RNA Calls the Shots

Courtesy of Vasudeva Mahavisno CANCER CARTOGRAPHY: Metastatic prostate cancer from a tissue array stained for EZH2 protein. The background represents gene expression signatures of prostate cancer as a heat map, which lead to the discovery of EZH2 as a prostate cancer biomarker. As soon as Watson and Crick deduced DNA's structure half a century ago, their thoughts turned to RNA. Arguably the most important molecule in the living world, RNA not only connects gene to protein, but its catalytic capability inspired the idea of an "RNA world," where the single-stranded nucleic acid, or something like it, enabled chemistry to become biology. Today, RNA is still revealing its secrets. Recent reports indicate that there's more to the informational content of a messenger RNA (mRNA) than the mere mirror of a gene's DNA base sequence. These eclectic molecules interact directly with metabolites, control the levels of other transcripts just by their own abundance, and even alert the translational machinery to block production of potentially toxic, stunted proteins. Deranged transcriptional controls even lie behind cancer. "I am one of the most faithful advocates of sophisticated function by RNA. But even I am astounded by what these RNAs are doing," says Yale University's Ronald R. Breaker, associate professor of molecular, cellular, and developmental biology. RNA's versatility may seem complex enough to fuel a creative design advocate's arguments, but it is rooted in molecular interactions--chemistry. ALTERNATE mRNA CONFORMATIONS REGULATE METABOLISM Because RNA is single-stranded, it can assume a secondary structure by base-pairing within itself. Some mRNAs can take alternate forms--what Breaker dubs 'riboswitches'--that modulate synthesis of certain vitamins so that they are made only when needed. Riboswitches precede the coding parts of large mRNA molecules, where they seemingly sense whether a certain metabolite is present. If it is, they shut off translation in the remaining mRNA, which encodes an enzyme necessary to synthesize the metabolite: It is classic negative feedback. So far, Breaker has found riboswitches in genes that control several coenzymes and vitamins, such as B1 (thiamine) synthesis and a B12 transporter, which are prime candidates for possible descendants of the RNA world's chemical residents. Breaker sought riboswitches as controls of vitamin synthesis because regulatory proteins with such roles have been predicted but never discovered. Another clue that RNA could be the regulator came from aptamers, which are short, synthetic RNAs selected for their catalytic activity. "Over the last several years, my lab has been engineering RNA switches that can sense various target compounds.1,2 We had become convinced that natural RNA switches would be present in modern organisms," he relates. And they are--"with actual complexity and performance characteristics that are entirely unexpected," he adds. The researchers created partial mRNAs, called thiM and thiC, for the gene that encodes an enzyme required for synthesizing thiamine. The RNAs included the portions that bind ribosomes, the organelles on which protein synthesis occurs. Then the researchers cut the mRNAs, either in the presence or absence of a thiamine derivative, TPP. Different-sized fragments resulted, depending on whether TPP was present, suggesting that the mRNA assumes alternate conformations, only one of which binds the vitamin. The mRNA-TPP complex yielded some larger pieces where the nucleic acid is blocked from the cutting enzyme. Next, the researchers tested the ability of the implicated mRNAs to affect expression of the gene that controls vitamin synthesis. "We created mutant riboswitches that either could or could not bind TPP," Breaker says. "We then fused each mutant riboswitch and the ß-galactosidase gene. The encoded reporter enzyme converted a colorless chemical substrate into a blue product: the darker the blue, the greater the gene expression. And they found that the vitamin indeed shuts off its own synthesis. The finding also reveals one function of the untranslated region that starts an mRNA. "These large noncoding regions are hardly a wasteland of unused RNA," says Breaker. Riboswitches for vitamins are just one type, the research-ers say. "We speculate that, for certain organisms, direct metabolite sensing by mRNAs will be a common form of genetic control," Breaker says. And they may be molecular fossils, survivors of a primordial biochemistry. "Modern riboswitches could be direct descendants of RNAs that existed nearly 4 billion years ago in the RNA world. I think that some organisms have thrown riboswitches, and a lot of other ancient RNAs, onto the evolutionary trash heap. However, riboswitches have great utility in some instances. Why make a protein when an RNA will do just fine?" he asks. Courtesy of Ben Boese IN A BIND: The sequence and secondary structure model of a riboswitch that binds thiamine pyrophosphate (space filling model). The magenta nucleotides are conserved through evolution; the green nucleotides designate the AUG start codon. The flare identifies the ribosome binding site. NONSENSE-MEDIATED mRNA DECAY RNA can also police itself. In some eukaryotic genes, the spatial relationship between a nonsense mutation (which