Researchers are exploring whether these ubiquitous fluorinated molecules might worsen infections or hamper vaccine effectiveness.
Image credit: Shutterstock/Dmitry Naumov.
Edited by Harry F. Noller, University of California Santa Cruz, Santa Cruz, CA, and approved March 16, 2017 (received for review October 30, 2016)
Significance
Ribosome assembly in Escherichia coli is an extremely efficient process owing to the existence of assembly factors. Recent work indicates that some of these factors aid in the folding of the decoding center. The cryo-EM structure presented here includes an assembly factor testing the decoding fidelity of the mature 30S subunit before the particle is released to the pool of actively translating ribosomes. This finding reveals that in addition to their role as an assembly factor, these factors also may have a checkpoint function in the context of the mature ribosomal subunit. Understanding their specific functions may help identify key steps of the ribosome assembly pathway that will serve as molecular targets for new antibiotics.
Abstract
Recent work suggests that bacterial YjeQ (RsgA) participates in the late stages of assembly of the 30S subunit and aids the assembly of the decoding center but also binds the mature 30S subunit with high affinity. To determine the function and mechanisms of YjeQ in the context of the mature subunit, we determined the cryo-EM structure of the fully assembled 30S subunit in complex with YjeQ at 5.8-Å resolution. We found that binding of YjeQ stabilizes helix 44 into a conformation similar to that adopted by the subunit during proofreading. This finding indicates that, along with acting as an assembly factor, YjeQ has a role as a checkpoint protein, consisting of testing the proofreading ability of the 30S subunit. The structure also informs the mechanism by which YjeQ implements the release from the 30S subunit of a second assembly factor, called RbfA. Finally, it reveals how the 30S subunit stimulates YjeQ GTPase activity and leads to release of the protein. Checkpoint functions have been described for eukaryotic ribosome assembly factors; however, this work describes an example of a bacterial assembly factor that tests a specific translation mechanism of the 30S subunit.
Understanding how the components of the bacterial ribosome come together and organize themselves remains a daunting challenge. The assembly of the simplest of its subunits, the 30S ribosomal subunit, is a multistep process that starts with transcription of the 16S ribosomal RNA (rRNA) and synthesis of the ribosomal proteins (r-proteins). Folding of the 16S rRNA starts before transcription is completed. This process is intimately coupled with modifications to the RNA and processing of the precursor sequences (1, 2). The binding of its 21 r-proteins to the 16 rRNA occurs in a hierarchical manner (3⇓⇓–6), and their binding stabilizes the 3D interactions encoded by the sequence of the rRNA and simultaneously suppresses RNA misfolding (7, 8).
Ribosome assembly is a much more efficient process in the cell than in in vitro reconstitution experiments. The entire process of synthesis, folding, assembly, and maturation occurs in less than 2 min in actively growing Escherichia coli (9). Cells achieve this higher efficiency owing to the existence of nonspecific RNA and protein chaperones that assist in the early stages of assembly and, more importantly, through the action of specific assembly factors that act mainly act at later stages of the process (1, 2). Three protein factors—YjeQ (RsgA), Era, and RbfA—are involved in the late stages of assembly of the 30S subunit, and their functions in ribosome maturation have been studied extensively over the last decade (10⇓⇓–13). Recent work indicates that these factors bind the 30S subunit at or near the decoding center and aid its folding (14⇓⇓–17); however, the precise mechanisms and the functional interplay among them remain unclear.
This study focuses on YjeQ, a protein that exhibits weak intrinsic GTPase activity (18). Association of YjeQ with the 30S subunit results in a 160-fold stimulation of its GTPase activity (19, 20). A defining feature of YjeQ, along with other GTPases involved in ribosome biogenesis (RbgA and YqeH), is that its GTPase domain presents a permutation in the order of the characteristic GTPase loops (21⇓–23). The canonical G motifs mediating the nucleotide binding and hydrolysis, (G1 (Walker A, P-loop)-G2 (T)-G3 (Walker B)-G4 (N/TKxD)-G5 [(T/G)(C/S)A]), are circularly permutated and adopt a G4–G5–G1–G2–G3 pattern. A N-terminal oligonucleotide/oligosaccharide binding (OB-fold) domain and a C-terminal zinc-finger domain flank the GTPase domain. The OB-fold domain consists of antiparallel β-sheets defining a β-barrel. The carboxyl-terminal zinc-finger domain in YjeQ comprises a 310-helix, a long loop mediating the tetrahedral coordination of a zinc metal ion and two additional α-helices. Two important functional GTPase elements of YjeQ are switch I (G2-loop) and switch II (G3-loop). Structural work in the ras GTPase has shown that nucleotide binding and hydrolysis in GTPases of this family causes conformational changes in these two switches. In YjeQ, switches I and II are ideally positioned to propagate conformational changes between the GTPase domain and the other two domains in YjeQ.
The precise function of YjeQ in assisting the late stages of maturation of the 30S subunit is largely unknown. Structural characterization of a late 30S assembly intermediate that accumulates in an E. coli yjeQ null (10) strain initially suggested that YjeQ may first bind to the 30S subunit when it is still in an immature state to act as a RNA chaperone assisting the folding of the upper region of helix 44—an essential component of the decoding center. However, recent experiments revealed that the binding affinity of YjeQ to these immature particles is weak, suggesting that they transition to a more thermodynamically stable assembly intermediate that is no longer recognized by YjeQ (24).
Unexpectedly for an assembly factor, YjeQ binds to mature 30S subunits with high affinity (24). This property hints at the possibility that YjeQ function in ribosomal assembly might not be limited to late assembly intermediates, but rather may act on both immature and mature 30S subunits. In this context, work from the Himeno group has suggested that one of the functions of YjeQ is assisting in the release of RbfA once maturation of the ribosomal particle is complete (25). Another possible checkpoint role of YjeQ could be in blocking the binding of initiation factors to premature 30S subunits and ensuring quality control of the 30S subunit production (15). The role of the GTPase activity in these functions of YjeQ remains unclear.
A structure of YjeQ bound to a bona fide immature 30S particle has yet to be obtained; however, two independently produced cryo-EM structures show that YjeQ binds to the decoding center of the mature 30S subunit (14, 15). Interestingly, these two cryo-EM structures propose different binding orientations for YjeQ. In one of the structures, the OB-fold domain interacts with the decoding center, and the zinc-finger domain contacts the head of the 30S subunit (15). The second structure suggests a different orientation, with the OB-fold interacting with the platform, the zinc-finger domain binding helix 44, and the GTPase domain largely covering the decoding center (14). The moderate resolution (10–16 Å) of both cryo-EM structures, however, precludes the identification of essential cues for understanding the function and molecular mechanisms of YjeQ, including specific amino acids and individual rRNA helices that mediate the interactions between the protein and ribosomal particle. Furthermore, the existing cryo-EM structures or any of the available X-ray structures of YjeQ do not provide an accurate description of important functional motifs of YjeQ. These motifs include the first 34 N-terminal amino acids of Escherichia coli YjeQ that are essential for binding to the 30S subunit (19), as well as switch I. Therefore, it is difficult to derive precise testable models of YjeQ function from the existing cryo-EM and X-ray structures.
