Mapping the Shh long-range regulatory domain

Abstract

Coordinated gene expression controlled by long-distance enhancers is orchestrated by DNA regulatory sequences involving transcription factors and layers of control mechanisms. The Shh gene and well-established regulators are an example of genomic composition in which enhancers reside in a large desert extending into neighbouring genes to control the spatiotemporal pattern of expression. Exploiting the local hopping activity of the Sleeping Beauty transposon, the lacZ reporter gene was dispersed throughout the Shh region to systematically map the genomic features responsible for expression activity. We found that enhancer activities are retained inside a genomic region that corresponds to the topological associated domain (TAD) defined by Hi-C. This domain of approximately 900 kb is in an open conformation over its length and is generally susceptible to all Shh enhancers. Similar to the distal enhancers, an enhancer residing within the Shh second intron activates the reporter gene located at distances of hundreds of kilobases away, suggesting that both proximal and distal enhancers have the capacity to survey the Shh topological domain to recognise potential promoters. The widely expressed Rnf32 gene lying within the Shh domain evades enhancer activities by a process that may be common among other housekeeping genes that reside in large regulatory domains. Finally, the boundaries of the Shh TAD do not represent the absolute expression limits of enhancer activity, as expression activity is lost stepwise at a number of genomic positions at the verges of these domains.

Keywords: Enhancers; Long-range regulation; Mouse; Sleeping beauty transposon; Sonic hedgehog (Shh); Topological domains.

Fig. 1.

Expression of genes within the Shh regulatory locus. (A) The genes within the interval from Nom1 to En2 are marked by the grey rectangles, shaded from dark to light in the 5′ to 3′ orientation. Known enhancers are shown as coloured bars. (B) Schematic illustration of the multiple sites of Shh expression in the E11.5 embryo; the colours used match the relevant enhancers shown in their genomic context in A. Expression in the ZLI is driven by an unknown enhancer. (C) Results of a study of the embryonic expression at E11.5 of those genes conducted by RT-PCR [in anterior and posterior halves of limb buds, in carcass (body) and isolated heads]. In situ hybridisation, in E11.5 embryos are shown for Shh in whole mount in a bisected head (D), limb (F) and isolated gut (G), and in tissue sections showing expression in the epithelial lining oral cavity (E). Expression of Cpny1 and En2 are shown in whole-mount bisected heads (H,J) and in sections (I,K). Expression within the midbrain-hindbrain boundary is marked by arrows.

Fig. 2.

Expression of SBLac insertions within the gene desert. (A-D) Embryos derived from the ES cells containing the SBLac re-insertions within the intergenic desert depicted in E, which were harvested at E11.5 and stained for expression of β-gal. The staining observed reflects the Shh expression pattern. The diagram in E depicts the genomic interval from Nom1 to En2. In addition to the genes (grey rectangles) and known enhancers (coloured bars), the sites of the original pHLED insertions are marked by the blue triangles and the positions of the mobilised and re-inserted SBLac insertions are marked with black arrowheads. The direction in which the arrowhead points depicts the 5′-to-3′ orientation of the reporter gene. Those SBLac insertions shown in A-D have been boxed.

Fig. 3.

Expression within guts and a timecourse for β-gal staining. (A) A series of guts, dissected at E11.5, from embryos derived from cells carrying the initial insertions called 5′ Insert (5′) and 3′ Insert (3′) and the mobilised SBLac containing ES cells throughout the entire interval and stained for β-gal expression. Insertions in the middle of the interval, from SBLac741 to SBLac(-15), reflect the Shh expression pattern, with expression being detected in the laryngotracheal tube [in panel labelled (-3)], oesophagus, the lung buds and stomach. The embryonic gut from either end of the interval, SBLac936, SBLac(-109) and SBLac(-290), show no detectable staining. Expression in the laryngotracheal tube, driven by the MACS1 enhancer, is not detected in SBLac855 and SBLac796 (arrowheads). MACS1 shows a directional preference within the regulatory domain, because SBLac695, which lies a similar distance 3′ of MACS1 as SBLac796 lies 5′, shows more activity (expression marked by arrowhead). SBLac855 also shows no expression in the developing lung buds (arrow). (B) The genomic interval, as in Fig. 2E. (C) The timecourse for β-gal staining in embryos from the SBLac796, SBLac695, SBLac526 and SBLac96 insertions, dissected at E11.5 and stained for 1, 2, 4 and 18 h. lb, lung buds; lt, laryngotracheal tube; o, oesophagus; s, stomach.