prematurely inserts a translation termination codon) and an exon/intron splice site located distal to it somehow signals the translational machinery to halt. This action, called nonsense-mediated mRNA decay (NMD), prevents production of truncated proteins, which can be useless or even toxic. "NMD was discovered in the late 1970s when it was observed that nonsense globin transcripts showed low abundance," relates professor Henry Dietz, a Howard Hughes Medical Institute investigator at Johns Hopkins University School of Medicine. His group recently showed that NMD is distinct from another protective process, nonsense-mediated altered splicing, which can excise the errant stop codon in the mRNA in the nucleus.3 NMD could happen when mRNA exits the nucleus. "Cells have evolved a way to define a termination codon as premature during a 'pioneer' round of translation, based on where within the precursor mRNA splicing took place," explains Lynne Maquat, professor of biochemistry and biophysics, University of Rochester School of Medicine and Dentistry.4 The relationship between the misplaced stop codon and the splice site is "intimate," says pathology professor Peter Byers, University of Washington in Seattle. "There must be a spliced intron downstream from the premature termination codon if the message is to be unstable. In the mRNA, the old exon/intron and now the exon/exon boundaries are marked by association of proteins about 20 to 25 nucleotides upstream. These protein complexes are removed in the course of translation. But if translation does not get to one of these sites because it is stopped by a premature termination codon, then the complex attracts a protein that triggers digestion of the mRNA by exonuclease." CANCER AS TRANSCRIPTION DISREGULATION DNA microarrays are providing unprecedented peeks into global gene expression, because they represent the mRNA transcripts in a particular cell type under specific conditions. To track the course of cancer, microarrays display the suites of genes that are under- or overexpressed over time. In this way, Arul Chinnaiyan, assistant professor of pathology and urology, University of Michigan Medical School, and colleagues used 10,000-gene microarrays to monitor prostate cancer progression. Courtesy of Harry C. Dietz PROMOTING EFFICIENCY: Data suggest that mRNAs form a closed loop conformation mediated by proteins bound at the 5' and 3' ends of the transcript. This conformation is believed to promote both translational efficiency through recruitment of the 40S ribosomal subunit, and mRNA stability through limitation of access by decapping enzymes and nucleases. In an initial round of experiments, the researchers discovered that 55 genes were overexpressed and 480 underexpressed in cells of metastatic tissue, compared to those from localized tumor specimens. The amount of EZH2 protein increased with each stage. The most overactive gene was the enhancer of zeste homolog 2 (EZH2).5 To isolate EZH2's effects, the investigators overexpressed it in cultured prostate cancer cells, which consistently dampened expression of 163 genes but did not activate any genes. EZH2, then, appears to be a powerful genetic off switch. If so, would its elevation in early, localized prostate cancer foretell spread? To find out, Chinnaiyan and coworkers measured EZH2 expression in tumor specimens stored from 64 individuals with known clinical outcome. "Localized specimens that exhibited high levels of EZH2 protein had a worse prognosis than those that had low or absent EZH2," he relates. As a result, Chinnaiyan dubbed EZH2 a "signature lethal biomarker." In fact, EZH2 level is a more reliable predictor of disease progression than standard measures, such as tumor size and prostate specific antigen (PSA) level before surgery. And that has practical implications. "We envision a test using prostate needle biopsies, in which EZH2 is measured, designating good prognosis or bad prognosis prostate cancer," Chinnaiyan concludes. It is intriguing, perhaps even ironic, that as science celebrates Watson and Crick's discovery, researchers are still exploring primordial ooze to understand its chemical cousin, RNA. Sums up Breaker: "Certainly, RNA will not cease to surprise us anytime soon!" Ricki Lewis (rlewis@nasw.org) is a contributing editor. References 1. W. Winkler et al., "Thiamine derivatives bind messenger RNAs directly to regulate bacterial gene expression," Nature, 419:952-6, Oct. 31, 2002. 2. W. Winkler et al., "An mRNA structure that controls gene expression by binding FMN," Proc Natl Acad Sci, 99:15908-13, Dec. 10, 2002. 3. J.T. Mendell et al., "Separable roles for rent1/hUpf1 in altered splicing and decay of nonsense transcripts," Science, 298:419-22, Oct. 11, 2002. 4. L.E. Maquat, "Nonsense-mediated mRNA decay," Curr Biol, 12:R196-7, 2002. 5. S. Varambally et al., "The polycomb group protein EZH2 is involved in progression of prostate cancer," Nature, 419:624-8, Oct. 10, 2002.

RNA Calls the Shots

Interested in reading more?

Become a Member of

Courtesy of Vasudeva Mahavisno
CANCER CARTOGRAPHY: Metastatic prostate cancer from a tissue array stained for EZH2 protein. The background represents gene expression signatures of prostate cancer as a heat map, which lead to the discovery of EZH2 as a prostate cancer biomarker.