A non-canonical role for a small nucleolar RNA in ribosome biogenesis and senescence

SUMMARY

Cellular senescence is an irreversible state of cell-cycle arrest induced by various stresses including aberrant oncogene activation, telomere shortening, and DNA damage. Through a genome-wide screen, we discovered a conserved snoRNA, SNORA13, that is required for multiple forms of senescence in human cells and mice. Although SNORA13 guides pseudouridylation of a conserved nucleotide in the ribosomal decoding center, loss of this snoRNA minimally impacts translation. Instead, we found that SNORA13 negatively regulates ribosome biogenesis. Senescence-inducing stress perturbs ribosome biogenesis, resulting in accumulation of free ribosomal proteins (RPs) that trigger p53 activation. SNORA13 interacts directly with RPL23, decreasing its incorporation into maturing 60S subunits and, consequently, increasing the pool of free RPs, thereby promoting p53-mediated senescence. Thus, SNORA13 regulates ribosome biogenesis and the p53 pathway through a non-canonical mechanism distinct from its role in guiding RNA modification. These findings expand our understanding of snoRNA functions and their roles in cellular signaling.

Graphical Abstract

In Brief

A small nucleolar RNA (snoRNA) promotes p53-mediated cellular senescence by limiting incorporation of ribosomal proteins into maturing ribosomal subunits; this in turn leads to accumulation of free ribosomal proteins that activate p53.

INTRODUCTION

Rather than promoting proliferation, aberrant oncogene activation in non-transformed cells often induces an irreversible state of cell cycle arrest termed oncogene-induced senescence (OIS)^1–3. This anti-proliferative response of cells to excessive oncogenic activity is widely believed to function as an important tumor suppressor mechanism that prevents pre-malignant cells with oncogenic mutations from expanding and initiating tumorigenesis. Indeed, many different types of pre-cancerous lesions show signs of senescence, including p16^INK4A expression and senescence-associated β-galactosidase activity (SA-β-gal)⁴, while more advanced tumors typically lose these markers^1,2,5. In addition to oncogene hyperactivity, senescence can be triggered by a variety of other stresses, including telomere shortening, DNA damage, oxidative stress, and several chemotherapeutic drugs^1,2,5,6. Along with cell cycle arrest, cellular senescence triggers a pro-inflammatory secretory phenotype that is thought to contribute to many aging-associated conditions, including neurodegeneration, cardiovascular disease, and diabetes^7–9. These broad effects of cellular senescence on human health and disease underscore the importance of achieving a comprehensive understanding of the molecular mechanisms that drive this critical stress response pathway.

The p53 and retinoblastoma protein (RB) pathways both play central roles in mediating cellular senescence, and impairment of either of these pathways is sufficient to prevent initiation of senescence in various model systems^10–12. Recently, oncogenic and replicative stress were shown to activate p53 and cellular senescence in part by perturbing ribosome biogenesis¹³. Imbalances between rRNA and ribosomal proteins (RPs) result in the accumulation of non-ribosome associated RPs, which bind and inhibit MDM2, a major negative regulator of p53¹⁴. This nucleolar stress response pathway has emerged as an important axis for activating p53 in the setting of diverse cellular stressors¹⁵.

While numerous protein-coding genes that contribute to the senescence program have been identified, the roles of noncoding RNAs in cellular senescence remain poorly understood. An intriguing class of noncoding RNAs with emerging links to senescence are the small nucleolar RNAs (snoRNAs), a family of noncoding RNAs that classically function as guides for chemical modification of other RNAs such as ribosomal RNAs (rRNAs) and small nuclear RNAs (snRNAs)^16–19. Most snoRNAs are encoded in introns of protein-coding or noncoding transcripts and are processed into their mature ~60-300 nucleotide forms after excision of the host intron during splicing. Although eukaryotic cells express hundreds of distinct snoRNAs, they can largely be classified into two groups: the H/ACA and C/D box snoRNAs. Each class of snoRNA assembles a distinct ribonucleoprotein (RNP) complex, with H/ACA box snoRNPs converting uridine to pseudouridine (Ψ) and C/D box snoRNPs depositing 2’-O-methylation at sites selected by snoRNA:target RNA base pairing interactions.

Beyond their canonical role as guides for RNA modifications, snoRNAs have been proposed to perform a diverse array of additional functions including regulation of pre-mRNA splicing²⁰ and regulation of target mRNAs through a microRNA-like mechanism²¹. Indeed, a large number of snoRNAs without identifiable RNA targets, referred to as orphan snoRNAs, are expressed in human cells²², highlighting the potential of these transcripts to perform yet-to-be characterized functions. Recently, the expression of H/ACA box snoRNAs was examined in human fibroblasts undergoing HRAS^G12V-induced senescence, revealing that SNORA24 is induced under these conditions²³. Depletion of SNORA24 in a mouse model of NRAS^G12V-driven liver cancer abrogated senescence and accelerated tumor progression. SNORA24 guides the pseudouridylation of two residues in 18S rRNA and loss of these modifications reduced translational fidelity, although how this effect impaired activation of the senescence program is not yet known. Other snoRNAs have been reported to bind directly to KRAS and negatively regulate downstream signaling²⁴, providing additional evidence of a functional link between snoRNAs and the Ras pathway.

Here we describe our discovery that SNORA13, a highly conserved H/ACA box snoRNA, is essential for multiple forms of senescence in human cells and in mice. Although this snoRNA guides the pseudouridylation of a highly conserved nucleotide in 18S rRNA located in the decoding center of the ribosome, we found that loss of SNORA13 had little effect on translation. Rather, loss of this snoRNA accelerated 60S ribosomal subunit biogenesis, thereby decreasing nucleolar stress and consequent p53 activation. Remarkably, these effects on ribosome biogenesis were genetically separable from SNORA13-guided pseudouridylation of 18S rRNA and could be attributed to a direct interaction of SNORA13 with RPL23, which regulated the rate of 60S subunit assembly. Together, these findings revealed a non-canonical function for an H/ACA box snoRNA as a regulator of ribosome biogenesis and the p53 pathway that potently impacts cellular senescence.

RESULTS

A CRISPRi screen for noncoding RNAs that are required for oncogene-induced senescence nominates EPB41L4A-AS1

To identify noncoding RNAs that are essential for OIS, we performed a genome-wide CRISPR interference (CRISPRi) screen using a well-established model in which TERT-immortalized human fibroblasts undergo senescence upon induction of a tamoxifen-inducible HRAS^G12V transgene (BJ-HRAS^G12V cells)^25,26. These cells were engineered to express a catalytically-inactive Cas9 protein fused to the KRAB transcription repressor domain (dCas9^KRAB), which directs transcriptional silencing when guided to target loci by single guide RNAs (sgRNAs)²⁷. We confirmed that CRISPRi-mediated silencing of p53 enabled escape from senescence in multiple independent dCas9^KRAB-expressing BJ-HRAS^G12V subclones that were used in the screen (Figure S1A).

BJ-HRAS^G12V-dCas9^KRAB cells were infected with a previously described lentiviral library consisting of 68,009 sgRNAs targeting 6701 fibroblast-expressed noncoding RNAs²⁸ (Figure 1A). Upon addition of tamoxifen, most cells entered a senescent state and stopped proliferating. Cells expressing sgRNAs targeting essential mediators of senescence, however, continued to proliferate under these conditions, resulting in sgRNA enrichment. High-throughput sequencing was used to determine sgRNA abundance in cells grown in the presence tamoxifen for three weeks and the MAGeCK algorithm²⁹ was used to identify significant hits (Figure 1B and Table S1). Only four hits reached a stringent enrichment and statistical significance threshold (log2 fold enrichment >2; p<10⁻⁶). Interestingly, for three of these hits, the nearest transcription unit to the enriched sgRNAs was a protein-coding gene (CCNC, CDKN1A, and FOXF1). All three of these genes have previously been reported to encode strong regulators of senescence or p53 activity^25,30,31, thus demonstrating that the screen successfully identified known mediators of senescence. The remaining highly-ranked hit was a poorly-characterized noncoding RNA known as EPB41L4A-AS1.

Figure 1. Identification of EPB41L4A-AS1 as a noncoding RNA required for oncogene-induced senescence.

(A) Overview of the genome-wide CRISPRi screen for noncoding RNAs required for OIS (created with BioRender.com).

(B) Plot of gene enrichment (tamoxifen-treated/untreated cells) versus p value calculated by MAGeCK.

(C) Schematic of EPB41L4A-AS1 locus showing location and conservation of SNORA13. UCSC Genome Browser PhastCons track shown in green.

(D-E) qRT-PCR analysis of EPB41L4A-AS1 expression relative to GAPDH (D), or northern blot of SNORA13 and U6 snRNA (loading control) (E), in BJ-HRAS^G12V-dCas9^KRAB cells after lentiviral expression of non-target (NT) or EPB41L4A-AS1-targeting sgRNAs. Indicated cells were treated with tamoxifen for 7 days.

(F) Growth of BJ-HRAS^G12V-dCas9^KRAB cells expressing the indicated sgRNAs in the absence (left) or presence of tamoxifen (right).

(G) Cell morphology, SA-β-gal staining, and EdU incorporation of BJ-HRAS^G12V-dCas9^KRAB cells expressing the indicated sgRNAs 15 days after addition of tamoxifen.

(H) Quantification of SA-β-gal activity and EdU incorporation in (G).

Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing sgNT group (sgNT1 and sgNT2) to sgEPB41L4A-AS1 group (EPB41L4A-AS1 sg1 and sg2) (panels F,H). n.s., not significant; ****p≤0.0001.

See also Figure S1.

We were intrigued by the finding that EPB41L4A-AS1 may be essential for OIS because it serves as a host transcript for a highly conserved H/ACA box snoRNA, SNORA13 (Figure 1C). To validate this result, we first knocked down EPB41L4A-AS1 using CRISPRi and grew cells in the presence or absence of tamoxifen. As expected, this resulted in a strong decrease in the levels of both EPB41L4A-AS1 and SNORA13 (Figures 1D and 1E). Under normal growth conditions (HRAS^G12V-off state), knockdown of EPB41L4A-AS1 had no effect on proliferation (Figure 1F, left). Upon induction of HRAS^G12V, however, cells with reduced EPB41L4A-AS1 expression failed to enter a senescent state and continued to proliferate (Figure 1F, right). In addition to sustained growth, loss of EPB41L4A-AS1 prevented the acquisition of phenotypic markers of senescence, including an enlarged and flattened cell morphology and the presence of SA-β-gal activity^3,32 (Figures 1G and 1H). EdU labeling further demonstrated that cells with knockdown of EPB41L4A-AS1 maintained active proliferation and DNA replication in the presence of HRAS^G12V. Importantly, we confirmed that induction of the HRAS^G12V transgene occurred normally in EPB41L4A-AS1 knockdown cells (Figure S1B). These results suggested that EPB41L4A-AS1, and/or the encoded snoRNA, are required for OIS.

SNORA13 is required for multiple forms of senescence

To determine whether the spliced EPB41L4A-AS1 transcript or the encoded snoRNA were required for senescence, we used CRISPR with dual sgRNAs to delete SNORA13, leaving the rest of the transcript intact (Figure 2A). Homozygous knockout clones exhibited complete loss of SNORA13 but maintained normal expression of the spliced EPB41L4A-AS1 transcript (Figures 2B and 2C). As observed in cells with knockdown of EPB41L4A-AS1, SNORA13 knockout cells proliferated normally in the absence of HRAS^G12V expression, but failed to enter a senescent state in the presence of oncogenic stress (Figures 2D, 2H, 2I, and S1C).

Figure 2. SNORA13 is required for oncogene-induced senescence.

(A) Schematic of dual guide CRISPR strategy for SNORA13 deletion.

(B-C) Northern blot of SNORA13 and U6 snRNA (B), or qRT-PCR analysis of EPB41L4A-AS1 expression relative to GAPDH (C), in wild-type BJ-HRAS^G12V cells or four independent SNORA13 knockout clones.

(D) Growth of wild-type or SNORA13 knockout BJ-HRAS^G12V cells in the absence (left) or presence of tamoxifen (right).

(E) Schematic of EPB41L4A-AS1 rescue constructs with or without SNORA13.

(F) Northern blot of SNORA13 and U6 snRNA in wild-type or SNORA13 knockout BJ-HRAS^G12V cells following stable transfection with empty vector or the indicated EPB41L4A-AS1 rescue constructs.

(G) Growth of BJ-HRAS^G12V cells expressing the indicated EPB41L4A-AS1 rescue constructs in the presence of tamoxifen.

(H) Cell morphology, SA-β-gal staining, and EdU incorporation of BJ-HRAS^G12V cells expressing the indicated EPB41L4A-AS1 rescue constructs 15 days after addition of tamoxifen.

(I) Quantification of SA-β-gal activity and EdU incorporation in (H).

Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing all groups to WT (panels D,G,I). n.s., not significant; ****p≤0.0001.

See also Figure S1.

We further verified the requirement for SNORA13 in OIS through a series of rescue experiments. The complete EPB41L4A-AS1 gene locus, including all introns and SNORA13, was introduced into SNORA13 knockout cells using a PiggyBac transposon expression vector (Figures 2E, 2F, and S1D). In parallel, we introduced a similar construct lacking only the SNORA13 sequence. While neither construct affected proliferation in the absence of oncogenic stress (Figure S1E), only the SNORA13-containing construct was able to rescue senescence upon induction of HRAS^G12V (Figures 2G–2I).

To determine whether SNORA13 is required for other forms of senescence, we depleted the snoRNA in primary human fibroblasts and subjected them to prolonged passaging. Loss of SNORA13 in this context delayed the onset of replicative senescence, as demonstrated by increased proliferation at late passage numbers, diminished SA-β-gal activity, and reduced expression of well-known senescence markers p21 and p16 (Figures S1F–S1M). Similarly, induction of senescence by DNA damage was significantly attenuated in SNORA13-deficient cells (Figures S1N–S1Q). Altogether, these results established that loss of SNORA13 impairs multiple senescence pathways in human cells.

SNORA13 guides the pseudouridylation of 18S:1248U

To begin to elucidate the mechanism through which SNORA13 participates in senescence, we first examined its subcellular localization. As demonstrated both by cellular fractionation and RNA fluorescence in situ hybridization (FISH), SNORA13 localized predominantly to the nucleolus, as expected for an RNA of this class (Figures S2A and S2B). Induction of oncogenic stress did not affect the localization or overall expression of this snoRNA (Figures S2A–S2C).

