An ultraconserved snoRNA-like element in long noncoding RNA CRNDE promotes ribosome biogenesis and cell proliferation

Abstract

Cancer cells frequently upregulate ribosome production to support tumorigenesis. While small nucleolar RNAs (snoRNAs) are critical for ribosome biogenesis, the roles of other classes of noncoding RNAs in this process remain largely unknown. Here we performed CRISPRi screens to identify essential long noncoding RNAs (lncRNAs) in renal cell carcinoma (RCC) cells. This revealed that an alternatively-spliced isoform of lncRNA Colorectal Neoplasia Differentially Expressed containing an ultraconserved element (UCE), referred to as CRNDE ^UCE, is required for RCC cell proliferation. CRNDE ^UCE localizes to the nucleolus and promotes 60S ribosomal subunit biogenesis. The UCE of CRNDE functions as an unprocessed C/D box snoRNA that directly interacts with ribosomal RNA precursors. This facilitates delivery of eIF6, a key 60S biogenesis factor, which binds to CRNDE ^UCE through a sequence element adjacent to the UCE. These findings highlight the functional versatility of snoRNA sequences and expand the known mechanisms through which noncoding RNAs orchestrate ribosome biogenesis.

Figure 1.. CRISPRi screening reveals that CRNDE is required for proliferation of RCC cells

(A) RNA-seq analysis of 48 matched tumor-normal pairs from RCC patients. The highlighted box indicates the lncRNAs whose expression level is at least 2-fold higher in tumors (FDR<0.05). (B) Schematic of CRISPRi drop-out screens performed in RCC cell lines. (C) MAGeCK analysis of CRISPRi screens. The common essential lncRNAs that were among the top 50 hits in all RCC cell lines are indicated. (D) TCGA (left) and UTSW RNA-seq data (right) showing elevated expression of CRNDE in RCC tumors. KIRC, Renal Clear Cell Carcinoma; KIRP, Renal Papillary Cell Carcinoma; RCC, all renal cell carcinoma subtypes. (E) qRT-PCR analysis of CRNDE expression relative to GAPDH in 786-O CRISPRi cells after lentiviral expression of non-target (sgNT) or CRNDE-targeting sgRNAs (left). Proliferation was measured by cell counts at the indicated timepoints after plating (right). (F) Fluorescence microscopy (left) and quantification (right) of DAPI and EdU incorporation in 786-O CRISPRi cells expressing the indicated sgRNAs. (G) qRT-PCR analysis of CRNDE expression relative to GAPDH in patient-derived primary RCC CRISPRi cells after lentiviral expression of non-target (sgNT) or CRNDE-targeting sgRNAs (left). Proliferation was measured by cell counts at the indicated timepoints after plating (right). Data are represented as mean ± SD (n=3 biological replicates). *p<0.05, **p<0.01, calculated by two-tailed t-test. See also Figure S1.

Figure 2.. A CRNDE isoform containing an ultraconserved element promotes RCC proliferation in trans

(A) Proliferation of 786-O CRISPRi cells expressing doxycycline (dox)-inducible unspliced (left) or fully-spliced (right) CRNDE rescue constructs after lentiviral expression of non-target sgRNA (sgNT) or sgRNA targeting endogenous CRNDE. (B-D) UCSC genome browser ENCODE transcription (B) and vertebrate conservation (PhastCons) (D) tracks (hg38), aligned with intron-containing CRNDE isoforms detected in the PacBio Universal Human Reference RNA dataset (C). (E) Northern blot analysis of total RNA from the indicated cell lines after lentiviral expression of non-target (sgNT) or CRNDE-targeting sgRNAs using a probe complementary to the UCE. (F) qRT-PCR analysis (bottom left) and proliferation (bottom right) of 786-O CRISPRi cells expressing a dox-inducible CRNDE UCE-containing isoform rescue construct after lentiviral expression of non-target sgRNA (sgNT) or sgRNA targeting endogenous CRNDE. The positions of qRT-PCR amplicons (A, B, C, D, and E) are indicated in the transcript schematic. Transcript abundance was normalized to GAPDH. (G-H) Schematic showing positions of UCE intron rescue constructs (G) and their ability to support proliferation upon depletion of endogenous CRNDE, as measured by cell counts six days after plating (H). Data are represented as mean ± SD (n=3 biological replicates). n.s., not significant; **p<0.01, calculated by two-tailed t-test. See also Figure S2.

Figure 3.. CRNDE^UCE localizes to the nucleolus and promotes 60S ribosomal subunit biogenesis

(A) Subcellular fractionation and qRT-PCR analysis of 786-O cells using primers that detect fully-spliced CRNDE (CRNDE^spliced), CRNDE UCE-containing isoforms (CRNDE^UCE), or all CRNDE isoforms (CRNDE^total). GAPDH and MALAT1 represent cytoplasmic and nuclear controls, respectively. (B) RNA FISH with a probe complementary to the UCE (red) in 786-O CRISPRi cells expressing the indicated sgRNAs. (C) Schematic of pre-ribosomal sequential extraction showing stages of ribosome biogenesis captured in each fraction (SN1, SN2, SN3). FC, fibrillar center; DFC, dense fibrillar component; GC, granular component; Nu, nucleoplasm; Cy, cytoplasm. (D) Western blot analysis of control proteins in nucleolar fractions. (E) qRT-PCR analysis of CRNDE^UCE in nucleolar fractions from 786-O cells. GAPDH and 32S pre-rRNA represent cytoplasmic and early nucleolar markers, respectively. (F-H) Schematic of polysome profiling by sucrose gradient ultracentrifugation (F) and results from 786-O (G) and HeLa (H) CRISPRi cells after lentiviral expression of non-target (sgNT) or CRNDE-targeting sgRNAs. Data are represented as mean ± SD (n=3 biological replicates). See also Figure S3.

