By Erika Check Hayden of Nature magazine

All science students learn the 'central dogma' of molecular biology: that the sequence of bases encoded in DNA determines the sequence of amino acids that makes up the corresponding proteins. But now researchers suggest that human cells may complicate this tidy picture by making many proteins that do not match their underlying DNA sequences.

In work published today in Science, Vivian Cheung at the University of Pennsylvania in Philadelphia and her team report that they have found more than 10,000 places where the base (A, C, G or U) in a cell's RNA messages is not the one expected from the DNA sequences used to make the RNA read-out. When some of these 'mismatched' RNAS were subsequently translated into proteins, the latter reflected the 'incorrect' RNA sequences rather than that of the underlying DNA.

It was already known that some cells 'edit' RNA after it has been produced to give a new coding sequence, but the new work suggests that such editing occurs much more often in human cells than anyone had realized, and that hitherto unknown editing mechanisms must be involved to produce some of the changes observed. If the finding is confirmed by other investigators -- and some scientists already say they see the same phenomenon in their own data -- it could change biologists' understanding of the cell and alter the way researchers study genetic contribution to disease.

Editing the central dogma

"The central dogma says that there is faithful transcription of DNA into RNA. This challenges that idea on a much larger scale than was known," says Chris Gunter, director of research affairs at the HudsonAlpha Institute for Biotechnology in Huntsville, Alabama.

The work suggests that RNA editing is providing a previously unappreciated source of human genetic diversity that could affect, for instance, how vulnerable different people are to disease.

Cheung does not know whether there are heritable changes, passed down from parent to child, that affect how much RNA editing occurs in different people. But scientists already know of a handful of RNA editing proteins that play a role in human health, such as the APOBEC enzymes, some of which have antiviral activity. Researchers investigating the connection between genetics and disease have been stymied by their inability to find strong connections between genetic variation and risk for most common diseases, leading researchers to wonder where the 'missing heritability' is hiding. The new study at least provides one place to look.

"These events could explain some of the 'missing heritability' because they are not present in everyone and therefore introduce a source of genetic variation which was previously unaccounted for," says Gunter.

Living with error

But because they do not know what mechanism might be responsible, most scientists contacted by Nature remained cautious about the significance of the finding and its possible impact on biology. Some say it is possible that technical errors could have caused the results. For instance, high-throughput sequencing machines can make systematic errors in DNA and RNA sequencing experiments.

And even if the findings hold up, it is still too early to know whether 'mismatching' plays an important role in human biology or not.

"The devil is in the details -- to determine if the results are caused by some unintended technical or computational flaw or are correctly describing a biological phenomenon," says Thomas Gingeras at the Cold Spring Harbor Laboratory in New York. "Assuming the latter, I would be encouraged to look at our own large data sets to see if we see similar phenomenona."

Other researchers, such as Manolis Dermitzakis at the University of Geneva in Switzerland, say they are seeing the phenomenon in their data. Indeed, Cheung's team drew in part on data generated by the 1000 Genomes project, of which Dermitzakis is a member. However, Dermitzakis says it is still unclear how important the phenomenon is for disease susceptibility.

Cheung's group attempts to address many of these concerns, some of which were raised when the preliminary work was presented last November (see 'DNA sequence may be lost in translation') at the annual meeting of the American Society for Human Genetics, in Washington DC. Since then, the team has been looking for possible errors that could have caused the results.

For example, the researchers first observed DNA-RNA 'mismatches' in data generated by next-generation sequencing technologies in the International HapMap Project and the 1000 Genomes project. They have now confirmed some of the putative DNA-to-RNA changes using traditional Sanger sequencing, and have found the same changes in different people, across different cell types, and reflected in proteins.

Cheung says that at first "we truly did not believe it". But after performing the additional experiments "we cannot explain this by any obvious technical errors, so we are pretty convinced that this is real," she says.

Researchers who study RNA editing, which up to now was known mostly from plants and some unicellular human parasites, are intrigued by the new finding.

Kazuko Nishikura of the Wistar institute in Philadelphia says she was sceptical at first, because some of the base changes could not be explained by previously identified mechanisms. But she was convinced once she saw Cheung's data.

"It's really exciting, because this study reports a different variety of RNA editing that is much more widespread than existing mechanisms," Nishikura says.

This article is reproduced with permission from the magazine Nature. The article was first published on May 19, 2011.