SNORA13 is a member of the H/ACA box family of snoRNAs, whose canonical function is to guide the pseudouridylation of target RNAs³³. Specifically, SNORA13 has been predicted to guide the pseudouridylation of nucleotide 1248 of human 18S rRNA (18S.1248U)³⁴. This nucleotide is located in the decoding center of the ribosome and, in addition to undergoing conversion to pseudouridine (Ψ), is known to be further modified by the enzymes EMG1 and TSR3 to produce 1-methyl-3-α-amino-α-carboxyl-propyl pseudouridine (m¹acp³Ψ)^35–40 (Figure 3A). We confirmed that SNORA13 is required for pseudouridylation of 18S.1248U using multiple assays. As described previously⁴¹, the bulky acp³ adduct on this nucleotide causes base misincorporation during reverse transcription, leading to disruption of a HinF1 restriction site at this position in PCR products that span this sequence. As expected, the majority of RT-PCR products encompassing this site are resistant to HinF1 cleavage in wild-type cells, but cleavage is increased in TSR3 knockout cells due to loss of the acp³ modification (Figure 3B, lanes 1 and 3). Note that pseudouridylation of this nucleotide is not required for addition of acp³ (Figure 3B, lanes 1 and 2). To detect pseudouridine at this position, RNA from TSR3 knockout cells was treated with N-cyclohexyl-N’-(2-morpholinoethyl)carbodiimide metho-p-toluenesulfonate (CMC), which adds a large, alkaline-resistant adduct specifically to pseudouridine residues^42,43. This again resulted in nucleotide misincorporation during reverse transcription and a consequent loss of cleavage by HinF1 (Figure 3B, compare lanes 3 and 7). Deletion of SNORA13 in TSR3 knockout cells resulted in complete digestion of the PCR product amplified from CMC-treated RNA (Figure 3B, lane 8), thus demonstrating that SNORA13 is required for pseudouridylation of 18S.1248U. We further confirmed this result using a primer extension assay that detects reverse transcriptase stalling due to the acp³ modification or the CMC modification of pseudouridine at this site (Figure S2D).

Figure 3. SNORA13 guides pseudouridylation of 18S rRNA and negatively regulates 60S ribosomal subunit biogenesis.

(A) Installation of m¹acp³Ψ on human 18S:1248U. Figure modified from Sloan et al.¹⁶.

(B) HinF1 digestion assay to monitor 18S.1248m¹acΨ.

(C) Anti-puromycin western blot of wild-type or SNORA13 knockout BJ-HRAS^G12V cells pulsed with puromycin in the absence of tamoxifen.

(D) Flow cytometry analysis of OP-Puro incorporation in wild-type or SNORA13 knockout BJ-HRAS^G12V cells in the absence of tamoxifen.

(E) Volcano plot of translation efficiency (TE; defined as ribosome occupancy/mRNA abundance), as determined by ribosome profiling and RNA-seq, in SNORA13 knockout relative to wild-type BJ-HRAS^G12V cells treated with tamoxifen for 7 days. Transcripts with a statistically-significant increase or decrease in TE (p<0.01) highlighted in red.

(F) Sucrose gradient fractionation of wild-type and SNORA13 knockout BJ-HRAS^G12V cells in the absence of tamoxifen (upper) or after 7 days of tamoxifen treatment (lower).

(G) Sucrose gradient fractionation showing total ribosome levels (upper) or fluorescently-labeled newly synthesized 60S ribosomal subunits (lower) in HEK293T RPL28-SNAP cells after lentiviral expression of Cas9 and non-target sgRNA (sgNT) or sgRNA targeting SNORA13.

(H) Images of pulse-labeled HEK293T RPL28-SNAP cells expressing Cas9 and the indicated sgRNAs immediately after labeling (0 hr) or 1 hour later. White arrowheads indicate nucleolar RPL28.

(I) Quantification of nuclear RPL28 signal in (H). Data are represented as mean ± SD with individual data points shown (n=20 cells per condition). p values were calculated by unpaired two-tailed student’s t-test. n.s., not significant; ****p≤0.0001.

See also Figures S2–S5.

Loss of SNORA13 has a minimal effect on translation

Since the m¹acp³Ψ modification of 18S:1248U is conserved in all eukaryotes, and the modified nucleotide is located in the decoding center of the ribosome where it makes contact with the P site tRNA^41,44, we initially hypothesized that loss of SNORA13 may impact the translation of mRNAs that encode regulators of senescence. To investigate this possibility, we first examined global translation rates by monitoring the incorporation of puromycin, or its analog O-propargyl-puromycin (OP-puro), into nascent peptides by western blotting or flow cytometry, respectively⁴⁵. We observed no detectable change in global translation upon loss of SNORA13 under normal growth conditions (Figures 3C and 3D). We did detect an increase in translation in SNORA13 knockout cells in the presence of oncogenic stress (Figures S2E–S2G), but this was likely a secondary consequence of the sustained proliferation of these cells relative to wild-type cells, since loss of p53 also led to an increase in translation under these conditions. Moreover, these data were consistent with the recent finding that p53 negatively regulates translation through transactivation of the translational repressor 4E-BP1⁴⁶. As shown below, p53 activation is impaired in SNORA13 knockout cells, which could lead to increased translation by abrogating this mechanism. We also showed that loss of SNORA13 had no overt effect on translation fidelity, as measured by reporter assays in which amino acid misincorporation allows stop codon readthrough or correction of a point mutation (Figures S2H and S2I).

We next considered the possibility that loss of SNORA13 affects the translation of specific mRNAs that regulate senescence. Ribosome profiling and RNA sequencing (RNA-seq) were performed on wild-type and SNORA13 knockout BJ-HRAS^G12V cells with or without oncogenic stress (Figure S3A, Tables S2 and S3). Surprisingly, only a small number of transcripts exhibited significantly altered translation efficiency in SNORA13-deficient cells (Figure 3E). Furthermore, CRISPR-mediated knockout of genes whose translation was decreased in SNORA13 knockout cells did not result in escape from senescence in wild-type cells (Figures S3B and S3C). Similarly, inactivation of genes whose translation was increased in SNORA13 knockout cells did not reactivate the senescence program in these cells (Figures S3D and S3E). Although it remains formally possible that knocking out these genes in combination would measurably impact senescence, these data argued against a major role for a SNORA13-dependent translation program as a driver of senescence.

Because the 18S:1248(m¹acp³Ψ) nucleotide directly contacts the P site tRNA^41,44, we also examined the translational efficiency of each codon in SNORA13 wild-type and knockout cells. In the presence of oncogenic stress, transcripts enriched for codons with an A or U at the third position (AU3 codons) exhibited an increase in translational efficiency in SNORA13 knockout cells, while transcripts enriched for codons with G or C at the third position (GC3 codons) were translated less efficiently (Figure S4A). However, several observations suggested that this effect was a secondary consequence of sustained proliferation of SNORA13 knockout cells in the presence of oncogenic stress. First, it has been reported that mRNAs enriched for AU3 codons are preferentially translated in proliferating cells^47,48. Indeed, we observed increased translation of transcripts enriched for AU3 codons in proliferating wild-type BJ-HRAS^G12V cells compared to senescent cells (Figure S4B). Moreover, analysis of a previously published ribosome profiling dataset from p53-deficient BJ-HRAS^G12V cells⁴⁹ revealed that loss of p53 also led to a similar pattern of codon usage upon activation of oncogenic stress (Figure S4C). Finally, we documented that loss of SNORA13 had minimal effects on codon usage under normal growth conditions (Figure S4D). Altogether, these results indicated that SNORA13 is unlikely to regulate senescence by altering global, transcript, or codon-specific translation rates. We note that these results are distinct from those reported in a recent study of the translational consequences of TSR3 knockout and the subsequent loss of the acp³ modification at this site in 18S rRNA, which impacts the translation of ribosomal protein-encoding mRNAs⁴¹. This difference is likely due to the fact that pseudouridylation of this nucleotide is not required for subsequent acp³ addition (Figure 3B).

SNORA13 negatively regulates 60S ribosomal subunit biogenesis

In addition to the aforementioned analyses of translation, we also examined the abundance of ribosomal subunits and translating ribosomes in SNORA13-deficient BJ-HRAS^G12V cells using sucrose gradient fractionation. Unexpectedly, this revealed an increase in the steady-state abundance of free 60S subunits and 80S monosomes in SNORA13 knockout cells with or without activation of oncogenic HRAS^G12V (Figure 3F). This effect was reversed by exogenous expression of SNORA13 (Figure S4E). While overall levels of rRNA were increased (Figure S4F), we did not detect an increase in the production of nascent rRNA (Figure S4G), suggesting that an excess of pre-rRNA may be transcribed in these cells and stabilized by assembly of new ribosomes upon loss of SNORA13. We also detected no increase in 28S pre-rRNA processing (Figure S4H), suggesting that a distinct step in 60S subunit assembly might be accelerated in SNORA13-deficient cells.

To monitor the rate of 60S subunit biogenesis, we took advantage of a HEK293T cell line expressing an endogenous SNAP-tagged large ribosomal subunit protein (RPL28) that enables selective labeling of newly-formed ribosomal subunits⁵⁰. Loss of SNORA13 resulted in elevated levels of total and newly-synthesized 60S subunits and 80S monosomes (Figures 3G, S5A, and S5B). Collapsing polysomes into free ribosomal subunits with EDTA confirmed that SNORA13 depletion increased the steady-state abundance of 60S subunits (Figure S5C, upper panel). A small increase in 40S subunit abundance was also detected under these conditions, possibly due to the role of SNORA13 in guiding pseudouridylation of 18S rRNA, which could impact 40S biogenesis. Analysis of newly synthesized 60S subunits under these conditions further verified a dramatic increase in SNORA13 knockout cells (Figure S5C, lower panel). In addition, we used fluorescent microscopy to monitor the rate of disappearance of labeled RPL28 from nucleoli, indicative of 60S assembly and export to the cytoplasm. Whereas most newly-labeled RPL28 was retained in nucleoli at one hour after labeling in wild-type cells, significantly less nucleolar RPL28 remained in SNORA13-depleted cells at this timepoint (Figures 3H and 3I). These experiments demonstrated that deficiency of SNORA13 resulted in accelerated 60S subunit biogenesis.

Our finding that SNORA13 functions as a strong negative regulator of ribosome biogenesis raised the question of whether this activity contributes to downregulation of ribosome production in physiologic settings. For example, in cells starved of amino acids, mTOR signaling is inhibited, resulting in downregulation of ribosome biogenesis at multiple levels including rRNA transcription and translation of mRNAs encoding ribosomal proteins⁵¹. Interestingly, we observed a strong upregulation of SNORA13 expression after amino acid starvation or rapamycin treatment (Figures S5D and S5E). Moreover, depletion of SNORA13 resulted in a relative preservation of 60S subunit production in amino acid-starved cells, as revealed by sucrose gradient analyses of steady-state and newly-synthesized 60S subunits (Figures S5F and S5G). Thus, SNORA13 activity is required for maximal repression of ribosome biogenesis after nutrient depletion.

SNORA13 regulates p53 activity through the nucleolar stress response pathway

Given our finding that altered translation was unlikely to explain the role of SNORA13 in senescence, we sought additional mechanistic insight by analyzing RNA-seq data from wild-type and SNORA13 knockout BJ-HRAS^G12V cells (Table S3). Interestingly, Gene Set Enrichment Analysis (GSEA) indicated that activity of the p53 pathway was decreased in SNORA13 knockout cells, with or without induction of oncogenic stress (Figure 4A). Accordingly, the key p53 target gene CDKN1A, which encodes the cell-cycle inhibitor p21, displayed reduced baseline expression, and significantly diminished induction after tamoxifen treatment, in SNORA13 knockout cells (Figures 4B and 4C). Although SNORA13-deficient cells did not exhibit a strong reduction in the overall abundance of p53, a large fraction of the protein localized to the cytoplasm after induction of oncogenic stress, in contrast to the strong nuclear accumulation of p53 that occurred in wild-type cells under these conditions (Figures 4C–4E). Consistent with these findings, nuclear p53 activity was reduced in SNORA13 knockout cells, as demonstrated by chromatin immunoprecipitation (ChIP), which documented significantly decreased binding of p53 to response elements (REs) in the CDKN1A promoter^52–54 (Figure 4F). Similarly, overall p53 levels were not decreased in SNORA13-deficient primary fibroblasts compared to wild-type cells after prolonged passaging or treatment with etoposide to induce DNA damage (Figures S1I, S1M, and S1Q), but a fraction of p53 aberrantly localized to the cytoplasm in SNORA13 knockout cells under these conditions (Figure S6A).

Figure 4. SNORA13 regulates p53 activity through the nucleolar stress response pathway.

(A) Gene set enrichment analysis (GSEA) showing downregulation of the p53 pathway in SNORA13 knockout BJ-HRAS^G12V cells at baseline or after tamoxifen treatment.

(B) qRT-PCR analysis of CDKN1A expression relative to GAPDH in wild-type BJ-HRAS^G12V cells or four independent SNORA13 knockout clones at baseline or after tamoxifen treatment.

(C-D) Western blots of total cell lysates (C) or nuclear and cytoplasmic fractions (D) in wild-type or SNORA13 knockout BJ-HRAS^G12V cells.

(E) Immunofluorescence of p53 in tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells.

(F) ChIP-qPCR of p53 binding at the CDKN1A promoter in tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells.

(G) Western blots showing ribosome and non-ribosome associated RP levels in wild-type or SNORA13 knockout BJ-HRAS^G12V cells with or without tamoxifen treatment. Bar graph shows western blot quantification of ribosome-free RP levels in tamoxifen-treated cells (n=3 biological replicates).

(H) Western blot of MDM2 co-immunoprecipitation with the indicated proteins in tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells.

(I-J) Growth of tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells after MDM2 inhibition using siRNA (I) or Nutlin-3 (J).

Tamoxifen treatment was carried out for 7 days for all experiments except panels I-J, where tamoxifen treatment was maintained throughout the experiment. Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing groups to WT (panels B,F,G), SNORA13 KO siNonTarget (panel I), or SNORA13 KO (DMSO) (panel J). **p≤0.01, ***p≤0.001, ****p≤0.0001.

