The binding of RbgA to a critical 50S assembly intermediate facilitates YphC function in bacterial ribosomal assembly

Abstract

The intricate process of 50S ribosomal subunit assembly in Bacillus subtilis involves multiple parallel pathways converging into a crucial intermediate known as the 45S particle. RbgA and YphC, play pivotal roles in completing the maturation of the functional sites in the 45S particle. In this work, we found that RbgA and YphC can independently bind the 45S particle with high affinity, but when RbgA binds first to the particle, it significantly increases the binding affinity of YphC. Using cryo-electron microscopy, we determined that the changes exerted by RbgA and YphC when binding independently closely resemble those observed when the two factors bind to the 45S particle simultaneously. However, the structural analysis revealed that RbgA binding causes a conformational change that uncovers the binding site for YphC, thus increasing its binding affinity. We concluded that the functional interplay between RbgA and YphC primarily revolves around one factor promoting the binding of the other, rather than the binding of the two factors inducing entirely new conformational changes compared with those induced by the factors individually. These results highlight the synergic mechanism between two essential assembly factors, underscoring the intricate mechanism bacteria use to maximize the efficiency of the ribosome assembly process.

Graphical Abstract

Graphical Abstract.

Introduction

During assembly of the 50S subunit in bacteria, the 23S and 5S ribosomal RNA (rRNA) molecules fold along funnel-shaped energy landscapes comprised of multiple parallel pathways (1). The built-in redundancy provided by these parallel pathways makes the ribosome assembly process robust, ensuring a steady supply of new ribosomes for the cell. The folding of the rRNA is co-transcriptional, and the transient folds adopted by the rRNA are stabilized by the binding of ribosomal proteins (r-proteins) (2). Upon binding, the r-proteins induce conformational changes in the rRNA, creating new binding sites for other r-proteins that keep driving the folding of the rRNA scaffold into the mature conformation (3,4). Assembly factors also bind the ribosomal particles and catalyze specific folding steps, mainly involving the maturation of the functional sites. These factors help the folding of the rRNA and orchestrate the recruitment of other factors, such as methyltransferases and pseudouridine synthases (5–7), that modify specific nucleotides within the rRNA molecules. Some assembly factors act as checkpoint proteins, performing functional checks in the assembling ribosomal particle before releasing them to the pool of actively translating ribosomes (8).

In Bacillus subtilis, the assembly pathways converge into a critical intermediate called the 45S particle (9,10). This convergent intermediate is in a ‘locked’ state, and its maturation is paused. Unlocking the 45S particle is particularly demanding on assembly factors, and it requires the coordinated action of at least two factors, RbgA (also called YlqF) and YphC (also known as EngA or Der). These factors license further maturation steps that transform the critical assembly intermediate into a functional 50S subunit. RbgA and YphC are essential GTPases in B. subtilis (11,12). Both factors are widely distributed in Gram-positive and Gram-negative bacteria (11). However, they are not always essential genes.

The role of RbgA in assembly is better understood than that of YphC. RbgA binds to the P-site in the 45S particle, inducing rRNA helices around its binding site and in the A-site to adopt their mature conformation (9,13). YphC, similar to the homolog protein in Escherichia coli (EngA) or other bacterial species (Der), contains two tandemly arranged GTPase domains, GD1 and GD2, followed by a C-terminal RNA-binding K Homology domain (KH domain) (14,15). A cryo-electron microscopy (cryo-EM) study showed that EngA binds the E- and P-sites in the mature 50S subunit in E. coli (16). However, the mechanism by which YphC assists the maturation of the 45S particle is still unknown.

Here, we studied the structural changes that YphC induces to assist the maturation of the 45S particles. Treating these particles with YphC showed that YphC stabilizes the L1 stalk and the peptidyl transferase center (PTC), prepares the A-site for the last maturation steps and influences other functional regions that RbgA does not stabilize. Overall, this demonstrates the complementary character of these two GTPases catalyzing some of the last maturation steps of the 45S particle.

In mitochondrial ribosomes of Trypanosoma brucei, homologs of RbgA and YphC (respectively called Mtg1 and mt-EngA) bind simultaneously to a large subunit assembly intermediate (7,17), suggesting that a functional interplay may also exist between RbgA and YphC. In our experiments, we observed that RbgA initial binding to the 45S particle significantly increases YphC binding affinity, and the two factors can bind simultaneously. Cryo-EM structures of the 45S particles with YphC and RbgA bound simultaneously showed that the functional interplay between the two proteins mainly involves one protein promoting the binding of the other, rather than causing entirely new changes when they bind together. This study describes the mechanism that implements the functional interplay between RbgA and YphC and highlights the intricate mechanism that exists in bacteria to maximize the efficiency of the ribosome assembly process.

Materials and methods

Bacterial strains

The ribosomal 45S_RbgA and 45S_YphC particles were purified from RbgA-depleted (RB301) (11) and YphC-depleted (RB290) (12) B. subtilis strains. In these strains, the gene ylqF (encoding for RbgA) and engA (encoding for YphC) were placed under the control of an isopropyl-β-D-thiogalactopyranoside (IPTG) inducible P_spank promoter. Mature 50S subunits were purified from a strain where the only copy of infB (encoding for IF2) (RB419) was also under a P_spank promoter. Without the inducer, these cells accumulate large amounts of immature and mature 50S subunits. The construction of these strains has been previously described (11,12).

Purification of ribosomal particles

Strain RB301 or RB290 were first plated on LB Agar supplemented with 5 μg/ml chloramphenicol and 1 mM IPTG and grown overnight at 37°C. To initiate protein depletion, cells were scraped and resuspended in 1 ml of pre-warmed LB media supplemented with 5 μg/ml chloramphenicol, which was then used to inoculate a volume of 500 ml of pre-warmed LB media until it reached an initial OD₆₀₀ of 0.05 for RB301 and 0.02 for RB290. Bacteria were grown at 37°C with agitation at 215 RPM until OD₆₀₀ of 0.5, and the culture was subsequently split into four larger flasks containing 800 ml of pre-warmed LB media. The doubling time of the culture was monitored, and the depletion was considered complete when the doubling time reached ≈140 min for RB301 and ≈120 min for RB290 (13). At that point, bacteria were pelleted by centrifugation at 4500 × g for 20 min, washed with phosphate-buffered saline (PBS) and stored at −80°C. Doubling times of cultures were calculated as DT = (t₂ − t₁) × [ln 2/(ln OD₆₀₀@t₂/ln OD₆₀₀@t₁)].

