MicroRNAs: New players in an old game

One of the cardinal steps in regulating gene expression is mRNA decay, and the numerous pathways and mechanisms that exist to regulate it underscore its importance. mRNA decay is regulated by trans-acting factors that assemble on cis-acting elements (1, 2). Together, they serve to up- or down-regulate a given mRNA. Some of the mechanisms that regulate mRNA levels involve surveillance pathways such as nonsense-mediated decay (NMD) and nonstop decay (NSD). The NMD pathway limits accumulation of mRNAs that contain a premature termination codon and whose translation would produce a truncated protein. In NSD, mRNAs that do not contain a termination codon because of improper poly(A) site selection within the coding region are rapidly degraded by the exosome, a complex of 3′→5′ exoribonucleases (3). Other pathways involve recognition of 3′ UTR sequences by specific RNA-binding proteins. For example, in AU-rich element (ARE)-mediated mRNA decay (AMD), binding of specific ARE recognition proteins to the 3′ UTR initiates mRNA degradation (1, 4). To one degree or another, all these decay pathways involve the step-wise deconstruction of a mRNA involving 3′→5′ trimming of the poly(A) tail, a process referred to as deadenylation; this is followed by removal of the 5′ m⁷GpppG cap and both 5′→3′ and 3′→5′ degradation of the mRNA body (5–7). This step-wise mechanism, first elucidated in Saccharomyces cerevisiae, has been recognized for some time now. Another mRNA decay pathway that has garnered much attention lately is RNA interference (RNAi). First discovered in Caenorhabditis elegans (8), RNAi has now been observed in several other multicellular organisms, including mammals. RNAi is triggered either by a small interfering RNA (siRNA) or, in some cases, by a microRNA (miRNA) that induces mRNA degradation via endoribonucleolytic cleavage within the site of si/miRNA–mRNA annealing. siRNAs derive from sources such as double-stranded RNA, transposons, and viruses and are perfectly complementary to their mRNA targets (9–11). miRNAs are ≈22 nt in length and are encoded within the genomes of both plants and animals. miRNAs contain regions possessing imperfect complementarity to 3′ UTRs of mRNA subsets to which they anneal. This leads to translational silencing to posttranscriptionally control gene expression (12). In this issue of PNAS, Wu et al. (13) demonstrate that a miRNA can also promote rapid mRNA degradation by accelerating the initial rate-limiting step, deadenylation.

The life of a miRNA begins as a miRNA precursor called primary miRNA (pri-miRNA). In metazoans, the enzyme Drosha catalyzes the first cleavage event that results in the production of a pre-miRNA intermediate. With the assistance of Exportin5, this intermediate travels to the cytoplasm, where it undergoes further cleavage by a second enzyme called Dicer. Cleavage by Dicer leaves an RNA duplex that is unwound, and the so-called guide strand, which contains complementarity to mRNA targets, assembles with Argonaute proteins and others to form the RNA-induced silencing complex (RISC) (14). A large body of evidence indicates miRNAs to be translational repressors and siRNAs to be purveyors of mRNA degradation. A few exceptions to translational silencing by miRNAs have begun trickling into the literature, however. For example, mammalian miR196a is perfectly complementary to one of its target transcripts, HOXB8, except for a single G:U wobble base pair (Fig. 1) (11). miR196a directs endoribonucleolytic cleavage of HOXB8 mRNA, which encodes one member of a group of related transcription factors involved in animal development. Although miR196a exhibits near-perfect complementarity to its target mRNA, miR125b and the miRNA known as let-7, do not (Fig. 1) (15). Although Wu and Belasco (15) demonstrated that miR125b target association did repress translation, this miRNA surprisingly led to reduced mRNA levels as well. In work presented here, they demonstrate that miR125b and let-7 increase mRNA decay rates upon association with their target mRNAs, not by endoribonucleolytic cleavage but rather by promoting rapid deadenylation.

