Some RNA polymerases run backwards. Credit: Laguna Design / Science Photo Library

Over the past decade, biologists have learned to credit RNA with more respect than it once garnered. Previously thought of simply as a chemical intermediate between DNA and protein, a host of RNA oddities that can switch genes off and on has revised that view.

Now, a suite of papers published in Science this week promises to add still more complexity by revealing several new classes of peculiar RNA molecules, many of which are created when proteins read DNA backwards.

DNA is transcribed into RNA by enzymes called RNA polymerases. Some of these RNA molecules will then be used as a template to create proteins, whereas others act directly to affect processes in the cell. RNA polymerases sometimes latch onto a bit of DNA called a promoter, which is generally found in front of genes and contains important sequence information that can influence when and where a gene is expressed. The polymerase then travels along the DNA strand as it transcribes RNA.

Sometimes, however, a polymerase will move in the opposite direction, creating what is known as an 'antisense' molecule of RNA. Some antisense RNAs can interfere with the function of its 'sense' partner, providing a way to regulate gene expression.

Promoting RNAs

Recently, scientists have noticed that some short RNA molecules that do not code for any proteins are being synthesized near the promoter region. Last year, Thomas Gingeras, now at Cold Spring Harbor Laboratory in New York, and his colleagues profiled the RNA population in human cells grown in culture and found evidence of short RNAs scattered near promoters, sometimes in an antisense orientation1.

This week in Science, several groups unveil results that elaborate on this finding by analysing patterns of RNA synthesis. John Lis and his colleagues at Cornell University in Ithaca, New York, used a method that allowed them to create a quantitative map of active polymerases throughout the genome2. The researchers then analysed the RNA made by those polymerases. Meanwhile, Phillip Sharp of the Massachusetts Institute of Technology in Cambridge and his colleagues focused on RNAs produced near the beginning of genes that code for proteins3.

Both teams uncovered a surprising trend. RNA polymerases indeed bound to the promoter and then proceeded to read the DNA from the beginning of a gene to the end. But another polymerase was often found just in front of the promoter, and headed in the opposite direction, effectively reading the DNA backwards.

In another study published this week, Torben Jensen of Aarhus University in Denmark and his colleagues found a similar class of RNAs that are formed further away from the promoter4. These RNAs ran either forwards or backwards, but were more likely to be destroyed by RNA-degrading proteins within cells than the RNAs discovered by the other two groups.

But why?

All these classes of RNAs are haunted by an important lingering question. "What are these things doing?" says Gingeras. "Why are promoters being criss-crossed with transcripts?"

At this stage, the answer is unknown. The "most tedious" hypothesis, says Jansen, is that the RNA molecules are simply a mistake. The DNA surrounding the promoter assumes a more open structure that allows RNA polymerases to sweep in and bind to the DNA strands. It may be that extraneous RNA molecules are produced near the promoter merely because the polymerase can bind at that site and start working.

But even if the process began as an accidental by-product, "evolution could have taken advantage of this", says Jensen. One popular hypothesis is that the presence of polymerases bound near promoters helps maintain the proper, more-open structure of the DNA surrounding the site. Alternatively, Jensen suggests that having lots of polymerases bound to DNA near the promoter may simply provide a local reservoir of RNA polymerase recruits to call on when needed.

Piero Carninci of RIKEN Yokohama Institute in Japan, says that one thing is clear: researchers have probably not yet identified all the different types of RNAs synthesized in a cell. "The differences among these findings suggest also that the methods we have are not yet comprehensive to identify all the types of RNAs that are produced," he says. "There is still a need for a full description of all RNAs."