By Rob Mitchum
In cells, DNA doesn’t often hang out in the long, stretched-out strings you see in science textbooks. Most of the time, it is stored tight in a package called a nucleosome, wound like a ball of yarn around a protein called chromatin. In order for a gene to be “activated,” the stretch of DNA where it resides must first be unspooled from the nucleosome, so that cellular factors can attach to the strand and begin making protein from the DNA recipe. In a new study published this week in Nature, a team of University of Chicago researchers took advantage of this connection between unspooling and activation to solve a mystery that haunts many a recent genetics study.
Genome-wide association studies, commonly called GWAS, look for genetic variants associated with diseases or other genetic traits such as height or hair color. Since the completion of the Human Genome Project and the development of gene chip technology, scientists have performed hundreds of these studies. But many of them offer a befuddling result, with some of the most significant GWAS “hits” coming from variants that lie in the spaces between the protein-encoding genes, regions once dismissed as “junk DNA.” Nevertheless, some of these variants have been observed to affect the expression of nearby genes by some unknown process, leading them to be named expression quantitative trait loci, or eQTLs.
But how do these “non-coding” variants exert their dramatic effects upon gene expression — and ultimately, upon diseases and traits? The Department of Human Genetics laboratories of Jonathan Pritchard and Yoav Gilad found one potential method by selectively targeting the unspooled segments of DNA.
“Much of the regulation is occurring in these regions where the DNA is unfolded, so it’s accessible for proteins to come in,” said Pritchard, professor of human genetics at the University of Chicago Biological Sciences. “What we were interested in was figuring out the main mechanisms by which variation is affecting regulation. We postulated that changes in these open regions would be a major mechanism.”
In cell cultures of B cells (a kind of white blood cell) from 70 West African individuals, researchers used an enzyme called DNaseI to cut the DNA into short segments. Because DNaseI can only work on segments that are unspooled from chromatin, the chopping process left the team with markers of DNA regions that are open for business – in this case the team measured a total of 2.7 billion DNaseI cut sites. The researchers could then use the DNaseI cut sites to create a detailed map and test for genetic variants that predict whether a given stretch of DNA was more likely to be open or closed in an individual, with open segments likely reflecting genes actively under transcription.
“Basically what we’re doing is mapping these locations,” Pritchard said. “The power of DNaseI is that it’s giving us a slightly indirect way of measuring transcription factor occupancy, but it’s giving us information about essentially all factors at once.”
The nearly 10,000 variants found in that test were dubbed “DNaseI sensitivity QTLs,” or dsQTLs for short. The naming similarity to eQTLs was no accident, as the researchers found a significant overlap between the two classes of genetic markers. Up to half of eQTLs were estimated to also be dsQTLs, meaning that the gene variant exerted its power to increase or decrease expression of its gene by affecting the probability of the DNA segment being opened or closed. “dsQTLs are therefore a major mechanism by which genetic variation may affect gene expression levels,” the authors write.
“I think one of the things this paper does is to clarify one of the main mechanisms by which eQTLs arise,” Pritchard said. “Many people measure eQTLs, but generally it has been very difficult to figure out what are the causal variants that drive them and how they act. This is kind of filling in the black box for perhaps as many as half the eQTLs.”
Decoding such a significant genetic trick is a valuable advance for genetics researchers, including those such as Gilad and Pritchard who study evolution. There is now growing evidence for the theory that the differences between closely related species such as chimpanzees and humans are less due to changes in the actual genes, but rather how those genes are regulated. The timing of when and where a particular gene is switched on — either through unspooling from the chromatin or another mechanism — could have dramatic effects upon brain structures, immune system function, or any of the other biological differences that separate Homo sapiens from its primate cousins.
“It is extremely difficult to figure out which non-coding variants in the genome are going to have a regulatory function. I’d like to have much better ways of figuring that out because I think it will tell us about the evolution of humans,” Pritchard said. “I imagine this mechanism will be kind of an obvious way that expression levels are changing over evolutionary time: You knock out one of these open regions, and that’s going to change expression.”
Degner, J., Pai, A., Pique-Regi, R., Veyrieras, J., Gaffney, D., Pickrell, J., De Leon, S., Michelini, K., Lewellen, N., Crawford, G., Stephens, M., Gilad, Y., & Pritchard, J. (2012). DNase I sensitivity QTLs are a major determinant of human expression variation Nature DOI: 10.1038/nature10808