To purify the 45S_RbgA and 45S_YphC particles, each cell pellet was resuspended with 10 ml of buffer A [20 mM Tris–HCl pH 7.5, 60 mM NH₄Cl, 10 mM MgCl₂, 0.5 mM ethylenediaminetetraacetic acid (EDTA), 3mM β-mercaptoethanol] supplemented with cOmplete™, Mini, EDTA-free Protease inhibitor cocktail (Roche) and 20 U of RNase-free DNase (Roche). Cell lysis was done by four consecutive passes through a cold French press at 20,000 psi. Cell debris was pelleted by centrifugation at 32 000 × g for 40 min. The supernatant was carefully deposited on a 1.1 M sucrose cushion in buffer A and centrifuged at 110 000 × g for 16 h. The ribosome pellet was resuspended in 5–10 ml of buffer C (20 mM Tris–HCl pH 7.5, 60 mM NH₄Cl, 10 mM MgCl₂, 0.5 mM EDTA, 7 mM β-mercaptoethanol) and further centrifuged at 110 000 × g for 16 h. The ribosome pellet was resuspended in 5–10 ml buffer E (10 mM Tris–HCl pH 7.5, 60 mM NH₄Cl, 15 mM MgCl₂, 3 mM β-mercaptoethanol) and carefully loaded on top of a 35 ml 18–43% sucrose gradient in buffer E. The gradients were centrifuged at 60 000 × g for 16 h. Gradients were fractionated by monitoring the A280, and peaks corresponding to 45S assembly intermediates were collected, pooled and pelleted by centrifugation at 110 000 × g for 16 h. The pellet was resuspended in 100 μl buffer E and validated for protein composition by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) on precast 12% polyacrylamide gels. The purified 45S assembly intermediates were quantified, aliquoted, flash-frozen in liquid nitrogen, and subsequently stored at −80°C for further use.

Protein overexpression and purification

Escherichia coli BL21 (DE3) cells were transformed with pET21b-ylqF (encoding RbgA with a C-terminal His6 tag) or with pET15b-yphC (encoding YphC with an N-terminal His6 tag). Bacteria were grown in LB media at 37°C with 215 RPM agitation and supplemented with 100 μg/ml ampicillin. Protein overexpression was induced at an OD₆₀₀ of 0.6 with 1 mM IPTG. After 3 h of induction, cells were pelleted by centrifugation at 4500 × g for 15 min, washed with PBS 1× buffer and stored at −80°C for further use. For purification, cells were resuspended in binding buffer (20 mM sodium phosphate pH 7.5, 500 mM NaCl, 20 mM imidazole) supplemented with cOmplete™, Mini, EDTA-free Protease inhibitor cocktail (Roche) and lysis was done by four consecutive passes through a cold French press at 20 000 psi. Cell debris was pelleted by centrifugation at 32 000 × g for 45 min and the supernatant was further cleaned by filtering through a 0.45- and 0.22-μm syringe filters. The resulting clarified and filtered lysate was manually loaded onto a 1 ml HisTrap™ column (Cytiva) pre-equilibrated with a binding buffer. The column was washed with binding buffer, and proteins were eluted using a 20–500 mM imidazole gradient.

Fractions containing the protein were pooled and dialyzed overnight against ion-binding buffer (20 mM sodium phosphate pH 7.5) supplemented with 100 mM NaCl and 25 U/ml of thrombin protease to cleave the His6 tag. Dialyzed RbgA was not supplemented with thrombin because the His6 tag is uncleavable. Dialyzed proteins were centrifuged at 12 000 × g for 10 min to remove any precipitated proteins and subsequently loaded onto a HiTrap™ Q HP (Cytiva) anion exchange column (YphC) or HiTrap™ SP HP (Cytiva) cation exchange column (RbgA) and eluted using a 100 mM to 1 M NaCl gradient.

Proteins purity was assessed by SDS–PAGE and were then labeled for microscale thermophoresis experiments (see ‘Microscale thermophoresis’ section), or buffer exchanged to buffer E supplemented with 250 mM NaCl and concentrated using 10-kDa cutoff centrifugal filter. Pure proteins were then quantified, aliquoted, flash frozen in liquid nitrogen and stored at −80°C for further use.

Microscale thermophoresis experiments

Prior to the labeling reaction, pure RbgA and YphC proteins were concentrated and subsequently spun at 14 000 × g for 10 min to remove any aggregates. The labeling reactions were performed according to the manufacturer’s instructions. The reaction mix contained 20 μM of protein, 30 μM of 2nd generation Red-NHS dye (NanoTemper Technologies) in labeling buffer (130 mM NaHCO₃ pH 8.2–8.3, 50 mM NaCl) supplemented with 1 mM 5′-Guanylyl-imidodiphosphate trisodium salt hydrate (GMPPNP). The Red-NHS dye labels proteins on their lysine residues or N-terminus of the protein. Proteins were incubated 30 min in the dark and subsequently loaded onto the provided desalting column equilibrated with microscale thermophoresis (MST) buffer A (10 mM Tris–HCl pH 7.5, 60 mM NH₄Cl, 15 mM MgCl₂, 250 mM NaCl, 1 mM 1,4-dithiothreitol (DTT), 0.05% Tween-20) to eliminate any free dye molecules. Labeled proteins were then quantified, and the degree of labeling was calculated. Proteins were subsequently aliquoted, flash-frozen in liquid nitrogen and stored at −80°C for further use.

To perform the MST measurements, labeled proteins were diluted to 120 nM in MST buffer B (10 mM Tris–HCl pH 7.5, 60 mM NH₄Cl, 15 mM MgCl₂, 1 mM DTT, 0.05% Tween-20) supplemented with 1 mM GMPPNP and ribosomal assembly intermediates were diluted to 4 μM in MST buffer. A serial dilution was performed with the 45 assembly intermediates over 16 tubes with a final volume of 10 μl each. A volume of 10 μl of labeled protein was added to each tube, yielding a final protein concentration of 60 nM and a concentration range of 2 μM to 61 pM for the assembly intermediates. The reaction was incubated at room temperature for 1 h before loading in premium coated glass capillaries (NanoTemper Technologies). The same procedure was performed to measure the affinity of YphC to the 50S subunits, but the range of concentration was 8 μM to 244 pM. Similarly, to measure the affinity of YphC with RbgA, the range of concentration was 100 μM to 3 nM of RbgA. For pre-binding experiments, a mix containing 4 μM of unlabeled protein and 4 μM of assembly intermediates was prepared in MST buffer supplemented with 1 mM GMPPNP and incubated for 1 h before serial dilution was performed. The labeled protein was then added to each tube for a final concentration of 60 nM and subsequently incubated for 1 h before loading into premium coated glass capillaries. Measurements were performed using a Monolith NT.115 instrument (NanoTemper Technologies) set at 25°C with LED power of 20% and medium MST power. The dissociation constants were obtained by plotting the difference in normalized fluorescence [ΔF_norm (‰) = F₁/F₀] versus the logarithm of the different concentrations of 45S assembly intermediates. F₁ and F₀ are the hot and cold regions of the thermophoresis traces. The obtained K_d values were calculated from three independently performed experiments using the NanoTemper MO. Affinity analysis software (version 2.3).

