleavage of ubiquitin from RPS27A (S31) is essential for 40S maturation (48). The observation that the ubiquitin domain of RPL40 also occupies a functionally important region of the 60S subunit may suggest a mechanism by which fused ubiquitin domains could prevent immature ribosomal subunits from entering the translational cycle until they are processed. However, the exact timing of this processing step is currently unclear. A superposition of the small and large ribosomal subunits from T. thermophila with the crystal structure of a prokaryotic ribosome with bound elongation factor Tu (EF-Tu) and tRNA (7) demonstrates that both N-terminal ubiquitin fusions are ideally positioned to block an important functional center in each of the ribosomal subunits (Fig. 5A).

Fig. 5

Positions of ribosomal ubiquitin fusion proteins. (A) Model of the T. thermophila 80S ribosome from a superposition of 40S (PDB code 2XZM) and 60S ribosomal subunits with EF-Tu–bound 70S (PDB codes 2WRN and 2WRO). EF-Tu is shown in yellow, bound tRNA in green, 40S rRNA and proteins in orange, and 60S rRNAs and proteins in light blue. RPS27A (S31) and RPL40 are displayed as red ribbons, with N termini of the processed proteins highlighted as red spheres to indicate the positions of the ubiquitin moieties. (B) Detailed view of RPL40 (red cartoon) with bound zinc (light green) in close proximity to the sarcin-ricin loop (SRL; pink spheres), with surrounding RNA and proteins in blue.

Eukaryotic initiation factor 6. eIF6 is conserved from archaea to eukarya and has been implicated in the inhibition of 80S complex formation as well as 60S maturation (15, 49). Studies in yeast demonstrated that the highly conserved Shwachman-Bodian-Diamond syndrome (SBDS) protein and the guanosine triphosphatase (GTPase) Efl1 are involved in the removal of eIF6 from late preribosomal 60S subunits (19). SBDS acts as an adaptor by physically coupling the GTP hydrolysis of Efl1 with the release of eIF6 (50). Because all three proteins act in the same pathway, the deletion of SBDS in yeast and the resulting localization of eIF6 to cytoplasmic pre-60S particles can be suppressed by a set of eIF6 mutants. Most mutations result in reduced affinity of the factor for the large ribosomal subunit and have been mapped to the putative 60S binding face of eIF6 (19).

In agreement with earlier low-resolution cryo-EM studies, we find eIF6 bound to RPL23 and in close proximity to the sarcin-ricin loop, where it would prevent binding of the 40S subunit (Fig. 6, A and B) (18). These results are in contrast to chemical probing results with Sulfolobus solfataricus 50S and aIF6 that suggested a contact between aIF6 and helix 69, which in our structure is positioned 75 Å away from eIF6 (17). As a member of the pentein protein family, eIF6 has a five-fold pseudosymmetry and two large flat surfaces, the larger of which commonly serves as a scaffold for active-site residues in other members of the protein family (51). The C terminus of RPL23 mediates the interaction with the major binding surface of eIF6, which is centrally positioned above it (Fig. 6B). Both hydrophobic and hydrophilic residues on the surface of eIF6 are involved in the interaction with RPL23, with a buried surface area of 870 Å2 between the two proteins.

Fig. 6

Eukaryotic initiation factor 6 bound to the 60S subunit. (A) Top view of eIF6 (orange surface) on the T. thermophila 60S large ribosomal subunit. (B) Side view of eIF6 bound to the C terminus of RPL23 (blue) in proximity to the sarcin-ricin loop (SRL; pink surface) and RPL24 (light green). N and C termini of RPL23 are indicated. (C) Surface view of RPL23 (blue) and the sarcin-ricin loop (pink) with residues within 3.7 Å of eIF6 highlighted as white spheres. (D) Surface view of eIF6 (orange) with color-coded mutations known to interfere with 60S binding. Hydrophilic residues (red), hydrophobic residues (light blue), and residues required for the structural integrity of eIF6 (pale yellow) are indicated. Single-letter abbreviations for residues: A, Ala; C, Cys; D, Asp; G, Gly; I, Ile; K, Lys; L, Leu; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; Y, Tyr.

With the crystal structure of the 60S ribosomal subunit in complex with eIF6, we can now rationalize previous biochemical data because the electron density for the contact region is of good quality, with many side-chain features clearly visible (fig. S15). Residues critical for the association of eIF6 with the 60S subunit fall into three classes: hydrogen-bonding residues, hydrophobic residues, and residues that most likely contribute to the structural integrity of eIF6 (19). The arrangement of residues on the surface of RPL23 in close proximity to eIF6 mirrors the positions of residues on the surface of eIF6, which have been shown to be critical for the interaction with the 60S subunit (Fig. 6, C and D) (19).

Conclusion. The crystal structure of the large eukaryotic ribosomal subunit provides a complete atomic description of this central assembly in eukaryotic cells. It reveals a high degree of conservation of the active site and small but important differences in the features of the ribosomal tunnel. Clustered in different regions of the 60S subunit, RNA expansion segments and eukaryotic ribosomal protein elements have coevolved to form architectural complexes that mediate long-distance tertiary interactions. The structure also offers insights into the regulatory mechanisms of the eukaryotic ribosome, the maturation of the large ribosomal subunit, principles of antibiotic specificity and of the mechanism of translation inhibition by cycloheximide, and the structural basis of extraribosomal roles of various eukaryotic-specific ribosomal proteins. As such, it provides a starting point for future biochemical, genetic, and structural studies of protein synthesis in eukaryotes.

Supporting Online Material

www.sciencemag.org/cgi/content/full/science.1211204/DC1

Materials and Methods

SOM Text

Figs. S1 to S21

Tables S1 and S2

References (5272)

References and Notes

  1. See supporting material on Science Online.
  2. Acknowledgments: All data were collected at the Swiss Light Source (SLS, Paul Scherrer Institut, Villigen). We thank T. Tomizaki, M. Müller, V. Olieric, G. Pompidor, and A. Pauluhn for their outstanding support at the SLS; J. Rabl for advice on cell growth and ribosome purification as well as a clone of eIF6; T. Maier for advice on data collection; T. Bucher for the preparation of crystals; J. Erzberger and T. Bucher for critically reading the manuscript; and all members of the Ban laboratory for suggestions and discussions. Supported by the Swiss National Science Foundation (SNSF), the National Center of Excellence in Research (NCCR) Structural Biology program of the SNSF, and European Research Council grant 250071 under the European Community’s Seventh Framework Programme (N.B.) and by EMBO and Human Frontier Science Program fellowships (S.K.). Coordinates and structure factors have been deposited in the Protein Data Bank (accession codes for molecule 1: 4A1E and 4A18; molecule 2, 4A17 and 4A19; molecule 3, 4A1A and 4A1B; molecule 4, 4A1C and 4A1D). ETH Zürich has filed a patent application to use the crystals and the coordinates of the 60S ribosomal subunit for developing compounds that can interfere with eukaryotic translation.
View Abstract

Stay Connected to Science

Navigate This Article