To gain key insights regarding the function of YjeQ in the context of the mature 30S subunit, we obtained the 3D structure of the mature 30S subunit in complex with YjeQ at 5.8-Å resolution using an electron microscope equipped with a direct electron detector. Consistent with previous structural work (14, 15), the cryo-EM map shows that YjeQ binds to the decoding center. The much higher resolution of this structure allows the identification of specific secondary structure elements of YjeQ. Consequently, the binding orientation of YjeQ to the 30S subunit can be precisely established. The structure shows that YjeQ anchors tightly to the three domains of the subunit, mainly through its N- and C-terminal domains. The OB-fold contacts the body of the 30S subunit, whereas the zinc-finger domain anchors the protein to both the head and platform domains of the 30S particle. The GTPase domain covers the decoding center almost completely and contacts the platform through a long loop. The 34 N-terminal amino acids of YjeQ are clearly visible in the EM map, and its location explains how this motif is essential for YjeQ binding to the 30S subunit. Switch I also is partially visible in the cryo-EM map, revealing a possible mechanism for the ribosome-activated GTPase activity of YjeQ.
We also obtained the cryo-EM map of the free 30S mature subunit, allowing us to visualize the structure of this particle when it is not confined in a crystal lattice. Surprisingly, we found that a long stretch of helix 44 is flexible and does not adopt the conformation described by the 30S subunit X-ray structure. However, binding of YjeQ to the free 30S subunit stabilizes helix 44 into a conformation similar to that on the X-ray structure. The specific interactions between the OB-fold of YjeQ and helix 44 indicate that YjeQ has a role as a checkpoint protein dedicated to test the decoding fidelity of the 30S subunit before the particle is released to the pool of actively translating ribosomes.
Results and Discussion
Cryo-EM Structure of the 30S+YjeQ Complex.
Previous biochemical work (13, 19, 20, 24) has revealed that the highest-affinity binding of YjeQ to the mature 30S subunit occurs in the presence of GMP-PNP. Thus, 30S+YjeQ complexes were assembled in the presence of this nucleotide analog and imaged by cryo-EM using a direct electron detector camera (Fig. S1A). This device allows for full correction of the beam-induced motion that the individual ribosomal particles experience during the image acquisition process (Fig. S1B). Performing a 3D classification of all particles in the dataset using the entire signal in the particles revealed the presence of both free 30S subunits and 30S+YjeQ complexes (Fig. S2A).
Fig. S1.
Cryo-EM images of the 30S+YjeQ complex and beam-induced motion correction. (A) A representative cryo-EM image of the 30S+YjeQ assembly reaction. This image was obtained using a Gatan K2 direct electron detector in counting mode, and the displayed image was produced after whole frame alignment (51) of the 30 frames contained in the original movies. (B) A vector map displaying the particle trajectories during the 15-s exposure used to collect the movies. Trajectories are exaggerated by a factor of 5 to allow visualization. Beam-induced motion correction was performed in individual particle images using the calculated trajectories (40).
Fig. S2.
3D particle classification. (A) A 3D classification of the dataset using the entire signal in the particles. The cryo-EM map obtained from all of the particles in the dataset is shown at the top, and the two classes obtained in this classification are at the bottom. One of the classes represents the 30S+YjeQ complex with a clear additional density attached to the decoding center. The other class represents free 30S subunits (30S subunit consensus structures) that are also present in the assembly reaction. (B) A 3D classification of the same dataset as in A but using a focused classification approach (43). During this classification, only the signal around the decoding center was considered. The consensus structure obtained from all of the particles in the dataset is shown at the top. The classification produces three classes, shown at the bottom. One of the classes has additional density in the decoding center and represents the 30S+YjeQ complex; the other two classes represent two different conformational subpopulations of the 30S subunit designated 30S subclass I and 30S subclass II.
The cryo-EM map of the 30S+YeQ complex was refined to a mean resolution of 5.8 Å (Fig. S3A) and revealed an additional density corresponding to YjeQ that covered the decoding center almost completely (Fig. 1 and Movie S1). Local resolution analysis revealed that the map exhibited isotropic resolution, with the areas representing the 30S subunit and YjeQ refining to resolution values around the mean value. None of the areas of the map refined to resolution values significantly lower than the mean value (Fig. S3 B and C), implying that YjeQ binds the 30S subunit tightly and adopts a single conformation.
Fig. 1.
Cryo-EM structure of the 30S+YjeQ complex. Shown are side (A), front (B), platform (C), and back (D) views of the 30S+YjeQ complex. Important landmarks of the 30S subunit and the three domains of the YjeQ protein are indicated. h44, helix 44.
Fig. S3.
Resolution analysis of the 30S+YjeQ structure. (A) Fourier shell correlation (FSC) plot for the 30S+YjeQ structure. We used a FSC value of 0.143 to report the resolution. (B) Local resolution analysis of the 30S+YjeQ structure performed with ResMap (46). (C) Histogram showing the number of voxels in the map at each resolution after local resolution analysis.
The entire 16S RNA fits perfectly in the electron density map (Fig. 2 A and B and Movie S2), albeit with minor rigid body movements of the head and platform domains with respect to the body, suggesting that binding of YjeQ causes small movements of the head backward and of the platform forward (Fig. S4). For 18 ribosomal proteins, the main chain, and even some side chains, were well defined in the electron density map (Fig. 2 C and D and Movie S2). The remaining three proteins (bS1, uS7, and bS21) showed poor or no electron density in the structure. uS7 shows good electron density at its N terminus, the region that anchors the protein to the head domain. The C terminus is highly exposed to the solvent, and thus the poor-quality electron density likely reflects the intrinsic flexibility of this region of uS7 (Fig. 2E and Movie S2). bS1 is normally lost during ribosome purification (26); therefore, we were not surprised to find only residual density for this protein (Fig. 2F). More surprising was the absence of bS21 (Fig. 2G). However, this protein is also absent in the two previously obtained cryo-EM structures of this complex (14, 15), as well as in some of the available X-ray structures of the mature 30S subunit (e.g., PDB ID codes 1FKA, 2UXD, and 2VQE).
Fig. 2.