Cryo-electron microscopy

To prepare the grids containing only 45S_YphC particles, we prepared a dilution of these particles at a concentration of 180 nM in buffer A (10 mM Tris pH 7.5, 15 mM magnesium acetate, 60 mM NH₄Cl, 3 mM β-mercaptoethanol) that was directly applied to the grid.

In the case of the 45S_YphC particles treated with YphC, we prepared a dilution of these particles at a concentration of 200 nM in buffer A supplemented with 1 mM GMPPNP. Then we added purified YphC to a final concentration 40-fold higher than the ribosomal particle concentration and the reaction mixture was incubated for 5 min at room temperature before the mixture was applied to the grid.

To assemble the double complex containing RbgA and YphC bound to the 45S_RbgA particle for cryo-EM analysis, we adapted the conditions determined by MST where binding was observed. The reaction and all dilutions were done in MST buffer B supplemented with 1mM GMPPNP and without 0.05% Tween-20. We initially prepared a 10 μl reaction mixture containing 1.8 μM 45S_RbgA particles and 18 μM of RbgA, and the mixture was subjected to 1-h incubation at room temperature. In a second step, this reaction was mixed with an equal volume of a solution containing YphC at a concentration of 18 μM, and the mixture was incubated for 1 h at room temperature. Before the reaction was applied to the grid, the assembly reaction was diluted 5-fold with a solution containing 1.8 μM of RbgA and YphC.

Cryo-EM grids (c-flat CF-2/1–3Cu-T) used for all samples were prepared by evaporating a continuous layer of carbon (5–10 nm) to reduce exposure of the ribosomal particles to the air-water interface. Right before applying the samples, grids were washed in chloroform for 2 h and treated with glow discharged in air at 5 mA for 15 s. Next, grids were vitrified in liquid ethane using a Vitrobot Mark IV (Thermo Fisher Scientific) using one blotting time for 3 s and a blot force +1. The Vitrobot chamber was set to 25°C and 100% relative humidity. All datasets were collected at the Facility for Electron Microscopy Research, McGill University, using a Titan Krios microscope at 300 kV equipped with a Gatan BioQuantum LS K3 direct electron detector. The software used for data collection was SerialEM (18). Images were collected in counting mode according to the parameters described in Supplementary Table S1.

Image processing

All image processing steps were done using CryoSPARC v4. Cryo-EM movies were corrected for beam-induced motion correction using Patch Motion Correction using default settings that included using information up to 5 Å resolution when aligning frames, a B-factor of 500 and a 0.5 calibrated smoothing constant applied to the trajectories. All frames in the movies were used to produce the merged micrograph. Constant Tranfer Function (CTF) parameter estimation was done using Patch CTF estimation using default settings in the program. The minimum and maximum resolution considered to estimate the CTF parameters were 25 and 4 Å, respectively, and the minimum and maximum defocus values were set up at 1000 and 40 000 Å, respectively. Images with an estimated resolution of 6 Å or better were kept for further processing.

In the RbgA + YphC-treated 45S_RbgA particles dataset, particles on the selected micrographs were initially picked using Blob Picker using a circular blob and a minimum and maximum particle diameter of 150 and 300 Å, respectively. The maximum resolution considered in the micrographs was 20 Å. The angular sampling used was 5 degrees, and the minimum particle separation distance was 1 (in units of particle diameters). Picked particles were extracted using a box size of 448 pixels, Fourier cropped to 360 pixels and subsequently curated using an ab initio 3D reconstruction and 2D classification routines. For the ab initio step, we selected 0.85 and 0.99 as inner and outer window radii, respectively, and requested three classes and a maximum and minimum resolution to consider of 35 and 12 Å, respectively. The number of iterations before and after annealing starts and ends was set to 200 and 300, respectively, and the increase of Fourier radius at each iteration was 0.04. All other parameters for this routine were used with the default settings and values. In the 2D classification step, we requested 50 classes, and we selected 0.85 and 0.99 as the inner and outer window radii, respectively. The maximum resolution considered in the images was 6 Å, and we used 2 for the initial uncertainty factor. All other settings were used with the default parameters. The particles from the best-aligned classes were selected and used to train a model using Topaz (19). The number of particles expected per micrograph to train the model was set to 50 and all the preprocessing, training and model parameters were set in the default values. This trained model was used to do a second round of particle picking in all the micrographs. This routine was run using default parameters, and the radius of extracted regions was set at 40. The picked particles were extracted with the original size and an extraction box of 432 pixels and were subsequently Fourier-cropped to 360 pixels. The extracted particles were curated using an ab initio reconstruction routine run requesting three classes and with the same parameters as described above. The curated set of particles was used to generate a consensus refinement cryo-EM map using nonuniform refinement under default settings with C1 symmetry and including optimized per-particle defocus and optimized per-exposure group CTF parameters. This 3D reconstruction was used to estimate the maximum resolution attainable with the data and to obtain a consensus cryo-EM map that was subsequently used as the initial model in the final refinements (see below).

In the case of the untreated 45S_YphC and YphC-treated 45S_YphC datasets, the Topaz step was not implemented, and particle picking was done in two steps: 500 randomly selected micrographs were first picked using Blob picker, and the selected particles were used to generate templates for subsequent template picking in all micrographs. Particle curation of the selected particles was done using 2D classification and ab initio reconstruction routines. Blob picker, class 2D and ab initio reconstruction routines were run using the same parameters as in the RbgA + YphC-treated 45S_RbgA datasets.

Particle heterogeneity was explored in all datasets through 3D variability analysis. To this end, the curated set of particles was re-extracted, binned to a pixel size of 3.42 Å/pixel and 112-pixel box size. The 3D variability analysis routines were always run requesting three orthogonal principal modes and in cluster mode with all settings at default values. The number of clusters requested varied between experiments and ranged from 4 to 10. Results were filtered at 10 Å. Resulting maps from the exhaustive 3D classification were visually inspected in Chimera (20,21), and clusters of particles representing similar assembly intermediates were merged. High-resolution refinements were performed in two stages: in the first step, particles from each class were re-extracted with the original pixel size (0.855 Å/pixel, 448-pixel box size) and subjected to a homogeneous refinement using as a 3D reference the initial consensus refinement cryo-EM map generated from the initial curated set of particles (after proper scaling and filtering using a 30 Å low-pass Fourier filter). These homogeneous refinements were run under default settings with C1 symmetry and included optimized per-particle defocus and optimized per-exposure group CTF parameters. The resulting maps were used as the initial model for a second refinement step that included a nonuniform refinement run under default settings with C1 symmetry, optimized per-particle defocus, optimized per-exposure group CTF parameters and options ‘Fit Spherical Aberration’, ‘Fit tetrafoil’ and ‘Fit anisotropic Magnification’ activated.