See also Figure S6.

Previous studies have shown that MDM2 can polyubiquitylate p53, resulting in its degradation by the proteasome. Alternatively, in settings of lower MDM2 activity, p53 monoubiquitylation leads to p53 nuclear export^55,56. Western blotting revealed the presence of an MDM2-dependent, slowly migrating p53 isoform in tamoxifen-treated SNORA13 knockout BJ-HRAS^G12V cells that matched the size of monoubiquitylated p53 (Figures S6B and S6C). Immunoprecipitation of p53 and immunoblotting with anti-p53 and anti-ubiquitin antibodies provided further evidence that this band represented monoubiquitylated p53 (Figures S6D and S6E). Together, our data suggested that MDM2, which is normally inhibited in the presence of senescence-inducing stress, remained partially active in SNORA13 knockout cells, preventing p53 nuclear accumulation and transactivation of p53 target genes.

Oncogenic stress is known to result in MDM2 inhibition and consequent p53 activation in part by inducing the nucleolar stress response in which free RPs, including RPL5, RPL11 and RPL23, directly bind and inhibit MDM2^{13,14,57–59}. We hypothesized that the acceleration of 60S ribosomal subunit biogenesis that occurs in SNORA13 knockout cells may diminish nucleolar stress by increasing the rate of incorporation of RPs into maturing ribosomes, thereby decreasing the concentration of free RPs that could otherwise inhibit MDM2. In keeping with this model, we observed a marked decrease in the accumulation of non-ribosome associated RPs in SNORA13-deficient BJ-HRAS^G12V cells after induction of oncogenic stress (Figure 4G). Furthermore, immunoprecipitation experiments confirmed the interaction of MDM2 with RPL5, RPL11, and RPL23 in wild-type cells under these conditions (Figure 4H). These interactions were greatly diminished in SNORA13 knockout cells, with a concomitant increase in binding of MDM2 to p53.

These observations suggested that in the presence of oncogenic stress, loss of SNORA13 prevented inhibition of MDM2 by free RPs. As a result, sustained MDM2-mediated inhibition of p53 enabled escape from senescence. We further tested this model by knocking down MDM2 using short-interfering RNA (siRNA) or inhibiting MDM2 with the small molecule Nutlin-3⁶⁰. Both strategies restored nuclear accumulation of p53 and rescued senescence in SNORA13 knockout BJ-HRAS^G12V cells (Figures 4I, 4J, S6B, and S6F). Collectively, these data demonstrated that loss of SNORA13 reduced the levels of free large ribosomal subunit proteins, thus impairing activation of the nucleolar stress response pathway and the p53-mediated senescence program.

Senescence and 18S rRNA modification are genetically separable functions of SNORA13

SNORA13 guides the pseudouridylation of a nucleotide in 18S rRNA, the core constituent of the 40S ribosomal subunit, yet loss of this snoRNA lead to an increase in the biogenesis of the 60S subunit. These observations suggested that SNORA13 may perform a non-canonical function, distinct from guiding pseudouridylation, that regulates 60S assembly. Alternatively, the effects on the large subunit could be a secondary consequence of guiding the modification of an early rRNA precursor, since 18S and 28S rRNA are derived from a common primary transcript. To investigate these possibilities, we performed rescue experiments with SNORA13 mutants that were unable to guide pseudouridylation. Wild-type SNORA13, or mutants with deletion or substitutions in the H and ACA boxes predicted to prevent assembly of the H/ACA box snoRNP required to guide nucleotide modification^61,62, were expressed in SNORA13 knockout cells using an RNA polymerase III promoter (Figures 5A and 5B). The RT-PCR/HinF1 cleavage assay described above (Figure 3B) confirmed that wild-type SNORA13 expressed in this manner appropriately guided pseudouridylation of 18S:1248U, whereas the H/ACA box mutants exhibited no detectable pseudouridylation activity (Figure 5C). Moreover, these mutant RNAs did not associate with the core H/ACA box snoRNP component dyskerin (DKC1), confirming that the mutations abolished snoRNP formation (Figure S6G). Despite their inability to guide the 18S rRNA modification, the SNORA13 mutants rescued senescence in SNORA13 knockout cells (Figures 5D–5F). Moreover, expression of the SNORA13 mutants restored 60S ribosomal subunit abundance to wild-type levels (Figure 5G).

Figure 5. Regulation of ribosome biogenesis and senescence by SNORA13 are genetically separable from pseudouridylation.

(A) Sequences of wild-type and mutant SNORA13 used for rescue experiments.

(B) Northern blots showing expression of wild-type and mutant SNORA13 constructs. U6 snRNA served as a loading control.

(C) HinF1 digestion assay for pseudouridylation of 18S:1248U in SNORA13 knockout cells reconstituted with wild-type or mutant SNORA13. Note that TSR3 was also knocked out in these cells to prevent the addition of acp, thereby enabling the specific detection of 18S:1248Ψ after treatment with CMC.

(D) Growth of tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells reconstituted with wild-type or mutant SNORA13.

(E) Cell morphology, SA-β-gal staining, and EdU incorporation of wild-type or SNORA13 knockout BJ-HRAS^G12V cells with or without rescue with SNORA13 mutants, 15 days after addition of tamoxifen.

(F) Quantification of SA-β-gal activity and EdU incorporation in (E).

(G) Sucrose gradient fractionation showing ribosomal subunit levels in SNORA13 mutant-expressing BJ-HRAS^G12V cells without tamoxifen treatment. Data representative of n=8 biological replicates.

Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing all groups to SNORA13 KO (panels D,F).

***p≤0.001, ****p≤0.0001.

See also Figure S6.

As an additional line of evidence that modification of 18S rRNA:1248U does not regulate senescence, we depleted EMG1 and TSR3, which catalyze N1 methylation and addition of acp to this nucleotide, respectively (Figures 3A and S6H). In contrast to cells lacking SNORA13, cells deficient in these factors entered senescence upon activation of HRAS^G12V expression. (Figure S6I). Overall, these data demonstrated that SNORA13 regulates ribosome biogenesis and senescence through a non-canonical mechanism that is distinct from guiding pseudouridylation.

SNORA13 directly interacts with ribosomal protein RPL23 and inhibits RPL23:28S rRNA interaction

To investigate potential non-canonical functions of SNORA13, the endogenous SNORA13 RNP was purified from UV-crosslinked cells using antisense oligonucleotides (ASOs) and subjected to mass spectrometry to identify interacting proteins⁶³ (Figures 6A, S7A, S7B, and Table S4). Recovery of DKC1 from wild-type cells, but not SNORA13 knockout cells, confirmed the successful retrieval of direct SNORA13 interactors (Figure 6B). Only a small number of significantly enriched candidate interacting proteins were identified, among which was the large ribosomal subunit component RPL23.

Figure 6. SNORA13 directly interacts with ribosomal protein RPL23 and inhibits RPL23:28S rRNA interaction.

(A) Schematic of UV-crosslinking and purification of the endogenous SNORA13 RNP to identify interacting proteins by mass spectrometry.

(B) Volcano plot showing fold enrichment and significance of putative SNORA13 interacting proteins detected by mass spectrometry. H/ACA box snoRNP component DKC1 and RPL23 highlighted in red text.

(C) qRT-PCR analysis of SNORA13 and SNORA25 in UV-RIP samples after pull-down of large and small subunit ribosomal proteins. Enrichment was normalized to input. Data are represented as mean ± SD with individual data points shown (n=3 biological replicates).

(D) Western blots of ribosomal proteins after UV crosslinking and pull-down of SNORA13 with ASOs under denaturing conditions. SNORA13 sense oligonucleotides served as a negative control.

(E) Coomassie stain of purified MBP-RPL23 protein used for in vitro binding experiments.

(F) In vitro binding assays with purified MBP-RPL23 and SNORA13 or SNORA25.

(G) Competitive binding assays with MBP-RPL23:28S rRNA^4389-5070 and increasing concentrations of unlabeled SNORA13. SNORA13 concentration is expressed as fold excess over the concentration of 28S rRNA^4389-5070 in the assay.

See also Figure S7.

UV crosslinking and RNA immunoprecipitation (UV-RIP) confirmed the highly-specific enrichment of SNORA13 upon immunoprecipitation of RPL23, but not other ribosomal proteins (Figures 6C and S7C). Likewise, pull-down of endogenous SNORA13 from UV crosslinked cells under denaturing conditions using ASOs demonstrated specific co-purification of RPL23 (Figure 6D). These results strongly suggested that SNORA13 interacts directly with RPL23, and that this interaction occurs outside the context of assembled ribosomal subunits.

Recombinant RPL23 formed a complex with SNORA13, but not SNORA25, in vitro (Figures 6E and 6F). Similar assays using only the first stem-loop of SNORA13 (nucleotides 1-60) revealed that RPL23 bound to this segment of the snoRNA (Figures S7D and S7E). This portion of SNORA13 contains the highly conserved sequences that base-pair with 18S rRNA to guide pseudouridylation (nucleotides 8-12 and 45-49). Interestingly, we found that these nucleotides were required for RPL23 binding both in vitro and in intact cells (Figures S7F–I). Moreover, a SNORA13 variant with mutations in these nucleotides was not able to rescue senescence in SNORA13 knockout cells nor restore the abundance of 60S subunits to wild-type levels (Figures S7J and S7K). In contrast, the SNORA13 H/ACA box mutants that rescued senescence but could not support pseudouridylation (Figure 5) were able to bind RPL23 in vitro and in cells (Figures S7H and S7I), further supporting the functional importance of this interaction.

We hypothesized that SNORA13 may competitively inhibit association of RPL23 with 28S rRNA, thereby slowing the rate of 60S subunit biogenesis. To test this model, competitive binding assays were performed using the 3’ terminal 682 nucleotide segment of 28S rRNA that extensively interacts with RPL23 in the intact 60S subunit (28S rRNA^4389-5070) (Figure S7L)⁶⁴. The RPL23:28S rRNA^4389-5070 complex formed readily in vitro and, consistent with our model, was disrupted by increasing concentrations of SNORA13 (Figure 6G). Specifically, we found that an approximately 15-fold excess of SNORA13 was required to titrate the majority of RPL23 from 28S rRNA^4389-5070. Copy number analysis in BJ-HRAS^G12V fibroblasts demonstrated that SNORA13 is expressed at approximately 1500 copies per cell (Figure S7M), while 32S pre-rRNA, the pre-rRNA intermediate in pre-60S subunits at the time of RPL23 incorporation⁶⁵, is present at approximately 110 copies per cell (Figure S7N). Thus, the endogenous stoichiometry of SNORA13 and 32S pre-rRNA is consistent with the results of our in vitro competitive titration assays. Based on these findings, we propose that SNORA13 negatively regulates ribosome biogenesis by inhibiting the incorporation of RPL23 into maturing 60S subunits, thereby enhancing activation of the nucleolar stress response in the presence of oncogenic stress.

Murine SNORA13 homologs are required for oncogene-induced senescence in vivo

To determine if SNORA13 is required for senescence in vivo, we identified the mouse homologs of this snoRNA. Interestingly, unlike humans, mice express three annotated SNORA13-like RNAs (GENCODE Gm25636, Gm23639, and Gm55482; Figure 7A). We generated sgRNAs that effectively depleted each SNORA13 homolog (Figure S8A) and examined the resulting effects on ribosomal subunit abundance in mouse embryonic fibroblasts (MEFs). Whereas single or double knockout of each SNORA13 homolog had no effect on the steady-state levels of ribosome subunits (Figures S8B and S8C), depletion of all three snoRNAs resulted in an increase in 60S subunits, as observed in human cells lacking SNORA13 (Figure 7B). These findings suggest that the murine SNORA13 homologs function redundantly to regulate 60S biogenesis.

Figure 7. Mouse SNORA13 homologs are required for oncogene-induced senescence in hepatocytes in vivo.

(A) Alignment of human SNORA13 with mouse homologs (GENCODE M33 annotation Gm25636, Gm23639, and Gm55482). H box (consensus ANANNA) and ACA box highlighted in gray.

(B) Sucrose gradient fractionation of MEFs infected with lentivirus expressing Cas9 and control sgRNA (sgNT) or sgRNAs targeting all three mouse SNORA13 homologs (Triple KO).

(C) Hydrodynamic transfection (HDT) was carried out by tail-vein injection of plasmids encoding a Sleeping Beauty transposon expressing NRAS^G12V, Cas9 and sgRNAs, and Sleeping Beauty transposase (SB100). Livers were analyzed 12 days post-injection. n=3 biological replicates for all tested conditions. Figure created with BioRender.com.

(D) qRT-PCR analysis of snoRNA and Cdkn1a (encoding p21) expression relative to Actb in mouse liver after HDT. p values were calculated by unpaired two-tailed student’s t-test comparing all groups to sgControl. n.s., not significant; **p≤0.01, *p≤0.05.

(E-G) Whole mount SA-β-gal staining (E), western blot analysis (F), and p21 immunohistochemistry and SA-β-gal staining (G) of liver after HDT of the indicated plasmids. See also Figure S8.

We next examined the effects of depleting murine SNORA13 homologs in a well-established in vivo model of oncogene-induced senescence^66,67. Hydrodynamic transfection (HDT) was used to deliver plasmids encoding i) a Sleeping Beauty transposon carrying oncogenic NRAS (NRAS^G12V); ii) Sleeping Beauty transposase; and iii) Cas9 along with control or SNORA13-targeting sgRNAs to mouse livers (Figure 7C). Two distinct pools of sgRNAs targeting all three SNORA13 homologs were used. As an additional control, some mice received a short-hairpin RNA targeting p53 (shp53) in lieu of CRISPR components. Twelve days after plasmid delivery, livers transfected with sgRNAs targeting SNORA13 homologs exhibited an approximately 50% reduction in snoRNA expression, suggesting that loss of function was achieved in a significant fraction of hepatocytes (Figure 7D). At this time-point, we observed robust activation of SA-β-gal activity accompanied by strong p53 and p21 expression in livers transfected with NRAS^G12V plus negative control sgRNAs (Figures 7E–7G). In contrast, depletion of SNORA13 homologs fully abrogated expression of senescence markers and phenocopied the effects of depleting p53. These data provide strong evidence that SNORA13 is required for oncogene-induced senescence in vivo and demonstrate that the role of this snoRNA in senescence is conserved between human and mouse.

