Small Worms, Small RNAs, Big Questions

Image: Courtesy of Frank Slack SMALL, YET POTENT: A new and intriguing class of small RNAs can regulate eukaryotic gene expression. But scientists are trying to understand which signals cause these molecules to repress translation (left), and which cause RNA degradation (right). The answer could pave the way for gene therapy advances, among others. Andrew Fire and colleagues first described RNA interference (RNAi) in Caenorhabditis elegans in 1998.1 Exogenous double-stranded RNA (dsRNA) injected into cells effectively induced silencing of an endogenous gene complementary to the injected RNA. Since then, scientists have used RNAi to silence genes in several organisms, and biotech companies have introduced products to assist them.2,3 In recent months, a number of groups have published data linking the RNAi pathway to a newly discovered class of small RNAs, called microRNAs (miRNAs), the functions of which are largely unknown. The discovery of these small RNAs in a variety of organisms has led researchers to speculate if and how the biologies of RNAi and miRNAs intersect. "How closely related are they, and what determines the distinction between them?" asks Victor Ambros of Dartmouth Medical School, relating two of the main questions currently under investigation. "Also, what pathways are these miRNAs involved in, and can we learn more biology by following their trails?" DICER IS THE COMMON LINK RNAi involves the cleavage of longer, exogenous dsRNA into small interfering RNAs (siRNAs), roughly 21-23 nucleotides (nt) long. Scientists believe the siRNAs then become incorporated into an RNA/protein complex known as RNA-induced silencing complex (RISC), which acts to recognize an mRNA target for subsequent cleavage.4 Generation of siRNA requires the Dicer enzyme, a ribonuclease whose biological function scientists are now beginning to understand. Two small temporal RNAs (stRNAs), products of lin-4 and let-7, regulate developmental timing in C. elegans.5,6 Both genes encode nontranslated precursor RNAs roughly 70 nt long with a characteristic stem-loop structure. These longer RNAs are processed into smaller stRNAs, which act to suppress translation of target genes by binding to complementary sequences in the 3' untranslated regions of the target mRNAs.5,6 Last year, three groups demonstrated that Dicer orthologs were required for processing of the let-7 precursors in Drosophila and C. elegans and for the activities of lin-4 and let-7 in C. elegans.7-9 The size, sequence, and temporal regulation of let-7 are conserved in organisms ranging from worms to humans.5,6 In addition, sequences complementary to let-7 were found in the Drosophila and zebrafish orthologs of one let-7 RNA target, lin-41. "This novel mechanism of gene regulation by small RNAs appears to be ancient, being present in the common ancestor of nematodes and vertebrates," says Frank Slack of Yale University. Interestingly, lin-4 and let-7 were recently found to be part of a broader class of noncoding miRNAs 22-24 nt long.10-13 Most of these small RNAs had not previously been identified, and their genes are not located on annotated parts of the genome. "When we cloned these RNAs, we were intrigued to find that they were coming from parts of the genome that shouldn't have any genes. In fact, there are genes there but people hadn't identified them, because they don't have the common features of a normal protein-coding gene," says David Bartel of the Whitehead Institute for Biomedical Research and Massachusetts Institute of Technology, whose group identified 55 miRNAs in C. elegans.10 Some of the miRNAs are also expressed in mammals and other organisms, and some, like lin-4 and let-7, are expressed at distinct developmental stages,10,14 but most of the miRNAs identified are of unknown function. "We already know that [lin-4 and let-7] are expressed only at particular times during development, whereas many of the other microRNAs are expressed at all times, so it's not clear yet whether these new microRNAs control other things, maybe metabolism or housekeeping gene expression," says Slack. Bartel adds that more miRNAs are likely to be discovered, perhaps representing more than 100 different nematode genes. PATHWAYS CONVERGE, THEORIES EMERGE It has been well established that the Dicer enzyme processes precursor RNAs to make stRNAs, miRNAs, and siRNAs. The siRNA pathway leads to cleavage of target mRNA, whereas the stRNA pathway leads to translational repression. How does the cell handle the different Dicer products and reach different functional outcomes? A single pathway may handle