Expanding the immune-related targetome of miR-155-5p by integrating time-resolved RNA patterns into miRNA target prediction

Abstract

The lack of a sufficient number of validated miRNA targets severely hampers the understanding of their biological function. Even for the well-studied miR-155-5p, there are only 239 experimentally validated targets out of 42,554 predicted targets. For a more complete assessment of the immune-related miR-155 targetome, we used an inverse correlation of time-resolved mRNA profiles and miR-155-5p expression of early CD4+ T cell activation to predict immune-related target genes. Using a high-throughput miRNA interaction reporter (HiTmIR) assay we examined 90 target genes and confirmed 80 genes as direct targets of miR-155-5p. Our study increases the current number of verified miR-155-5p targets approximately threefold and exemplifies a method for verifying miRNA targetomes as a prerequisite for the analysis of miRNA-regulated cellular networks.

Keywords: CD4+ T cell activation; miR-155-5p; miRnomics; target prediction; time-resolved RNA patterns; transcriptomics.

Figure 1.

Implementation of time-resolved, inverse-correlated miRNA and mRNA expression data in prediction and validation by HiTmIR. Diener et al. investigated the time resolved mRNA and miRNA expression profiles of early CD4+ T cell activation [15]. We predicted the miR-155-5p target genes using miRwalk2.0 and used inverse correlation of the predicted target genes and the time resolved mRNA data for target prediction. For further validation by HiTmIR, we randomly chose 103 potential target gene 3’UTR sequences.

Figure 2.

Analysis of the predicted inverse correlated miR-155-5p target genes by HiTmIR. (A): target genes with an RLU ≤ 70% and (B) target genes with an RLU ≥ 70%: HEK 293T cells were transfected with 50 ng/well of either reporter plasmid pMIR-RNL-TK, with or without insert, and 200 ng/well of miRNA expression plasmid containing either the respective miRNA or no insert. The RLU of the miR1555p transfected samples were normalized to the RLU measured with empty reporter vector. Four independent experiments were conducted in duplicates. Columns coloured in magenta show a significant reduction of the luciferase activity with a p-value ≤ 0.001. Columns coloured in green show a significant reduction of the luciferase activity with a p-value ≤ 0.01 and ≥0.001. Columns coloured in violet show a significant reduction of the luciferase activity with a p-value ≤0.05. Columns coloured in light blue show a non-significant reduction of the luciferase activity with a p-value ≥0.05. Data are shown as mean±sem. (C): distribution of the miR-155-5p effect on reporter plasmids. The analysed reporter plasmids were categorized by the detected RLU. (D): distribution of all positive target genes by RLU and p-value. Turquoise: highly significant reduction of RLU ≤ 85% with a p-value ≤ 0.001, magenta: significant reduction of RLU ≤ 85% with a p-value ≤ 0.01 and ≥0.001, Black: negative target genes. (E) Validation rate miR-155-5p. A positive target gene is defined by an RLU ≤ 85% with a p-value ≤ 0.01 (turquois). (F) Comparison of MTIs of miR-155-5p deposited in miRtarbase with positive MTIs from this study. Turquois: number of MTIs deposited in miRtarbase, Beige: number of positive MTIs from this study.

Figure 3.

(A) validation of miR-155-5p target genes by mutagenesis HiTmIR. HEK 293T cells were co-transfected with miR-155-5p expression plasmids and wild-type reporter vectors of the respective target genes (light grey) or mutated reporter vectors (−mut) of the respective target genes (blue) as shown in the diagram. Three independent experiments were conducted in duplicates. Three asterisks represent a significant reduction of the RLU with a p-value ≤0.001. Two asterisks represent a significant reduction of the RLU with a p-value ≤0.01 and ≥ 0.001. One asterisk represents a significant reduction of the RLU with a p-value ≤0.05. (B) Association between binding site type and RLU reduction. The box coloured in brown represents 6-mer binding sites. The box coloured in magenta represents 7-mer-A1 binding sites. The box coloured in green represents 7-mer-m8 binding sites. The box coloured in turquoise represents 8-mer binding sites.

Figure 4.

Protein–protein interaction networks of 535 inverse-correlated miR-155-5p targets using the STRING database version 11.5. The miR-155-5p targets genes which are associated with the displayed pathways are coloured as indicated in the diagram.