Constricting restricted transcription: the (actively?) shrinking web

It is now quite generally accepted that the mammalian genome is, in essence, littered with functionally critical regulatory elements that can be located anywhere in relationship to the gene that they control: internally, 5′ or 3′, sometimes even more than a megabase away (Lettice et al. 2002). Furthermore, seemingly straightforward activities of a gene can be controlled by multiple spatially or temporally distinct enhancers, even for eliciting, in a single developing tissue, functions that never could have been identified as distinct by conventional methods (Khandekar et al. 2004). The discovery and initial exploration of the activity of globin gene enhancers has served as a well-documented paradigm for how such activities are mechanistically interpreted. Following discovery of the delocalized β-type globin gene enhancers (that later became known as the locus control region, the LCR), decades of experiment and speculation followed (Choi and Engel 1988; Enver et al. 1990; Tuan et al. 1992; Wijgerde et al. 1995; Martin et al. 1996; Bulger and Groudine 1999) that addressed the mechanisms underlying enhancer/LCR function. The mechanism of long-range enhancer function was at least partially resolved in 2002, thereby ending (at least one aspect of) this ancient debate. Using novel techniques capable of measuring the spatial proximity of distal genomic elements it was shown that the LCR and actively transcribed globin genes were in close physical proximity inside erythroid cells, looping out the sequences lying between the LCR and globin gene promoter (Carter et al. 2002; Tolhuis et al. 2002). The results strongly implied that a direct functional interaction between the LCR and globin gene was necessary for high-level transcription. A report in this issue of Genes & Development by Ragoczy et al. (2006) now shows that the LCR directs the nuclear relocation of the globin locus during erythroid …