Average resolution estimation and local resolution analysis were done with cryoSPARC using the gold-standard approach (22). Cryo-EM map visualization was performed in UCSF Chimera and Chimera X (20,21).

Molecular model building

All atomic models were manually built using WinCoot v0.9.8.1 (23) using the following procedure. The mature 50S subunit (PDB 3J9W) was used as an initial model and fitted by rigid-body docking in UCSF Chimera (20). This model was extensively refined using the real-space refine tool and protein/RNA segments where no corresponding cryo-EM was found. These initial models were improved by rounds of real-space refinement in Phenix (24) and manual model building in Coot. The final models were validated using Phenix cryo-EM comprehensive validation tool and the Molprobity server (25,26) (Supplementary Table S1). Figures were prepared using UCSF Chimera (20), Chimera X (21,27) and Photoshop (Adobe). To produce the molecular models of the 45S_RbgA particle bound to RbgA and YphC, RbgA was added from the previously published 45SRbgA + RbgA model (PDB 6PPK) and the molecular model for YphC was taken from AlphaFold (AF-P77360-F1). Both proteins were initially fitted by rigid-body docking in UCSF Chimera and real-space refined in WinCoot (23). Figures showing cryo-EM density and molecular models were prepared using UCSF ChimeraX.

Multiple sequence alignment

Protein sequences for YphC (also called Der or EngA) were retrieved from NCBI database. A total of 1000 sequences were extracted from diverse bacteria, excluding phylum Bacillota (Firmicutes). Sequences were analyzed in UGENE (28) using MAFFT (29) multiple sequence alignment with default parameters. The initial 1000 sequences were curated to exclude sequences with extension in the N-terminus of the proteins and gaps in the alignment. The aligned and curated 590 sequences (Supplementary Table S4) were used to generate logos using the WebLogo 3 application (30).

Results

YphC and RbgA have a permissive binding behavior against 45S ribosomal particles

Depletion of RbgA or YphC in B. subtilis cells causes a slow growth phenotype (Supplementary Figure S1A and B) and accumulation of 45S particles (Supplementary Figure S1C) (12). These particles are named 45S_RbgA or 45S_YphC particles depending on whether the depleted gene is the one expressing RbgA or YphC, respectively (11,31). However, the two 45S particles contain an identical protein complement (Supplementary Figure S1D) and are structurally similar (13,31), suggesting that these 45S particles constitute a critical node where multiple parallel assembly pathways converge (10).

YphC and RbgA bind the 45S_RbgA and 45S_YphC particles (31). Here, we measured the affinity of each one of the factors for the two 45S particles using MST (32). The thermophoretic mobility of fluorescently labeled YphC or RbgA was measured at increasing concentrations of either 45S_RbgA or 45S_YphC particles. We observed that the affinity of the two factors was slightly higher for the 45S_RbgA (Figure 1A) than for the 45S_YphC particles (Figure 1B) and that the overall affinity of RbgA for the two 45S particles was higher than that exhibited by YphC. However, the measurements in all binding reactions indicated that in the presence of GMPPNP, YphC and RbgA have a permissive behavior and bind both 45S particles with much higher affinity than that observed against the mature 50S subunit (Supplementary Figure S2) (13).

Figure 1.

Binding affinity of RbgA and YphC to the 45S particle measured by MST. The panels show the MST experiments measuring the binding of RbgA (top panels) and YphC (bottom panels) to the 45S_RbgA (A) and 45S_YphC (B) particles. PB means ‘prebound’. All MST reactions contained 60 nM of fluorescently labeled assembly factor (RbgA or YphC) and increasing concentrations of the immature 45S particle. When RbgA or YphC was prebound to the 45S particle before the affinity of the other factor was measured, the concentration of the prebound factor was equimolar with the 45S particle for each concentration tested. Plots on the left show thermophoretic mobility traces of the MST reactions depicting individual traces for each ribosomal particle concentration and highlight the F₀ (blue) and F₁ (red) regions used to calculate binding. RbgA and YphC binding plots (right panel) depict ΔF_norm (F₁/F₀) versus particle concentration. The F_norm curves were fit using the law of mass action to derive K_d values, which are reported with a 68% confidence interval derived from the variance of the fitted parameter. Dots represent the average from the three replicates at each concentration; error bars denote standard deviation.

YphC binds the functional sites in the 45S_YphC particles

To determine the specific maturation steps that YphC catalyzes upon binding to the 45S particle, we used cryo-EM to study the structural variability of the 45S_YphC particles before and after treatment with YphC. Before YphC treatment, we found that the 45S_YphC particles coexist as a heterogeneous mixture of two assembly intermediates that mainly differ in the presence of the central protuberance (CP; Figure 2A and B). The cryo-EM map for Class 45S_YphC_1 refined to 2.9 Å resolution (Supplementary Figure S3A and C). It lacked densities representing H80-88 and the 5S rRNA forming the CP and the long H75 at the base of the L1 stalk (Figure 2B). These particles represented 65% of the population. The remaining 35% were Class 45S_YphC_2 particles. The cryo-EM map for Class 45S_YphC_2 was refined to 2.8 Å resolution (Supplementary Figure S3B and C) and exhibited CP and density for H75 (Figures 2A and B). Both Class 45S_YphC_1 and Class 45S_YphC_2 showed a mature body, but densities representing functionally essential rRNA helices in the A (H38, H89 and H91-92), P (H69, H71 and H93) and E-site (H68) were still absent in the cryo-EM maps. The L1 stalk comprised of H76-H78 was also not visible in any of the two classes.

Figure 2.