Structural details of the 30S+YjeQ cryo-EM map. Densities of several structural elements of the 30S+YjeQ cryo-EM map with the derived atomic model are shown as color-coded ribbons. (A and B) Regions of the cryo-EM map showing the quality of the density around the indicated 16S rRNA helices. (C and D) Close-up views of specific regions of the 30S+YjeQ cryo-EM map with side chains of amino acids visible. (E and F) Zoom-in view of the density representing r-proteins uS7 and uS2. The C-terminal helix of uS2 (labeled) is not present in the X-ray crystal structure (38) but is visible in the cryo-EM map. The residual density observed for bS1 is labeled. (G) A density representing bS21 is not present in the cryo-EM map. (H) Density corresponding to YjeQ.
Fig. S4.
Binding of YjeQ changes the relative orientation of the domains in the 30S subunit. Overlap of the atomic model of the 30S+YjeQ complex overlaps with the X-ray structure of the free 30S subunit (PDB ID code 2AVY) (38), showing the displacement of the head (A) and platform (B) domains of the 30S subunit on binding of YjeQ.
The power of this structure is reflected in regions of the ribosomal proteins that were missing in previous crystal structures but are visible in this one. For instance, the C-terminal helix of uS2 was missing in previous ribosome structures, but it could readily be traced in our structure using the structure of the S2-S1 complex from E. coli (27) (Fig. 2F and Movie S2). Furthermore, r-protein uS5, which could be traced only up to residue Lys159 in previous crystal structures, has additional electron density at its C terminus, and residues Ser160-Leu165 are well defined in our map.
YjeQ Binds to the Decoding Center of the 30S Subunit in a Single Orientation.
The electron density for YjeQ was of excellent quality and allowed us to trace almost the entire polypeptide chain, including an N-terminal helix spanning residues 7–28 that is disordered in all other YjeQ crystal structures (21⇓–23) (Fig. 1, Fig. 2H, and Movie S1). The cryo-EM structure also showed that the three structural domains (OB-fold, GTPase, and zinc-finger domains) that are present in all YjeQ homologs (21⇓–23) exist in E. coli YjeQ as well.
The high resolution of the cryo-EM map allowed us to unambiguously determine the relative binding orientation of YjeQ to the 30S subunit (Fig. 3 and Movie S1). The orientation in this structure is consistent with that in a previously published cryo-EM structure (15). This structure clearly shows that the N-terminal OB-fold of YjeQ interacts with the body of the 30S subunit mainly through helix 18 (Fig. 3A), the GTPase domain contacts helix 44 and helix 24 (Fig. 3B) in the platform and the C-terminal zinc finger domain anchors the protein to the head through helices 29 and 30 and to the platform by contacting helix 45 (Fig. 3C). The relative orientation of YjeQ in this structure is different from that proposed in the cryo-EM map previously reported by our group (14), which places YjeQ with a 180° rotation around an axis perpendicular to the interface surface of the 30S subunit with the large subunit. With the limited resolution of our previous structure (16 Å), rigid body docking of the YjeQ X-ray structure in the cryo-EM map produced multiple solutions with similar fitting scores preventing the unequivocal positioning of YjeQ in the map.
Fig. 3.
The cryo-EM structure and atomic model of YjeQ bound to the 30S subunit. (A, Left) Zoom-in view of the density representing YjeQ in the cryo-EM map of the 30S+YjeQ complex. The atomic model of YjeQ is shown superimposed in the cryo-EM map. (A, Right) The equivalent close-up view of the atomic model of the 30S+YjeQ complex with the assembling factor binding to the decoding center of the 30S subunit. The rRNA and r-proteins interacting with YjeQ are labeled. (B) Front view of density representing YjeQ in the cryo-EM map and the atomic model. (C) Platform view of the YjeQ density in the cryo-EM map and the corresponding atomic model. All panels show equivalent orientations to Fig. 1.
To explore the possibility that YjeQ may be binding the 30S subunit in more than one orientation, we performed focused classification with the entire dataset (Fig. S2B). To this end, we kept the signal in the particle images corresponding to YjeQ and its binding area in the decoding center. The remaining part of the 30S subunit was masked out and subtracted from the particle images. A 3D classification with the signal-subtracted particle dataset produced a similar result as the classification performed with the entire signal in the particles (Fig. S2A). One class exhibited an extra density identical to that described above for YjeQ, whereas the other two classes lacked any additional density and were consistent with free 30S subunits. However, no class presented an additional density that could be consistent with alternative binding orientations for YjeQ.
Therefore, we conclude from this analysis that YjeQ binds to the mature 30S subunit in only one orientation (Fig. 1). This orientation is consistent with one of the previously published cryo-EM structures (15) and chemical modification experiments (28) showing enhanced protection of 16S rRNA bases localized at the interface between YjeQ and the 30S ribosomal subunit.
The Structure of the Mature 30S Subunit in Solution Differs from That Described by X-Ray Crystallography.
The structure of the entire 30S subunit at atomic resolution was first described using X-ray crystallography (29). This structure demonstrates that the 16S rRNA largely determines the overall shape of the 30S subunit, which folds into four different domains. In the canonical front view of the 30S subunit (Fig. 4A), the 5′ domain forms the body of the particle with the shoulder at the top left side and the spur at the lower left. The central domain of the 16S rRNA forms the platform that occupies the top right side of the particle. The 3′ major domain constitutes the bulk of the head, and the 3′ minor domain is also an integral part of the body. In this structure, the 3′ minor domain of the 16S rRNA consisting of helices 44 and 45 occupies the subunit interface with the 50S subunit, and both helices are perfectly structured. Helix 44 stretches from the bottom of the body to the bottom of the head, and helix 45 is connected to the top end of helix 44, adopting an orientation almost perpendicular to the preceding helix.
Fig. 4.
Cryo-EM structure of the 30S subunit. (A) Density map of the 30S subunit obtained by applying a low pass-filter at 7-Å resolution to the X-ray structure (PDB ID code 2AVY). Structural landmarks of the 30S subunit are labeled. (B) Consensus cryo-EM structure obtained for the 30S subunit. (C) Cryo-EM structure of the 30S subunit subclass I obtained from focused classification. (D) Cryo-EM structure of the 30S subunit subclass II obtained from focused classification. In B–D, the region of helix 44 showing missing density or protruding outward is indicated by an arrow.
The presence of both 30S+YjeQ complexes (48% of the particles) and free 30S subunits (52% of the particles) in our sample provided an opportunity to obtain the structure of the free 30S subunit (Fig. S2A). This cryo-EM structure showed no substantial differences from the crystal structure in the body, platform, and head domains; however, helix 44 in the 3′ domain showed significant discrepancies. We designated this cryo-EM map the 30S subunit consensus structure. In the X-ray structure, helix 44 represents the single longest helix in the subunit and runs in a defined single conformation from the bottom of the body to the neck region of the 30S subunit (Fig. 4A). A defined density representing the upper region of helix 44 is completely absent in the cryo-EM map, however, although helix 45 still adopts similar conformations in both the cryo-EM map and the X-ray structure of the 30S subunit (Fig. 4B and Fig. S2A).
