Retro-miRs: novel and functional miRNAs originating from mRNA retrotransposition

Abstract

Background: Reverse-transcribed gene copies (retrocopies) have emerged as major sources of evolutionary novelty. MicroRNAs (miRNAs) are small and highly conserved RNA molecules that serve as key post-transcriptional regulators of gene expression. The origin and subsequent evolution of miRNAs have been addressed but not fully elucidated.

Results: In this study, we performed a comprehensive investigation of miRNA origination through retroduplicated mRNA sequences (retro-miRs). We identified 17 retro-miRs that emerged from the mRNA retrocopies. Four of these retro-miRs had de novo origins within retrocopied sequences, while 13 retro-miRNAs were located within exon regions and duplicated along with their host mRNAs. We found that retro-miRs were primate-specific, including five retro-miRs conserved among all primates and two human-specific retro-miRs. All retro-miRs were expressed, with predicted and experimentally validated target genes except miR-10527. Notably, the target genes of retro-miRs are involved in key biological processes such as metabolic processes, cell signaling, and regulation of neurotransmitters in the central nervous system. Additionally, we found that these retro-miRs play a potential oncogenic role in cancer by targeting key cancer genes and are overexpressed in several cancer types, including liver hepatocellular carcinoma and stomach adenocarcinoma.

Conclusions: Our findings demonstrated that mRNA retrotransposition is a key mechanism for the generation of novel miRNAs (retro-miRs) in primates. These retro-miRs are expressed, conserved, have target genes with important cellular functions, and play important roles in cancer.

Keywords: Cancer; Gene expression; Gene origination; Retrocopies; miRNAs.

Fig. 1

Screening microRNAs (miRNAs) originated by retrotransposition of protein-coding genes. A Summary of databases, datasets, and strategy used to identify retrocopy-derived miRNAs (retro-miRs). B Schematic representation of a parental gene, exonic miRNAs, retrocopies and retro-miRs. All three classes of retro-miRs, exon-junction retro-miRs (EJR-miR, in red), novel retro-miRs (NR-miR, in green), and retroposed retro-miRs (RR-miR, in orange) are represented. Tiny purple bars in retro-miR representations and harpins represent mismatches between them and their respective region in the parental gene

Fig. 2

Retrocopy-derived miRNAs (retro-miRs) originated in the primate lineage and have evolved under a functional constraint. A We identified 17 retro-miRs that originated from 11 parental genes (left side). On the right side, we show the conservation of these retro-miRs in human, chimpanzee, gorilla, orangutan, rhesus, and marmoset genomes. Gray markings indicate the absence of the sequence in that organism. B An alignment between primate, mouse, and rat genomes of a retro-miR (miR-4468, a retro-miR conserved among all primates) sequence. The regions conserved are highlighted in green (matches in black; mismatches in light gray), and “-” indicates a lack of conservation. The red boxes represent the mature mRNA sequence, and the bold sequences are the miRNA seed region (2nd to 8th nt). C Alignment of a human-specific retro-miR (miR-4426). Lack of conservation is presented by “-”. The red boxes represent the mature sequence, and the bold sequences are the miRNA seed region (2nd to 8th nt). D Retro-miRs in retrocopies with a dN/dS value significantly different than 1 (p-value <  = 0.05, likelihood ratio test) under purifying (dN/dS < 1) or positive (dN/dS > 1) selection are highlighted in black, while those with non-significant (or non-evaluated “NA”) dN/dS values are shown in gray

Fig. 3

Retrocopy-derived miRNAs (retro-miRs) are expressed in normal tissues and cell lines. A Expression (median per tissue) of retro-miRs in normal samples (404 individuals, 32 tissues) and cell lines (48 samples of stem cells) from the FANTOM phase 5 [29, 30]. B Table showing the presence (in green) or absence (in red) of a transcription start site (TSS) upstream of the retro-miR, a bubble plot representing the intensity of expression of retro-miRs, as well as the number of tissues in which they were expressed and presence of ENCODE Candidate Cis-Regulatory Elements (cCREs) nearby retrocopies hosting the retro-miRs. C Expression level per number of tissues of all known human miRNAs. Retro-miRs are marked in colors, while all other miRNAs are represented by gray dots (expression based on samples from FANTOM phase 5 [29, 30]). Only expressed retro-miRs are shown in A and C

Fig. 4

Retrocopy-derived miRNAs (retro-miRs) target genes are involved in key cell functions, including regulation of neural systems and have oncogenic roles. A Number of predicted (7mer-m8 and 8mer-1a) and experimentally validated target genes for retro-miRs. B Gene Ontology enrichment (Biological Processes) analysis of gene targets from retro-miRs. C Gene Ontology enrichment (Biological Processes) analysis of target genes of miR-572 and miR-4788. D Cancer-related KEGG pathways enriched in target genes of retro-miRs. Only enriched (> 1.5) and significant processes and pathways (FDR < 0.05) are shown

Fig. 5

Retrocopy-derived miRNAs (retro-miRs) are highly expressed in cancer and some retro-miRs present a cancer-specific expression. Expression level of (mature) retro-miRs in different types of tumors (BRCA = Breast cancer, COAD = Colon adenocarcinoma, KIRC = Kidney renal clear cell carcinoma, LIHC = Liver hepatocellular carcinoma, LUSC = Lung squamous cell carcinoma, LUAD = Lung adenocarcinoma, PAAD = Pancreatic adenocarcinoma, PRAD = Prostate adenocarcinoma, STAD = Stomach adenocarcinoma, THCA = Thyroid carcinoma) and their respective normal samples. Wilcoxon test Tumor x Healthy samples: p-value <  = 0.05 and >  = 0.01 (*); < 0.01 and >  = 0.0001 (**); < 0.0001 (***)

Fig. 6

Prognostic value of retrocopy-derived miRNA (retro-miR) expression in liver hepatocellular carcinoma and stomach adenocarcinoma. A Kaplan–Meier plot showing the overall survival differences (log-rank test) between two groups of patients with liver hepatocellular carcinoma, stratified based on expression levels of retro-miRs: miR-4444–2 and miR-3654. B Violin plot of the expression of miR-4444 and miR-3654 in the groups with better and worse overall survival probability presented in A. C Kaplan–Meier plot showing the overall survival differences (log-rank test) between two groups of patients with stomach adenocarcinoma, stratified based on expression levels of retro-miRs: miR-4444, miR-4788, and miR-622. D Violin plot of the expression of miR-4444, miR-4788, and miR-622 in the groups with better and worse overall survival probability presented in C