One way to determine whether something has value is to do without it and see what happens. The same principle holds in biology, but the 'doing without' part can be tricky to accomplish. Bradley Bernstein and his team at Harvard Medical School adapted genome-engineering tools to eliminate the function of enhancers.

Enhancers are regulatory DNA elements that bind transcription factors and other trans-acting proteins to augment gene expression. Unlike promoters—regulatory regions immediately upstream of the transcription start site of a gene—enhancers can be located distantly from the genes they regulate.

Recent work showed that enhancers are marked by characteristic chromatin signatures such as a mono- or dimethylated lysine 4 on histone H3 (H3K4me1 or H3K4me2, respectively) or an acetylated lysine 27 on H3 (H3K27ac). In the course of the Encyclopedia of DNA Elements (ENCODE) project and the Roadmap Epigenomics Project, researchers have identified close to 1 million putative enhancer elements. “But,” notes Bernstein, “that does not tell you whether they are really functional; there is a lot of noncoding genome out there and a lot of potential elements to test.”

Traditional reporter assays, in which an enhancer element is cloned in front of a reporter gene, can show which elements can be active, but they don't demonstrate which endogenous genes are being regulated.

Bernstein wanted to look at enhancers in their endogenous context. He reasoned that specific removal of the histone modifications that mark enhancers would lead to an enhancer's inactivation. He and Eric Mendenhall, now an Assistant Professor at the University of Alabama, collaborated with Keith Joung, also from Harvard Medical School, to design TALEs fused to a lysine-specific demethylase that target candidate enhancers in a leukemia cell line. The demethylase also recruited deacetylases, and both histone marks were removed.

The team members targeted 40 enhancers in a leukemia cell line then sequenced the transcriptome to identify genes whose expression changed in response to the inactivation of nine different enhancers. In four of the cases, they detected a clear downregulation of genes in proximity to the enhancer, but the other five did not show a detectable effect.

Although this shows that the principle of inactivating an enhancer via its chromatin signature works, the rules underlying the approach are not yet fully understood, and not all of the engineered reagents were effective.

Interestingly, the researchers found effects only in genes close to the enhancer. Bernstein speculates that to capture enhancers working over longer distances, as some developmentally important enhancers are known to do, they would have to use a different cell model.

Another exciting possibility for Bernstein is to combine this approach with chromosome confirmation capture methods and quantitative trait loci mapping. This would allow further validation and could improve methods for predicting enhancer targets.

The applications go beyond inactivation. By recruiting methyltransferases, enhancers can also conceivably be activated. “This might be a way to achieve specificity in regulation,” says Bernstein, “to engineer highly specific regulators that only act on a given gene in a very specific context.”

For Bernstein and his colleagues to reach the ultimate goal that prompted them to start this work—the detailed characterization of thousands of enhancers—the method will have to be scaled up. The newest genome engineering tool, CRISPR (clustered, regularly interspaced, short palindromic repeats) systems, might; but more work may be needed to improve the efficiency with which these can change the epigenome.

The hope is that before long these tools will allow systematic probing and modulation of the flood of cis-regulatory elements emerging from genome-wide mapping studies.