High-throughput RNA structure probing reveals critical folding events during early 60S ribosome assembly in yeast

Abstract

While the protein composition of various yeast 60S ribosomal subunit assembly intermediates has been studied in detail, little is known about ribosomal RNA (rRNA) structural rearrangements that take place during early 60S assembly steps. Using a high-throughput RNA structure probing method, we provide nucleotide resolution insights into rRNA structural rearrangements during nucleolar 60S assembly. Our results suggest that many rRNA-folding steps, such as folding of 5.8S rRNA, occur at a very specific stage of assembly, and propose that downstream nuclear assembly events can only continue once 5.8S folding has been completed. Our maps of nucleotide flexibility enable making predictions about the establishment of protein-rRNA interactions, providing intriguing insights into the temporal order of protein-rRNA as well as long-range inter-domain rRNA interactions. These data argue that many distant domains in the rRNA can assemble simultaneously during early 60S assembly and underscore the enormous complexity of 60S synthesis.Ribosome biogenesis is a dynamic process that involves the ordered assembly of ribosomal proteins and numerous RNA structural rearrangements. Here the authors apply ChemModSeq, a high-throughput RNA structure probing method, to quantitatively measure changes in RNA flexibility during the nucleolar stages of 60S assembly in yeast.

Fig. 1

Affinity purification of 27SA₂ and 27SB pre-rRNA. a Schematic representation of the assembly steps of the nucleolar pre-60S particles. The 35S pre-rRNA is processed to 27SB pre-rRNA in various stages, which involves binding of assembly factors required for 27SA₃ processing (A₃ factors), assembly factors required for cleavage in ITS2 (B factors) and r-proteins. b Primer extension analysis of 27S pre-rRNAs isolated from the Rrp5-TAP and Nsa2-TAP pre-60S particles. The cleavage sites are indicated on the right side of the gel image. Lanes labeled with G, A, T and C represent dideoxy sequencing ladders. c, d Violin plots showing the mass spectrometry results for ribosome assembly factors. The plot shows the distribution of A₃ and B factors based on the ESCs from the Rrp5 (c) or Nsa2 (d) pre-60S complexes. The p-value was calculated using a t-test. e Ball plots showing the mass spectrometry results (ESC) for the r-protein composition of the complexes purified by Rrp5-TAP and Nsa2-TAP. The results were compared to the r-protein composition of the Fun12-TAP purified 80S particles. The larger the ball, the more enriched the protein was in the purification

Fig. 2

ChemModSeq experimental workflow and data analysis. a Particles containing the 35S, 27SA₂, and 27SB pre-rRNAs were purified using TAP-tagged Mrd1, Rrp5, or Nsa2 as baits, respectively. Chemical probing was performed on the purified particles as well as PK-treated particles. As controls for natural primer extension stops, we also incubated purified particles and PK-treated particles with the 1M7 solvent (DMSO). The rRNAs were gel purified and their integrity was assessed on a bioanalyzer chip. b–e Many conformational changes take place between the 27A₂ and 27SB particles. Shown are scatter plots that compare normalized SHAPE reactivity values of the 35S data to 27SA₂ and 27SB data (b) and (c) and 27SA₂ compared to 27SB data (d). Data generated from deproteinized particles is shown in e. The similarity between the data sets was determined by calculating Pearson correlation coefficients

Fig. 3

Folding of 25S rRNA domains. a Secondary structure model for the 25S and 5.8S rRNAs generated using RiboVision. The individual domains are colored. b Differential SHAPE (ΔSHAPE) analysis of 35S, 27SA₂, and 27SB particles. The histograms show a comparison between 35S and 27SB data and 27SA₂ compared to 27SB data. If a region is highlighted in green, then it is more reactive in the 35S (top) or 27SA₂ (bottom) data. If a region is highlighted in purple, it is more reactive in the 27SB data. ΔSHAPE reactivities were calculated as using code developed by the Weeks lab. c The heat map indicates the percentage of nucleotides in each 25S rRNA domain (indicated on the left side of each domain) that showed a decreased or an increase reactivity when comparing the various particles (35S vs. 27SB and 27SA₂ vs. 27SB) as well as whether these changes were sensitive or insensitive to PK treatment. Shown are the results for all the nucleotides as well as the nucleotides that overlapped with r-protein-binding sites in the 60S crystal structure. d Overview of the PKSRP sites in the three-dimensional structure of the 25S rRNA. Most of the proposed protein–rRNA interactions are concentrated on the solvent interface, with fewer PKSRP sites identified on the subunit interface, central protuberance (CP), and feet. Protein-binding sites were defined according to the crystal structure of the ribosome