DISCUSSION

snoRNAs represent a highly abundant and diverse class of noncoding RNAs that play central roles in ribosome biogenesis. Two major activities are known to be carried out by snoRNAs during this process: i) guiding the installation of nucleotide modifications (pseudouridylation and 2’-O-methylation) in rRNA; and ii) directing specific steps in pre-rRNA processing and folding⁶⁸. Here we discovered a snoRNA that performs a third function in ribosome biogenesis: the regulation of incorporation of a ribosomal protein into maturing ribosomal subunits. We propose that, as a result of its ability to bind directly to RPL23 and slow the rate of an early step in ribosome biogenesis, SNORA13 increases the concentration of non-ribosome associated RPL23 and additional RPs that are incorporated at later steps in 60S subunit assembly. Ultimately, this accumulation of non-ribosome associated RPs enhances activation of the nucleolar stress response, an important mechanism of inducing p53 in the presence of various forms of stress, including aberrant oncogene activation^13,15 (Figure S8D). This model provides a parsimonious link between the primary function of SNORA13 as a negative regulator of 60S ribosomal subunit biogenesis and its role in promoting p53 activation and senescence.

Although SNORA13 potently regulates 60S ribosomal subunit assembly, its target for pseudouridylation is 18S rRNA, the core constituent of the 40S subunit. Both of these functions appear to be deeply conserved in eukaryotes, as deletion of the Saccharomyces cerevisiae ortholog of SNORA13, snR35, results in loss of pseudouridylation of the corresponding nucleotide in 18S rRNA as well as a dramatic increase in the steady-state abundance of 60S subunits³⁶. These observations raise the question of why this dual function of SNORA13 evolved. One possibility is that SNORA13 plays a role in coordinating biogenesis of the 40S and 60S subunits. For example, high levels of 18S rRNA precursors undergoing SNORA13-guided pseudouridylation might sequester this snoRNA, preventing it from interacting with RPL23 and, as a result, enhancing 60S biogenesis. On the other hand, reduced levels of 18S precursors could release a pool of SNORA13 that could delay 60S assembly. In this manner, SNORA13 could balance ribosomal subunit abundance. In keeping with this model, we found that the nucleotides in SNORA13 that base pair with 18S rRNA are also required for RPL23 interaction, suggesting that interaction of SNORA13 with rRNA and RPL23 might be mutually exclusive.

Regulation of SNORA13 abundance could provide a mechanism to tune ribosome production to meet cellular demand under various growth or stress conditions. Indeed, we found that nutrient starvation or mTOR inhibition strongly induced SNORA13 expression, which contributed the inhibition of ribosome biogenesis that occurs under these conditions⁵¹. Additionally, EPB41L4A-AS1, the SNORA13 host gene, was reported to be strongly induced in human fibroblasts upon serum starvation or contact-induced growth arrest, conditions where SNORA13-mediated inhibition of ribosome biogenesis might be advantageous⁶⁹. Beyond SNORA13, these observations also hint at the possibility of a more widespread role for snoRNAs in the regulation of ribosome assembly through interactions with RPs. As a highly abundant and diverse family of noncoding RNAs localized at the site of ribosome assembly, snoRNAs are well-positioned to directly regulate all aspects of ribosome biogenesis through their canonical roles as guides for nucleotide modifications and pre-rRNA processing, their ability to control RP incorporation into maturing subunits as reported here, and perhaps through additional yet-to-be discovered functions.

Cellular senescence is a well-established tumor suppressor mechanism^1,2,5. Given the requirement for SNORA13 in this pathway, downregulation or deletion of this snoRNA might be expected to be a recurrent feature in human malignancies. Interestingly, SNORA13 is located on human chromosome 5q22.1, within 1 MB of the well-known tumor suppressor APC. Accordingly, deletion and loss-of-heterozygosity of this genomic locus is a common feature of many tumor types and decreased expression of EPB41L4A-AS1 is associated with reduced survival in many cancers⁷⁰. These observations raise the possibility that co-deletion of SNORA13 and APC provides a selective advantage in tumors. Intriguingly, loss of the acp modification on 18S:1248m¹acp³Ψ was also recently reported to be a common feature of human tumors⁴¹. Like SNORA13, TSR3, the enzyme that installs acp on this nucleotide, is non-essential and its deletion does not affect bulk translation rates in cell lines^38,41. Ribosome profiling of TSR3-deficient cells, however, revealed an increase in translation of RPs⁴¹. Notably, addition of acp³ at this position does not require prior SNORA13-guided pseudouridylation, likely explaining why loss of SNORA13 does not impact translation of RP-encoding mRNAs. Nevertheless, together with our finding that SNORA13 deficiency increases 60S ribosome assembly, these observations suggest that the hypermodified 18S:1248m¹acp³Ψ nucleotide represents a nexus for pathways that regulate distinct steps in ribosome biogenesis. Coordination of the activity of TSR3 and SNORA13, which would be reflected in the modification status of 18S:1248m¹acp³Ψ, provides a means to congruently orchestrate RP production and RP incorporation into maturing ribosomes. The frequent loss of the 18S:1248m¹acp³Ψ modification in human tumors may result from reduced activity of TSR3 and SNORA13 that is driven by an enhanced demand for ribosomes in rapidly proliferating cancer cells.

As a non-essential gene product whose loss of function increases ribosome biogenesis, SNORA13 may represent an attractive therapeutic target in conditions where increasing ribosome biogenesis would be advantageous. Ribosomopathies, for example, represent a spectrum of diseases caused by haploinsufficiency for ribosomal proteins or loss-of-function of ribosome biogenesis factors. Reduced levels of functional ribosomes, leading to impaired translation of tissue-specific mRNAs, as well as cell type-specific sensitivity to activation of p53 through the nucleolar stress response, are believed to underlie the tissue specificity of ribosomopathy phenotypes^71–73. Since inhibition of SNORA13 has the dual effect of increasing ribosome production while simultaneously reducing nucleolar stress and p53 activation, targeting SNORA13 may be an effective approach for reversing the molecular defects present in these disorders. ASOs that guide RNase-H-dependent cleavage of targeted RNAs have entered clinical use and could be leveraged for the development of specific SNORA13 inhibitors⁷⁴. Thus, further study of SNORA13 promises to illuminate our understanding of the biological roles of snoRNAs in normal physiology and the potential utility of targeting this class of noncoding RNAs in disease.

Limitations of the study

The CRISPRi screen performed in this study targeted a set of 6701 fibroblast-expressed lncRNAs that included few snoRNA host transcripts. It is therefore possible that additional lncRNAs and snoRNAs that regulate ribosome biogenesis and senescence remain to be discovered. Furthermore, we have not yet directly examined the sequestration of RPL23 and the regulation of ribosome biogenesis by SNORA13 in vivo. Thus, additional studies are needed to establish the cell-types and tissues in which this mechanism actively controls ribosome assembly.

STAR METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Joshua T. Mendell (Joshua.Mendell@UTSouthwestern.edu).

Materials availability

All unique reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• High-throughput sequencing data from the CRISPRi screen, RNA-seq, and ribosome profiling have been deposited in GEO and are publicly available as of the date of publication. This study also analyzed existing, publicly available ribosome profiling data from p53-deficient BJ-HRAS^G12V cells⁴⁹. Accession numbers for all datasets are listed in the key resources table. Original northern blot, western blot, and microscopy images will be shared by the lead contact upon request.

• No custom code was generated in this study.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-Ras	BD Bioscences	Cat#610001; RRID:AB_397424
Anti-NRAS	Proteintech	Cat#10724-1-AP; RRID:AB_2923963
Anti-GAPDH	Cell Signaling Technology	Cat#2118; RRID:AB_561053
Anti-Puromycin	Sigma	Cat#MABE343; RRID:AB_2566826
Anti-p21 Waf1/Cip1	Cell Signaling Technology	Cat#2946; RRID:AB_2260325
Anti-Mouse p21	BD Biosciences	Cat#556431; RRID:AB_396415
Anti-p53	Santa Cruz	Cat#sc-126; RRID:AB_628082
Anti-p16	Abcam	Cat#ab108349; RRID:AB_10858268
Anti-gH2AX	Cell Signaling Technology	Cat#9718; RRID:AB_2118009
Anti-Ubiquitin	Cell Signaling Technology	Cat#58395; RRID:AB_3075532
Anti-Dyskerin	Invitrogen	Cat#PA5-28922; RRID:AB_2546398
Anti-hnRNP C1/C2	Santa Cruz	Cat#sc-32308; RRID:AB_627731
Anti-RPL5	Bethyl Laboratories	Cat#A303-933A; RRID:AB_2620282
Anti-RPL11	Proteintech	Cat#16277-1-AP; RRID:AB_2181292
Anti-RPL23	Bethyl Laboratories	Cat#A305-010A; RRID:AB_2621204
Anti-α-Tubulin	Sigma	Cat#T5168; RRID:AB_477579
Anti-MDM2	Sigma	Cat#OP115; RRID:AB_564806
Anti-RPL8	Bethyl Laboratories	Cat#A305-059A; RRID:AB_2631454
Anti-RPL37A	Proteintech	Cat#14660-1-AP; RRID:AB_2238668
Anti-RPS5	Bethyl Laboratories	Cat#A304-011A; RRID:AB_2620359
Anti-RPS14	Bethyl Laboratories	Cat#A304-031A; RRID:AB_2621280
Anti-Nucleolin	Abcam	Cat#ab22758; RRID:AB_776878
Anti-Digoxigenin	Roche	Cat#11333062910; RRID:AB_2313639
Normal mouse IgG	Santa Cruz	Cat#sc-2025; RRID:AB_737182
Normal rabbit IgG	Cell Signaling Technology	Cat#2729; RRID:AB_1031062
Anti-Mouse secondary for Western blotting	Licor	Cat#926-68072; RRID:AB_10953628
Anti-Rabbit secondary for Western blotting	Licor	Cat#926-32213; RRID:AB_621848
Anti-Mouse secondary for FISH	Millipore	Cat#AP124C; RRID:AB_92459
Anti-Rabbit secondary for FISH	Invitrogen	Cat#A-21206; RRID:AB_2535792
Anti-Mouse secondary for Immunofluorescence	Invitrogen	Cat#A-11005; RRID:AB_2534073
Bacterial and virus strains
Endura electrocompetent cells	Lucigen	Cat#60242-2
E.coli Stbl3	Invitrogen	Cat#C7373-03
DH10Bac Competent Cells	Invitrogen	Cat#10361012
Biological samples
Chemicals, peptides, and recombinant proteins
Dulbecco’s Modified Eagle’s Medium (DMEM)	Invitrogen	Cat#11995-073
Medium 199	Gibco	Cat#11150067
Eagle’s Minimum Essential Medium (EMEM)	ATCC	Cat#30-2003
Fetal Bovine Serum (FBS)	Sigma	Cat#F2442
Amino acid free Dulbecco’s modified Eagle medium powder	USBiological	Cat#D9800-13
Dialyzed FBS	ThermoFisher Scientific	Cat#A3382001
Antibiotic-antimycotic	Gibco	Cat#15240112
FuGENE HD Transfection Reagent	Promega Corporation	Cat#E2311
Polybrene	Millipore	Cat#TR-1003-G
Puromycin	ThermoFisher Scientific	Cat#A11138-03
(Z)-4-Hydroxytamoxifen	Sigma	Cat#H7904
TRIzol	Invitrogen	Cat#15596026
NuPAGE LDS Sample Buffer (4X)	Invitrogen	Cat#NP0007
Albumin, Bovine Fraction V (BSA)	Fisher	Cat#A30075-100.0
Lipofectamine 3000 Transfection Reagent	Invitrogen	Cat#L3000015
Hygromycin B	Invitrogen	Cat#10687010
ULTRAhyb-Oligo	Invitrogen	Cat#AM8663
N-Cyclohexyl-N’-(2-morpholinoethyl)carbodiimide methyl-p-toluenesulfonate	Sigma	Cat#C106402
Cycloheximide	Sigma	Cat#C1988
Protease Inhibitor Cocktail Set III, EDTA-Free	Millipore	Cat#539134
RNasin Ribonuclease Inhibitors	Promega	Cat#N2515
SlowFade Diamond Antifade Mountant with DAPI	Invitrogen	Cat#S36964
SNAP-Cell Block	New England Biolabs	Cat#S9106S
SNAP-Cell TMR-Star	New England Biolabs	Cat#S9105S
SNAP-Cell Oregon Green	New England Biolabs	Cat#S9104S
Poly-L-lysine	Sigma	Cat#P4832
Lipofectamine RNAiMax	Invitrogen	Cat#13778150
Proteinase K	New England Biolabs	Cat#P8107S
Cellfectin II Reagent	Gibco	Cat#10362100
Sf-900 II SFM	Gibco	Cat#10902096
Bio-11-UTP	Invitrogen	Cat#AM8450
RNase A/T1 Mix	Thermo Scientific	Cat#EN0551
IRDye 800CW Streptavidin	Licor	Cat#92632230
DAPI	Invitrogen	Cat#D1306
Rapamycin	Sigma	Cat#553210
EDTA	Invitrogen	Cat#AM9260G
Etoposide	Sigma	Cat#341205
Critical commercial assays
NEBuilder HiFi DNA Assembly Master Mix	New England Biolabs	Cat#E2621S
MasterPureTM Complete DNA Purification Kit	Lucigen	Cat#MC85200
Herculase II Fusion DNA Polymerase	Agilent	Cat#600679
Agencourt AMPure XP beads	Beckman Coulter Life Sciences	Cat#A63882
Phusion High-Fidelity DNA Polymerase	New England Biolabs	Cat#M0530S
Qubit dsDNA HS Assay Kit	ThermoFisher Scientific	Cat#Q32851
miRNeasy mini	QIAGEN	Cat#217004
PrimeScript RT Master Mix	TaKaRa	Cat#RR036A
SYBR Green PCR Master Mix	Applied Biosystems	Cat#4309155
Senescence β-Galactosidase Staining Kit	Cell Signaling Technology	Cat#9860
Click-iT Plus EdU Cell Proliferation Kit for Imaging	Invitrogen	Cat#C10639
Superscript IV Reverse Transcriptase	Invitrogen	Cat#18090010
AMV Reverse Transcriptase	New England Biolabs	Cat#M0277S
Protein Synthesis Assay Kit	Cayman Chemical	Cat#601100
TURBO DNA-free kit	Invitrogen	Cat#AM1907
DIG RNA labeling mix	Roche	Cat#11277073910
Click-iT Nascent RNA Capture Kit	Invitrogen	Cat#C10365
Superscript VILO cDNA Synthesis Kit	Invitrogen	Cat#11754050
Dual-Luciferase Reporter Assay System	Promega	Cat#E1980
Dynabeads Protein G	Invitrogen	Cat#10004D
Dynabeads MyOne Streptavidin T1	Invitrogen	Cat#65601
Amylose Resin	New England Biolabs	Cat#E8021S
HiScribe T7 High Yield RNA Synthesis Kit	New England Biolabs	Cat#E2040S
Deposited data
CRISPRi screen	This paper	GSE232859
Ribosome profiling	This paper	GSE232859
RNA-sequencing	This paper	GSE232859
Ribosome profiling and RNA-sequencing from p53-deficient BJ-HRASG12V cells	Loayza-Puch et al.49	GSE45833
Experimental models: Cell lines
BJ-HRASG12V	Laboratory of Reuven Agami (Voorhoeve et al.26)	N/A
BJ-HRASG12V-dCas9KRAB cells	This paper	N/A
BJ-HRASG12V SNORA13 knockout cells	This paper	N/A
HEK293T	ATCC	CRL-3216; RRID:CVCL_0063
HEK293T-RPL28-SNAP	Laboratory of Michael Buszczak (Ni et al.50)	N/A
Sf9 cells	Gibco	Cat#11496015
CCD-1070Sk	ATCC	CRL-2091; RRID:CVCL_2332
IMR-90	ATCC	CCL-186; RRID:CVCL_0347
MEF	Laboratory of Joshua T. Mendell (Kopp et al.76)	N/A
Experimental models: Organisms/strains
FVB/NJ mice	The Jackson Laboratory	Cat#001800; RRID:IMSR_JAX:001800
Oligonucleotides
See Table S5	N/A	N/A
Recombinant DNA
pHR-SFFV-KRAB-dCas9-P2A-mCherry	Gilbert et al.77	Addgene Plasmid #60954; RRID:Addgene_60954
CRISPRi non-coding library (CRiNCL) - Common to all 7 cell lines	Liu et al.28	Addgene Plasmid #86538; RRID:Addgene_86538
CRISPRi non-coding library (CRiNCL) - iPSC & HFF	Liu et al.28	Addgene Plasmid #86550; RRID:Addgene_86550
CRISPRi non-coding library (CRiNCL) - Unique to HFF	Liu et al.28	Addgene Plasmid #86541; RRID:Addgene_86541
pCRISPRia-v2	Horlbeck et al.79	Addgene Plasmid #84832; RRID:Addgene_84832
lentiCRISPR v2	Sanjana et al.81	Addgene Plasmid #52961; RRID:Addgene_52961
lentiCRISPR v2-Blast	Laboratory of Mohan Babu	Addgene Plasmid #83480; RRID:Addgene_83480
LcV2-Hygro	Golden et al.80	Addgene Plasmid #91977; RRID:Addgene_91977
pT/Caggs-NRASV12	Wiesner et al.87	Addgene Plasmid #20205; RRID:Addgene_20205
pT2/shp53/GFP4	Wiesner et al.87	Addgene Plasmid #20208; RRID:Addgene_20208
pCMV-SB100	This paper	N/A
pX333	Maddalo et al.86	Addgene Plasmid #64073; RRID:Addgene_64073
pSpCas9(BB)-2A-GFP (PX458)	Ran et al.82	Addgene Plasmid #48138; RRID:Addgene_48138
XLone-GFP	Randolph et al.83	Addgene Plasmid #96930; RRID:Addgene_96930
XLone-EPB41L4A-AS1 with SNORA13	This paper	N/A
XLone-EPB41L4A-AS1 without SNORA13	This paper	N/A
XLone-EPB41L4A-AS1 with SNORA13 Mut(8–12,45–49)	This paper	N/A
pCMV-hyPBase	Laboratory of Nancy L. Craig (Yusa et al.84)	N/A
PLKO.1-Scrambled	Fischer et al.85	Addgene Plasmid #136035; RRID:Addgene_136035
PLKO.1-SNORA13 WT	This paper	N/A
PLKO.1-SNORA13 H/ACA Mut	This paper	N/A
PLKO.1-SNORA13 H/ACA Del	This paper	N/A
MTTH-SNAP	Laboratory of Michael K. Rosen (Lin et al.95)	N/A
pFastBac Dual	Gibco	Cat#10712024
pGL3-Renilla-Ter-Firefly	Laboratory of Joshua T. Mendell (Zhu et al.78)	N/A
pGL3-Renilla-noTer-Firefly	Laboratory of Joshua T. Mendell (Zhu et al.78)	N/A
pcDNA3 Renilla-FLAG-firefly	This paper	N/A
pcDNA3 Renilla-FLAG-firefly K529E	This paper	N/A
psPAX2	Laboratory of Didier Trono	Addgene Plasmid #12260; RRID:Addgene_12260
pMD2.G	Laboratory of Didier Trono	Addgene Plasmid #12259; RRID:Addgene_12259
Software and algorithms
MAGeCK	Li et al.29	N/A
STAR	Dobin et al.97	N/A
RiboDiff	Zhong et al.105	N/A
GraphPad Prism	GraphPad	N/A
FlowJo	FlowJo	N/A
Other

