Collision-induced ribosome degradation driven by ribosome competition and translational perturbations

Abstract

Individual stalling of catalytically inactive ribosomes at the start codon triggers ubiquitination of ribosomal protein uS3 and subsequent 18S rRNA decay. While collisions between ribosomes during translation elongation represent a more widespread form of translation perturbation, their impact on ribosome stability remains unknown. Here, we clarify a bifurcation in ubiquitination-mediated ribosome turnover, identifying a collision-induced branch of uS3 ubiquitination and small subunit destabilization in yeast. This pathway eliminates not only non-functional ribosomes but also translationally active ones with a prokaryotic-like decoding center, driven by competition with wild-type ribosomes due to differing translation rates. We further show that endogenous ribosomal subunit stoichiometry shifts toward a small-subunit-shortage state via ubiquitination upon perturbed translation triggered by the anti-cancer drug cisplatin and the growth phase transition. These findings reveal a mechanism by which ribosome dynamics generally affects ribosome stability, implicating ribosome dysfunction, heterogeneity, and stress-related translational disturbances in small subunit degradation.

Fig. 1. Mag2-Fap1- and Mag2-Hel2-mediated branches of 18S rRNA decay.

A, B Schematic of the secondary structure of the yeast decoding center (A) and northern blots analyzing the impact of uS3-related E3 ubiquitin ligase deletion on the stability of 18S rRNAs with conserved bases substituted in or near the decoding center (B). Northern blot analysis was performed using three biological replicates, and representative results are shown. Time after transcription shut-off is indicated in hours (hr). Mutants are grouped and color-coded based on changes in their stability upon E3 ubiquitin ligase deletion. WT: wild-type; Endo: endogenous. C Comparison between the secondary structures of yeast 18S rRNA and bacterial 16S rRNA helix 44 (h44). The two non-conserved bases in yeast 18S rRNA are highlighted in cyan. D Spot growth assay of endogenous rDNA-deleted (rdnΔΔ) cells complemented with either the 18S: WT or 18S: GA/AG plasmid on YPD plates under the indicated conditions. E Northern blots assessing the stability of plasmid-derived 18S: GA/AG rRNA in the indicated strains. Half-lives were estimated by fitting the band intensities and time points from three independent experiments to a single-exponential decay model using least-squares fitting. Replicates and/or samples from different strains were run on separate gels when needed due to lane number limitations. All gels and blots were processed and analyzed using the same method. The mean half-lives ± standard deviation are shown. F Western blot analysis detecting 3×HA-tagged uS3 in lysates (input) or MS2-purified ribosomes containing plasmid-derived 18S: WT or 18S: GA/AG with an MS2 binding sequence (MS2 bs). Ub: ubiquitin. Enrichment of plasmid-derived 18S rRNA by the purification was confirmed by RT-PCR using primers flanking the MS2 binding sequence, comparing fragment lengths between input and MS2-purified fractions. Experiments were performed using two biological replicates. Representative results are shown. Source data are provided as a Source Data file.

Fig. 2. Ribosomes with a prokaryotic-like decoding center are translationally active but suboptimal in yeast.

A Length distribution of ribosome-protected mRNA fragments (RPFs) in ribosome profiling libraries from MS2-purified ribosomes containing either 18S: WT or 18S: GA/AG. WT: wild-type. B Metagene plots showing A-site positions of MS2-purified ribosomes around the start and stop codons. RPM: reads per million; n = 1503 transcripts (also for other metagene analysis data shown in this figure). C Scatter plots showing the log2 pause scores for P-site (left) or A-site (right) codons in MS2-purified ribosomes, comparing 18S: GA/AG to 18S: WT. D Histogram showing the distribution of the ratio of RPFs mapped to 3’ untranslated region (UTR) relative to all RPFs. E Averaged distribution of RPFs from MS2-purified 18S: WT or 18S: GA/AG ribosomes across open reading frames (ORFs). The fraction of RPFs mapped within ORF regions, excluding the first and last five codons, is shown. Data represent the mean of two replicates. The cartoon illustrates a possible interpretation of the loss of 18S: GA/AG RPFs at start and stop codons. F Distribution of RPFs on two representative ORFs showing a depletion of 18S: GA/AG-derived RPFs at start and stop codons compared to 18S: WT. Results from one replicate are shown. G Western blot assessing puromycin incorporation in rdnΔΔ cells complemented with 18S: WT or 18S: GA/AG plasmids. The duration of puromycin (Puro) treatment is indicated in minutes (min). Ponceau S staining was used as a loading control. A representative blot from four biological replicates is shown. H Ribosome run-off assay of rdnΔΔ cells complemented with 18S: WT or 18S: GA/AG plasmids. Time after lacmidomycin (LTM) addition is indicated in minutes (min). Average polysome-to-80S ratios at each time point are shown, with error bars indicating standard deviation (n = 3 biological replicates). A two-sided Welch-Aspin test was used to compare 18S: GA/AG and 18S: WT. P-values (p) are shown. Source data are provided as a Source Data file.

Fig. 3. Competition between two active ribosome species drives collision-induced clearance of the slower ribosomes.