To investigate whether the upper region of helix 44 is present in several discrete conformations or rather is a continuous of multiple conformations, we analyzed the two classes representing free 30S subunits in the focused classification performed with all particles in the dataset (Fig. S2B). One of the classes (subclass I) was very similar to the 30S consensus structure described above, with most of the density representing the upper motif of helix 44 missing (Fig. 4C and Fig. S2B). Interestingly, in the second class (subclass II), the entire length of helix 44 was visible; however, the upper region of helix 44 was not latched to the decoding center as described by the crystal structure. Instead, this entire section of the helix protruded from the surface of the 30S subunit, thereby distorting the interface with the 50S subunit (Fig. 4D and Fig. S2B). We found no class in which helix 44 adopted the conformation displayed in the X-ray crystal structure (Fig. 4A and Fig. S2B).
Local resolution analysis of the consensus 30S subunit, subclass I, and subclass II maps (Fig. S5 B and C) revealed that an isotropic resolution of the structures, with most areas exhibiting resolution values similar to the mean resolution value measured for the entire map (Fig. S5A). Only isolated areas on the platform, back of the head, and neck showed slightly lower resolution. More importantly, in the consensus and subclass I structures, the part of helix 44 that was visible showed a resolution value close to the mean resolution. However, in the map obtained for subclass II, with the entire helix 44 defined by density, its upper domain refined to a slightly lower resolution, suggesting that this region is partially flexible.
Fig. S5.
Resolution analysis of the 30S subunit structure. (A) FSC plots for the 30S subunit consensus, 30S subunit subclass I, and 30S subunit subclass II structures. A FSC value of 0.143 was used to report the resolution. (B) Local resolution analysis of the three structures performed with ResMap (46). (C) Histogram showing the number of voxels in the maps at each resolution after local resolution analysis.
To establish whether any of the helix 44 conformations observed in the free 30S particles were induced by the presence of YjeQ, we imaged a sample containing purified mature 30S subunits in the absence of YjeQ. Using all particle images in the dataset produced a cryo-EM map similar to the 30S subunit consensus structure described above, which we designated the 30S subunit average structure (Fig. S6). Image classification revealed the existence of three subpopulations. One class (subclass Ia), representing a 22% of the population, was equivalent to the 30S subclass I (Fig. 4C and Fig. S2B) and lacked the density for the upper section of helix 44 (Fig. S6). The second class (subclass IIa) had the upper region of helix 44 unlatched from the decoding center (Fig. S6) and thus was similar to the 30S subclass II (Fig. 4D and Fig. S2B); 26% of the particle images belonged to this class. The third subpopulation was unique to this sample and produced a cryo-EM map lacking density for the entire helix 44 (subclass IIIa) (Fig. S6). Interestingly, these particles represented 51% of the population, similar to the percentage of particles representing 30S+YjeQ complexes in the sample containing both 30S subunits and YjeQ (48%). This result suggests that YjeQ may predominantly bind 30S subunits with helix 44 in a flexible conformation. Intriguingly, these 30S particles are the ones most resembling immature assembly intermediates (10⇓–12).
Fig. S6.
Cryo-EM structures of 30S subunit subpopulations in a sample containing mature 30S subunits. A sample containing purified mature 30S subunits was imaged by cryo-EM and subjected to image classification. The 3D reconstruction at the top, designated the 30S subunit average structure, was obtained by combining all of the particle images in the dataset. The cryo-EM maps at the bottom represent the three conformational subpopulations identified in this sample. The conformation of helix 44 is different in the three cryo-EM maps. Labels indicate the numbers of particles used for each structure and the percentage of the total population that they represent.
We concluded that when the 30S subunit is in solution and free of the stabilizing contacts provided by a crystal lattice or the 50S subunit, helix 44 adopts multiple conformations. These conformations do not seem to be induced by YjeQ, because they are also present in the absence of the factor. Surprisingly, the specific conformation described by the crystal structure of the 30S subunit was not observed among the population of free 30S subunits in our sample.
YjeQ Stabilizes Helix 44 in a Conformation Suggesting a Checkpoint Role in Ribosome Fidelity.
Previous structural work has established an essential role of the universally conserved A1492 and A1493 nucleotides in the decoding process and proofreading mechanisms of the ribosome (30). Decoding the mRNA requires the correct recognition of each A-site codon by the anticodon of the corresponding aminoacyl-tRNA (aa-tRNA). Interestingly, the ribosome performs this step with a much higher level of accuracy than can be explained by the difference in binding energy between the cognate and noncognate codon-anticodon pairing (31). This observation suggested early on that the ribosome has built-in proofreading capabilities. Nucleotides A1492 and A1493 in helix 44 mediate one of these mechanisms. During the decoding process, these two bases swivel out and their N1 atoms form hydrogen bonds with the 2′ OH groups on both sides of the codon-anticodon helix. Through the simultaneous interaction with both strands of the codon-anticodon helix, these two adenines can monitor the base-pairing geometry of the codon-anticodon and sense the distortions arising from mispairing (30, 32).
Interestingly, when we compared the cryo-EM structures of the 30S+YjeQ complex with the structure of the free 30S subunit obtained from the same set of images, we found that the binding of YjeQ stabilizes helix 44 in a conformation similar to that observed in the X-ray structure (Fig. 4A). The assembly factor sits over the decoding center, interacting mainly through the GTPase and OB-fold domains (Fig. 3). The GTPase domain interacts with helix 44 mainly through switch I, switch II, and the loop connecting the β6 and β7 strands in this domain (Fig. 5 A and B). These contacts occur in the last turn of helix 44, immediately before the rRNA continuous toward helix 45. In the OB-fold, the loop connecting strands β1 and β2 impinges on the final turn of helix 44, causing reorganization of the electron density in the area surrounding A1492 (Fig. 5 A and C). Analysis of this region on the cryo-EM map revealed that the density corresponding to the adenine moiety of A1492 lies outside the helical axis (Fig. 5 C and D), indicating that A1492 flips out from the helix. This conformation is stabilized through hydrophobic interactions between the adenine moiety and the OB-fold of YjeQ, as well as π-stacking interactions between the side chain of Phe48 and A1408 (Fig. 5D). The Phe-mediated stabilization of the unpaired nucleotide resembles the structure of the mismatch repair protein MutS in complex with a DNA including a one-nucleotide insertion, where the side chain of Phe36 promotes extrusion of the unpaired nucleotide through an aromatic ring stacking with the cDNA strand (33).