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell lines

BJ-HRAS^G12V cells (male) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Invitrogen) supplemented with 20% Medium 199 (Gibco), 10% fetal bovine serum (Sigma-Aldrich), and 1X Antibiotic-Antimycotic (Gibco). HEK293T (female) and HEK293T-RPL28-SNAP (female) cells were cultured in DMEM medium supplemented with 10% fetal bovine serum and 1X Antibiotic-Antimycotic. CCD-1070Sk (male) and IMR-90 (female) cells were cultured in Eagle’s Minimum Essential Medium (EMEM) (ATCC) supplemented with 10% fetal bovine serum and 1X Antibiotic-Antimycotic. Mouse embryonic fibroblasts (MEFs) (male and female) were generated and cultured as previously described⁷⁶. All cells were grown at 37°C with 5% CO₂. BJ-RAS^G12V cells were a gift from Reuven Agami. HEK293T, CCD-1070Sk, and IMR-90 cells were obtained from ATCC. Cell lines were confirmed to be free of mycoplasma contamination.

Animals

All mice were handled in accordance with and with the approval of the Institutional Animal Care and Use Committee (IACUC) at UT Southwestern. Male 6-week old FVB/NJ mice were obtained from the Jackson Laboratory and randomly assigned to experimental groups. Mice were housed in groups in the UT Southwestern barrier facility.

METHOD DETAILS

Generation of BJ-HRAS^G12V-dCas9^KRAB cells

BJ-HRAS^G12V-dCas9^KRAB cells were generated by infecting BJ-HRAS^G12V cells with a dCas9^KRAB-mCherry lentivirus (Addgene #60954)⁷⁷ and sorting for single cell clones stably expressing high mCherry. Two independent clones were chosen for screening.

Genome-wide CRISPRi screening

CRISPRi-based genome-scale screening of functional noncoding RNAs was performed using a previously described library.²⁸ A lentiviral library with 68,009 guides targeting 6701 noncoding RNAs expressed in human fibroblasts was assembled by combining CRiNCL sublibraries Addgene #86538, Addgene #86550, and Addgene #86541. The screen was performed in replicate using two independent BJ-HRAS^G12V-dCas9^KRAB cell lines. Previously reported methods⁷⁸ with the following modifications were used. For each replicate, 8x10⁷ cells were seeded into forty 15 cm dishes in medium containing 8 μg/mL of polybrene (Millipore) and lentiviral library at a multiplicity of infection of ~0.3. Cells were cultured for 2 days after infection, then treated for 4 days with 2 μg/mL puromycin, and then allowed to recover for 2 days. At least 2.1x10⁷ cells were then plated for each replicate, representing ~300x coverage of the library, and treatment with 100 nM (Z)-4-Hydroxytamoxifen (Sigma) was initiated. This timepoint was designated “day 0”. Cells were maintained in continuous tamoxifen for 21 days, with passaging every 3 days, each time plating at least 2.1x10⁷ cells to maintain a minimum of 300x library coverage. Cells were harvested at day 0 and day 21 for genomic DNA isolation.

Genomic DNA was extracted using the MasterPure Complete DNA Purification kit (Lucigen). Sequencing libraries were generated with two sequential rounds of PCR using Herculase II Fusion DNA polymerase (Agilent). 6.6 μg of genomic DNA (representing ~10⁶ cells) was used in each 100 μl reaction for the first round of PCR. A total of 20 first-round PCR reactions were performed per replicate with 18 cycles of amplification (primer sequences provided in Table S5). All first-round PCR reactions from each replicate were then pooled. 5 μl from the pooled reaction product and primers containing barcodes and Illumina sequencing adaptors (sequences provided in Table S5) were used in the second round of PCR (11–13 cycles). Agencourt AMPure XP beads (Beckman Coulter Life Sciences) were used to purify PCR products. The amplicons were sequenced on an Illumina NextSeq500 with 75 bp single-end reads. Approximately 2.3x10⁷ reads were obtained from each replicate. sgRNA sequences were extracted from fastq files through an in-house Galaxy script and normalized read counts were calculated. MAGeCK analysis²⁹ was used to identify genes targeted by sgRNAs enriched in the day 21 versus day 0 populations.

CRISPRi-mediated gene knockdown

sgRNAs targeting genes of interest or control sgRNAs (sequences provided in Table S5) were cloned into pCRISPRia-v2 (Addgene #84832) as previously described.⁷⁹ For validation of hits in the CRISPRi screen, BJ-HRAS^G12V-dCas9^KRAB cells transduced with the pCRISPRia-v2 vectors were selected in medium containing 2 μg/mL puromycin for 4 days. Cells were then allowed to recover for 2 days before tamoxifen treatment to induce HRAS^G12V expression and subsequent phenotypic analyses.

CRISPR-Cas9 mediated gene knockout

sgRNAs targeting genes of interest or control sgRNAs (sequences provided in Table S5) were cloned into lentiCRISPR_v2 (Addgene #52961), lentiCRISPR v2-Blast (Addgene #83480), or LcV2-Hygro (Addgene #91977)⁸⁰ as previously described.⁸¹ Knockout pools were generated by transducing cells with lentiCRISPR_v2 vectors. Cells were selected in medium containing 1-2 μg/mL puromycin, 2 μg/mL Blasticidin, or 100 μg/mL hygromycin for 7 days before further analysis. To generate clonal SNORA13 KO cell lines, BJ-HRAS^G12V cells were transfected with pX458 (Addgene #48138)⁸² expressing Cas9, GFP, and sgRNAs targeting the genomic sequence flanking SNORA13. 48 hours after transfection, the brightest 10% of cells were collected using FACS. Single cell clones were then expanded and screened for deletion at the targeted locus. For generating MEFs with knockout of SNORA13 homologs, single knockout pools were generated using lentiCRISPR_v2. Subsequent infection with LcV2-Hygro was used to generate double knockout pools and further infection with lentiCRISPR_v2-Blast produced triple knockout pools.

Cloning and expression of rescue constructs

PiggyBac expression vector XLone-GFP (Addgene #96930)⁸³ was modified by replacing the blasticidin resistance gene with a hygromycin resistance gene. Unspliced EPB41L4A-AS1 with or without SNORA13 in the intron was then inserted into this vector between the KpnI and SpeI sites using NEBuilder HiFi DNA Assembly Master Mix (NEB) according to the manufacturer’s instructions. EPB41L4A-AS1 was amplified from genomic DNA as two separate fragments with overhangs for HiFi cloning. Stable expression of EPB41L4A-AS1 with or without SNORA13 was achieved by transfecting WT or SNORA13 KO BJ-HRAS^G12V cells with XLone-EPB41L4A-AS1 (+SNORA13) or XLone-EPB41L4A-AS1 (−SNORA13) plasmids together with pCMV-hyPBase plasmid⁸⁴ which encodes the PiggyBac transposase. Transfected cells were selected in 100 μg/mL hygromycin and expanded for 1 month before analysis.

For expression of SNORA13 with mutations in the H/ACA box, a lentiviral pol III expression vector was used. DNA oligos corresponding to the sequence of WT or mutant SNORA13 was cloned into AgeI- and EcoRI-digested pLKO.1-scrambled (Addgene #136035)⁸⁵ using NEBuilder HiFi assembly. WT or SNORA13 KO BJ-HRAS^G12V cells were transduced with pLKO.1-SNORA13 (WT or mutant) lentivirus followed by selection in 2 μg/mL puromycin for 7 days. Clonal cell lines were then derived and screened for SNORA13 expression by northern blot. Sequences of all primers used for cloning are provided in Table S5.

Hydrodynamic transfection of mouse liver

To generate sgRNA pools targeting mouse SNORA13 homologs, pairs of sgRNAs were sequentially cloned into pX333 (Addgene #64073)⁸⁶. Each SNORA13 homolog targeting pool consisted of: i) a pX333 plasmid expressing two sgRNAs targeting Gm23639 and ii) a pX333 plasmid expressing a guide targeting Gm25636 and a guide targeting Gm55482. pX333 expressing sgRNAs targeting GFP and Gal4 served as a negative control (sgControl). sgRNA sequences are provided in Table S5. Male FVB/NJ mice were injected at 7 weeks of age with pT/Caggs-NRASV12 (Addgene #20205)⁸⁷, SB100 transposase plasmid (pCMV-SB100), and either 1) sgControl plasmid (sgGFP/sgGal4), 2) SNORA13 homolog targeting pool 1 plasmids, 3) SNORA13 homolog targeting pool 2 plasmids, or 4) a Sleeping Beauty plasmid expressing short-hairpin RNA targeting p53 (pT2/shp53/GFP4; Addgene #20208)⁸⁷. Plasmids were resuspended in 2 mL of saline and administered via tail vein injection in 7 seconds. A 10:1 mass ratio of transposon plasmids to transposase plasmid was used (10 μg pT/Caggs-NRASV12 + 10 μg pX333 or pT2/shp53/GFP4 + 2 μg pCMV-SB100). Injected mice were sacrificed at 12 days after HDT for analysis.