Wu et al. (13) conducted the bulk of their experiments using chimeric reporter β-globin genes containing two copies of either the lin-28 miRNA responsive element (miRE) for miR125b or the distinct lin-28 miRE for let-7. These chimeric mRNAs decayed much faster in cells expressing the respective miRNAs than in those that did not express them. RT-PCR analysis of reporter mRNAs containing either the miREs or a synthetic element (referred to as element P), to which miR125b bore perfect complementarity, nicely eliminated the possibility that endoribonucleolytic cleavage was responsible for initiating degradation of miRE-containing mRNAs, because no internal cleavage product could be identified. By contrast, element P did direct endoribonucleolytic cleavage of the control mRNA. Next, the authors (13) determined that rapid deadenylation was the initiating event in mRNA degradation conferred by the miREs. Removal of the poly(A) tail, rather than decapping, preceded the mRNA’s decay, because RNA purified from cells was resistant to 5′-exoribonucleolytic digestion in vitro; this demonstrated the 5′ cap structure to be in place on purified mRNAs. Microarray experiments identified 21 mRNAs besides lin-28 whose levels were reduced by miR125b expression and that had the potential to interact productively with this miRNA (e.g., Ajuba, which controls entry of vertebrate cells into mitosis and meiosis and MKK7, mitogen-activated protein kinase kinase 7; Fig. 1). Assays with predicted miREs from these transcripts revealed miRE-dependent deadenylation as well, demonstrating that miR125b (and probably let-7) promotes deadenylation-dependent degradation of a number of transcripts. Importantly, experiments using three cell lines yielded consistent results, suggesting generality of this miRNA function. For brevity, we will hereafter refer to this entire decay process as miRNA-mediated mRNA degradation, or MMD.

Many studies over the years have suggested links between translation and mRNA decay (16). Thus, Wu et al. (13) next investigated whether the increase in deadenylation promoted by miR125b and let-7 was the result of decreased translation, a simultaneous effect also controlled by these miRNAs. To accomplish this, translation of the reporter miRNA-containing mRNA was abrogated by inserting a strong stem–loop structure in the 5′ UTR. Deadenylation of this mRNA was only marginally impaired, however. Interestingly, the authors also found that a poly(A) tail was a requisite feature for mRNA to undergo degradation but not translation arrest, because reporter mRNAs containing a histone stem–loop in place of a poly(A) tail were stable but translationally repressed. Finally, they observed (13) that repression of translation and reduction in mRNA levels conferred by miR125b were additive. This finding has important implications for miRNA-controlled gene expression, because highly effective silencing of some genes could result from the combined actions of translational repression and mRNA decay (and, of course, transcriptional silencing as well).

The results of Wu et al. (13) raise a number of important questions. For example, how does an imperfectly complementary miRNA direct mRNA degradation at all, whether it be deadenylation or endoribonucleolytic cleavage? Clues may come from the study of Jing et al. (17), who demonstrated that miR16 participates in degradation of TNF-α mRNA by the AMD pathway.

miRNA can promote rapid mRNA degradation by accelerating the initial rate-limiting step, deadenylation.

Apparently, the actions of miR16 require the ARE-binding protein tristetraprolin, which acts to recruit a deadenylase to its target mRNAs (18). Thus, by virtue of their interactions within RISC or with other factors, miRNAs could recruit either the deadenylase itself or an intermediate player (e.g., tristetraprolin), which in turn recruits the deadenylase. Another question one is forced to ask is, why are the mRNAs examined deadenylated and not simply cleaved by an endoribonuclease? The most conspicuous difference between deadenylation and endoribonucleolytic cleavage is that the latter is irreversible, whereas the former is not. By evolving deadenylation over cleavage for some transcripts, the cellular gene expression machinery is provided an additional branch within a decision tree that presents an opportunity to either re-add the poly(A) tail to permit translation or degrade the mRNA. [In fact, cytoplasmic polyadenylation is a major regulator of gene expression during development (19).]

The profile of mRNA decay products observed by Wu et al. (13) unequivocally points to a deadenylase. Is the deadenylase one of the enzyme complexes we already know (20–23), or will it be a new addition to this family of 3′→5′ poly(A)-specific exoribonucleases? Perhaps a protein recruited by miR125b and let-7 unleashes an intrinsic deadenylation activity of RISC itself. After deadenylation, do the mRNAs undergo decapping followed by 5′→3′ and 3′→5′ exoribonucleolytic degradation? Many, if not all, of the enzymes required for these processes are known as well, so it should be fairly straightforward to examine the effects of their knockdown on MMD.

Finally, where in the cell does deadenylation-dependent MMD occur? Processing bodies (P-bodies) are discrete cytoplasmic foci that are centers for mRNA decay (24). They house mRNA decapping enzymes, exoribonucleolytic enzymes, and components of RISC (12, 24, 25). So far, the only deadenylase found in these structures is Ccr4 (26). Because P-bodies contain many of the enzymes of mRNA decay, they appear to be the cell’s way of concentrating these factors in the same location. This undoubtedly provides an enzymatic assembly line that orchestrates the step-by-step deconstruction process. Nonetheless, although it is perhaps unlikely, we should not exclude the possibility that RISC may not be involved in deadenylation-dependent MMD at all; maybe there is a novel sister complex that is necessary and sufficient for deadenylation and translational suppression of mRNA subsets. Regardless of the outcomes, fasten your seat belts, because answers to these and other exciting questions are sure to arrive fast and furious in the coming months.