An updated human snoRNAome

Abstract

Small nucleolar RNAs (snoRNAs) are a class of non-coding RNAs that guide the post-transcriptional processing of other non-coding RNAs (mostly ribosomal RNAs), but have also been implicated in processes ranging from microRNA-dependent gene silencing to alternative splicing. In order to construct an up-to-date catalog of human snoRNAs we have combined data from various databases, de novo prediction and extensive literature review. In total, we list more than 750 curated genomic loci that give rise to snoRNA and snoRNA-like genes. Utilizing small RNA-seq data from the ENCODE project, our study characterizes the plasticity of snoRNA expression identifying both constitutively as well as cell type specific expressed snoRNAs. Especially, the comparison of malignant to non-malignant tissues and cell types shows a dramatic perturbation of the snoRNA expression profile. Finally, we developed a high-throughput variant of the reverse-transcriptase-based method for identifying 2'-O-methyl modifications in RNAs termed RimSeq. Using the data from this and other high-throughput protocols together with previously reported modification sites and state-of-the-art target prediction methods we re-estimate the snoRNA target RNA interaction network. Our current results assign a reliable modification site to 83% of the canonical snoRNAs, leaving only 76 snoRNA sequences as orphan.

Figure 1.

Schematic overview of structural types of snoRNAs. (A) Canonical C/D box snoRNAs have a C box and a D box motif located close to the terminal stem, and additional internal C' and D’ boxes. Canonical H/ACA box snoRNAs are composed of two stem loop structures with an internal H box motif and an ACA box motif at the 3′ end. (B) Cajal body-associated snoRNAs additionally have specific localization motifs, which are the CAB box in the case of H/ACA box snoRNAs, and a G/U rich sequence in the case of C/D box snoRNAs. (C) SnoRNAs with hybrid structure that consist of both a C/D box and an H/ACA box domain have been identified. Recent studies have also uncovered extremely short C/D box-like snoRNAs (D) as well as long (several hundred nucleotides) noncoding RNAs with snoRNA ends (E and F).

Figure 2.

Outline of the snoRNA annotation strategy used in this study. We combined de novo search on ENCODE sRNA-seq expressed regions with snoRNA genes and predictions from various databases. All predicted candidate sequences were checked for a supportive sRNA-seq read pattern to identify high confidence, currently not annotated snoRNA genes. Finally, snoRNAs from all sources were merged and filtered for redundancy to establish a comprehensive map of human snoRNA loci.

Figure 3.

Distribution of orphans, single guides (sg), and double guides (dg) among known and novel snoRNAs based on our target predictions. (A) Of the 275 canonical C/D box snoRNAs, 48 are orphan, 38 are double guides and 189 are single guides. Of the latter, 93 (56 D'+D box, +37 no D' box) have a functional ASE adjacent to the D-box and 96 adjacent to the D'-box. (B) Of 190 canonical H/ACA box snoRNA sequences 30 remain orphan (of which SNORA73A/B have a non-canonical role in 18S rRNA maturation (83)), 89 are double guides and 71 are single guides. Of the latter 45 have a functional ASE in the 5′ stem, and 26 in the 3′ stem.

Figure 4.

Visualization of predicted snoRNA–target RNA interactions. Each row displays binding properties of one snoRNA sequence. The columns represent: ICI: Interaction Conservation Index of selected target (scaled to [−1,1]); rank: rank of selected interaction within set of predictions for the snoRNA ASE in the human genome (1: respective interaction is best for this ASE, else:1/(log of rank) (scaled)); levelC: level of conservation of the interaction among deuterostomes (1: primates, 2: eutherians, 3: therians, 4: mammals, 5: amniotes, 6: tetrapodes, 7: tetrapodes and teleosts, 8: vertebrates, 9: deuterostomes, scaled to [−1,1], gray denotes human-specific); MFE: minimum free energy (MFE) of the interaction (scaled to [-1,1]); reported: target reported in the literature (1: yes, −1: no). For each cell, the value is also illustrated by the position of the vertical black line relative to the 0-value line, located in the middle of the cells. Supplementary Figure S6 provides the heatmaps in higher resolution with snoRNA names next to the rows. (A) C/D box snoRNAs (not including multi-copy and C/D-like snoRNAs). The left set of columns refer to the ASE upstream of D box; the right set of columns the ASE adjacent to D' box (a grey line if the D' box and the particular ASE could not be annotated). (B) H/ACA box snoRNAs (not including the ALUACA class). The left set of columns refer to the 5′ stem-associated ASE, whereas the right set of columns refer to the 3′ stem-associated ASE.

Figure 5.

Structure of the elongated SCARNA21. The snoRNA-characteristic sequence motifs are enclosed in a black frame. The C/D box domain folds into the characteristic terminal stem and the obligatory kink-turn motif. The H/ACA domain forms the typical double-hairpin structure. Predicted target sites for the ASEs are displayed in the grey boxes. From 5′ to 3′, the predicted functions are: guide1: U12–17, guide2l&r: U12–18 (102), guide3l&r: U6atac-83, guide4: 28S-4426. See Supplementary Text S2 for details about these interactions. The figure was produced with R2R (103).

Figure 6.

Expression profiling of snoRNA genes in ENCODE sRNA-seq data. (A) The pool of human snoRNA genes is dominated by a few abundantly expressed snoRNA genes. (B) Evaluation of tissue specific expression of snoRNA genes. The top panel shows values for C/D box snoRNAs, while the bottom panel does for H/ACA box snoRNAs. The higher the specificity score is the more biased is the expression to a specific tissue or cell type. MFOCP is an acronym for melanocytes, fibroblasts, osteoblasts, chondrocytes and placental tissue. The red stars mark the 20 most highly expressed C/D box (upper panel) and H/ACA box (lower panel) snoRNAs in the entire data set (further details in Supplementary Figure S2).