Immunohistochemistry

Immunohistochemistry was performed on paraffin sections using standard protocols. Isolated livers were fixed in 10% formalin before paraffin embedding and sectioning at the UTSW Histopathology core. Expression of p21 was detected using a mouse monoclonal antibody (BD Biosciences, 556430, 1:50 dilution). Images were taken on a Zeiss AxioObserver Z1 with a 40X objective.

RNA extraction and qRT-PCR

Total RNA was isolated from cells or liver tissues using Trizol (Invitrogen) and further purified using the miRNeasy Mini kit (Qiagen) with on-column DNase digestion. cDNA was generated from reverse transcription of 1 μg total RNA using the Primescript RT Master Mix (Takara) according to the manufacturer’s instructions. SYBR Green PCR master Mix (Applied Biosystems) was used for qPCR. Expression levels were normalized to GAPDH. Primer sequences used for qRT-PCR are provided in Table S5.

Northern blot analysis

For snoRNA northern blots, total RNA (10-15 μg) was separated on 8% TBE-Urea polyacrylamide gels and transferred to BrightStar-Plus nylon membranes (Invitrogen). For pre-rRNA northern blots, total RNA (10-15 μg) was separated on 1.2% formaldehyde-agarose gels and transferred using the iBlot Dry Blotting System (Invitrogen). Membranes were crosslinked using a 254 nm UV crosslinker at 120 mJ/cm² and prehybridized with ULTRAhyb-Oligo hybridization buffer (Invitrogen). Blots were hybridized overnight with near-infrared (IR) fluorescent probes (IDT) and imaged using a LI-COR Odyssey imager. U6 snRNA or ACTB was used as a loading control. For SNORA13 and 32S pre-rRNA copy number analysis, total RNA from a defined number of cells was run alongside a standard curve of in vitro transcribed SNORA13 or a fragment of 32S pre-rRNA containing the probe sequence, respectively. In vitro transcription was performed with the HiScribe T7 High Yield RNA Synthesis Kit (NEB). Probe sequences are provided in Table S5.

Cell proliferation assay

BJ-HRAS^G12V cells were seeded at a density of 6x10⁵ cells/10 cm dish in 10 mL medium containing vehicle or 100 nM (Z)-4-Hydroxytamoxifen (Sigma). BJ-HRAS^G12V cells were passaged every 3 days and re-seeded at 6x10⁵ cells/10 cm dish after each passage. Lentivirally-infected CCD-1070Sk and IMR-90 cells were seeded at a density of 6x10⁵ cells/10 cm dish in 10 mL medium containing 1 μg/mL puromycin. Cells were passaged every 2 days and re-seeded at 6x10⁵ cells/10 cm dish after each passage. Cell number at each passage was determined using a Countess cell counter (Invitrogen) and cumulative cell numbers were plotted. All growth curves were performed with 3 biological replicates.

Senescence-associated β-galactosidase stain

SA-β-gal activity in cells was assayed 15 days after induction of HRAS^G12V, 4 days after etoposide treatment for DNA damage induced senescence, and at the indicated passage numbers for replicative senescence using a SA-β-gal staining kit (Cell Signaling) according to the manufacturer’s instructions. Briefly, cells were seeded in 4-well chamber slides (50,000 cells/well). After 48 hours, cells were fixed and incubated with β-galactosidase staining solution at pH 6.0, 37°C, overnight. The cells were then washed twice with 1× PBS and imaged. Detection of S-β-gal activity in dissected mouse liver lobes or frozen sections was carried out as described previously⁸⁸. Samples were fixed in 0.5% glutaraldehyde in PBS for 15 min at RT, followed by incubation in β-gal staining buffer (pH 5.5) for 8 hours.

EdU incorporation assay

EdU incorporation assays were performed using the Click-iT Plus EdU Alexa Fluor 594 Imaging Kit (Invitrogen) according to the manufacturer’s instructions. After 15 days of tamoxifen treatment, BJ cells were seeded in 4-well chamber slides (50,000 cells/well) and cultured for 48 hours. Half of the medium was then replaced with fresh medium containing 20 μM EdU and cells were incubated for 3 hours. Cells were then fixed with formaldehyde and permeabilized with 0.5% Triton X-100. A reaction cocktail and Hoechst were used to detect EdU and nuclei. Cells were imaged on a Zeiss LSM700 microscope.

Western blotting

Cells were lysed in 1× NuPAGE LDS Sample Buffer (Invitrogen) for 10 minutes at 70°C. Liver tissue was lysed in Laemmli buffer. Proteins were separated on 4-12% Bis-Tris NuPAGE gels (Thermo Fisher Scientific) and wet transferred onto nitrocellulose membranes (0.45 μm, ThermoFisher). Blocking was performed in TBST containing 5% non-fat milk for 1 hour at room temperature. Membranes were incubated with primary antibodies overnight at 4°C, followed by incubation with IR dye-labeled secondary antibodies for 1 hour at 4 °C. A LI-COR Odyssey imager was used for imaging blots. All antibodies are listed in the key resources table.

Detection of 18S:1248m¹acp³Ψ

Detection of 18S:1248m¹acp³Ψ by RT-PCR/HinFI cleavage was performed as previously described⁴¹ with minor modifications. Briefly, total RNA was treated with freshly made 0.2 M CMC in BEU buffer (50 mM bicine, pH 8.3, 4 mM EDTA, and 7 M urea) or BEU buffer only. Reverse transcription was carried out with 200 ng CMC-treated or mock-treated RNA using Superscript IV reverse transcriptase (Invitrogen). cDNA was diluted and used as template for PCR. Amplicons were digested with HinFI and separated on a 2% agarose gel. Primer sequences are provided in Table S5.

Primer extension was performed with AMV reverse transcriptase (NEB). 1 μg CMC-treated or mock-treated total RNA and 2 pmol IR labeled primer were mixed in the supplied AMV reverse transcriptase buffer and denatured at 65°C for 5 minutes. RT reactions were incubated at 42°C for 1 hour. Reaction products were separated on a 12.5% polyacrylamide gel at 4°C and visualized with a LI-COR Odyssey imager. Primer sequence provided in Table S5.

Analysis of global translation rate

The SUnSET method was used as previously described⁴⁵ to monitor global translation rates. Briefly, 8x10⁵ cells were seeded in a 10 cm dish 48 hours before adding media containing 10 μg/mL puromycin. Cells were collected 1 hour later. Puromycin incorporated into nascent peptides was detected by western blotting with anti-puromycin antibody (Sigma).

The O-propargyl-puromycin (OP-puro) incorporation assay was performed using the Protein Synthesis Assay Kit (Cayman Chemical) according to the manufacturer’s instructions. For each sample, 1x10⁶ cells were resuspended in 0.5 mL of OP-puro working solution for 1 hour at 37°C. Cells were then stained with 1 mL of 5 FAM-Azide staining solution in the dark at room temperature for 30 minutes followed by flow cytometry analysis.

RPL28-SNAP labeling and pulse-chase analysis

A previously described HEK293T cell line expressing an endogenous RPL28-SNAP fusion protein⁵⁰ was used. Cells were incubated with SNAP-Cell Block (NEB) at a final concentration of 10 μM in growth medium for 30 minutes to block pre-existing ribosomes. Cells were then washed twice with complete medium and cultured for 12 hours (for imaging studies) or 18 hours (for polysome analyses). To label newly synthesized ribosomes, cells were incubated with 3 μM SNAP-Cell TMR-Star (NEB) for imaging or SNAP-Cell Oregon Green (NEB) for polysome fractionation in growth medium for 30 minutes. Cells were then washed three times with complete medium and incubated in fresh medium for 30 minutes. The medium was then replaced one more time to remove unreacted SNAP-tag substrate that diffused out of the cells. For microscopy studies, cells were imaged at 0 and 1 hour after the final medium change on a Zeiss LSM980 confocal microscope. For polysome fractionation, cells were incubated with 100 μg/mL cycloheximide (Sigma) immediately after final medium change for 5 minutes and then collected.

Polysome fractionation

Polysome fractionation was performed as described⁸⁹ with the following modifications. 1.5x10⁷ cells per sample were seeded 24 hours before lysis. Cells were incubated with 100 μg/mL cycloheximide (Sigma) in growth medium for 5 minutes and then washed twice with ice-cold 1x PBS containing 100 μg/mL cycloheximide. Cells were scraped in 5 mL of ice-cold 1x PBS containing 100 μg/mL cycloheximide and collected by centrifugation. Cell pellets were lysed in hypotonic buffer (5 mM Tris-HCl, pH 7.5, 2.5 mM MgCl₂, 1.5 mM KCl, 1x EDTA-free protease inhibitor cocktail, 100 μg/mL cycloheximide, 2 mM DTT, 200 U/mL RNAse inhibitor, 0.5% Triton X-100, and 0.5% sodium deoxycholate). Lysates were centrifuged at 16,000 g for 7 minutes at 4°C. Supernatants (~500 μL) were collected, diluted with lysis buffer as needed so all samples had an equivalent OD₂₆₀, and then loaded onto a 5–50% sucrose gradient. Gradients were ultra-centrifuged at 222,228 g (36,000 rpm) for 2 hours at 4°C using a TH-641 rotor (Thermo Scientific). Samples were fractionated on a Piston Fractionator (BioComp) to evaluate polysome profiles. For total ribosome analysis, UV absorbance at 260 nm was used. For analysis of newly synthesized ribosomes in HEK293T RPL28-SNAP cells, a fluorescence detector was used. To dissociate ribosomal subunits (Figures S5C and S5G), 50 mM EDTA was added to cell lysates and sucrose gradient. For amino acid starvation experiments, HEK293T RPL28-SNAP cells were cultured in amino acid free DMEM media (USBiological) supplemented with 10% dialyzed fetal bovine serum (Thermo Scientific) and 1X Antibiotic-Antimycotic for 3 days. Quantification was performed by baselining the gradients and calculating the area under the curve for the 60S subunit. For detection of ribosome-associated and ribosome-free ribosomal proteins, proteins were concentrated from sucrose gradient fractions corresponding to 60S ribosome fractions and ribosome-free fractions using trichloroacetic acid (TCA) precipitation and resuspended in 1× NuPAGE LDS Sample Buffer (Invitrogen) for western blotting.

Subcellular fractionation

Cell pellets were resuspended in 1 mL ice-cold Buffer A (10 mM HEPES, pH 8.0, 1.5 mM MgCl₂, 10 mM KCl, and 0.5 mM DTT) and incubated on ice for 5 minutes, followed by lysis with a Dounce homogenizer. After centrifugation at 1000 rpm for 5 minutes, the supernatant was collected as the cytoplasmic fraction. Pellets were resuspended in 750 μL Buffer S1 (250 mM sucrose and 10 mM MgCl₂) and overlaid onto a 1 mL cushion of Buffer S2 (350 mM sucrose and 0.5 mM MgCl₂), followed by centrifugation at 2500 rpm for 10 minutes. The supernatant was discarded and pellets (nuclei) were resuspended in 1 mL Buffer S2 and sonicated for 60 seconds total, with 10 second pulses. Sonicated samples were overlaid onto 1 mL Buffer S3 (880 mM sucrose and 0.5 mM MgCl₂) and centrifuged at 13,000 rpm for 20 minutes. The supernatant was collected as the nucleoplasmic fraction and the pellet contained the nucleolar fraction. 1x Protease Inhibitor Cocktail and 200 U/mL RNase inhibitor were added freshly to all buffers before use. All centrifugations were performed at 4°C. RNA was extracted from each fraction using Trizol in combination with the miRNeasy mini kit (Qiagen) with an on-column DNase digestion. For western blotting, sonicated samples (nuclei) were used without further fractionation.

RNA FISH

RNA FISH was performed as previously described,⁹⁰ with minor modifications. DIG-labeled anti-sense RNA probes for human SNORA13 were synthesized by in vitro transcription with a DIG-labeling mix (Roche) and purified with Micro Bio-Spin P-30 chromatography columns (Bio-Rad). Primers used for amplification of the DNA template for in vitro transcription are provided in Table S5. 2x10⁵ cells were grown on poly-L-lysine (Sigma) coated coverslips in 6-well plates for 48 hours. Cells were washed twice with DEPC-treated PBS and fixed with 4% paraformaldehyde for 10 minutes. Cells were then washed once with DEPC-treated PBS and permeabilized with 0.5% Triton X-100 for 10 minutes. After washing twice with DEPC-treated PBS, cells were incubated with pre-hybridization buffer (50% formamide, 2X SSC, 1X Denhardt’s solution, 10 mM EDTA, 0.1 mg/mL yeast tRNA, 0.01% Tween-20) for 1 hour. 10 ng/μL DIG-labeled RNA probes in hybridization buffer (prehybridization buffer with 5% dextran sulfate) were denatured at 75°C for 10 minutes and added to slides for hybridization at 37°C for 18 hours. Samples were washed, treated with RNase A, and blocked with blocking reagent (Roche) after hybridization. DIG-labeled probes were detected by incubation with mouse monoclonal anti-DIG primary antibody (Roche) followed by incubation with Cy3-labeled goat anti-mouse IgG secondary antibody (EMD Millipore). The nucleolus was labeled using anti-nucleolin antibody (Abcam). Samples were mounted with SlowFade Diamond Antifade Mountant with DAPI (Thermo Fisher Scientific) and imaged on a Zeiss LSM700 microscope.

Measurement of nascent rRNA

Nascent RNA was pulse labeled for 30 minutes with media containing 0.5 mM 5-ethynyluridine (EU). Total RNA was isolated and labeled RNA was purified using the Click-iT Nascent RNA Capture kit (Invitrogen) according to the manufacturer’s instructions. Superscript VILO cDNA synthesis kit (Invitrogen) was used for cDNA synthesis and qRT-PCR was performed (primer sequences provided in Table S5).

