Enhancer elements are a somewhat mysterious feature of higher eukaryotic genomes, predominantly because how they enhance the expression of genes that lie far away from them remains unknown. Do they, for example, make direct contact with their target gene by looping out the intervening DNA or do they act indirectly by producing a transcriptionally favourable environment?

Answers to these questions now come from a recent paper by David Carter and colleagues who have developed a new technique — RNA TRAP — to investigate enhancer elements. They show, for the first time, that long-range enhancer elements very likely come into physical contact with the genes they regulate — results that shed doubt over non-contact models of enhancer function and demonstrate the usefulness of this technique for exploring transcription-regulating elements.

Step one of RNA TRAP involves localizing horseradish peroxidase (HRP) to oligos that are targeted to an RNA as it is being transcribed. In this study, oligos were directed against two genes that lie in the mouse β-globin cluster, downstream of a locus control region (LCR). This LCR contains six DNase-I hypersensitive sites (HS1–6) and is required for the high-level expression of β-globin locus genes in erythroid cells. In step two, the localized HRP catalyses the covalent deposition of biotin onto chromatin proteins in close proximity to the transcribed gene. The labelled chromatin is then purified by affinity chromatography and the DNA sequences bound to it are identified by PCR.

In their study, Carter et al. used mouse E14.5 fetal liver cells, which express only two of the four genes at the Hbb locus, Hbb-b1 and Hbb-b2 . In their first RNA TRAP experiment, probes were targeted to the 3′ intron of Hbb-b1, and the enrichment of sequences across the Hbb locus was measured. The sequence around the targeted region was most greatly enriched, as expected, with enrichment dropping off sharply over the silenced regions of the locus. The enrichment picked up again around the LCR, especially at HS2, and to a lesser extent at HS1 and HS3. A similar enrichment pattern was detected when a 3′ intron of Hbb-b2 was targeted with oligos. Again, HS2 was highly enriched; as was HS4, but to a lesser extent. These findings indicate that certain regions of the LCR, especially HS2, come into close physical proximity to the active Hbb-b1 and Hbb-b2 genes. Moreover, these results tie in nicely with previous Hbb-locus deletion studies in mice that have shown that gene expression in this region is most drastically reduced by the deletion of HS2.

To rule out the possibility that these results might be caused by the preferential deposition of biotin in certain chromatin regions, Carter et al. also ran control experiments in which they omitted the intronic probes at step one, causing biotin to be randomly deposited across the genome. No preferential labelling of Hbb-locus sequences occurred as a result, lending further weight to their findings. The authors' planned improvements to this assay should shed more light on the exact nature of the interaction that occurs between enhancers and the genes they regulate.