Fig. 5.
YjeQ checkpoint role in ribosome fidelity. (A) Interactions of YjeQ switches I and II with helix 44 of the 16S rRNA. (B) Loop between strands β6 and β7 in the GTPase domain of YjeQ interacting with helix 44. (C) Density representing helix 44 in the 30S+YjeQ cryo-EM map showing a disruption of the rRNA ribbon in this region. (D) Atomic model fitted into the cryo-EM density map showing that A1492 on the 16S rRNA adopts a flip-out conformation. A1493 adopts the standard configuration seen in the X-ray structure of the free 30S subunit. Phe48 in the loop connecting strands β1 and β2 in the OB-fold domain of YjeQ inserts itself into helix 44 and stabilizes adenosine 1408 through π-stacking interactions.
The conformation of A1492 resembles that exhibited by nucleotides A1492 and A1493 during proofreading of the codon-anticodon pairing (34); however, in our structure, base A1493 did not swivel out of the helix. This finding suggests that YjeQ plays a role as a checkpoint protein in the mature 30S subunit by testing the ability of the 30S subunit to perform proofreading before the subunit is released to the pool of actively translating ribosomes. Simultaneously, the binding location of YjeQ covering essential intersubunit bridges also prevents premature association with the 50S subunit.
This function of YjeQ suggests that, similar to the eukaryotic ribosome, bacteria also have specific quality control mechanisms to survey nascent ribosomes and test their functionality. Quality control mechanisms in eukaryotic cells during ribosome maturation have been studied extensively (35). In these cells, some factors simply prevent nascent subunits from entering the translation by sterically and allosterically blocking binding of initiation factors, tRNA, or 60S subunits (36), but other factors perform functional checks in the assembling ribosomes by mimicking elements or steps of the translational cycle. The best-characterized example of this type of assembly factor is the pre-40S subunit in yeast (37). During the late stages of assembly of this ribosomal particle, seven different factors (Tsr1, Rio2, Dim1, Nob1, Pno1, and Enp1/Ltv1) bind to the subunit interface of the 40S particle and cooperate to block each step in the translation initiation pathway. Interestingly, release of these assembly factors requires a translation-like cycle, during which translation factors must function. This cycle probes key functional properties on the pre-40S particle, including the ability to associate and correctly position the 60S subunit and the ability to bind important elements in the translation cycle (eIF5B, Fap7, Rli1, and Dom34).
Previous studies (14, 15) have suggested a checkpoint function for YjeQ or other factors (11, 12, 16) in aiding the assembly of the bacterial ribosome. However, none of these studies was able to pinpoint a specific checkpoint function in which the factor performs a functional check in assembling the ribosomal particle. Previous structural work has suggested a general checkpoint role for YjeQ by sterically blocking the binding of initiation factors to the 30S subunit or its association with the large 50S subunit (15). The cryo-EM structure presented here provides evidence of a bacterial assembly factor testing a specific translation mechanism of the 30S ribosomal subunit before the particle is released to the pool of actively translating ribosomes. The remaining question is how deleterious are the effects of eliminating these quality control mechanisms on bacterial fitness, pathogenicity, and survival. It will be exciting to test whether their suppression or inhibition leads to large populations of malfunctioning ribosomes, affecting the bacterial proteome and in turn creating an observable phenotype in these bacterial cells.
The Cryo-EM Structure of the 30S+YjeQ Complex Suggests a Role for the N-Terminal Region of YjeQ in Promoting the Release of RbfA.
One specific function that has been attributed to YjeQ during ribosome biogenesis is assisting the release of RbfA from the mature 30S subunit once the maturation of the particle is complete (13, 25). How this functional interplay between YjeQ and RbfA is implemented remains unclear, however. The RbfA protein binds to the small ribosomal subunit at the junction between the body and head and dramatically alters the positions of helices 44 and 45 (Fig. S7), placing this region in a conformation unsuitable for the subunit’s participation in protein synthesis (17).
Fig. S7.
Structure of the mature 30S subunit in complex with RbfA. RbfA binds to the small ribosomal subunit at the junction between the body and head and dramatically alters the position of helices44 and 45 (17).
The cryo-EM structure of the 30S+YjeQ complex presented here suggests a dual mechanism through which YjeQ may facilitate the release of RbfA from the 30S subunit. We find that binding of YjeQ to the 30S subunit has a significant stabilizing effect in the upper region of helix 44 (Figs. 1 and 4). This is the same rRNA motif that appears severely disrupted on RbfA binding (Fig. S7). Therefore, it is likely that binding of YjeQ to the 30S subunit forces helix 44 back into the normal decoding position and triggers the release of RbfA. In addition, E. coli YjeQ includes a 34-aa N-terminal extension immediately preceding the OB-fold domain. In the cryo-EM map, this N-terminal region is visible and defines an α-helix that points into the neck region in the same area that has been described as the binding site for RbfA (Fig. 3A). We concluded that the stabilizing effect of YjeQ in the conformation of helix 44, combined with the insertion of the N-terminal α-helix of YjeQ in the binding site of RbfA, likely creates the necessary conditions to force the removal of RbfA factor from the mature 30S subunit.
Interestingly, this N-terminal stretch of amino acids is not conserved among all bacterial species. It is present in E. coli and Salmonella typhimurium YjeQ proteins. In S. typhimurium, the N-terminal region is eight amino acids longer (23); however, this region is disordered in the X-ray structure of S. typhimurium YjeQ, and thus the conformation and relative orientation with respect to the other protein domains remain unknown. YjeQ orthologs from bacterial species including Thermatoga maritima, Bacillus subtilis (called YloQ), Pseudomonas putida, Mycoplasma pneumoniae, and others lack this N-terminal motif (21, 22). In these species, whether YjeQ can remove RbfA from the 30S subunit is unclear, but the absence of this N-terminal motif may suggest that it is not essential for this function.
The C-terminal helix in the zinc-finger domain of YjeQ is also essential for its ability to remove RbfA from the 30S subunit (13). Interestingly, when YjeQ is bound to the 30S subunit, this C-terminal helix lies far away from the RbfA-binding site (Fig. 3 B and C). Nevertheless, our cryo-EM map shows that the zinc-finger domain of YjeQ constitutes a major anchor point of the protein to the 30S subunit head and platform. In this domain, the long loop that mediates the tetrahedral coordination of the zinc ion mediates strong interactions with helices 29 and 30 in the 16S rRNA head domain (Fig. 3C). The interaction with the platform occurs mainly through the C-terminal helix of the zinc-finger domain. The main interactions between this motif of YjeQ and the platform include the side chain of Arg331 interacting with the backbone of helix 24 and the C terminus of the protein interacting with the rRNA loop connecting helices 44 and 45 (Fig. 3C). These interactions, along with the results of a previous mutational study (13), suggest that the zinc-finger domain anchors YjeQ on its binding site, allowing for proper positioning of the OB-fold and GTPase domains. In turn, the OB-fold and GTPase domains function as effector elements for the release of RbfA from its binding site.
