MiRNAs originate from longer precursor RNAs, called primary miRNAs (pri-miRNA), which are regulated by conventional transcription factors and transcribed by RNA polymerase II. pri-miRNAs are hundreds to thousands of nucleotides long and are processed in the nucleus into a ~70-100 nucleotide hairpin-shaped precursor miRNA (pre-miRNA) by the RNase III enzyme Drosha and the double-stranded RNA binding protein DGCR8. The pre-miRNA is then transported into the cytoplasm by the nuclear export factor exportin 5 and further processed into an ~19-25 nucleotide double stranded RNA by the RNaseIII enzyme Dicer. This duplex miRNA is then incorporated into the RNA-induced silencing complex (RISC). One strand remains in the RISC and becomes the “mature” miRNA, while the other strand is often rapidly degraded and is called the “star” strand (miRNA*). Upon being loaded into RISC, the mature miRNA associates with target mRNAs and acts as a negative regulator of gene expression by promoting mRNA degradation or inhibiting translation3
. Translational inhibition seems to be the predominant mechanism in mammals, however target genes that are strongly downregulated on the protein level often show a reduced mRNA level4
, suggesting mRNA destablization is a major contributor to gene silencing.
A mature miRNA typically regulates gene expression via an association with the 3’UTR of an mRNA with complimentary sequence, although emerging evidence suggests miRNAs may also target 5’UTRs or exons, and may potentially even undergo base pairing with regulatory DNA sequences to regulate transcription. Upon miRNA binding to a 3’ UTR, the degree of transcriptional degradation and/or translational repression is affected by multiple mechanisms, including the overall complimentarity between the miRNA and target mRNA, the secondary structure of the adjacent sequences, the distance of the miRNA binding site to the coding sequence of the mRNA, and the number of target sites within the 3’UTR5
. Complimentarity between nucleotides 2 through 8 of the miRNA, termed the “seed” region, appears to be essential for 3’UTR identification. Therefore, miRNAs with high sequence homology and identical seed region are commonly grouped into miRNA families that are likely to target similar sets of mRNAs6
Up to 1000 miRNAs are predicted to exist in the human genome, each of which could potentially target hundreds of mRNAs. Most 3’UTRs contain potential binding sites for a large number of individual miRNAs, allowing for redundancy or cooperative interactions between various seemingly unrelated miRNAs. Furthermore, the targets of many miRNAs can modulate the expression of additional miRNAs or groups of miRNAs, generating positive or negative feedback loops. Finally, miRNA maturation seems to be post-transcriptionally regulated in a sequence specific manner7
, potentially explaining why genetically clustered and co-transcribed miRNAs are often expressed at different levels.
Multiple miRNA target prediction tools are now available (summarized in Supplemental Table
). Generally, in silico target prediction algorithms use a standard scheme to identify and rank potential targets8
. Briefly, potential targets are ranked based on the complimentarity between miRNA and 3’UTR and the degree of conservation of the miRNA and the 3’ UTR target sequence across species. A particular miRNA target is considered to be more meaningful if the sequence is evolutionarily conserved.
Identification and validation of miRNA targets remains a major hurdle in the study of miRNA function since many putative targets display little or no detectable regulation when tested in vitro. This is likely due, at least partially, to the relatively modest effect of any single miRNA on the translational output of the target mRNA. Therefore, many predicted miRNA binding sites are probably not true targets and experimental validation is essential for confirming target genes.
A surprising number of published miRNA targets do not conform to the traditional rules of target prediction outlined above. For example, in several instances, cross-species conservation is not observed in the target sequence, even between rodents and humans. Of course, for a miRNA target to be therapeutically viable, the miRNA target sequence must be conserved in humans, and not simply present in the model organism studied. Therefore, this review primarily focuses on published miRNAs that target 3’UTR binding sites that are conserved in humans.
A key insight into the mechanism of miRNA action has been that a large number of miRNAs apparently target multiple functionally related mRNAs. The coordinated regulation of multiple steps in a complex physiological process by one miRNA or a group of similarly expressed miRNAs is an important characteristic of miRNA biology that lends itself to therapeutic applications. In contrast to therapeutically modulating a single target with a conventional drug, miRNA biology can, in principle, modulate multiple levels of a pathological process by targeting a single nucleic acid molecule.