Researchers identify the genome's controlling elements
Scientists have churned out genome sequences for everything from fungi to dogs to chimps, and they won't be letting up any time soon. However, because a genome sequence is little more than a static list of chemicals — like, say, a parts list for a 747 airplane — scientists are increasingly turning their attention to figuring out how living organisms put their genes to work. Using yeast as a testing ground, researchers at Whitehead Institute for Biomedical Research, along with researchers from MIT and the Broad Institute, have for the first time revealed all the "controlling elements" of an entire genome — findings that may soon contribute to a new way of understanding human health and disease.
"This is really the next stage in human genome research," says Whitehead member Richard Young, who headed the project together with Whitehead fellow Ernest Fraenkel and MIT computer scientist David Gifford. Young and Gifford are also associate members at Broad.
Decoding the genome. By combining microarray data (far left) with other published yeast genome sequences (far right), Whitehead and MIT researchers were able to locate genome-wide landing points of gene regulators, or transcription factors. Image courtesy of Chris Harbison, Nature.
Key to understanding how the genome is controlled are gene regulators, also known as transcription factors. These small molecules intermittently land on a region of DNA, close to a particular gene, and then switch that gene on. They can also influence the amount of protein that the gene will produce. Many diseases, such as diabetes and cancer, are associated with mutated gene regulators, which is one reason why scientists are so interested in them.
The problem is that very few of these regulators have been identified in any organism. Locating their landing sites is essential to identifying their function, and therein lies the rub: Gene regulators are hard to find. They typically just land on a small stretch of DNA, do their job, and then take off again. And owing to the vastness of the genome, locating just one gene regulator with conventional lab tools can take many years. The Whitehead/MIT/Broad team, in the Sept. 2 issue of Nature, report developing a method for scanning an entire genome and quickly identifying the precise landing sites for these regulators.
This work builds upon research reported by Young in Science in Oct. 2002, in which he mapped the general locations of approximately half the gene regulators in yeast.
"The results of the Science paper were pretty low-resolution," says Young. "We were only able to identify in a general way regions where these gene regulators landed. In this paper, we've located all 203 regulators in yeast," and using tools developed in Fraenkel's lab have also been able to nail down the exact landing points. As a result, scientists now can begin to understand how genes and their regulators "talk" to each other. According to Fraenkel, knowing these communication patterns ultimately will have a profound influence on our understanding of everything from infectious disease to cloning.
To eavesdrop on these cellular conversations, graduate student Chris Harbison from Young's lab and postdoctoral researcher Ben Gordon from Fraenkel's lab combined the latest biological tools with new computational methods.
Harbison took yeast cells and subjected them all to a dozen nutritional, chemical, and temperature environments. "We tried to come up with different conditions that a yeast cell would encounter in its natural habitat," says Harbison.
Gene regulators come out of hiding and do their job in response to environmental conditions, but they don't all respond to the same kinds of predicaments. Running the cells through a wide spectrum of stimuli was a way of waking up all the regulators — in a sense, shaking the bushes and then nabbing them once they're out.
Next, Harbison placed gene fragments associated with these regulators onto a series of microarrays — small dime-sized silicon or glass chips that contain thousands of pieces of DNA — which allowed him to come up with a list of approximate locations. Gordon and Fraenkel created computer algorithms that fused Harbison's data with data from other yeast species in order to find the exact landing points.
"The microarray information is like a noisy telephone call," says Fraenkel. "We needed to find the words in all the static before we could understand what they meant."
The next challenge is to scale the platform so it can tackle human cells, something that the researchers are gearing up to do. In fact, a paper from Young's lab in the journal Science last Feb. anticipated this possibility. The team reported locating in human cells all the genome-wide landing sites of a handful of gene regulators associated with type 2 diabetes, revealing some surprising mechanisms of the disease. "That paper is just the beginning of what we'll soon be doing in human cells," Young declares.
Even though the yeast genome's 203 regulators are a far cry from the roughly 2,000 in human cells, Young explains, "now we have the technology and the concepts to get started on decoding the human genome."
The study team also included Broad founding director Eric Lander and former postdoctoral fellow Manolis Kellis.
Adapted from a release written by David Cameron, Whitehead Institute. To receive a copy of this paper, please contact firstname.lastname@example.org
Nature, 430 (7004), pp. 99-104
"Transcriptional regulatory code of a eukaryotic genome"
Authors: Christopher T. Harbison(1,2), D. Benjamin Gordon(1), Tong Ihn Lee(1), Nicola J. Rinaldi(1,2), Kenzie D. Macisaac(3), Timothy W. Danford(3), Nancy M. Hannett(1), Jean-Bosco Tagne(1), David B. Reynolds(1), Jane Yoo(1), Ezra G. Jennings(1), Julia Zeitlinger(1), Dmitry K. Pokholok(1), Manolis Kellis(1,3,4), P. Alex Rolfe(3), Ken T. Takusagawa(3), Eric S. Lander(1,2,4), David K. Gifford(3,4), Ernest Fraenkel(1,3), Richard A. Young(1,2,4)
1. Whitehead Institute for Biomedical Research, Nine Cambridge Center, Cambridge, Massachusetts
2. Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts
3. MIT Computer Science and Artificial Intelligence Laboratory, 32 Vassar Street, Cambridge, Massachusetts
4. Broad Institute, 320 Charles St., Cambridge, Massachusetts
This work was supported by a grant from the National Institutes of Health.