small RNAs, with the outcome of the pathway depending on the structure of the RNA and/or the amount of sequence homology it has to its target. Alternatively, two separate pathways may be involved. If that is the case, asks Phillip Zamore of the Department of Biochemistry and Molecular Pharmacology, University of Massachusetts Medical School, "How does the cell know the difference between an siRNA and a microRNA? Is it because small temporal RNAs are single-stranded, and siRNAs are double-stranded, or is it because microRNAs have special sequences and we just don't know the hallmarks of those sequence features?" Zamore notes that a significant difference between the two types of small RNAs occurs at the end of the pathway, as siRNAs are nearly perfectly complementary to their targets, while stRNAs pair to their targets with bulges and mismatches. This difference could be sufficient to distinguish the two outcomes. The counterargument is that the siRNA and stRNA pathways in C. elegans require different members of the Argonaute/PPD (PAZ/Piwi domain) protein family. For example, Hirokai Tabara and colleagues demonstrated, using RNAi, that null mutations of argonaute family member rde-1 eliminated RNAi function in C. elegans but did not negatively affect growth or development.15 Furthermore, Alla Grishok and colleagues showed last year that inactivation of the related genes alg-1 and alg-2 were necessary for proper functioning of the lin-4 and let-7 stRNAs in C. elegans, but that inactivation of these genes had no affect on RNAi function in worms.9 "These are proteins that presumably act very early in the pathway," says Zamore. "It could be that very early in the pathway, the choice of PPD protein determines the outcome." Photo: Courtesy of Marlena Emmons Gregory Hannon Several researchers favor this model. Gregory Hannon of Cold Spring Harbor Laboratory, NY, suggests that siRNAs and miRNAs may preferentially associate with different Argonaute proteins to form complexes that have specific cellular functions. "There could be a core complex design--an Argonaute protein and a small RNA--that acts as a platform for finding complementarity," Hannon says. Craig Mello of the University of Massachusetts Program in Molecular Medicine adds that the presence of 26 rde-1 homologs in C. elegans12 points to the possibility that each forms a distinct complex with a specific small RNA. "One interesting model would be that ... multiple different types of complexes form, and the rde-1 homologs interact with the small RNAs that guide these distinct complexes in their different activities." RNAi is an effective means of targeted gene silencing, but the intersection of the miRNA and siRNA processing mechanisms raises questions about the technique's appropriateness for applications such as gene therapy. Says Mello, "One concern is that the machinery is shared by several different mechanisms; if you start to overload the system you could end up shutting down important developmental mechanisms." He notes, however, that RNAi functions very precisely and with no apparent side effects in nematodes. Zamore points out another potential problem. If the difference between the two pathways stems from differences in complementarity to the target mRNA, then any small RNA has the potential to either destroy a target mRNA or to turn off translation of that target, depending on the presence or absence of mismatches. "If you design a small interfering RNA that can destroy a mutant gene but not a wild-type gene, you have to ask whether the mutant-specific small interfering RNA--which has a small mismatch when it binds to the wild-type gene and no longer destroys that mRNA--will it nonetheless function to turn off translation? If it does, your strategy is not going to work," he explains. The mutant RNA will be destroyed, but neither gene will be expressed. THE BEST OF BOTH WORLDS In small-RNA research, new developments are a daily occurrence. Mello's group, for example, has a paper in press exploring the roles of rde-1 and rde-4, which are required for RNAi but not for miRNA processing. These proteins, Mello says, "look like they're involved upstream of Dicer in the recognition of foreign dsRNA. Clearly, [the mechanisms] are distinct upstream of Dicer in that there are genes in C. elegans that are required for RNAi but not for microRNA function." Other experiments currently underway include a collaborative effort between the labs of Bartel, Ambros, and MIT biology professor H. Robert Horvitz, to generate C. elegans deletion mutants by knocking out miRNA genes. "With these mutants we'll be able to explore the