RNA Sequences Don’t Predict In Vivo Transcript Structure

RNA Sequences Don’t Predict In Vivo Transcript Structure

Eukaryotes prevent secondary RNA structures called G-quadruplexes, commonly observed in vitro, from forming in the cell. By Catherine Offord | January 1, 2017

(http://www.the-scientist.com/?articles.view/articleNo/47741/title/RNA-Sequences-Don-t-Predict-In-Vivo-Transcript-Structure/)

IN A TWIST: Certain guanine-rich regions of the transcriptome can fold up into four-stranded structures called RNA G-quadruplexes that are thought to have harmful effects in living cells. Both eukaryotes and bacteria have evolved mechanisms to keep these structures unfolded, according to a new study. Eukaryotes likely have protein machinery—possibly helicases to unravel the structure (1) and proteins that bind to the transcripts to keep them from reforming (2). Bacteria appear to have depleted quadruplex-forming sequences from their genomes during their evolution.

EDITOR'S CHOICE IN GENETICS & GENOMICS

The paper J.U. Guo, D.P. Bartel, “RNA G-quadruplexes are globally unfolded in eukaryotic cells and depleted in bacteria,” Science, doi:10.1126/science.aaf5371, 2016. RNA doesn’t lie flat. Interactions between nucleotides can turn sections of transcripts into loops, bends, and knots, some of which have regulatory functions in the cell. It was with these tangles in mind that Junjie Guo, a postdoc in David Bartel’s lab at MIT, developed an in vitro chemical probe to detect folded regions of RNA. “We were trying to identify all the structures that can stably form in vitro,” he says.

Testing the technique on transcripts extracted from mouse embryonic stem cells, Guo found that one particular conformation was unexpectedly abundant: RNA G-quadruplexes—stable, guanine-rich regions folded into four-stranded structures. “Only dozens of [these] regions have been studied previously,” Guo notes. “But we identified thousands of endogenous RNA regions that can form these structures in vitro.”

RNA G-quadruplexes, first identified more than five decades ago, were for a long time mysterious in terms of function, Guo says. There’s evidence now that they influence translation, and they may be involved in diseases including cancer and neurodegeneration. But although quadruplexes are thought to form readily in the chemical environment of the cytoplasm, they have not been studied in action because they are difficult to identify inside living cells. So Guo and Bartel used their new probe to do just that.

They didn’t find quite what they were looking for. It turned out that the quadruplex-forming regions detected in vitro weren’t actually folding up inside mouse cells, Guo explains. “We found that most of these structures are in an open state.” And it wasn’t just mammalian cells. Using the technique in yeast yielded similar results: the transcriptome contained regions that formed quadruplexes in vitro, but these regions were not folded in vivo. The results suggest eukaryotes possess machinery to prevent quadruplex formation.

Looking for further insights, the researchers turned to bacteria. Unlike eukaryotic RNA, Escherichia coli’s transcriptome exhibited no obvious quadruplex-forming regions. However, when Guo introduced specific guanine-rich regions to cells via a plasmid, he found the mRNA easily folded into quadruplexes, triggering abnormal protein translation and reduced growth rates. The results “could partly speak to why eukaryotes have come up with machinery that suppresses formation of these structures,” Guo says. The findings also imply that, instead of developing similar machinery, bacteria have eliminated quadruplex-forming sequences over the course of evolution.

Stephen Neidle, an emeritus professor of chemical biology at University College London, says Guo and Bartel’s eukaryotic results have particular relevance for the study of disease. “[They’re] saying that the prevalence of sequences does not directly map onto the prevalence of stable structures,” he notes. “There’s significant literature discussing isolated RNA quadruplexes as therapeutic targets—this paper is putting much of that work on its head and saying, actually, those targets [may be] illusory.”

Guo says he and Bartel are now looking for components of the eukaryotic unfolding machinery by identifying proteins that bind quadruplex-forming regions of RNA. “We’re getting some very interesting candidates,” he says, and adds that while “some have previously been shown to bind G-rich RNA, some have never been characterized.”