In conclusion, this structure of the 30S+YjeQ complex provides a detailed specific testable model describing the functional interplay between YjeQ and RbfA.
Role of the GTPase Activity in the YjeQ Function.
Although the GTPase activity of YjeQ was characterized early on (18⇓–20), the role of this activity in the overall function of YjeQ and its regulation remains unclear. The present cryo-EM structure of the 30S+YjeQ complex shows that through extensive interactions involving the OB-fold and zinc finger domains, YjeQ interacts with three of the four domains of the 30S subunit (body, head, and platform) (Fig. 1 and Movie S1). These interactions are important for the functionality of YjeQ (19). For example, deletion of the first N-terminal 20 amino acids of YjeQ significantly decreases the binding of YjeQ to the 30S subunit, whereas removal of the OB-fold domain completely suppresses any association with the ribosomal particle. Similarly, partial or complete removal of the C-terminal zinc finger domain also abolishes YjeQ binding (13).
The interaction of YjeQ with the 30S subunit places the GTPase domain in direct contact with the upper part of helix 44 (Fig. 3B). This area of helix 44 constitutes the ribosomal motif undergoing the largest conformational change on YjeQ binding. Consequently, in the 30S+YjeQ complex, the GTPase domain of YjeQ is ideally placed to monitor these changes. The atomic model of the 30S+YjeQ complex describes the structural determinants of YjeQ that are likely probing the conformation of helix 44. On YjeQ binding, switch I and switch II lie at the interface with the 30S subunit (Fig. 5A). The tip of the loop between strands β6 and β7 in the GTPase domains reaches helix 44 and also may be involved in the monitoring function (Fig. 5B).
The GTPase activity of YjeQ is stimulated by >160-fold in the presence of mature 30S subunits (18⇓–20). Although the structure does not reveal what triggers this stimulation, it is plausible that the restrained conformation of helix 44 may stimulate the GTPase activity in YjeQ. If this were the case, then the conformational change would be first sensed by switch I. Consistently, the crystal structure of YjeQ implies that the flexible switch I region of YjeQ must experience conformational changes to stimulate the GTPase activity (21). Once GTP hydrolysis has occurred, additional conformational changes in switch I and switch II trigger a reorganization of their interactions with the 30S subunit, causing an overall decrease in the binding affinity of YjeQ (19, 24) and its release from the 30S subunit.
This model is supported by the conformational differences existing between YjeQ in our cryo-EM structure (GMP-PNP-bound form that mimics the GTP state) and that observed in the YjeQ crystal structure of T. maritima YjeQ (GDP-bound form) (21). Superimposition of the GTPase domain of these two structures yielded a 1.6-Å rmsd for more than 500 atoms and revealed the reorientation of both the OB-fold and zinc-finger domains (Fig. 6). The large contact area and weak interactions among YjeQ domains (22) allow for ready propagation of the conformational changes from the GTPase domain to neighboring domains.
Fig. 6.
Conformational change of YjeQ on GTP hydrolysis. Opposite views of YjeQ from T. maritima (PDB ID code 1u0l; white; GDP bound) superimposed onto the structure of E. coli YjeQ bound to the 30S ribosome (blue; GMP-PNP bound) from the cryo-EM structure. The zinc metal ions are shown as spheres, and the GMP-PNP molecules bound to YjeQ are shown as color-coded sticks.
Therefore, we concluded that the GTPase activity of YjeQ functions as a sensor to facilitate the release of the protein factor from the 30S subunit once YjeQ has performed its functions. Our structure suggests that a specific conformation of helix 44 stimulates the GTPase activity of YjeQ. Hydrolysis of GTP and the associated conformational changes in YjeQ then lead to reorganization of the interface of the complex and, in turn, release of the factor. Our structure also rules out the possibility that the energy from GTP hydrolysis is necessary to place helix 44 in its normal decoding position.
Conclusion
Biochemical and genetic studies over the last decade have suggested that YjeQ is an assembly factor that assists in the late stages of assembly of the 30S subunit (10, 19, 20). The specific roles of YjeQ at those critical stages of the assembly process remain elusive, however. The present evidence suggests that YjeQ is an important factor in the assembly of the functional core of the subunit (10, 14). Recent work (24) measuring the binding affinity of YjeQ to ribosomal particles revealed that YjeQ also binds the mature 30S subunit with high affinity. The cryo-EM structure of the 30S+YjeQ complex suggests a specific mechanism through which YjeQ tests the decoding fidelity of the 30S subunit and promotes the release of RbfA (Fig. 7). Both functions occur in the context of the mature 30S subunit instead of the immature ribosomal subunit. Taken together, these results imply that the function of YjeQ is not restricted to the immature 30S particles and some of the roles of YjeQ during ribosomal assembly occur in the context of mature 30S subunit. The data presented here do not provide information on how YjeQ may be exerting other functions in preribosomal particles. Additional work and future structures of YjeQ in complex with bona fide preribosomal particles should clarify the molecular mechanisms of how this protein assists the maturation of the functional core of the 30S subunit.
Fig. 7.
Diagram illustrating the function and mechanism of YjeQ. YjeQ likely binds to the pre-30S subunit when it is still bound to RbfA. RbfA binding dramatically alters the positions of helices 44 and 45 (stage 1). Binding of YjeQ to the 30S subunit forces helix 44 back into the normal decoding position and triggers the release of RbfA. YjeQ also causes A1492 to flip out from helix 44 to test the proofreading ability of the 30S subunit (stage 2). This conformation of helix 44 may trigger the ribosome-activated GTPase activity in YjeQ. Hydrolysis of GTP causes a reorganization of the interactions between YjeQ and the 30S subunit that leads to the release of the factor. The 30S subunit is then free to join the pool of actively translating ribosomes (stage 3).
Materials and Methods
Protein and 30S Ribosomal Subunit Purification.
Detailed descriptions of all experimental procedures used for protein expression and purification, as well as the purification of 30S ribosomal subunits, are provided in SI Materials and Methods.
Cryo-EM, Image Processing, and Model Building.
The procedures used for cryo-EM, image processing, and model building are described in detail in SI Materials and Methods.
SI Materials and Methods
Cell Strains and Protein Overexpression Clones.
