Mapping methylation's mysterious background

April 29, 2010

Alla Katsnelson

By chemically modifying and then sequencing the DNA of 17 different species, researchers in California say that they have come closer to understanding the mysteries surrounding DNA methylation. This is an essential process for the regulation of many cellular events in mammalian development.

Many plants, animals and fungi have their DNA altered by chemical modifications. Among these is methylation, the addition of a methyl group to cytosine, one of the four bases that make up DNA. But some organisms, such as flies, budding yeast and nematode worms, live perfectly well without this mechanism.

"One of the big mysteries of DNA methylation has been how come you have some organisms with DNA methylation and some without," says Daniel Zilberman, the senior author of a paper published in Science today1. "I wouldn't say we have provided the answer, but I would say we have provided a plausible version of an answer."

To do this, his team at the University of California, Berkeley, has determined the DNA methylation patterns of 17 organisms – five plants, seven animals and five fungi – choosing species in different parts of the evolutionary tree to reconstruct how methylation might have evolved. From this, the authors predict what the process may have looked like around 1.6 billion years ago in the last common ancestor of plants, animals and fungi.

They created genome-wide methylation maps for each species, using high-throughput sequencing in conjunction with a technique that chemically converts each cytosine base in a genome to another base, uracil, unless that cytosine is methylated – a process known as bisulphite sequencing.

"Previously, people had glimpses of DNA methylation in a variety of organisms, but what this [study] does is try to give a more global view, which is really valuable," says Eric Selker, a biologist at the University of Oregon in Eugune who was not involved in the study.

Mobile-DNA control

In mammals and plants, methylation has been widely observed in transposons – mobile pieces of DNA that can cause mutations in the genome – suggesting that the process may function to keep transposons in check. But in a recent study in plants, Zilberman's group also identified methylation in the middle of active genes2. The latest analysis extends the finding to other organisms, ranging from rice and the fungus Phycomyces to puffer fish to anemones1.

The fact that methylation occurs within active genes in many organisms suggests that it is an ancient phenomenon, says Zilberman. While DNA methylation in plants, fungi and vertebrates was concentrated in transposons, invertebrates showed the opposite pattern, with modifications occurring mainly in active genes.

What might explain that disparity, he says, is sex. In organisms that reproduce sexually, transposons – essentially genomic parasites – tend to be more aggressive in moving about the genome and wreaking mutational havoc. Conversely, in asexual organisms, transposons are generally quite tame; if they reduce the fitness of their host too much, they risk becoming extinct.

Ancestral enzymes

Zilberman proposes that the common ancestor of plants, animals and fungi carried enzymes that methylated both transposons and gene bodies. When animals split off from fungi, they were probably single-celled, asexually reproducing organisms with no need for a mechanism to control their transposons, so the enzyme that methylated transposons was lost. Vertebrates re-evolved it, but invertebrates did not – instead they developed other mechanisms to deal with their transposons.

However, not everyone agrees with this explanation.

"It is kind of a fun paper, but it has a lot of issues," says Timothy Bestor, a developmental geneticist at Columbia University in New York.

For a start, he says, the group's analysis doesn't distinguish between transposons that actually jump around the genome, as in mammals, and those that are so heavily mutated that they are pretty much stable, as in the Neurospora fungi. These two transposon types may well have different methylation patterns. Furthermore, says Bestor, the ancestral state suggested by Zilberman and his colleagues "is a fairly recent ancestral state" and doesn't explain where DNA methylation came from in the first place.

Another mystery remaining to be solved is why DNA methylation occurs in gene bodies at all, as it does not seem to interfere with gene expression, says Selker.

"I think the question of the future is, what the heck is this methylation doing?" he says.