Heterogeneity analysis and multiple classes present within the 45S_YphC particle population before and after YphC treatment. (A) Density map of the mature 50S subunit for reference purposes showing the ribosomal particles' main landmarks and functional sites (A-, P- and E-sites). This map was created from PDB 3J9W using the molmap command in Chimera and was low pass filtered to 3 Å resolution. (B) Cryo-EM maps (top panel) of the two classes observed for the purified 45S_YphC particles before YphC treatment. The 23S rRNA and 5S rRNA are colored in light gray, and the r-proteins are shown in red. The diagrams in the bottom panel show the rRNA helices lacking or having a highly fragmented density in the cryo-EM maps, indicating these helices have yet to adopt the mature conformation. Each helix missing the rRNA is colored differently. (C) Cryo-EM maps of the five classes observed in the purified 45S_YphC particles after they have been treated with YphC. The 23S rRNA and 5S rRNA are colored in light cyan and the r-proteins are shown in red. In Class 45S_YphC+ YphC_5, the YphC protein bound to the particle is colored in blue, and the distorted helix H89 is colored in green. The diagrams in the bottom panel show the rRNA helices still in an immature conformation and lacking a density in the cryo-EM maps. These rRNA helices were colored using the same color code as in panel (B).

Consistent with previous quantitative mass spectrometry (qMS) data (31), we did not find density in any of the classes for uL16, bL28, bL33, bL35 and bL36, while bL27 was missing only in Class 45S_YphC_1. We concluded that these proteins have not been incorporated into these assembly intermediates (Supplementary Figure S4A). Other proteins, including uL5, uL18 (only missing in Class 1) and bL31 in the CP and uL10 and uL11 in the bL12 stalk, did not appear in the cryo-EM maps, suggesting that these proteins are bound in a flexible manner. Finally, uL6 at the base of H42 in the bL12 stalk was present. Still, the density representing this r-protein was highly fragmented, suggesting it is not firmly attached to its binding site as in the mature 50S subunit.

Adding purified YphC (Supplementary Figure S1D) to the 45S_YphC particles and incubating the reaction increased the heterogeneity of the sample. This allowed us to observe up to five different subpopulations of assembly intermediates. Incubation with YphC increased the percentage of particles with CP from 35 to 44% (Figure 2C). The remaining 56% of the particles still did not show CP. Most of the particles lacking the CP (Class 45S_YphC+ YphC_1) (37%) were structurally identical to the Class 45S_YphC_1 particles in the untreated dataset, but a new group appeared (Class 45S_YphC+ YphC_2) (19%) that also lacked the CP but had clear density for H75 at the base of the L1 stalk. Within the group of particles exhibiting CP, we identified three types, Class 45S_YphC+ YphC_3 (20%) and Class 45S_YphC+ YphC_4 (21%), that were structurally identical to Class 45S_YphC_2 in the untreated sample. Class 45S_YphC+ YphC_3 and 45S_YphC+ YphC_4 only differed on the density representing H75, which was complete in Class 45S_YphC+ YphC_4, but fragmented in Class 45S_YphC+ YphC_3. The last class observed in this sample (Class 45S_YphC+ YphC_5) represented 3% of the population, exhibited CP and showed an additional density snuggled in the E-site that was assigned to YphC since part of the density closely resembled YphC X-ray structure (14). The cryo-EM maps for Class 45S_YphC+ YphC_1 to Class 45S_YphC+ YphC_5 were refined to resolutions between 3 and 3.3 Å (Supplementary Figures S5 and S6).

These results suggest that when treated with YphC, the 45S_YphC particles continue to mature. Particles without the CP appear to stably fold this structural domain – a process that likely occurs independently of YphC binding. The emergence of Classes 45S_YphC+ YphC_2 and 45S_YphC+ YphC_3 in the treated sample also suggests that H75 is not fully stable and may undergo cycles of folding and unfolding until surrounding rRNA helices on the E-site are properly formed. The observed classes show that YphC binds in the E- and P-sites of 45S_YphC particles with developed CP. However, it is possible that some of the particles in the dataset may have initially bound YphC before the CP was formed and subsequently developed this structural domain.

YphC binding to the 45S particle induces large conformational changes in 23S rRNA helices of critical functional importance

We constructed a molecular model (Figure 3B) for Class 45S_YphC+ YphC_5 showing YphC bound to the 45S_YphC particle (Figure 3A) to examine the interaction between YphC and the assembly intermediate and the conformational changes induced upon binding. YphC binds inside the transfer RNA passage on the 50S subunit in an extended conformation, with the two N-terminal GTPase domains (GD1 and GD2) occupying the E-site and its C-terminal KH domain located in the P-site (Figure 3C). The three domains of YphC make extensive contact with the 45S_YphC particle, particularly with the rRNA helices near the binding site. Only the GD1 domain establishes protein–protein interactions with r-proteins bL33 and uL1 (Figure 3C). The back of the GD1 domain interacts with H88, located right below bL33. At the front, the GD1 domain is stabilized through interactions with the top of the uL1 stalk. The GD2 domain sits on the location usually occupied by the proximal part of the long H68, which in this cryo-EM map is displaced and sits on top of the GD2 domain (Figure 3D; top panel). The back of the GD2 domain contacts with parts of H88, and the bottom of the domain interacts with the base of the uL1 stalk, which is formed by H75 (Figure 3D; top panel). The KH domain is tightly snuggled inside the P-site, preventing H69 and H71 from adopting their mature conformation. The back of the KH domain interacts with the single-stranded rRNA strand connecting H88 and H89 (Figure 3D; bottom panel), and the bottom of the domain contacts H93 (Figure 3D; bottom panel).

Figure 3.

Cryo-EM structure of YphC bound to the 45S_YphC particle. (A) Cryo-EM map of YphC bound to Class 45S_YphC+ YphC_5. YphC is shown in blue, and H89 in green. The rRNA is colored in light cyan, and the r-proteins are in red. The main landmarks of the ribosomal particle are labeled. The functional core of the particle is framed, and the locations of the A-, P- and E-sites are indicated. Details from this area are shown in panels (C) and (D) below. (B) Molecular model of the Class 45S_YphC+ YphC_5. Components of the complex are colored as in the cryo-EM map. The interaction area between YphC and H89 is framed, and details about this interaction are shown below in panels (E) and (F). (C) Details of YphC in the molecular model adopting an extended conformation. YphC is colored in blue, and the GD1, GD2 and KH domains are labeled. Proteins interacting with YphC are labeled and colored in red. H89 also interacts with YphC and is colored in green. (D) Front (top panel) and top (bottom panel) view of the interactions of YphC with rRNA helices in the A- and P-sites. The rRNA helices establishing interactions with YphC are labeled and colored using different colors. (E) Electrostatic potential surface representation of the YphC KH domain calculated according to Coulomb potential. Red represents areas of negative potential, white near neutral and blue positive potential. On the right panel, H89 is shown as a ribbon representation to better display the electrostatic potential surface of the YphC KH domain. U2521 and U2522 adopting the flip-out conformation are also shown. (F) These two panels highlight the main amino acids from the YphC KH domain interacting with U2521 and U2522, adopting the flip-out conformation.