Translation fidelity luciferase assays

The dual luciferase assay for measuring stop codon readthrough or amino acid misincorporation was performed as previously described.⁷⁸ Briefly, pGL3-Renilla-Ter-firefly or pGL3-Renilla-noTer-firefly plasmids were generated by cloning a Renilla luciferase ORF with or without a TAA termination codon upstream of a firefly luciferase ORF in pGL3-Control (Promega). pcDNA3-Renilla-FLAG-firefly was generated by replacing the IRES sequence in pcDNA3 RLUC POLIRES FLUC (Addgene #45642)⁹¹ with a Flag-encoding sequence. pcDNA3-Renilla-FLAG-firefly K529E (AAA to GAA) was then generated by site-directed mutagenesis. Untreated or tamoxifen-treated BJ-HRAS^G12V cells (WT or SNORA13 KO) were seeded into 6-well plates (1.3x10⁵ cells/well). After 24 hours, cells were transfected with 2.5 μg of reporter plasmids using Lipofectamine 3000 transfection reagent (Invitrogen). 72 hours after transfection, the Dual-Luciferase Reporter Assay System (Promega) was used to determine luciferase activity. For normalization of stop codon readthrough, firefly activity was first normalized to Renilla activity in each sample and then firefly/Renilla activity in pGL3-Renilla-Ter-firefly transfected cells was normalized to firefly/Renilla activity in pGL3-Renilla-noTer-firefly transfected cells. For normalization of amino acid misincorporation, firefly activity was first normalized to Renilla activity in each sample and then firefly/Renilla activity in pcDNA3-Renilla-FLAG-firefly K529E transfected cells was normalized to firefly/Renilla activity in pcDNA3-Renilla-FLAG-firefly transfected cells.

Immunolocalization of p53

BJ-HRAS^G12V cells (WT or SNORA13 KO) were treated for 5 days with 100 nM (Z)-4-Hydroxytamoxifen and then seeded on poly-L-lysine coated coverslips in 6-well plates (1.3x10⁵ cell/well). CCD-1070Sk and IMR-90 cells were seeded at the indicated passage numbers or 4 days after etoposide treatment. Immunofluorescence for p53 was performed 48 hours later using previously-described methods.⁹² Briefly, cells were fixed with 4% paraformaldehyde for 15 minutes and permeabilized with 0.3% Triton X-100 for 15 minutes. Samples were washed with PBS and blocked in blocking buffer (5% FBS and 0.3% Triton X-100 in PBS) for 30 minutes followed by p53 primary antibody (Santa Cruz) incubation in blocking buffer overnight at 4°C. Samples were washed three times with PBS and incubated with goat anti-mouse IgG cross-adsorbed secondary antibody (Invitrogen) in blocking buffer for 1 hour. Nuclei were stained with DAPI for 1 minute and samples were mounted with SlowFade Diamond Antifade Mountant (Thermo Fisher Scientific). Cells were imaged on a Zeiss LSM980 confocal microscope.

Chromatin immunoprecipitation (ChIP)

ChIP was performed as described.⁹³ Briefly, 1x10⁷ cells per IP were crosslinked with 37% formaldehyde. Cells were lysed in 1 mL lysis buffer (150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris, 5 mM EDTA, pH 8.0, and 1x protease inhibitor cocktail). DNA was sheared by sonication to yield fragments of 300-1,000 bp. Lysates were cleared by centrifugation at 16,000 g for 10 minutes. Protein G Dynabeads (Invitrogen) were preincubated with p53 mouse monoclonal antibody (Santa Cruz) for 1 hour and added to lysates for overnight rotation at 4°C. Samples were eluted with elution buffer (50 mM Glycine, pH 2.8), treated with RNAse A, and reverse crosslinked at 65°C overnight. DNA was purified with the PCR Purification Kit (Qiagen) and used as a template for qPCR (primer sequences provided in Table S5).

MDM2 and p53 immunoprecipitation

5x10⁶ tamoxifen-treated BJ-HRAS^G12V cells (WT or SNORA13 KO) were lysed in 1 mL IP buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA and 0.5% NP-40) for 30 minutes on ice. Lysates were centrifuged at 16,000 g for 10 minutes, and supernatant was collected. The cleared lysates were incubated with 2 μg MDM2 mouse monoclonal antibody (Sigma) or p53 mouse monoclonal antibody (Santa Cruz) for 4 hours at 4°C and rotated at 4°C overnight after adding Protein G Dynabeads (Invitrogen). Samples were washed four times with IP buffer, eluted with NuPAGE LDS Sample Buffer (Invitrogen), and analyzed by western blotting.

MDM2 inhibition

For siRNA-mediated MDM2 inhibition, BJ-HRAS^G12V cells (WT or SNORA13 KO) were reverse transfected with 20 nM control siRNA or pre-designed siRNAs targeting MDM2 (Invitrogen) using Lipofectamine RNAiMAX (Thermo Fisher Scientific) according to the manufacturer’s instructions. MDM2 mRNA levels were analyzed by qRT-PCR 72 hours post-transfection. For MDM2 inhibition with Nutlin-3, cells were treated with vehicle or 1 μM Nutlin-3. For cell proliferation assays, cells were transfected with siRNA or treated with fresh Nutlin-3 every 3 days when passaged.

SNORA13 purification and mass spectrometry

RNA antisense purification and mass spectrometry (RAP-MS) was performed as described in detail previously,⁶³ with the following modifications:

1. 3 biological replicates were performed for both wild-type and SNORA13 knockout BJ-HRAS^G12V cells grown in the absence of tamoxifen. Pull-downs from SNORA13 knockout cells served as the negative control.

2. A pool of five 60-nucleotide 5’ biotinylated antisense oligos (IDT) were used to pull-down SNORA13. 10 μg of pooled oligos per 2x10⁸ cells were used for each pull-down.

3. Cells were cultured in normal growth medium and crosslinked using a Spectrolinker XL-1500 (Spectronics) at 254 nm with 0.4 J/cm².

4. For cell lysis, cell suspensions were sonicated with a Bioruptor (Diagenode) for 15 cycles (30 sec ON and 45 sec OFF) and treated with TurboDNase (Invitrogen) for 30 minutes.

5. On-bead peptide digestion was performed for mass spectrometry without elution.

6. Label-free semi-quantitative mass spectrometry analysis was performed at the UT Southwestern Proteomics Core using reverse-phase LC-MS/MS and an Orbitrap Fusion Lumos mass spectrometer. Raw MS data files were analyzed using Proteome Discoverer v2.4 SP1 (Thermo), with peptide identification performed using Sequest HT searching against the human protein database from UniProt. Only proteins detected in all 3 wild-type replicates were considered as candidate interactors. p values were calculated by unpaired two-tailed student’s t-test.

UV crosslinking and RNA immunoprecipitation

UV crosslinking and RNA immunoprecipitation (UV-RIP) was performed as previously described⁹⁴ with the following modifications. 2x10⁷ BJ-HRAS^G12V cells were washed with cold PBS and UV crosslinked on ice at 254 nm with 0.4 J/cm². Cells were scraped in cold PBS and pelleted by centrifugation. Cell pellets were lysed in 1 mL cold lysis buffer (50 mM Tris-HCl, 100 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate, 1x Protease Inhibitor Cocktail, 400 U/mL RNase inhibitor) for 30 minutes on ice. Lysates were centrifuged at 14,000 g for 10 minutes at 4°C. 3.75 mg pre-washed Protein G Dynabeads (Invitrogen) coupled with antibody was added to cleared lysate and rotated for 2 hours at 4°C. Beads were washed three times with 900 μL cold High Salt Wash Buffer (50 mM Tris-HCl, 1M NaCl, 1 mM EDTA, 1% NP-40, 0. 1% SDS, 0.5% sodium deoxycholate) and then washed three times with 500 μL wash buffer (20 mM Tris-HCl, 10 mM MgCl₂, 0.2% Tween-20). After the final wash, each sample was resuspended in 100 μL wash buffer. 30 μL of the sample was used for western blotting. The remainder was treated with proteinase K (NEB) followed by RNA extraction with Trizol. snoRNA levels were determined by qRT-PCR (primer sequences provided in Table S5). All antibodies are listed in the Key Resources Table.

Western blot detection of SNORA13 interactors

Endogenous SNORA13 was purified with antisense probes after UV crosslinking as described above for RAP-MS, except 1 μg of pooled oligos and 2x10⁷ cells were used. After final wash, proteins were eluted from streptavidin-coated magnetic beads by incubation in NuPAGE LDS Sample Buffer (Invitrogen) for 10 minutes at 70°C and analyzed by western blotting. Sense probes and SNORA13 KO cells served as negative controls. All antibodies are listed in the Key Resources Table.

Purification of MBP-RPL23 from insect cells

A maltose binding protein (MBP)-RPL23 fusion protein sequence was cloned as two separate fragments into the EcoRI site of pFastBac Dual expression vector (Gibco). The full-length human RPL23 sequence was amplified by PCR from cDNA. The MBP sequence was PCR amplified from plasmid MTTH-SNAP⁹⁵. Recombinant viruses were generated and amplified using Bac-to-Bac Baculovirus Expression Systems (Gibco) according to the manufacturer’s instructions. Sf9 cells and Cellfectin II transfection reagent (Gibco) were used to produce baculovirus. 5 mL of P2 viruses were used to infect 500 mL of Sf9 cells at a density of 2x10⁶ cells/mL for protein expression. After 72 hours, cells were collected by centrifugation, resuspended in lysis buffer (20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1 mM PMSF, 1 mM DTT, 1 mM EDTA, 2 μg/mL RNase A, 5 U/mL RNase T1, and 1x EDTA-free Protease Inhibitor Cocktail) and sonicated on ice. Lysates were centrifuged at 22,000 g for 30 minutes at 4°C. The cleared lysates were incubated with Amylose Resin (NEB) and washed twice with 10 column volumes of wash buffer (20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1 mM DTT, and 1 mM EDTA). Protein was eluted with elution buffer (wash buffer containing 10 mM maltose) and centrifuged in a bench-top centrifuge at 15,000 RPM to remove aggregates. The concentration of protein was measured by absorbance at 280 nm on Nanodrop.

Electrophoretic mobility shift assay (EMSA)

EMSAs were performed using purified MBP-RPL23 protein and biotin-labeled RNAs were synthesized by in vitro transcription with HiScribe T7 High Yield RNA Synthesis Kit (NEB) or obtained as RNA oligonucleotides (IDT). DNA templates for in vitro transcription were PCR amplified from cDNA (primer sequences provided in Table S5). 50 ng biotin-labeled snoRNA was incubated with increasing concentrations of MBP-RPL23 (0-2 μM) in EMSA buffer (10 mM HEPES, pH 7.3, 20 mM KCl, 1 mM MgCl₂, 1 mM DTT, 5% glycerol, and 100 ng/μL tRNA) with or without 500 ng unlabeled snoRNA for 30 minutes at room temperature. Following incubation, 5% Ficoll-400 was added, and samples were separated on a native 6% TBE-polyacrylamide gel (Invitrogen). Samples were transferred to a BrightStar-Plus nylon membrane (Invitrogen) and UV crosslinked at 254 nm with 120 mJ/cm². Membrane was blocked in blocking buffer (0.5% SDS and 1% BSA in 1x PBS) for 30 minutes. Biotin-labeled RNA was detected using streptavidin-conjugated IR Dye (LI-COR) diluted in blocking buffer (1:300) for 1 hour. A LI-COR Odyssey imager was used for imaging.

RPL23:28S rRNA competitive binding assay

The biotin-labeled 28S rRNA fragment (nucleotides 4389-5070 of human 28S rRNA) was in vitro transcribed from a PCR-generated template using HiScribe T7 High Yield RNA Synthesis Kit (NEB) and gel purified. Primers for amplification are provided in Table S5. 2 μM RPL23, 0.5 μM labeled 28S rRNA fragment, and increasing concentrations of unlabeled SNORA13 were mixed in EMSA buffer and incubated for 30 minutes at room temperature. After adding 5% Ficoll-400, samples were separated on a 1.5% agarose gel, transferred to a nylon membrane using the iBlot Dry Blotting System (Invitrogen), and UV crosslinked at 254 nm with 120 mJ/cm². Membrane was blocked in blocking buffer (0.5% SDS and 1% BSA in 1x PBS) for 30 minutes. Biotin-labeled RNA was detected using streptavidin-conjugated IR Dye (LI-COR) diluted in blocking buffer (1:300) for 1 hour. A LI-COR Odyssey imager was used for imaging.

RNA-seq and Ribosome Profiling

For RNA-seq and ribosome profiling, BJ-HRAS^G12V cells (WT and SNORA13 KO) were harvested after treatment with vehicle or 100 nM (Z)-4-Hydroxytamoxifen for 7 days. For RNA-seq, total RNA was isolated and sequencing libraries were generated with the Stranded Total RNA Library Prep Kit (Illumina) and sequenced on a HiSeq 4000 (Illumina) with 150 bp paired-end reads. Reads were mapped to GRCh38⁹⁶ using STAR⁹⁷, read counts were generated using featurecounts⁹⁸, and edgeR⁹⁹ was used to analyze differential gene expression.

Ribosome profiling was performed as described,¹⁰⁰ with the following modifications:

1. Ribosome footprints were size selected by gel electrophoresis as described.¹⁰¹ For library preparation, fragments between 24–35nt were excised.

2. After linker ligation, samples were not pooled. The subsequent steps were performed individually for each sample.

3. 3’ linker-ligated RNA fragments were separated from unligated linker by gel electrophoresis and purified by gel extraction as previously described.¹⁰¹

4. Ribosomal RNA was depleted as described previously.¹⁰¹

For analysis of reads from ribosome profiling, the adaptor, inline barcode, and the random-mer incorporated into the library amplicon were trimmed. Then, using STAR⁹⁷, reads mapping to known rRNA¹⁰², tRNA¹⁰³, and snRNA¹⁰⁴ sequences were removed, and reads that did not map to these noncoding RNA databases were subsequently mapped to GRCh38⁹⁶. To analyze changes in mRNA translation efficiency, a statistical test for differential translation efficiency using both mRNA abundance and ribosome occupancy was performed using RiboDiff¹⁰⁵. Only transcripts with more than 50 reads in both ribosome profiling and RNA-seq were analyzed in RiboDiff.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analysis

All experiments, including those presented as images such as western blots and microscopy, were repeated with a minimum of three biological replicates. All values are reported as mean ± SD in each figure. Numbers of replicates and statistical tests performed are reported in each figure legend. Statistical analysis of CRISPRi screen was performed using MAGeCK²⁹. For mass spectrometry experiments, p values were calculated by unpaired two-tailed student’s t-test. For RNA-seq, edgeR⁹⁹ was used to analyze differential gene expression. For ribosome profiling, a statistical test for differential translation efficiency using both mRNA abundance and ribosome occupancy was performed using RiboDiff¹⁰⁵.

