The CFTR (cystic fibrosis transmembrane conductance regulator) gene, which when mutated causes cystic fibrosis, encompasses nearly 200 kb of genomic DNA at chromosome 7q31.2. It is flanked by two genes ASZ1 [ankyrin repeat, SAM (sterile α-motif) and basic leucine zipper] and CTTNBP2 (cortactin-binding protein 2), which have very different expression profiles. CFTR is expressed primarily in specialized epithelial cells, whereas ASZ1 is transcribed exclusively in the testis and ovary, and CTTNBP2 is highly expressed in the brain, kidney and pancreas, with lower levels of expression in other tissues. Despite its highly regulated pattern of expression, the promoter of the CFTR gene apparently lacks the necessary elements to achieve this. We previously suggested that cis-acting regulatory elements elsewhere in the locus, both flanking the gene and within introns, were required to co-ordinate regulated, tissue-specific expression of CFTR. We identified a number of crucial elements, including enhancer-blocking insulators flanking the locus, intronic tissue-specific enhancers and also characterized some of the interacting proteins. We recently employed a high-resolution method of mapping DHS (DNase I-hypersensitive sites) using tiled microarrays. DHS are often associated with regulatory elements and use of this technique generated cell-specific profiles of potential regulatory sequences in primary cells and cell lines. We characterized a set of cis-acting elements within the CFTR locus and demonstrated direct physical interaction between them and the CFTR promoter, by chromosome conformation capture (3C). These results provide the first insight into the three-dimensional structure of the active CFTR gene.
The CFTR (cystic fibrosis transmembrane conductance regulator) gene encompasses 189 kb at chromosome 7q31.2 . CFTR is flanked on the 5′ side by ASZ1 [ankyrin repeat, SAM (sterile α-motif) and basic leucine zipper] and 3′ by CTTNBP2 (cortactin-binding protein 2). These neighbouring genes have very different expression profiles: CFTR is expressed primarily, although not exclusively, in specialized epithelial cells [2–4], whereas ASZ1 is transcribed solely in the testis and ovary , and CTTNBP2 is highly expressed in the brain, kidney and pancreas, with lower levels of expression in other tissues .
CFTR exhibits tightly regulated expression, both temporally during development and spatially in different tissue types [2,7,8]. However, the CFTR promoter resembles that of a housekeeping gene, in that it is CpG-rich, contains no TATA box, has multiple transcription start sites and has several putative binding sites for the transcription factor Sp1 (specificity protein 1) . Consistent with promoters of this type, the CFTR promoter demonstrates no apparent tissue-specificity, suggesting the involvement of distal regulatory elements in control of CFTR expression. It is probable that these elements are associated with DHS (DNase I-hypersensitive sites) within the genomic region encompassing the CFTR locus [10–15].
Although 20 years have elapsed since the identification of CFTR, the regulatory mechanisms for this gene have not been fully elucidated. The reasons for this slow progress include the paucity of appropriate primary cell types to evaluate CFTR expression and technical challenges arising in the global analysis of large genes that are regulated by multiple interacting cis-acting elements separated by tens to hundreds of kilobases of genomic DNA. The recent wave of new technologies that have been facilitated by the ENCODE project  overcame several of these difficulties. They not only provide novel methods to interrogate the tissue-specific regulation of a gene such as CFTR, but simultaneously enable data to be generated from much smaller amounts of biological material than are required for older techniques. Thus primary cell cultures with limited in vitro replication potential can be used for reliable experimentation.
We applied several of these recently developed protocols in combination with classical methods to elucidate CFTR regulatory elements. Moreover, we examined CFTR expression in primary human airway epithelial cells and genital duct cells to elucidate the critical regulatory elements in normal cells in addition to the cancer cell lines that have been examined to date. These experiments generated structural and functional evidence for a CFTR transcriptional hub in which intronic enhancer elements are brought into close proximity to the CFTR promoter to activate transcription.
Identification of DHS in the CFTR gene
We used a high-resolution tiled microarray-based assay, DNase-chip , to map DHS across the CFTR locus in a number of cell types. To perform DNase-chip, chromatin-bound DNA is digested with a small amount of DNase, so that DHS are preferentially digested. DNase-digested ends are made blunt and ligated to biotinylated linkers. DNA is then sheared and DHS ends are enriched on a streptavidin column. A second linker is added before the DNA is amplified, labelled with dyes, and hybridized to tiled microarrays. Using this technique, we identified a number of cell-type-specific DHS within and flanking the CFTR locus (Figure 1A). Moreover, since the microarrays used in this analysis were the Nimblegen ENCODE arrays, we simultaneously generated data on other genes within the 30 Mb covered by the ENCODE pilot project . These include a number of other genes involved in differentiated functions of the airway epithelium. More recently, the development of genome-wide approaches to evaluate regulatory elements, such as DNase-chip, DNase-seq and FAIRE [17–20], has enabled simultaneous analysis across the whole genome. The data generated by these experiments have tremendous power to dissect transcriptional regulatory pathways in specific cell lineages, such as primary airway epithelial cells.