The parental strain E. coli K-12 (BW25113) use to produce the mature 30S subunits was obtained from the Keio collection, a set of E. coli K-12 in-frame, single-gene knockout mutants (39). The pDEST17-yjeQ plasmid used to overexpress YjeQ protein with an amino-terminal His6 tag cleavable by tobacco etch virus (TEV) protease was generated as described previously (19).
Protein Overexpression and Purification.
YjeQ protein was overexpressed as an amino-terminal His6-tag protein by transforming E. coli BL21-A1 with the pDEST17-yjeQ plasmid described above. Typically, 1 L of LB medium was inoculated with 10 mL of saturated overnight culture, and cells were grown to an OD600 of ∼0.6 by incubation at 37 °C. Expression was induced with 0.2% l-arabinose. Cells were then induced for 3 h at 37 °C and harvested by centrifugation at 3,700 × g for 10 min. Cell pellets were washed with 1× PBS buffer (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4 at pH 7.4) and resuspended in 20 mL of buffer A [50 mM Tris⋅HCl pH 8.0, 500 mM NaCl, 5% (vol/vol) glycerol] containing a protease inhibitor mixture (cOmplete Protease Inhibitor Mixture Tablets; Roche). Cells were lysed by sonication, the lysates were spun at 39,200 × g for 45 min to clear cell debris, and the supernatant was collected. The lysate was then filtered with a 0.45-μm syringe filter (Millipore) and loaded onto a HiTrap Nickel Chelating Column (GE Healthcare Life Sciences) that had been equilibrated with 50 mM Tris⋅HCl pH 8.0, 0.5 M NaCl, and 5% (vol/vol) glycerol. Nonspecifically bound proteins were washed with 45 mM to 90 mM imidazole. YjeQ was eluted with 180 mM imidazole. Purity of the fractions was monitored by SDS/PAGE, and fractions containing each respective protein were collected and pooled together to dialyze overnight in buffer containing 50 mM Tris⋅HCl pH 8.0, 200 mM NaCl, and 5% (vol/vol) glycerol.
The amino-terminal His6-tag was removed by digestion with TEV protease at a ratio of 10:1 (YjeQ:TEV). Following digestion, the reaction was loaded onto a HiTrap metal chelating column that had been equilibrated with 50 mM Tris⋅HCl pH 8.0, 0.5 M NaCl and 15 mM imidazole. Fractions were collected, and their purity was evaluated by SDS/PAGE and staining with Coomassie Brilliant Blue. Fractions containing pure untagged YjeQ were pooled and dialyzed against 50 mM Tris⋅HCl pH 8.0, 150 mM NaCl, and 5% (vol/vol) glycerol overnight. The protein was concentrated using a 10-kDa cutoff filter (Amicon), and the purified YjeQ was frozen in liquid nitrogen and stored at −80 °C.
Purification of 30S Ribosomal Subunits.
The parental strain (BW25113) was used for purification of the mature 30S subunits. Here 1 L of LB media was grown at 37 °C to an OD600 of 0.6. Cells were harvested by centrifugation at 3,700 × g for 10 min. Cell pellets were resuspended in 7 mL of buffer containing 20 mM Tris⋅HCl at pH 7.5, 10 mM magnesium acetate, 100 mM NH4Cl, 0.5 mM EDTA, 3 mM 2-mercaptoethanol, and a protease inhibitor mixture (cOmplete Protease Inhibitor Mixture Tablets; Roche) and DNaseI (Roche). Each of the subsequent steps was performed at 4 °C. The cell suspension was passed through a French pressure cell at 1,400 kg/cm2 three consecutive times to lyse the cells. The lysate was spun at 30,000 × g for 40 min to clear cell debris. The clarified lysate was collected and spun down at 138,488 × g for 132 min to pellet the ribosome. The pellet was resuspended in buffer containing 10 mM Tris⋅HCl pH 7.5, 10 mM magnesium acetate, 500 mM NH4Cl, 0.5 mM EDTA and 3 mM 2-mecaptoethanol. The resuspended solution was spun down at 30,000 × g for 15 min, after which the supernatant was loaded onto a sucrose buffer containing 30% sucrose, 20 mM Tris⋅HCl pH 7.5, 10 mM magnesium acetate, 500 mM NH4Cl, 0.5 mM EDTA, and 3 mM 2-mecaptoethanol, and then spun down for 16 h at 100,000 × g. The washed ribosome pellet was resuspended in buffer F containing 10 mM Tris⋅HCl pH 7.5, 1.1 mM magnesium acetate, 60 mM NH4Cl, 0.1 mM EDTA, and 3 mM 2-mecaptoethanol, which caused dissociation of subunits. Approximately 120 A260 units of resuspended crude ribosomes were then applied to 34 mL of 10–30% (wt/vol) sucrose gradients prepared with buffer F. The gradients were centrifuged for 16 h at 40,000 × g on a Beckman Coulter SW32 Ti rotor. Gradients were fractionated using a Brandel fractionator apparatus and an AKTAprime FPLC system (GE Healthcare). The elution profile was monitored by UV absorbance at A260, and fractions corresponding to the 30S subunit peak were pooled and spin down for another 16 h at 40,000 × g on a Beckman SW32 Ti rotor. Pellet containing the purified 30S subunits were resuspended in buffer E containing 10 mM Tris⋅HCl pH 7.5, 10 mM magnesium acetate, 60 mM NH4Cl, and 3 mM 2-mecaptoethanol and stored at −80 °C.
Cryo-EM.
The entire dataset of images was collected over 10 cryo-EM sessions. Assembly of the 30S+YjeQ complexes for these experiments was done in 20-μL reactions containing assembly buffer (10 mM Tris⋅HCl pH 7.5, 10 mM magnesium acetate, 150 mM NH4Cl, 3 mM 2-mercaptoethanol, and 2 mM GMP-PNP). The concentration of the 30S subunits in the assembly reactions for the multiple cryo-EM sessions was always maintained at 1 μM, whereas the concentration of YjeQ was either 5 μM or 7 μM, depending on the reactions. The assembly reaction was incubated for 30 min at 37 °C and then diluted in the same buffer between 10 and 20 times before the reaction was applied to the grid. Using these assembly and dilution conditions, depending on the cryo-EM session, the concentration of 30S subunits in the sample applied to the grid ranged from 50 nM to 100 nM, and that of YjeQ ranged from 300 nM to 700 nM. Concentrations of YjeQ >700 nM in the reaction applied to the grid caused a significant drop in the number of 30S subunit particles absorbed to the grid. Approximately 4 μL of the diluted sample was applied in the holey carbon grids (c-flat CF-2/2–2C-T) with an additional layer of continuous thin carbon (5–10 nm).