Supplementary Material

1

Figure S1. SNORA13 is required for multiple forms of senescence, related to Figure 1 and Figure 2

(A) Growth of BJ-HRAS^G12V-dCas9^KRAB clones used in the screen expressing the indicated sgRNAs in the presence of tamoxifen.

(B-C) Western blot showing induction of HRAS^G12V after 7 days of tamoxifen treatment in BJ-HRAS^G12V-dCas9^KRAB cells with lentiviral expression of non-target (NT) or EPB41L4A-AS1-targeting sgRNAs (B) or in wild-type and SNORA13 knockout BJ-HRAS^G12V cells (C).

(D) qRT-PCR analysis of EPB41L4A-AS1 expression relative to GAPDH in wild-type or SNORA13 knockout BJ-HRAS^G12V cells following stable transfection with empty vector or the indicated EPB41L4A-AS1 rescue constructs.

(E) Growth of BJ-HRAS^G12V cells expressing the indicated EPB41L4A-AS1 rescue constructs in the absence of tamoxifen.

(F, J, N) Growth of CCD-1070Sk human fibroblasts (F), IMR-90 human fibroblasts (J), or etoposide-treated IMR-90 human fibroblasts (N). Prior to monitoring growth, cells were infected with lentivirus expressing Cas9 and non-target (NT) or SNORA13-targeting sgRNA. For etoposide experiments, day 0 represents the timepoint at which drug was removed and proliferation measurements were initiated (after prior treatment with 50 μM etoposide for 48 hours). Validation of SNORA13 sg1 is shown in Figure S5A.

(G, K, O) Cell morphology and SA-β-gal staining of CCD-1070Sk (G), IMR-90 (K), or etoposide-treated IMR-90 (O) cells at the indicated timepoints.

(H, L, P) Quantification of SA-β-gal activity in (G, K, and O).

(I, M, Q) Western blots of total cell lysates in CCD-1070Sk (I), IMR-90 (M) or etoposide-treated IMR-90 (Q) cells at the indicated timepoints.

Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing all groups to WT (panel E) or sgNT (panels A,F,H,J,L,N,P). n.s. not significant; *p≤0.05, ****p≤0.0001.

2

Figure S2. SNORA13 localization and impact on translation, related to Figure 3

(A) qRT-PCR analysis of SNORA13 subcellular localization in the absence of tamoxifen (upper) or after 15 days of tamoxifen treatment (lower).

(B) SNORA13 RNA FISH in the absence of tamoxifen (upper) or after 15 days of tamoxifen treatment (lower). Red, SNORA13; green, nucleolar marker Nucleolin; blue, DAPI.

(C) Northern blot of SNORA13 and U6 snRNA (loading control) after treatment with tamoxifen for the indicated number of days.

(D) Primer extension assay for detection of 18S.1248U m¹acp³Ψ.

(E) Flow cytometry analysis of OP-Puro incorporation in BJ-HRAS^G12V cells of the indicated genotypes after 7 days of tamoxifen treatment.

(F) Time-course analysis of global translation (as measured by OP-Puro incorporation and flow cytometry) and cell proliferation after tamoxifen addition in BJ-HRAS^G12V cells of the indicated genotypes. MFI, mean fluorescence intensity.

(G) Anti-puromycin western blot of BJ-HRAS^G12V cells of the indicated genotypes pulsed with puromycin after 7 days of tamoxifen treatment.

(H-I) Stop codon readthrough (H) or amino acid misincorporation assay (I) for translational fidelity. Firefly/Renilla activity in reporter-transfected cells was normalized to firefly/Renilla activity in control-transfected cells. Cells were treated with tamoxifen for 7 days where indicated. Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing groups to WT (panels F,H,I). n.s., not significant; **p≤0.01, ***p≤0.001, ****p≤0.0001.

3

Figure S3. Altered translation of specific mRNAs is unlikely to explain senescence escape in SNORA13 knockout cells, related to Figure 3

(A) Metagene plots showing the periodicity of ribosome profiling reads and the distribution of reads across 5’ UTRs, coding sequences, and 3’ UTRs.

(B) qRT-PCR analysis of transcripts whose translation was significantly decreased (p<0.01) in both tamoxifen-treated and untreated SNORA13 knockout cells, after lentiviral expression of Cas9 and the indicated sgRNAs.

(C) Growth of BJ-HRAS^G12V cells expressing Cas9 and the indicated sgRNAs in the presence of tamoxifen.

(D) qRT-PCR analysis of transcripts whose translation was significantly increased (p<0.01) in both tamoxifen-treated and untreated SNORA13 knockout cells, after lentiviral expression of Cas9 and the indicated sgRNAs.

(E) Growth of SNORA13 knockout BJ-HRAS^G12V cells expressing Cas9 and the indicated sgRNAs in the presence of tamoxifen.

For qRT-PCR experiments and growth curves, data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test (panels B,D) or unpaired one-tailed student’s t-test (panels C,E) comparing all groups to sgNT. n.s., not significant; **p≤0.01, ***p≤0.001, ****p≤0.0001.

4

Figure S4. SNORA13 does not regulate codon-specific translation rates, rRNA synthesis, or rRNA processing, related to Figure 3

(A-D) Spearman correlations between the frequency of each codon in transcripts and their relative translational efficiency in tamoxifen-treated SNORA13 knockout versus wild-type BJ-HRAS^G12V cells (A), wild-type BJ-HRAS^G12V cells minus tamoxifen (normal growth) versus plus tamoxifen (senescence) (B), tamoxifen-treated TP53 knockout versus wild-type BJ-HRAS^G12V cells (C), or untreated SNORA13 knockout versus wild-type BJ-HRAS^G12V cells (D).

(E) Sucrose gradient fractionation showing that expression of SNORA13 normalizes levels of 60S subunits in SNORA13 knockout BJ-HRAS^G12V cells.

(F-G) qRT-PCR analysis of total RNA (F) or EU-labeled nascent rRNA (G) in wild-type and SNORA13 knockout BJ-HRAS^G12V cells.

(H) Northern blot analysis of rRNA precursors with ITS1 and ITS2 probes. ACTB served as a loading control.

Data are represented as mean ± SD with individual data points shown (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test (panels F,G). n.s., not significant; ***p≤0.001.

5

Figure S5. SNORA13 negatively regulates ribosome biogenesis under normal growth conditions and after amino acid starvation, related to Figure 3

(A) Northern blot analysis of SNORA13 and U6 snRNA (loading control) in HEK293T RPL28-SNAP cells after lentiviral expression of Cas9 and the indicated sgRNAs.

(B) Sucrose gradient fractionation showing total ribosome levels (upper) or fluorescently-labeled newly synthesized 60S ribosomal subunits (lower) in HEK293T RPL28-SNAP cells after lentiviral expression of Cas9 and non-target sgRNA (sgNT) or sgRNA targeting SNORA13. Bar graph shows quantification of 60S peak area normalized to sgNT.

(C) Sucrose gradient fractionation performed as in (B) after treatment of lysates with 50 mM EDTA to collapse polysomes into individual ribosomal subunits.

(D) qRT-PCR analysis of SNORA13 in HEK293T RPL28-SNAP cells after amino acid starvation for the indicated number of days.

(E) qRT-PCR analysis of SNORA13 in HEK293T (left) or BJ-HRAS^G12V (right) cells after 3 days of 50 nM rapamycin treatment.

(F) Sucrose gradient fractionation of total (upper) or newly-synthesized ribosomes (lower) as in (B) after culturing cells for 3 days in amino acid-free media.

(G) Sucrose gradient fractionation of total (upper) or newly-synthesized ribosomes (lower) as in (B) after culturing cells for 3 days in amino acid-free media and treating lysates with 50 mM EDTA to collapse polysomes into individual ribosomal subunits.

Data are represented as mean ± SD with individual data points shown (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing groups to Day 0 (panel D), DMSO (panel E), or sgNT (panels B,C,F,G). **p≤0.01, ***p≤0.001, ****p≤0.0001.

6

Figure S6. Analysis of the roles of MDM2 and 18S.1248m¹acp³Ψ in senescence escape in SNORA13 knockout cells, related to Figure 4 and Figure 5

(A) Immunolocalization of p53 in CCD-1070Sk and IMR-90 fibroblasts undergoing replicative or DNA damage-induced senescence after lentiviral expression of Cas9 and the indicated sgRNAs.

(B) qRT-PCR analysis of MDM2 expression following siRNA-mediated knockdown.

(C) Western blot analysis of total cell lysates from tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells following transfection with control siRNA (siNT) or siRNA targeting MDM2 (siRNA2 from panel B used in this experiment).

(D-E) p53 was immunoprecipitated from tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells and analyzed by western blot with anti-p53 (D) or anti-ubiquitin antibodies (E).

(F) Immunolocalization of p53 in tamoxifen-treated BJ-HRAS^G12V cells following MDM2 knockdown with siRNA2 from (B) or Nutlin-3 treatment.

(G) qRT-PCR analysis of wild-type or mutant SNORA13 or GAPDH in UV-RIP samples after pull-down of Dyskerin. Enrichment was normalized to input and then normalized to GAPDH enrichment for each antibody.

(H) qRT-PCR analysis of EMG1 and TSR3 following lentiviral expression of Cas9 and sgRNAs targeting the indicated genes.

(I) Growth of BJ-HRAS^G12V cells expressing Cas9 and the indicated sgRNAs in the presence of tamoxifen.

Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired one-tailed student’s t-test comparing all groups to sgNT1 (panel I). n.s., not significant; ****p≤0.0001.

7

Figure S7. Interaction of SNORA13 with RPL23, related to Figure 6

(A) qRT-PCR analysis of SNORA13 recovery after pull-down with ASOs.

(B) Silver-stain of recovered proteins after pull-down of SNORA13 from UV-crosslinked lysates.

(C) Western blot of ribosomal proteins in UV-RIP samples used for qRT-PCR analysis in Figure 6C.

(D) Secondary structure of SNORA13. Sequences that base pair with 18S rRNA and guide pseudouridylation are highlighted in yellow (nucleotides 8-12, 45-49).

(E) In vitro binding assay demonstrating interaction of RPL23 with the first stem loop of SNORA13 (nucleotides 1-60).

(F) Sequences of wild-type and mutant SNORA13 used for binding assays and rescue experiments.

(G) Northern blots showing expression of wild-type and mutant SNORA13 constructs. U6 snRNA served as a loading control.

(H) In vitro binding assays with RPL23 and the indicated SNORA13 mutants.

(I) qRT-PCR analysis of SNORA13 (wild-type and mutants) and SNORA25 in UV-RIP samples after pull-down of RPL23. Enrichment was normalized to input.

(J) Growth of tamoxifen-treated wild-type or SNORA13 knockout BJ-HRAS^G12V cells reconstituted with wild-type or mutant SNORA13.

(K) Sucrose gradient fractionation showing 60S ribosomal subunit levels in wild-type, SNORA13 knockout, or SNORA13 Mut^{(8–12,45–49)}-expressing BJ-HRAS^G12V cells without tamoxifen treatment.

(L) Structure of human 28S rRNA in 60S ribosomal subunit (grey and green) with RPL23 (red) (PDB ID: 6OM7)⁶⁴. Other ribosomal proteins not shown. The 3’ fragment of human 28S rRNA used for binding assays (nucleotides 4389 to 5070) is highlighted in green. Structure rendered by UCSF ChimeraX⁷⁵.

(M-N) Northern blot analysis of SNORA13 (M) or 32S pre-rRNA (N) and quantification of RNA copy number in BJ-HRAS^G12V cells.

Data are represented as mean ± SD (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing all groups to SNORA13 KO (panel J). n.s., not significant; ****p≤0.0001.

8

Figure S8. Depletion of SNORA13 homologs in mouse embryonic fibroblasts and model of SNORA13-mediated regulation of ribosome biogenesis and the nucleolar stress response, related to Figure 7

(A) Identification of sgRNAs that effectively deplete mouse SNORA13 homologs. MEFs were infected with lentivirus expressing Cas9 and individual sgRNAs. Northern blotting (Gm25636 and Gm23639) or qRT-PCR (Gm55482) were used to monitor snoRNA levels. For each northern blot, all lanes came from the same blot but irrelevant lanes were removed where indicated with vertical lines. U6 served as a loading control. For qRT-PCR experiments, Gm55482 expression was normalized to Actb. Data are represented as mean ± SD with individual data points shown (n=3 biological replicates). p values were calculated by unpaired two-tailed student’s t-test comparing groups to sgNT. ***p≤0.001.

(B,C) Sucrose gradient fractionation of MEFs with single (B) or double (C) knockout of mouse SNORA13 homologs.

(D) Proposed model for the regulation of ribosome biogenesis and the nucleolar stress response by SNORA13. (Left panel) Through a direct interaction with RPL23, SNORA13 slows the rate of assembly of pre-60S subunits (1). Under conditions of nucleolar stress, as triggered by aberrant oncogene activation for example, inhibition of ribosome biogenesis by SNORA13 increases the concentration of non-ribosome associated RPs (2) that can interact with and inhibit MDM2 (3), thereby promoting p53 activation (4). (Right panel) In SNORA13-deficient cells, accelerated pre-60S assembly (5) reduces the concentration of free RPs (6), preventing activation of the nucleolar stress response (7) and enabling MDM2 to diminish p53 activity (8). Figure created with BioRender.com.

9

Supplemental Table S1. MAGeCK analysis of CRISPRi screen for noncoding RNAs required for OIS, related to Figure 1

10

Supplemental Table S2. Ribosome profiling analysis, related to Figure 3.

11

Supplemental Table S3. RNA sequencing analysis, related to Figure 3 and Figure 4.

12

Supplemental Table S4. Mass spectrometry analysis of SNORA13 interactors, related to Figure 6

13

Supplemental Table S5: Oligonucleotide sequences, related to STAR Methods

Highlights.

• A CRISPR screen identifies a H/ACA box snoRNA, SNORA13, that is required for senescence

• SNORA13 is required for multiple forms of senescence in human cells and in mice

• SNORA13 interacts directly with RPL23 and slows the rate of 60S ribosome biogenesis

• SNORA13 increases levels of free ribosomal proteins, activating p53 in senescent cells