question of what the function of these microRNAs may be," says Ambros. Yet to be explored is the possibility that miRNAs play a role in chromatin silencing. "Some of the mechanisms that are involved in transcriptional gene silencing and chromatin silencing are exactly the same," says Mello, "so it makes sense to me that this type of mechanism, perhaps involving small-RNA guides to remodel and to silence chromatin, are important in development. And it's also possible that there are small RNAs that guide complexes that can desilence genes or maybe activate the translation of a target gene." MicroRNAs may also be involved in medically significant regulatory mechanisms. "We may find that many of the miRNA genes in mammals or vertebrates turn out to be involved in regulating gene expression in a cell cycle context or tissue homeostasis context ... some of these might be tumor suppressors or oncogenes. And they've been previously misidentified because the genes were so small and people haven't traditionally been looking for RNA genes," says Ambros. The collaborative gene knockout project will help elucidate the functions of these genes. An interesting prediction regarding the role of Dicer in RNAi and miRNA expression is that small RNAs engineered to resemble stRNA precursors would be able to induce RNAi.8 Several groups recently demonstrated this in Drosophila and mammalian cells, solidifying the link between the miRNA and siRNA pathways.16,17 These results suggest that the different functions of stRNAs and siRNAs may depend on the relationships between the small RNAs and their targets, and that the miRNA and siRNA pathways may in fact "be interchangeable, from biogenesis of the RNA itself, all the way to interaction with its target," says Ambros. Cellular processing of small RNA represents a new, highly specific mechanism for regulating gene expression--one that can be exploited as a research tool and possibly as a therapeutic method. The genes involved appear to be widely conserved among a variety of organisms and may be part of an ancient mechanism. Zamore suggests that gene regulation via small RNAs demonstrates how cells solved the problem of molecular specificity, using RNAs to provide specificity at the nucleic acid level and proteins to provide specificity at the functional level. "In each case [stRNA and siRNA function], you have this small piece of RNA ... that lets these small molecular machines be incredibly specific, despite the fact that they're composed of proteins with no intrinsic specificity for what they destroy and what they translationally regulate," says Zamore. "By combining [RNAs and proteins], cells have really exploited the best of both molecules." Aileen Constans (aconstans@the-scientist.com) is a contributing editor. References 1. A. Fire et al., "Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans," Nature, 391:806-11, 1998. 2. H.E. Sussmann, J.L. Pierce, "New tools aid RNA interference studies," The Scientist, 16[2]:44, Jan. 21, 2002. 3. A. Constans, "RNAi for the masses," The Scientist, 16[9]:36, April 29, 2002. 4. S.M. Hammond et al., "An RNA-directed nuclease mediates post-transcriptional gene silencing in Drosophila cells," Nature, 404:293-6, 2000. 5. A.E. Pasquinelli et al., "Conservation of the sequence and temporal expression of let-7 heterochronic regulatory RNA," Nature, 408:86-9, 2000. 6. D. Banerjee, F. Slack, "Control of developmental timing by small temporal RNAs: A paradigm for RNA-mediated regulation of gene expression," BioEssays, 24:119-29, February 2002. 7. R.F. Ketting et al., "Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans," Genes and Development, 15:2654-9, Oct. 15, 2001. 8. G. Hutvagner et al., "A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA," Science, 293:834-8, Aug. 3, 2001. 9. A. Grishok et al., "Genes and mechanisms related to RNA interference regulate expression of the small temporal RNAs that control C. elegans developmental timing," Cell, 106:23-34, July 13, 2001. 10. N.C. Lau et al., "An abundant class of tiny RNAs with probable regulatory roles in Caenorhabditis elegans," Science, 294:858-62, Oct. 26, 2001. 11. R.C. Lee, V. Ambros, "An extensive class of small RNAs in Caenorhabditis elegans," Science, 294:862-4, Oct. 26, 2001. 12. M. Lagos-Quintana et al., "Identification of novel genes coding for small expressed RNAs," Science, 294:853-8, Oct. 26, 2001. 13. H. Tabara et al., "The rde-1 gene, RNA interference, and transposon silencing in C. elegans," Cell, 99:123-32, 1999. 14. M. Lagos-Quintana et al., "Identification of tissue- specific microRNAs from mouse," Current Biology, 12:735-9, April 30, 2002. 