Exceptional stories of microRNAs

Introduction

MicroRNAs (miRNAs) are small, non-coding, endogenous RNAs that repress translation of genes post-transcriptionally. ^1,2 Most miRNAs are transcribed as large primary transcripts (pri-miRNAs) by RNA polymerase II. ³ These structured RNAs are then processed into 60–70 nt hairpin loop precursors (pre-miRNAs) by the nuclear RNase III protein, Drosha, in concert with the enzyme called DGCR8. ⁴ Pre-miRNAs are exported to the cytoplasm by Exportin-5. The pre-miRNAs are then processed again by another RNase III enzyme, Dicer, to generate duplex forms of miRNAs. Following the maturation of miRNAs, one strand of the duplex is incorporated into the RNA-induced silencing complex containing Argonaute 2 (Ago2), creating the inhibitory complex on the target mRNA (Figure 1). ^5–8 However, there are exceptions to conventional or canonical processing of miRNAs. This review focuses on those exceptions. As regulators of gene expression, miRNAs are involved in variety of cellular and developmental events and diseases. Therefore, it is extremely important to know how miRNAs exactly control their targets. OPEN IN VIEWER

In this review, six questions are chosen to illuminate non-traditional functions of miRNAs and the author summarizes recent, including bacteria, studies that might herald a new era of functional studies of miRNA.

Do miRNAs only bind to 3′ UTR of target mRNAs?

miRNAs have been extensively studied since their discovery in 1993. ⁹ Most miRNA functional studies are related to miRNAs that bind to the 3′ UTR of their target transcripts. More recent investigations have provided evidence for different sites of action. miRNA lin-4 was first discovered in Caenorhabditis elegans as a regulator of cell fate determination by its binding to lin-14 mRNA during development. ^9,10 Ever since, the target search for miRNA genes has generally focused on the 3′ UTR of the target gene. When a miRNA pairs extensively with its target, it directs cleavage of the target transcript (this occurs mostly in plants) while partial pairing between the target and miRNA inhibits protein synthesis. ¹¹ For example, murine miR-7b, which responds to osmotic stress, inhibits Fos (known as pivotal regulator of major biological events) translation, but not transcription, by binding to the 3′ UTR of the Fos gene. ¹² Several reports, however, have suggested that binding site of the target gene for miRNA is not restricted to the 3′ UTR: other regions of the mRNA and its gene are targeted, including the 5′ UTR, ¹³ promoter ¹⁴ and open reading frames. ^15–17 Conventionally, targeting to destroy mRNAs (siRNAs) was accomplished by the selection of the coding regions free of potentially interfering translational or regulatory proteins; i.e. anywhere between 100 nucleotides upstream and downstream of the start and the stop codons, respectively, to avoid interference by RNA-regulatory proteins that bind the 5′ or 3′ UTRs. ¹⁸ Since siRNAs and miRNAs share the same machinery, it is not surprising, therefore, that miRNAs bind to the coding region of the target mRNAs in addition to other areas. However, this presents difficulties for researchers when they attempt to predict miRNA target sites and emphasizes the need for more efficient computer algorithms.

Do miRNAs only have inhibitory functions?

RNA activation (RNAa) is a newly found mechanism of gene regulation directed by miRNAs ^14,19 and double-stranded RNAs (dsRNAs). ^19,20 In contrast to RNA inhibition (RNAi), RNAa is relatively uncharacterized phenomenon. Synthetic dsRNAs targeting promoter regions of human genes caused sequence–sequence specific induction of targeted genes. ¹⁹ Place et al. ¹⁴ also found that miR-373 could induce gene expression by binding to its promoter and competing with the promoter binding repressor. This RNAa function seems likely conserved across mammals including some primates, mouse and rat. ²⁰ It is postulated that miRNAs would need to translocate into the nucleus in order to target the genomic DNA region. Therefore, it is important to note that the hexanucleotide element at the 3′ end of miRNA can direct their nuclear localization (see below). ²⁰

Do miRNAs exist only in the cytoplasm of the cell?

The conventional function of miRNAs is focused on endogenous translational regulation in the single cell, especially in the cytoplasm. However, there is increasing evidence that miRNAs may play important roles by returning to the nucleus.

miR-29b is predominantly translocated into the nucleus and the distinctive hexanucleotide terminal motif of miR-29b acts as a transferable nuclear localization element. ²¹ miR-206 and a number of miRNAs are concentrated in the nucleus of rat myoblasts. ^22,23 Moreover, one of the small nucleolar RNAs (snoRNA) that processes ribosomal RNAs in the nucleolus, ACA45, also harbors miRNA-like functions. ²⁴

It is not clear whether mitochondria have their own miRNAs that are transcribed from the mitochondrial genome. A number of small non-coding RNAs with sizes similar to miRNAs have been identified. ²⁵ Indeed, miRNAs that were found in liver mitochondria may be involved in the regulation of apoptotic genes there. ²⁶

There is increasing evidence that miRNAs are also secreted and taken up by other cells.

miRNAs have been found in body fluids including serum and saliva and may be useful as biomarkers for diseases. ^27,28 Despite the high RNase activity in those fluids, endogenous serum and salivary miRNAs are more stable than exogenous miRNAs perhaps because the former, extracellular miRNAs are protected within lipoprotein vesicles. ^27,29 miR-150, which is abundantly expressed in monocytes, is secreted after being packaged into microvesicles. It can then be delivered to other cell types (endothelial cells) and effectively reduce c-Myb gene expression and enhance cell migration. These results suggested that miRNAs could serve as intercellular signaling molecules. ³⁰

Are Drosha and Dicer absolutely necessary for miRNA processing?