The more intriguing feature of this model is the interaction of the KH domain with helix H89 (Figure 3; Supplementary Figure S7). H89 adopts its functional conformation in one of the latest steps of the 50S assembly process in a maturation event assisted by assembly factor ObgE and r-proteins bL36 and uL16 (5). At the time of YphC binding, ObgE (Obg in B. subtilis), bL36 and uL16 are absent, and H89 is free to move. The interaction of the YphC KH domain with H89 brings the entire middle section of this helix closer to the P-site (Figure 3B). Analysis of the surfaces in H89 and the KH domain mediating this interaction revealed that the first two strands in the β-sheet in the YphC KH domain create a hydrophobic platform that opens the H89 by pushing the negatively charged phosphate backbone outwards and allowing the interaction between the KH domain and the base moieties in this segment of H89 (Figure 3E). The side chains of Tyr 383, Tyr 384, Thr 386 and Val 388 in the first strand of the β-sheet and Val 396 and Phe 398 in the second strand define the hydrophobic surface buttressing H89 (Figure 3F). This interaction extrudes U2521 and U2522 outside the helical axis, making both nucleotides adopt a flip-out conformation while the side chain of Tyr 384 is inserted inside the helical axis (Figure 3E and F). This conformational change is stabilized by a π-stacking interaction between the guanidium moiety of Arg 353 and U2521 (Figure 3F; top panel), as well as hydrophobic interactions between Val 388, Val 396 and Phe 431, and U2522 (Figure 3F; bottom panel). Arg 433 and Lys 429 also provide favorable electrostatics to maintain U2522 in the flip-out conformation. To investigate if this mechanism could exist throughout different species of bacteria, we computed sequence alignments of YphC of >500 bacterial species (Supplementary Table S4). We found that many residues of the KH domain that contact H89 are highly conserved (Supplementary Figure S8). Notably, at position 353, the most common residues are Arg (52%) and Lys (35%), both positively charged sidechains that can stabilize flipped-out nucleotide moieties. A striking feature is at position 384. In our model, the sidechain of Tyr 384 is inserted in the helical axis, presumably helping to force out the two nucleotides. This residue is found in 82% of the aligned sequences, and in the remaining 18%, Tyr is replaced by Phe.

In summary, our analysis showed that YphC binds to the 45S particle, adopting an extended conformation and causing drastic conformational changes in functionally critical rRNA helices.

YphC binding to the 45S particle causes a drastic movement of its GD1 domain and induces intradomain conformational changes in both GTPase domains

X-ray crystallography determined the structures of YphC from B. subtilis (14) and Thermotoga maritima (called Der) (15). Unlike the 45S-bound YphC (Figure 4A), these structures representing the free protein showed that the three domains adopt a nonextended conformation when they are not bound to the ribosomal particles. The orientation of the GD2 domain with respect to the KH domain is the same in the bound and free structures. However, the orientation of the GD1 domain changes between structures (Figure 4B). In B. subtilis YphC, the GD1 folds back into the KH domain, and the three domains arrange themselves into a ‘C’ shape (Figure 4B; top panel). This movement is even more prominent in T. maritima Der, which places the GD1 domain over the KH domain (Figure 4B; middle panel).

Figure 4.

Comparison of the unbound and ribosome-bound YphC structures. (A) Ribbon representation of YphC. The arrow diagram at the top describes the extended conformation adopted by the three domains of the protein. The GMPPNP molecules bound to the GD1 and GD2 domains are shown in orange. (B) Ribbon representation of the B. subtilis YphC (top panel) and T. maritima Der (middle panel) crystal structures and E.coli EngA cryo-EM structure. The arrow diagram in each panel shows the arrangement of the domains in each structure. (C) Overlap of the GD1 (left panel) and GD2 domains (right panel) of the YphC cryo-EM structure (colored in blue) and the X-ray YphC structure (colored in orange), providing a front view of switch 1 in both domains, which are visible in the cryo-EM structure. However, in the X-ray structure, the switches move upward, and the density for most of the amino acids forming the switch disappears. (D) Panels showing the orientation of the a-helix forming switch II in the GD1 domain in the YphC cryo-EM structure, the B. subtilis YphC (top panel), T. maritima Der (middle panel) crystal structures and E. coli EngA cryo-EM structure. (E) This panel shows the orientation of the a-helix forming switch II in the GD2 domain in the same structures.

The drastically different orientation of the GD1 domain in the free versus bound proteins suggests that the interaction with the 45S particle causes YphC to unfold its domains into the extended conformation observed in our cryo-EM map (Figure 4A). This is consistent with a previous cryo-EM structure of E. coli EngA (YphC homolog) bound to the mature 50S subunit (Figure 4B; bottom panel), where the protein also adopts the extended conformation (16).

Next, we analyzed the conformational differences within each domain between the 45S-bound and free YphC protein. The GD1 domain in the free B. subtilis YphC represents the GDP bound or ‘off state’ of the enzyme (14), whereas our cryo-EM structure shows clear density of the GMPPNP molecule, and thus, this domain is adopting a conformation equivalent to the GTP bound or ‘on state’ (Figure 4A). Aligning the GD1 domain in these two structures shows a close overlap of most structural elements, except for switches I and II. These two switches stabilize the γ-phosphate of the bound nucleotide, and during GTP hydrolysis, they undergo large conformational changes.

The switch I in the GD1 is visible in the cryo-EM structure (‘on state’) and positioned near the γ-phosphate of the nucleotide bound to this domain. However, this loop is not visible in the crystal structure (‘off state’) (14), suggesting that switch I becomes flexible and moves away upon GTP hydrolysis to facilitate the release of GDP (Figure 4C; left panel). Switch II, in the GD1 domain, adopts different conformations in all four structures (Figure 4D). This motif comprises a loop region, only partially visible, and a long α-helix. We observed that the C-terminal end of this α-helix pivots and moves its N-terminal end toward the uL1 stalk. In the 45S-bound YphC cryo-EM structure, the α-helix resembles the conformation of EngA when bound to the mature 50S subunit (16), and it is positioned further away from uL1 compared with the position adopted in the free YphC and Der proteins.

Similarly, the GD2 domain in the free YphC is bound to GDP and represents the ‘off state’, whereas in the 45S-bound YphC cryo-EM structure, this domain represents the ‘on state’. Switch I in the GD2 domain of the 45S-bound YphC structure also shows a defined conformation, but it becomes flexible in the free protein in the ‘off state’ (Figure 4C; right panel). In switch II, the position of the long α-helix is also variable between structures, but the most similar positions are those in the 45S-bound YphC and 50S-bound EngA cryo-EM structures (Figure 4E).