The dataset from the sample containing only free 30S subunits was obtained in a single cryo-EM session from cryo-EM grids prepared in the same manner as those used in the 30S+YjeQ sample. Purified 30S subunits were diluted in assembly buffer to a concentration of 50 nM and applied directly to the grid as described above.
Before the samples were applied, grids were glow-discharged in air at 5 mA for 15 s. Vitrification of samples was performed in a Vitrobot (FEI) by blotting the grids once for 15 s with an offset of 0 before plunging them into liquid ethane. Grids were loaded into an FEI Tecnai F20 electron microscope operated at 200 kV using a Gatan 626 single-tilt cryoholder. Images were collected in a Gatan K2 Summit direct detector camera. This detector was used in counting movie mode with 5 electrons per pixel per second for 15-s exposures and 0.5 s per frame. This method produced movies containing 30 frames with an exposure rate of 1 electron per Å2. Movies were collected with a defocus range of 1–2.5 μm and a nominal magnification of 25,000×, which produced images with a calibrated pixel size of 1.45 Å.
Image Processing.
The 30 frames in each movie were aligned using the program alignframesleastsquares_list (40) and averaged into a single micrograph with the shiftframes_list program (40). The averaged frames were used for determination of the transfer function (CTF) parameters with CTFFIND3 (41). Coordinates of each particle image were selected from these averaged frames and extracted as 220 × 220-pixel particle images. The motion of the individual particles in the frames was tracked and corrected using the alignparts_lmbfgs algorithm (40). This procedure produced a stack of particle images fully corrected from beam-induced motion from the first 20 frames of each movie. The total accumulated dose to produce each particles image was 20 electrons per Å2. From here, all processing was done with Relion 1.4.
The initial dataset from the grids containing 30S+YjeQ mixture after particle extraction contained 417,018 particles (“dirty” dataset). These images were subjected to 2D and 3D classification (Class3D first iteration), which resulted in a “clean” dataset comprising 273,407 particles. The clean dataset was run through a second cycle of 3D classification (Class3D second iteration) using the entire signal in the particle images (Fig. S2A). This classification produced two 3D classes, representing the 30S+YjeQ complex and the 30S subunit. Each class was refined separately using a binary mask with a soft edge. This approach produced the best map for the 30S+YjeQ complex (displayed in the figures) and a map for the 30S subunit that we called the 30S subunit consensus structure. These cryo-EM maps were produced from 130,462 and 142,945 particle images, respectively.
In a parallel approach (Fig. S2B), the clean dataset plus some additional particle images not initially selected in the approach described below (total number of particles, 299,825) was subjected to focused classification with subtraction of the residual signal using Relion (42) following a previously described approach (43). To this end, we applied a low-pass filter to the atomic model of the 30S+YjeQ complex (PDB ID code 2YKR) (15), after which we removed the head domain and lower part of the body. This procedure produced a density map that was used to create a soft-edge mask for the focused classification as well as for the signal subtraction in the experimental particles. The newly created stacks of particles after signal subtraction and the mask served as input for the focused classification run. During the classification step, all orientations were kept fixed at the values determined in the refinement of the consensus maps. The classification produced three distinct classes that were refined independently using a binary mask with a soft edge. One of the classes represented the 30S+YjeQ complex (167,299 particles). The remaining classes were two distinct subpopulations of the 30S subunit that we designated 30S subclass I (78,440 particles) and 30S subclass II (54,086 particles).
In the case of the grids containing only free 30S subunits, the initial dataset after particle extraction contained 90,417 particles (dirty dataset). These images were subjected to 2D and 3D classification (Class3D first iteration), which resulted in a clean dataset comprising 66,493 particles. The clean dataset was run through a second cycle of 3D classification (Class3D second iteration) using the entire signal in the particle images. This classification produced three 3D classes (Fig. S6). Each class was refined separately using a binary mask with a soft edge. This approach produced three maps for the free 30S subunit that we designated subclass Ia, subclass IIa, and subclass IIIa. These cryo-EM maps were produced from 15,088, 17,443, and 33,962 particle images, respectively
Sharpening of the cryo-EM maps was done by applying a negative B-factor estimated using automated procedures (44). Relion processes were calculated using the SciNet cluster (45). We used the program ResMap (46) to estimate the local resolution of the structures.
Map Analysis and Atomic Model Building.
To build the atomic model for the 30S+YjeQ complex, we first obtained the atomic model of E. coli YjeQ using I-TASSER (47). To this end, we used the YjeQ primary sequence (GenBank accession no. BAE78165) and the crystal structure of S. typhimurium as a template. Using Chimera (48), the resulting YjeQ atomic model was docked as a rigid body into the 30S+YjeQ cryo-EM map along with one of the available structures for the 30S subunit from E. coli (PDB ID code 2AVY) (38). Docking was optimized by a flexible fitting approach based on molecular dynamics simulation (49). The flexible fitted model was examined and used as the starting point for manual model building in Coot (50).
Acknowledgments
We thank the SciNet General Purpose Cluster of Compute Canada for providing computer resources for this project, Dr. Brandon Aubie for providing technical assistance with the MSLR computing cluster, Dr. John Rubinstein (Hospital for Sick Children) for granting access to the EM facility, and Dr. Samir Benlekbir for assisting with EM. This work was supported by National Science and Engineering Research Council of Canada Grant RGPIN288327-07 (to J.O.) and Canadian Institutes of Health Research Grants MOP-82930 (to J.O.) and MOP-67189 (to A.G.).
Footnotes
- ↵1To whom correspondence should be addressed. Email: ortegaj{at}mcmaster.ca.
Author contributions: A.R. and J.O. designed research; A.R. and A.G. performed research; A.R., A.G., and J.O. analyzed data; and A.R., A.G., and J.O. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Data deposition: Structures have been deposited in the Electron Microscopy Data Bank (ID nos. 8621 for 30S+YjeQ, 8626 for the 30S subunit consensus structure, 8627 for 30S subunit subclass I, and 8628 for 30S subunit subclass II). The coordinates for the atomic model built for the 30S+YjeQ complex have been deposited in the Protein Data Bank, www.pdb.org (PDB ID code 5UZ4).
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1618016114/-/DCSupplemental.
References
Cryo-EM structure of the 30S+YjeQ complex
Article Classifications
- Biological Sciences
- Biochemistry
Sign up for Article Alerts
Jump to section
You May Also be Interested in
Recent experiments and simulations are starting to answer some fundamental questions about how life came to be.
Image credit: Shutterstock/Radoslaw Lecyk.
Archaeologists have long tried to define the transition between the two time periods.
Image credit: Ceri Shipton.
Adam Mastroianni and Daniel Gilbert explore why conversations almost never end when people want them to.
A study finds that giant pandas roll in horse manure to increase their cold tolerance.
Image credit: Fuwen Wei.