Figure 4.. CRNDE^UCE interacts directly with the 60S biogenesis factor eIF6

(A) Schematic of UV-crosslinking and purification of endogenous CRNDE^UCE to identify interacting proteins by mass spectrometry. (B) Western blot of eIF6 after UV crosslinking and pull-down of CRNDE^UCE with ASOs under denaturing conditions. Scrambled ASO served as a negative control. (C) qRT-PCR analysis of CRNDE^UCE or GAPDH after UV-crosslinking and immunoprecipitation with anti-eIF6 antibody or control IgG. Fold enrichment over IgG plotted. Data are represented as mean ± SD (n=3 biological replicates). **p<0.01, calculated by two-tailed t-test. (D-E) Schematics of CRNDE UCE intron fragment (D, upper) or sub-fragments (E, upper) used in EMSAs with eIF6 (lower panels). See also Figure S4.

Figure 5.. CRNDEUCE promotes the loading of eIF6 onto pre-60S subunits in the nucleolus

(A) Immunolocalization of eIF6 in 786-O CRISPRi cells expressing the indicated sgRNAs. (B-C) Sucrose gradient fractionation of nucleoli from 786-O CRISPRi cells expressing sgRNAs targeting eIF6 (B) or CRNDE (C, upper). Western blot analysis of eIF6 in nucleolar fractions (C, lower). (D) Western blot analysis of input and eIF6 immunoprecipitates from 786-O CRISPRi cells expressing the indicated sgRNAs. (E) Schematic of pre-rRNA processing in human cells (left). Northern blot analysis of rRNA precursors with ITS1 and ITS2 probes in total RNA from 786-O cells (right). EtBr, ethidium bromide. (F) Schematic of SNAP labeling approach to detect newly synthesized large ribosomal subunits. (G) Fluorescence microscopy of newly-synthesized RPL23a-SNAP in 786-O CRISPRi cells expressing the indicated sgRNAs. (H) Western blot analysis of 786-O CRISPRi cells expressing RPL23a-SNAP and the indicated sgRNAs. (I) Western blot analysis of NMD3 or control IgG immunoprecipitates from 786-O CRISPRi cells expressing the indicated sgRNAs. See also Figure S5.

Figure 6.. The CRNDE UCE is a C/D box snoRNA-like element that base pairs with 32S pre-rRNA

(A) Schematic of AMT crosslinking experiment to detect RNA:RNA base-pairing interactions. (B-C) qRT-PCR analysis of the indicated transcripts after AMT crosslinking and pull-down of CRNDE^UCE with ASOs under denaturing conditions. Experiments were performed in parental 786-O CRISPRi cells (B) or in 786-O CRISPRi cells expressing non-target (sgNT) or eIF6-targeting sgRNAs (C). Scrambled ASO served as a negative control. Enrichment was normalized to input. (D) The UCE of CRNDE resembles a C/D box snoRNA. Putative C box, D box, and complementarity to 32S pre-rRNA indicated. (E) qRT-PCR analysis of negative control transcripts (ACTB, NORAD, H/ACA box snoRNA SNORA13), 32S pre-rRNA (positive control), or CRNDE transcripts (CRNDE^UCE or CRNDE^LastExon) in FBL immunoprecipitates from UV-crosslinked 786-O cells. Fold enrichment over IgG plotted. (F) Western blot analysis of C/D box snoRNP components FBL and 15.5K, or H/ACA box snoRNP component DKC1, after UV crosslinking and pull-down of CRNDE^UCE with ASOs under denaturing conditions. Scrambled ASO served as a negative control. (G) Schematic of CRNDE UCE intron fragments and sequence of 32S binding site mutant used in rescue experiments. (H-I) qRT-PCR analysis of the indicated transcripts after AMT crosslinking and pull-down of CRNDE^UCE from cells expressing wild-type (H) or mutant (I) UCE-containing intron fragments after depletion of endogenous CRNDE using CRISPRi. (J) Proliferation of 786-O CRISPRi cells expressing wild-type or mutant UCE-containing intron fragments after lentiviral expression of non-target sgRNA (sgNT) or sgRNA targeting endogenous CRNDE. (K) Regulation of 60S ribosomal subunit biogenesis by CRNDE^UCE. The UCE functions as an unprocessed C/D box snoRNA that base-pairs with 32S pre-rRNA. This facilitates delivery of eIF6, bound to an adjacent sequence element, to maturing pre-60S subunits, thereby promoting subsequent steps in 60S biogenesis. Data are represented as mean ± SD (n=3 biological replicates). n.s., not significant; **p<0.01, calculated by two-tailed t-test. See also Figure S5 and S6.