Based on these results, it appears that the conformation of switch I of both GTPase domains is largely influenced by the nucleotide bound to the active sites. Meanwhile, the position of switch II is primarily determined by whether or not the protein is bound to the ribosomal particle.

RbgA increases the binding affinity of YphC to the 45S assembly intermediates

To investigate whether a functional interplay between RbgA and YphC exists, we measured the effect of pre-binding one of the factors to the 45S assembly intermediate in the binding affinity of the second factor. Using MST, we found that the affinity of RbgA against the 45S_RbgA particle did not change after pre-binding YphC to the ribosomal particle (Figure 1A; top panel), and it only slightly increased when binding affinity was measured against the 45S_YphC particle (Figure 1B; top panel). Instead, the affinity of YphC increased about 3-fold upon pre-binding of RbgA to either the 45S_RbgA (Figure 1A; bottom panel) or 45S_YphC particles (Figure 1B; bottom panel). We also measured that the binding affinity of YphC to the 45S_RbgA particle is around 100 times greater than its affinity to RbgA alone, with a K_d value of 37 μM (Supplementary Figure S9). This finding suggests that YphC and RbgA interacting with each other in solution and then binding to the 45S particle as a heterodimer is unlikely. Instead, we conclude that binding first of RbgA to the 45S particle favors the subsequent binding of YphC and makes this sequence of events the preferred maturation pathway.

RbgA facilitates access of YphC to its binding site in the 45S assembly intermediate

To mechanistically understand how RbgA increases the binding affinity of YphC toward the 45S particle, we took purified 45S_RbgA particles initially pre-incubated with a molar excess of purified RbgA and added purified YphC (Supplementary Figure S1D). The assembly reaction was then analyzed using cryo-EM. Image classification approaches were necessary to sort out the various classes present in the sample. We found that the 45S_RbgA particles existed in the two classes, Class 45S_RbgA+ RbgA + YphC_1 and Class 45S_RbgA+ RbgA + YphC_2 (Figure 5A). Class 45S_RbgA+ RbgA + YphC_1 accounted for 46% of the particles, while Class 45S_RbgA+ RbgA + YphC_2 represented 54%. The two classes mainly differed in the presence of the CP. The cryo-EM maps refined to 3.1 Å resolution (Supplementary Figure S10), and showed molecular resolution details according to this resolution (Supplementary Figure S11). The two maps showed RbgA and YphC bound in the functional sites of the ribosomal particle.

Figure 5.

Cryo-EM structures of YphC and RbgA bound to two different classes of 45S_RbgA particle. (A) Cryo-EM maps of YphC (colored in blue) and RbgA (colored in yellow) bound to two different classes of the 45S_RbgA particle. The rRNA is shown in light blue, and the r-proteins are colored red. The H89 displaced out of its mature position is shown in green. RbgA promotes H91 and H92 to adopt the mature conformation and their location is indicated. The uL1 and uL12 stalks, which are major landmarks of the 45S_RbgA particle, are also indicated. (B) Molecular model of the YphC and RbgA bound to Class 45S_RbgA+ RbgA + YphC_2. Components of the model are colored as in panel (A). (C) Zoomed view of the framed area in panel (B) highlighting YphC and RbgA bound to the P- and E-sites. This panel shows the interactions of the two factors with each other and with r-proteins around their binding sites, as well as the interaction of H89 with the YphC KH domain. (D) Ribbon representation of YphC and RbgA colored according to the RMSD values obtained upon overlapping each protein in the double and single complex with the 45S particle. Main domains of YphC and RbgA are indicated.

After building molecular models for Classes 45S_RbgA+ RbgA + YphC_1 (Supplementary Figure S12) and 45S_RbgA+ RbgA + YphC_2 (Figure 4B), we found that in the context of these double complexes, the C-terminal KH domain of YphC contacts the N-terminal GTPase domain of RbgA (Figure 5B and C; Supplementary Figure S13), contributing to the simultaneous binding to the ribosomal particle. The two factors bound to the 45S_RbgA particles in identical conformation and through the same interactions with the ribosomal particle as in the single complexes. The YphC and RbgA proteins bound in the single (Figure 3) (13) and double complexes (Figure 5B; Supplementary Figure S12) overlapped with root mean square deviation (RMSD) average values of 0.74 and 1.55 Å, respectively (Figure 5D; Supplementary Figure S12).

The binding of RbgA to the P-site in the double complex triggers similar conformational changes in the 45S_RbgA particle than those observed in the single complex, including delocalization of H68 in the E-site and H69 and H71 in the P-site, and the stabilization of H91, H92 in the A-site and H93 in the P-site (Figure 5A). YphC also induced similar conformational changes in the single and double complexes, including bringing the middle section of H89 closer, preventing this helix from adopting its mature conformation (Figure 5A). However, we also observed two differences between the single and double complexes. First, the H38, whose stabilization is promoted by RbgA in the single complex (13), was not apparent in the double complexes. This was likely a consequence of the YphC displacement of H89 (Figure 5A and B). Second, H68 did not wrap around the GD2 domain of YphC as observed in the YphC single complex (Figure 3), likely because this large helix is displaced at its origin in the P-site by the prior binding of RbgA (Figure 5A and B).

Overall, the cryo-EM structures of the double complexes explain how the pre-binding of RbgA facilitates the subsequent binding of YphC (Figure 1) and how the simultaneous binding of RbgA and YphC promotes subsequent maturation of the A-site. The initial binding of RbgA forces H68 to remain flexible and outside the E-site, which is required by YphC to bind the ribosomal particle in an extended conformation (Figure 5A, B and C). RbgA binding also stabilizes H93 (Figure 3D), providing YphC with a first point of contact to initiate its binding through its KH domain. Subsequent binding of the YphC KH domain to the P-site maintains H38 and H89 flexible, facilitating the subsequent final maturation steps in the A-site.

An opposed model suggesting that YphC and RbgA interact with each other in solution and bind to the 45S_RbgA particle as a heterodimer is unlikely, given that the affinity of YphC for free RbgA is much lower (K_d value of 37 μM) (Supplementary Figure S9) than the affinity of each factor to the ribosomal particles (Figure 1).

Discussion

The 45S particle is a hub where multiple assembly pathways converge before the assembly intermediate undertakes the last maturation steps involving folding the functional sites (9). At least two assembly factors, YphC and RbgA, are essential for these maturation steps. However, whether these two factors act separately or in conjunction is still unknown. In this study, we first analyzed the role of YphC individually in assisting the maturation of the 50S subunit functional sites. Upon binding to the E- and P-sites, YphC displaces H89 and H68 outside their mature position and maintains the functional sites in an immature state (Figure 6; top diagram). We also observed that prior binding of RbgA to the P-site increases the binding affinity of YphC (Figure 1), thus facilitating its role in the ribosome assembly process and suggesting a functional interplay between both factors.