Identification of DHS within part of ENCODE region 1
Figure 1(A) shows the identification of DHS by DNase-chip within approx. 1.5 Mb spanning the CFTR locus and including the Met proto-oncogene, CAPZA2 [capping protein (actin filament) muscle Z-line α2], ST7 (suppressor of tumorigenicity 7 isoform B), WNT2 [Wingless-type MMTV (murine-mammary-tumour virus) integration site 2], ASZ1, CFTR and CTTNB2. A number of DHS are observed, some of which are apparently common to many cell types, such as the DHS associated with the promoter of the CAPZA2 gene (Figure 1A, arrow a) and others are cell-type-specific, an example being the DHS at the CFTR promoter in 16HBE14o- cells  (Figure 1A, arrow b), the only cell type shown that expresses CFTR. Ubiquitous DHS are also evident, for example, arrow c in Figure 1(A) marks a ubiquitous site in the last intron of the CTTNB2 gene.
Within the CFTR gene itself, many of the DHS identified by DNase-chip correspond to ones that we saw previously by classical methods of DHS mapping, but a significant number are novel and some are only evident in primary airway cells that we have now evaluated for the first time (C.J. Ott, N.P. Blackledge, J.E. Kerschner, G.E. Crawford, C.U. Cotton and A. Harris, unpublished work). Figure 1(A) shows the DHS profile in skin fibroblasts that do not express CFTR, Beas2B cells  that express very low levels of the transcript and 16HBE14o- that express high levels of CFTR mRNA. The airway lines show very few DHS within the CFTR locus, in contrast with intestinal and genital duct cells (; C.J. Ott, N.P. Blackledge, J.E. Kerschner, G.E. Crawford, C.U. Cotton and A. Harris, unpublished work). Of interest is a prominent DHS in intron 10 of the gene (Figure 1A, arrow d) that warrants further investigation.
Moving from the DHS to the regulatory element
Not all DHS contain cis-acting regulatory elements; some may be associated with structural elements that function in chromatin organization. However, for each DHS, we pursue a number of experimental methods to investigate whether they are associated with important regulatory elements. These are well illustrated by our analysis of a DHS in the first intron of CFTR at 185+10 kb (where 185 is the last base in CFTR exon 1) and of DHS flanking the locus at −20.9, +6.8 and +15.6 kb. For all DHS, an initial bioinformatics search for cross-species homologies can often be informative and reinforce predictions for functional importance. However, since there is some divergence in the patterns of regulation of expression of the CFTR gene in different species, particularly rodents and human, extensive cross-species conservation is not a pre-requisite for further investigation of a DHS.
A cis-acting element in the first intron of the CFTR gene acts as an HNF1 (hepatocyte nuclear factor 1)-dependent enhancer of the CFTR promoter
Using DNAse-chip, we detected a cell-type-specific DHS within the first intron of CFTR (Figure 1B, arrow) that corresponds to a regulatory element that we identified in our earlier work [11,24] and has been confirmed by others . This element (known as 7/8 based on primer sets used to amplify the region ) was shown to positively regulate CFTR promoter activity specifically in intestinal cells both in vitro and in vivo: removal of the element from a YAC (yeast artificial chromosome) containing human CFTR reduced expression levels of the human gene by approx. 60% in transgenic mice carrying the YAC, but only within the epithelium of the small intestine . More recently, we showed that this element functions as a classical, tissue-specific enhancer in transient transfection experiments and can also independently recruit general factors necessary for transcription initiation . To determine the nuclear factor(s) interacting with the regulatory sequence, we performed an in vitro DNase I footprinting analysis, which revealed a significant protected sequence encompassing a conserved HNF1-binding site. Expression of HNF1α correlates with CFTR expression and this transcription factor binds in vitro to another cluster of intronic DHS in CFTR . In vivo, HNF1α contributes to the maintenance of normal mouse Cftr expression levels in the small intestine . More recently, we showed that HNF1 binds to the intron 1 enhancer both in vitro, by EMSA (electrophoretic mobility-shift assay), and in vivo, by ChIP (chromatin immunoprecipitation) . We have now demonstrated that a number of other intronic DHS within the CFTR locus are associated with enhancer elements that co-operate to activate CFTR expression (C.J. Ott, N.P. Blackledge, J.E. Kerschner, G.E. Crawford, C.U. Cotton and A. Harris, unpublished work).