The canonical miRNA biogenesis pathway uses Drosha and Dicer RNase III enzymes to govern the maturation of miRNAs. ^5,31 Studies using knockouts of Dicer and Drosha genes in mice show that they are critical for normal development, differentiation and physiology. ³² In mice, loss of Dicer is lethal early in development ³³ due to abnormal functions of various tissues including cardiac ³⁴ and neural tissues. ³⁵

Since not all cells respond in the same way to loss of Drosha and Dicer, it is postulated that there might be a diversity of alternative miRNA biogenesis mechanisms that bypass the core enzymes Drosha or Dicer (see the review ³² ).

A dicer-independent miRNA biogenesis pathway that requires catalytic activity of Ago2 was recently discovered. ^36,37 Chong et al. ³¹ also found that a subset of miRNAs are generated by a Dicer-dependent, Drosha-independent mechanism. It has also been suggested that certain introns, called ‘mirtrons’, have structural features that resemble pre-miRNAs and may be processed into miRNAs without Drosha-mediated cleavage. ³⁸

Small nucleolar RNA-derived RNAs (sdRNAs) and tRNA-derived RNAs (tdRNAs) also contribute to the miRNA pool. Both sdRNAs and tdRNAs have alternative biogenesis pathways ³² suggesting this might be a more common process than was previously thought. Indeed, miRNA-size small RNAs in bacteria (discussed below) might be processed by other factors because no known homologues of Drosha or Dicer have been found in bacteria. These findings suggest miRNA biosynthesis cannot be restricted to the canonical pathway.

Are miRNA genes located only in intronic or intergenic regions of genomic DNA?

Many miRNAs are found in clusters and transcribed as polycistronic primary transcripts. These clustered miRNAs are often related to each other as a result of evolutionary duplication. ³⁹ miRNAs are usually found in three genomic locations: introns of protein-coding genes, introns of non-coding genes and exons of non-coding genes. ^39,40

Analysis of miRNA genes has shown that a large number of miRNAs are transcribed within introns of protein-coding genes. miRNAs expression is generally correlated with their neighbor or host genes within a distance of 50 kb. Intronic miRNAs are co-expressed with their host gene mRNA by sharing the same promoter. ⁴¹ This is an important location to consider, because deletion of intronic miRNAs along with the targeted genes when knockout mice are generated may lead to phenotypes erroneously attributed to the targeted gene. ^42,43

In the early days of miRNA research, it was reported that more than half of miRNAs were located within introns of either protein-coding or non-coding transcription units and less than 10% were present within exons of long non-protein-coding transcripts as well as mRNA-like non-coding RNAs. ⁴⁴ Ro et al. ⁴⁵ found 30% of miRNAs expressed in mouse testis were derived from exonic sequences.

Nevertheless, it is still possible that protein-coding exonic miRNAs were excluded bioinformatically when researchers analyzed cloned miRNAs simply because it is hard to differentiate between mature miRNAs and degraded forms of mRNA.

Are there miRNAs made in other than eukaryotic cells?

Certain fungi like budding yeasts have a homolog of Ago and miRNA-size small RNAs. ⁴⁶ Small RNAs with the same size as miRNAs in bacteria have not been reported yet, but small regulatory RNAs (sRNAs) 50–200 nt in length have been identified in a variety of bacterial species. ⁴⁷ Like miRNAs, sRNAs act as post-transcriptional regulators by interacting with the target transcripts through a variety of mechanisms, such as by changing RNA conformation and modulating the stability of mRNAs. These small RNAs regulate responses to changes in environmental stresses. ^48,49 Before miRNAs were known, RNA I (or species I, approximately 108 nt) was discovered in Escherichia coli to block ColE1 plasmid replication by base paring with the target, ^50,51 which resembles the mechanism of RNAi in miRNA functions.

Recently, my lab found a number of small RNAs from oral bacteria that have sizes similar to miRNAs and that fulfill the potential characteristics of miRNAs. ⁵² However, the function of these small RNAs remains to be elucidated. ⁵³ This research suggests miRNA machinery is not limited to eukaryotes and may be one of the most well-conserved biological systems for regulating gene expression from bacteria to animals and plants.

In addition to miRNAs in eukaryotic cells, the current miRNA registry (miRBase ver.18, www.mirbase.org) lists 406 mature viral miRNAs. ⁵⁴ Viral miRNAs play a key role in inhibiting antiviral immunity and can also promote cell transformation in culture. ⁵⁵ However, it has been suggested that the majority of viral miRNAs may use a different target strategy from host miRNAs, i.e. they may have evolved to regulate viral-originated transcripts rather than host transcripts. ⁵⁶

Conclusions

We are still in the early stages of miRNA research, so there are possibly more roles for miRNA than we originally thought. Currently, many studies focus on the emerging links of miRNAs to stem cell research and human diseases including cancer. Further investigations that aim at finding and elucidating the mechanisms involved in miRNA synthesis and function are important not only for basic research but for developing diagnostic tools and therapeutic strategies. In order to reach a full understanding of how miRNAs are generated and function, we must consider many aspects of miRNAs. To more thoroughly understand the miRNA world, researchers should not be constrained by current dogma about miRNAs.