Figure 6.

Model describing the implementation of the functional interplay between RbgA and YphC. YphC is able to bind the 45S particle individually. However, this binding has a lower affinity than if RbgA binds to the 45S particle first. Without RbgA (top diagram), YphC must displace the long H68 from the E-site to access its binding site in the ribosomal particle. Upon binding, the displaced H68 wraps around YphC and YphC positions H89 in a distinct conformation that prepares the 45S particle for its last assembly step involving the maturation of the A-site. When RbgA binds first to the 45S particle (bottom diagram), the factor displaces H68, and the YphC binding site becomes fully accessible, increasing YphC binding affinity. RbgA also stabilizes H93 in the P-site, providing a first point of contact for YphC and further facilitating its binding. In this double complex, YphC also relocates H89 in the distinct conformation, prepping the 45S particle for its last assembly step involving the maturation of the A-site.

The obtained cryo-EM structure of the 45S particle in complex with RbgA and YphC provided a structural explanation for how this functional interplay is implemented. When RbgA binds first to the 45S particles, it increases YphC binding affinity for the ribosomal particle by binding to the P-site and preventing H68 in the E-site from adopting its mature conformation (Figure 6; bottom diagram). This makes the A- and P-sites fully accessible, facilitating the subsequent binding of YphC. In addition, RbgA plays a crucial role in stabilizing H93 in the P-site (13), which serves as a first point of contact to initiate the interaction between the YphC KH domain and the ribosomal particle. These events combined provide a functional model that explains the binding cooperativity between RbgA and YphC and how the two factors work in conjunction during the last steps of maturation of the 50S subunit. Our results do not exclude the possibility of the 45S particle maturing through an assembly pathway where YphC binds first. However, the ∼30-fold higher affinity of RbgA against the 45S particle and the increased affinity exhibited by YphC upon pre-binding of RbgA to the 45S particle suggest that this last assembly pathway (Figure 6; bottom diagram) is a more likely sequence of events for the maturation of the 50S subunit to proceed.

The observation that the binding of RbgA creates a high-affinity binding site for YphC represents a close parallelism with the mechanisms by which r-proteins assist the folding of the rRNA. As the ribosome is assembled, the initial folds adopted by the rRNA are rapidly bound hierarchically by r-proteins that stabilize these local RNA structures and induce conformational changes that create binding sites for other r-proteins that subsequently enter the assembling particle (3,4). RbgA helps YphC in a similar way, revealing that some assembly factors at the mechanistic level function like r-proteins. However, the main difference is that r-proteins stay bound and become part of the mature ribosome, whereas assembly factors are released once their function is completed.

A striking aspect of the cryo-EM structures of the 45S particles with YphC bound is the alternative conformation that H89 is forced to adopt. The KH domain of YphC appears to unwind the middle section of H89 and forces U2521 and U2522 outside the rRNA helix, possibly poising these two bases for modifications. Most rRNA modifications that have been characterized cluster around functionally important sites in the ribosome, such as the PTC or A-site (33), where H89 is located in the mature 50S subunit. The most frequent modification in rRNA affecting uridines is their isomerization to pseudouridines (typically abbreviated with Ψ). In bacteria, enzymes called pseudouridine synthases recognize specific uridines in the rRNA and catalyze the isomerization (34). The number and location of Ψ residues have been mapped in the rRNA of various bacterial species. Bacillus subtilis contains one Ψ residue in the 16S rRNA in the 30S subunit and five Ψ residues in the 23S rRNA in the 50S subunit. The five modified nucleotides in the 23S rRNA are U1940, U1944, U1946, U2521 and U2634 (33–37). Interestingly, U2521 is one of the two nucleotides in H89 that we observed to adopt a flip-out conformation in our cryo-EM structure. It is still unknown what pseudouridine synthase modifies U2521. A plausible candidate is Ylml (BSU15460) (35). However, this prediction and the exact mechanism of how YphC may assist the modification of U2521 after the nucleotide is forced out of the helix still need to be determined.

In addition, the discovery that YphC causes a displacement of H89 and that RbgA assists YphC binding suggests a maturation timeline that places these two factors ahead of ObgE. This is another assembly factor that has been involved in completing the maturation of the A-site, mainly by properly positioning H89 in its final position in close contact with the bL12 stalk and r-proteins uL16 and bL36. The cryo-EM structure of a 50S assembly intermediate from E. coli with multiple assembly factors bound, including ObgE (5), showed that this factor initially binds the assembly intermediates in the A-site, introducing its N-terminal domain as a wedge between H89 above and H91 and H92 below. H89 at this stage is already formed, but it is displaced upward and progressively transitions into its mature position through a conformational change supported by ObgE. Once in its mature position, H89 is an important structural element that facilitates the entry of the late r-proteins bL36 and uL16. These two proteins are crucial in linking H89 with other functionally critical rRNA helices (H38, H42, H91 and H97) and complete the maturation of the A-site. These are part of the last steps of assembly before the particle reaches maturity. Our results in this study suggest that this final step in the maturation process only occurs after YphC and RbgA have exerted their function. It is possible that YphC, by maintaining H89 outside of its normal position in the A-site, prevents the binding of ObgE. Once YphC is released and H89 is free to move, ObgE can bind and reposition H89 into its mature conformation in the A-site. However, this possibility remains to be tested.

Overall, our results provide new insights into the coordinated sequence of events leading to the maturation of the 50S subunit functional sites and how assembly factors perform separate functions in this process. However, functional interplays between these factors facilitate their activities, boosting the overall efficiency of the ribosome assembly process.

Data availability

The cryo-EM maps obtained in this study and the derived molecular models have been deposited in the Electron Microscopy Data Bank (EMDB) and the Protein Data Bank (PDB) with accession codes detailed in Supplementary Tables S1–S3. Sequences used for multiple sequence alignments are available in Supplementary Table S4.

Supplementary data

Supplementary Data are available at NAR Online.

Funding

Canadian Institutes of Health Research [PJT-180305 to J.O.]; McGill University [to D.A.]; Canada Foundation for Innovation; Fonds de Recherche du Québec – Santé (FRQS). Funding for open access charge: Canadian Institutes of Health Research.

Conflict of interest statement. The authors declare no competing financial interests. The funders had no role in study design, data collection and analysis, decision to publish or manuscript preparation.