A Western blot detecting 3×HA-tagged uS3 of MS2-purified ribosomes from rdnΔΔ cells expressing the POL1p- and GAL7p-rRNA plasmids in the indicated combinations. RT-PCR using primers flanking the MS2 binding sequence shows relative abundance of POL1p- and GAL7p-derived rRNA. Experiments were performed using two biological replicates. A representative result is shown. WT: wild-type; Ub: ubiquitin. B Northern blot investigating the stability of GAL7p-derived 18S rRNA from rdnΔΔ cells expressing the indicated plasmid combinations. Time after transcription shut-off is indicated in hours (hr). Data from three independent clones are shown. C Sucrose density gradient profiles of hel2Δ cells expressing POL1p-18S: WT or 18S: GA/AG, treated with (lower) or without (upper) MNase. Absorbance was normalized by the 80S peak height. A representative replicate is shown. The box plot shows changes in the polysome-to-monosome ratio upon MNase treatment, calculated from areas under the 80S and polysome (di-some and higher) peaks of six experiments using three clones. Box plot elements: center line, median; box limits, quartiles; whiskers, 1.5× interquartile range; points, individual data values; triangles, mean. Statistical significance was assessed using a linear model including batch as a covariate. The p-value (p) is indicated. Endo: endogenous. D Western blot analysis detecting 3×HA-tagged uS10 of MS2-purified ribosomes with the indicated 18S rRNA. uL14 serves as a ribosome abundance control. A representative blot from two independent experiments is shown. E Western blot detecting 3×HA-tagged uS10 from cells overexpressing (OE) HEL2, MAG2, or both, treated with anisomycin or the vehicle (DMSO). eEF3 serves as a loading control. Experiments were performed using three biological replicates. A representative blot is shown. F Western blot detecting protein products from the GFP-R12-FLAG-HIS3 reporter in cells overexpressing MAG2 or the empty vector (EV). eEF3 serves as a loading control. A representative result from two experiments is shown. Source data are provided as a Source Data file. G Model of ribosome competition leading to small subunit destabilization.

Fig. 4. Collision-induced ribosome degradation responds to stress-related translation perturbations.

A Western blot detecting 3×HA-tagged uS3 in cells treated with cisplatin (+) or the vehicle DMSO (-). eEF3 is a loading control. A representative result from two experiments is shown. WT: wild-type; Ub: ubiquitin. B Sucrose density gradient profiles of cisplatin or vehicle-treated cells. Absorbance was normalized by aligning the height of the 80S peak across all samples. A representative replicate is shown. Red arrows indicate cases where the 40S peak height is lower than the 60S peak; gray arrows indicate the opposite. C Quantification of the 40S-to-60S ratio based on the area under the curve in sucrose density gradient profiles in Fig. 4B. Mean ratios in each condition are shown. Error bars represent standard deviation, n = 6 (WT) or 3 (others) biological replicates. Statistical significance was assessed using a two-sided paired t-test. p: p-value. D Scatter plot comparing pause scores for A-site codons from ribosome profiling of cisplatin-treated cells and the vehicle control. n = 1294 transcripts; first and last five codons excluded. E Metagene plots depicting average pause scores in a ±50-codon window centered around the tryptophan (Trp) UGG (TGG as DNA) codon. F, G Sucrose density gradient profiles of cells harvested at the indicated time points after the optical density at 600 nm (OD600) reached 0.6 to 0.8 (mid-log phase) (F), using normal or low Mg²⁺ in lysis buffer and gradients (G). Time is indicated in hours (hr). Representative results from two experiments are shown. H Western blot detecting 3×HA-tagged uS3 and Ponceau S staining. A representative blot from two biological replicates is shown. I, J Sucrose density gradient profiles of cells harvested at 0 hr or 18 hr after mid-log phase (I) and quantification of the 40S-to-60S ratio (J). Absorbance was normalized to align the height of the 80S peak within each time point. Average ratios are shown. Error bars represent standard deviation, n = 5 (WT), 2 (fap1Δ 0 hr), or 3 (others). No error bar is shown for fap1Δ 0 hr due to only two replicates. Source data are provided as a Source Data file.

Fig. 5. Model of a branched enzymatic pathway of uS3 ubiquitination linking translational perturbations to ribosomal small subunit degradation.

During translation, stalled or slow ribosomes first receive a yellow card from Mag2, marked by mono-ubiquitination at uS3 as potential degradation substrates. These ribosomes remain stable unless further stalling or collisions occur. However, persistent translational disturbances trigger additional recognition steps and extension of the ubiquitin chain on uS3: Individual stalled ribosomes are poly-ubiquitinated by Fap1, while collided ribosomes are poly-ubiquitinated by Hel2. Poly-ubiquitination of uS3, along with ubiquitination of uS5 (and, depending on the context, uS10), acts as a red card committing ribosomes to small subunit degradation. We refer to the Mag2-Fap1-mediated pathway as stalling-induced ribosome destabilization (SRD) and the Mag2-Hel2-mediated pathway as collision-induced ribosome destabilization (CoRD). Ribosomes moving smoothly in translation but encounter abrupt stalling and collisions, e.g., at mRNA sequences with strong road-blocking effects, may bypass Mag2 binding, leading Hel2 to ubiquitinate uS10 in the absence of uS3 mono-ubiquitination and triggering RQC/NGD rather than CoRD. Ub: ubiquitination.