Fig. 4.

Expression of SBLac insertions at the extreme ends of the Shh regulatory locus. (A,B) Embryos derived from ES cells carrying the initial insertion sites and (C-F) from SBLac-carrying ES cells from the 5′ end of the regulatory domain. SBLac936 shows faint Shh like expression in the limb buds but the staining down the back is detected in two stripes (C′), which does not reflect the Shh pattern in the floor plate and notochord (E′), but more likely reflects expression of Mnx1, the next gene past Nom1. Expression of Mnx1 is shown by in situ hybridisation (C″). In situ hybridisation for Shh expression in the floor plate and notochord is shown in E″. Expression in SBLac855, SBLac796 and SBLac741 does reflect the Shh pattern; however, expression is lost preferentially from the developing forebrain in SBLac855 and SBLac796, whereas expression in the midbrain is maintained. [Compare expression marked by arrows (forebrain), with that marked by arrowheads (midbrain) in E,F.] (G) The regulatory locus with the relevant SBLac insertions (black triangles) boxed. At the 3′ end of the region (H-K), highest levels of expression are detected in SBLac(-3), which has integrated within the coding region of Shh. Expression in the next insertion, SBLac(-15), is maintained in most of the sites of Shh expression but at much lower levels, and expression is completely missing from the forebrain (arrow). However, in the last two insertions, SBLac(-109) and SBLac(-290), expression is detected only at the midbrain-hindbrain junction, mirroring expression of Cnpy1 and En2, and suggesting these SBLac insertions have integrated within a different topological domain.

Fig. 5.

OPT analysis of SBLac insertions. Outputs of OPT analysis of E11.5 embryos derived from SBLac insertions across the regulatory locus [SBLac855 (A,E), SBLac526 (B,F), SBLac96 (C,G) and SBLac(-15) (D,H)]. Sagittal views are shown in A-D and frontal views in E-H. The anatomy of the samples is translucent, whereas the β-gal expression is coloured red. Expression is confirmed to be missing from the forebrains of SBLac8555 and SBLac(-15) (yellow arrows), whereas midbrain expression is maintained.

Fig. 6.

Summary of the limits of topological domains and enhancer activity. (A) The Hi-C analysis from mouse ES cells (taken from the Mouse ENCODE website), and the boundaries of the topological domain determined by the directionality index and marked by dotted lines. The relationship between the positions of these features and the genes and insertions within the Nom1 to En2 region are indicted in B. Below the genes is the track showing the positions of RNA polymerase II (Pol2 peaks marked with black lines) in E14.5 limb buds taken from ENCODE. (C) A summary of relative expression activity driven by individual enhancers in particular tissues (ZRS driving expression in the limb bud, SBE2 in the forebrain, SBE1 in the midbrain and MACS1 in the laryngotracheal tube) at each of the SBLac insertions. The solid vertical lines show an estimate of the relative expression levels and the dotted lines the predicted levels throughout the genomic interval. Expression at the 5′ end shows that reduction in expression occurs over an interval of greater than 100 kb, whereas limb bud expression is still detectable at the furthest insertion site. Positions of CTCF peaks (bright blue lines) in E14.5 limb buds taken from ENCODE are shown below. There are no appreciable peaks in the middle of the interval, the majority lying at either end; however, there is no relationship between position of the peaks and the position at which expression is reduced.