Genomic analysis suggests that mRNA destabilization by the microprocessor is specialized for the auto-regulation of Dgcr8

Abstract

Background: The Microprocessor, containing the RNA binding protein Dgcr8 and RNase III enzyme Drosha, is responsible for processing primary microRNAs to precursor microRNAs. The Microprocessor regulates its own levels by cleaving hairpins in the 5'UTR and coding region of the Dgcr8 mRNA, thereby destabilizing the mature transcript.

Methodology/principal findings: To determine whether the Microprocessor has a broader role in directly regulating other coding mRNA levels, we integrated results from expression profiling and ultra high-throughput deep sequencing of small RNAs. Expression analysis of mRNAs in wild-type, Dgcr8 knockout, and Dicer knockout mouse embryonic stem (ES) cells uncovered mRNAs that were specifically upregulated in the Dgcr8 null background. A number of these transcripts had evolutionarily conserved predicted hairpin targets for the Microprocessor. However, analysis of deep sequencing data of 18 to 200nt small RNAs in mouse ES, HeLa, and HepG2 indicates that exonic sequence reads that map in a pattern consistent with Microprocessor activity are unique to Dgcr8.

Conclusion/significance: We conclude that the Microprocessor's role in directly destabilizing coding mRNAs is likely specifically targeted to Dgcr8 itself, suggesting a specialized cellular mechanism for gene auto-regulation.

Figure 1. Transcripts differentially regulated in Dgcr8 KO relative to WT and Dicer KO ES cells.

The sets of genes differentially up- and down- regulated in Dgcr8 KO relative to Dicer KO and WT ES cells were determined based on a cutoff of FDR <5%. Data are represented as a mean of 3 biological replicates of WT, Dgcr8 KO and Dicer KO arrays. Transcripts positive for EvoFold hairpin predictions and transcripts with 5 or more small RNAs mapping to their exons are shown (see legend). Arrow points to Dgcr8 expression levels, which, as expected, is down in Dgcr8 KO (exon 3 deletion results in premature termination codon and, hence, non-sense mediated RNA decay .

Figure 2. The distribution of reads across hairpins in the first exon of Dgcr8 in mES cells.

The location of each small RNA read relative to the exon is represented by a grey bar and was generated using the custom tracks feature on the UCSC genome browser. For each RNA species, the number of reads that were obtained with that sequence is indicated at the left. The predicted secondary structure is represented below the genomic sequence. Genomic coordinates are based on UCSC Known Genes annotations (mm8) (A) 5′UTR hairpin (B) CDS hairpin. Small RNA reads in WT cells are represented by black bars and small RNA reads in Dicer KO cells are represented by a grey bar.

Figure 3. Representative examples of read distribution in exons with >5 reads in WT cells.

The location of unique small RNA reads from WT (black bars), Dgcr8 KO (dark grey bars) and Dicer KO (grey bars) are represented. For each RNA species, the number of reads that were obtained with that sequence is indicated at the left. Genomic coordinates are based on UCSC Known Genes annotations (mm8) (A) Example showing reads that are localized to a small window consistent with Microprocessor cleavage but are not Dgcr8-dependent (B) Reads are distributed across the range of the exon and most likely represent degradation products.

Figure 4. Read distribution across hairpins in the first exon of Dgcr8 in <200nt small RNA sequencing data from HeLa and HepG2 cells.

Small RNA locations are presented as in Figure 2. Genomic coordinates are based on UCSC Known Genes annotations (hg18) (A) 5′UTR hairpin (B) CDS hairpin.

Figure 5. Read distribution across hairpins positive for >5 small RNA reads in HeLa cell <200 nt small RNA sequencing data.

Small RNA locations are presented as in Figure 2. Genomic coordinates are based on UCSC Known Genes annotations (hg18) Reads across hairpins in (A) RPS3 (B) HIST1H4C (C) RHOB and (D) RPS8.