Promoter-based identification of novel non-coding RNAs reveals the presence of dicistronic snoRNA-miRNA genes in Arabidopsis thaliana

Abstract

Background: In the past few decades, non-coding RNAs (ncRNAs) have emerged as important regulators of gene expression in eukaryotes. Most studies of ncRNAs in plants have focused on the identification of silencing microRNAs (miRNAs) and small interfering RNAs (siRNAs). Another important family of ncRNAs that has been well characterized in plants is the small nucleolar RNAs (snoRNAs) and the related small Cajal body-specific RNAs (scaRNAs). Both target chemical modifications of ribosomal RNAs (rRNAs) and small nuclear RNAs (snRNAs). In plants, the snoRNA genes are organized in clusters, transcribed by RNA Pol II from a common promoter and subsequently processed into mature molecules. The promoter regions of snoRNA polycistronic genes in plants are highly enriched in two conserved cis-regulatory elements (CREs), Telo-box and Site II, which coordinate the expression of snoRNAs and ribosomal protein coding genes throughout the cell cycle.

Results: In order to identify novel ncRNA genes, we have used the snoRNA Telo-box/Site II motifs combination as a functional promoter indicator to screen the Arabidopsis genome. The predictions generated by this process were tested by detailed exploration of available RNA-Seq and expression data sets and experimental validation. As a result, we have identified several snoRNAs, scaRNAs and 'orphan' snoRNAs. We also show evidence for 16 novel ncRNAs that lack similarity to any reported RNA family. Finally, we have identified two dicistronic genes encoding precursors that are processed to mature snoRNA and miRNA molecules. We discuss the evolutionary consequences of this result in the context of a tight link between snoRNAs and miRNAs in eukaryotes.

Conclusions: We present an alternative computational approach for non-coding RNA detection. Instead of depending on sequence or structure similarity in the whole genome screenings, we have explored the properties of promoter regions of well-characterized ncRNAs. Interestingly, besides expected ncRNAs predictions we were also able to recover single precursor arrangement for snoRNA-miRNA. Accompanied by analyses performed on rice sequences, we conclude that such arrangement might have interesting functional and evolutionary consequences and discuss this result in the context of a tight link between snoRNAs and miRNAs in eukaryotes.

Fig. 1

Computational analysis pipeline for the identification of ncRNA genes containing TeloSII elements in Arabidopsis

Fig. 2

Identification of the clustered scaRNAs ncR20 and ncR21. (a) Schematic representation of scaRNA cluster 7. The scaRNAs are denoted by gray pentagons. The TeloSII and the TATA elements are represented by boxes: Telo (open box), Site II (gray box), and TATA (black box). The primers designed for cluster 7 and the expected amplicon sizes are indicated by arrows and a solid line. Additional primers corresponding to the adjacent genes designed for evaluation of independency of transcription of cluster 7 are shown as dashed lines. (b) The sequence features of ncR20 and ncR21: motifs C/D and the target antisense region are indicated in red and green colors, respectively; double-stranded-forming regions at the termini are underlined. (c) Predicted targets of ncR20 and ncR21. The positions of predicted methylated residues are indicated bold letters and red color. (d) RT-PCR analysis of cluster 7 precursor expression in 2-week-old seedlings. RT+ and RT- indicate the presence or absence of reverse transcriptase in the reaction, respectively. Molecular markers are shown in lane labeled with M. Genomic DNA (lane labeled with G) was used as positive control. The expected RT-PCR product sizes are provided below the lanes with product amplified on genomic DNA. The absence of amplification signals in “Upstream” and “Downstream” panels indicates the independent from neighboring genes character of cluster 7 transcription

Fig. 3

Schematic representation and experimental validation of the sno-miR775 gene. (a) Gene organization of sno-miR775. snoR775 and miR775 are denoted by a gray pentagon and an open pentagon, respectively. The TeloSII elements and TATA box are represented by different boxes: Telo-box (open box), Site II (gray box), and TATA box (black box). Both snoR775 and pre-miR775 are covered by cDNA BX818024. The transcription start site (TSS) and transcription termination site (TTS) were determined by 5’ RACE and 3’ RACE, respectively. (b) The centroid secondary structure of snoR775, drawn by using the RNAfold program. The H and ACA boxes are denoted by green color. The complementary region to the predicted target 25S rRNA is represented by blue color, and details are shown in the blue dashed box. Ψ: pseudouridylation site. (c) Secondary structure of the sno-miR775 precursor sequence. Mature miR775 and the predicted snoRNA are in purple and green colors, respectively. (d) Northern blot analysis of sno-miR775. The hybridization was carried out in 3 different tissues: 2-wk-old seedlings, 3-wk-old leaves and 5-wk-old flowers. Lanes in the blots represent the following samples: dcl1-7, dcl3, dcl4, rdr2, and rdr6as well as wild-type Col0. U6 was used as the loading control

Fig. 4

Schematic representation and experimental validation of the sno-miR779 gene. (a) Gene organization of sno-miR779. The snoRNAs and miR779 are denoted by gray pentagon and open pentagon, respectively. The TeloSII elements and TATA box are represented by different boxes: Telo-box (open box), Site II (gray box), and TATA box (black box). The transcription start site (TSS) was determined by 5’ RACE. (b) 5’ RACE amplification of snoR779. A single major signal in 5’ RACE PCR indicates the transcription start site (TSS); the product size, confirmed by sequencing, was 815 nt. (c) Northern blot analysis of snoR129. The hybridization was carried out in 3 different tissues: 2-wk-old seedlings, 3-wk-old leaves and 5-wk-old flowers. The lines in the blots include the following samples: dcl1-7, dcl3, dcl4, rdr2, and rdr6 mutants and wild type Col0. U6 was used as the loading control

Fig. 5

The identification of ncR40. (a) Schematic representation of ncR40. The termini were examined by small RNA fragments enrichment. The TeloSII elements and TATA box are represented by different boxes: Telo-box (open box), Site II (gray box), and TATA box (black box). The primers designed for ncR40 and their expected sizes are indicated by arrows and solid lines. (b) RT-PCR validation of ncR40 in 2-week-old seedlings. RT+ and RT- refer to the presence and absence of reverse transcriptase, respectively. Molecular markers are shown as M. Genomic DNA, G, was used as a positive control. The expected RT-PCR product size is indicated below the lanes