15. Z. Mourelatos et al., "miRNPs: A novel class of ribonucleoproteins containing numerous microRNAs," Genes and Development, 16:720-8, March 15, 2002. 16. T. Tuschl, "Expanding small RNA interference," Nature Biotechnology, 20:446-448, May 2002. 17. J.-Y. Yu et al., "RNA interference by expression of short-interfering RNAs and hairpin RNAs in mammalian cells," Proceedings of the National Academy of Sciences, 9:6047-52, April 30, 2002. A CELLULAR CPU? Small interfering RNAs (siRNAs) may represent the tip of a gene-expression iceberg, according to John Mattick, University of Queensland, Brisbane, Australia. Noncoding RNAs could regulate phenotypic diversity in higher eukaryotes, serving as a "higher order control architecture" similar to complex information systems--like computers--that orchestrate protein expression during differentiation and development.1 Mattick argues that what is striking about the evolution of higher organisms is not the increased numbers of protein-coding genes but the expansion of noncoding transcription. "You've either got to assume that the genomes of complex organisms are replete with useless transcription or that there's something else going on," Mattick says. Mattick says it is likely that RNA signals are acting as cellular "memory." "When a gene is expressed, that event is relatively blind to the rest of the system," he says. Bacteria generally detect such events through feedback loops. "But what if you had a system that evolved where, at the same time that the protein-coding sequences are being transcribed and translated, there were secondary signals produced that ... could tell the rest of the system what had just happened," thus providing molecular memory of the event? Such a system would act as a communication network, conveying information to cellular components in real time and providing a means to coordinate and integrate this information for the orchestration of future gene-expression events. Mattick's theory is based largely on circumstantial, but compelling, evidence of an informational role for noncoding RNA. "Nobody would disagree that RNA-based gene regulation is far more widespread and sophisticated in the higher organisms than anyone expected," Mattick says. "And all of the complex genetic phenomena in those organisms have now been shown to be RNA-mediated." Mattick has yet to discover data that contradicts his hypothesis. "I think John's ideas are good speculations at this point, but not yet supported by the data," says Sean Eddy, Washington University School of Medicine, St. Louis. Both researchers agree that the role of noncoding RNAs has been largely unexplored, primarily because functional genomics studies are biased towards proteomics. "[T]here are genes that don't make proteins," explains Eddy, "but people assume that it's just ... ribosomal and transfer RNA." The absence of noncoding RNA in gene-discovery experiments stems in part from the way these studies are done: cDNA libraries exclude noncoding RNAs. "This selection for polyadenylated messengers takes over 95% of the cell's RNA and throws it away," says Eddy, regardless of the RNA's functionality. And, protein-coding gene identification is favored because such genes have unique statistical signatures. Though the functions of most noncoding RNAs remain elusive, Eddy says researchers can make a pretty good guess. RNA, he notes, is very good at recognizing other RNA molecules by base pairing. Proteins, on the other hand, require complex RNA-binding domains in order to recognize a single RNA transcript in the cell. "In tasks where all you really need to do is recognize something--another RNA--RNA is probably the material that evolution could take advantage of the fastest," says Eddy. He suggests, for instance, a role in splicing decisions. "RNA could very easily affect the choice of splice site just by binding to particular splice sites and blocking them." The functions of noncoding RNAs will be better understood as new tools for discovering and analyzing RNA transcripts are developed. But also, functional genomicists must look beyond proteomics. Mattick concludes, "It escaped our notice that the important part of the evolution of complex organisms was the evolution of a complex control architecture to orchestrate the expression of [proteins]." 1. J.S. Mattick, M.J. Gagen, "The evolution of controlled multitasked gene networks: The role of introns and other noncoding RNAs in the development of complex organisms," Molecular Biology and Evolution, 18:1611-30, 2001.

Small Worms, Small RNAs, Big Questions

Interested in reading more?

Become a Member of