Fig. 4

Models of inter-domain interactions formed during nucleolar stages of 60S assembly. The yellow spheres indicate r-proteins. Predicted protein–rRNA interactions formed during the conversion of 35S of 27SA₂ and the conversion of 27SA₂ to 27SB are indicated with red and blue dots, respectively. Protein-binding sites were defined according to the crystal structure of the ribosome. a Binding of L17/uL22 L4/uL4 and L42/eL42 to domains I (purple), V (wheat), and VI (green) of the 25S during the conversion of 35S to 27SB pre-rRNA. b Sequential binding of L43/eL43, L2/uL2, and L34/eL34 brings domains III (pink) and V (wheat) in close proximity

Fig. 5

The 5.8S region undergoes a major restructuring event during the conversion of 27SA₂ to 27SB pre-rRNA. a Heat map representing the ChemModSeq 1M7 reactivities. Each block indicates a single nucleotide, the darker the block the higher the 1M7 reactivity. The blue blocks in lane 6 indicate r-protein-binding sites. Lane 7 represents the nucleotides that are base-paired (blue blocks) or single-stranded (white blocks) in the mature 5.8S rRNA crystal structure. b Secondary structure of the 5.8S rRNA. Black lines indicate base-pairing interactions observed in the yeast 80S crystal structure. The red nucleotides are single-stranded in the crystal structure. c Overview of ΔSHAPE analysis in the 5.8S region. Yellow-colored nucleotides showed significantly higher SHAPE values in 35S compared to the 27SB particle data

Fig. 6

ITS2 forms a very compact structure. a Overview of the various proposed secondary structure models for ITS2^{, ,} . Roman numerals indicate the regions discussed in the main text. Dots of different colors represent nucleotides with SHAPE reactivity ≥0.4 in ITS2 primer extension data (red) and 35S and 27SA₂ particles (dark blue), 27SB particles (light blue), and deproteinized samples (magenta) based on ChemModSeq data. Note that the primer extension results for the 3′ end are not shown as this region showed high variability in reactivity. The known binding sites for assembly factors that interact with ITS2 are highlighted in the ring-pin model. b Heat map representing the ChemModSeq 1M7 reactivities. Each block indicates a single nucleotide. The darker the block the higher the 1M7 reactivity. c Primer extension analysis of the 5′ end of ITS2 and the 3′ end of 5.8S region and line scans of the signal intensities of each lane. d RNA structure probing data agrees well with the cryo-EM structure of ITS2 fragments. Shown is the secondary structure of parts of ITS2 that were resolved by cryo-EM. The black nucleotides in the sequence are nucleotides that are either base-paired or bound by assembly factors in the structure. Gray nucleotides are predicted to be single-stranded as they are not involved in Watson–Crick or Hoogsteen base-pairing interactions and are not predicted to be bound by proteins

Fig. 7

Model of rRNA structural changes and pre-60S remodeling steps that take place during the nucleolar stages of 60S synthesis. ITS2 forms a highly compact structure in the nascent transcript. We predict that base-pairing interactions between 25S and 5.8S have already been formed co-transcriptionally. Major rRNA restructuring events occur in pre-ribosomal complexes during the conversion of 27SA₂ pre-rRNA to 27SB pre-rRNA. These include the compaction of the 5.8S region and formation of long-range interactions in the 25S sequence. All these changes occur after Rrp5 dissociates and in the presence of A₃ and early B factors. After C₂ cleavage, the 5S rRNA rotates to adopt its final structure