The CFTR locus is flanked by enhancer-blocking insulators
The ASZ1 and CTTNBP2 genes, which flank CFTR, are located within approx. 50 kb of the gene and due to their highly divergent patterns of expression we investigated whether there were functional elements preventing transcriptional interference between these loci. By conventional DHS mapping, we initially identified two enhancer-blocking insulators 5′ and 3′ to the CFTR gene that had distinct properties. First, a DHS located at −20.9 kb with respect to the translation start site was associated with a classical CTCF (CCCTC-binding factor)-dependent insulator element . CTCF, a ubiquitously expressed, zinc finger DNA-binding protein [28,29], often establishes independently regulated domains of gene activity. A second element, located 3′ to CFTR, within a DHS at +15.6 kb (with respect to the translational end point) also demonstrated enhancer-blocking activity but this was independent of CTCF binding. The +15.6 kb DHS was marked by a peak of euchromatin-specific histone modifications, unlike the −20.9 kb DHS , supporting the hypothesis that these elements function by different mechanisms.
In addition to the prominent site at +15.6 kb other DHS were evident 3′ to the coding region of the gene, in particular, a complex cluster of sites at +5.4, +6.8, +7.0 and +7.4 kb from the CFTR translation end point . The DHS at +5.4 and +7.0 kb were observed in a variety of cell types irrespective of CFTR expression. However, the DHS at +6.8 and +7.4 kb were only found in a restricted number of CFTR-expressing cell types, including primary epididymis cells , suggesting that they may contain tissue-specific regulatory elements that participate in controlling CFTR expression . We subsequently demonstrated by ChIP that the +6.8 kb DHS is associated with a tissue-specific CTCF recognition site that binds CTCF in vivo (Figure 2) . Both Caco2 cells and primary epididymis cells express CFTR but the +6.8 kb DHS is evident only in primary epididymis cells. Enrichment of the +6.8 kb DHS after ChIP with an antibody specific for CTCF is evident in primary epididymis cells, but not Caco2. The +6.8 kb DHS core also displays enhancer-blocking activity comparable with that of other known insulator elements, including the one at the CFTR −20.9 kb DHS [27,31].
In vivo binding of CTCF at the +6.8 kb DHS region
The three-dimensional structure of the active CFTR locus
A number of techniques have been developed over the last several years to interrogate the three-dimensional structure of genes within the nucleus [32–36]. These methods, which usually capture the chromatin structure in vivo by formaldehyde cross-linking of proteins and associated DNA, enable the mapping of regions of a locus that are in close physical proximity, despite their linear separation on the chromosome. These close interactions are often associated with transcriptionally active genes and presumably facilitate the co-operation of cis-acting regulatory elements and gene promoters. They may also delineate the boundaries of an active locus and compartmentalize it away from neighbouring genes with different cell-type-specific expression patterns.
Since we identified enhancer-blocking insulators flanking the CFTR locus that bind CTCF, and this protein is thought to be involved in regulating nuclear organization, we next evaluated the three-dimensional structure of the CFTR locus in a number of cell types that express CFTR or in which the gene is silent. This also enabled us to determine whether the intron 1 enhancer element, which is located 10 kb distal to the CFTR promoter, was brought into its close proximity by a looping mechanism to augment gene expression. We employed the technique of chromosome conformation capture (3C), which enables the identification of direct interactions between different parts of a locus in three dimensions . In 3C experiments, DNA–protein interactions in intact nuclei are fixed by formaldehyde. The cross-linked chromatin is then digested with a restriction enzyme followed by ligation. If there is a protein-mediated interaction between a remote regulatory element, for example in an intron, and the gene promoter, new chimaeric DNA molecules are formed that contain both elements. These chimaeric fragments can be determined by PCR with carefully designed primers. A fixed Taqman probe and reverse primer were designed within a HindIII fragment at the CFTR promoter, and multiple forward primers were generated within distal regions across the CFTR locus (Figure 3A). These forward primers were located within HindIII fragments encompassing the −20.9 kb, +6.8 kb and +15.6 kb DHS, and within specific intronic HindIII fragments. The assay fragments were positioned approx. 25–50 kb apart, such that they would give a good overall representation of the structure of the CFTR locus (Figure 3A). Real-time PCRs using the ‘fixed’ reverse probe/primer and each of the ‘variable’ forward primers enabled quantification of ligation events (subsequently referred to as ‘interaction frequency’) between the CFTR promoter and specific distal regions within each 3C sample. 3C was performed using chromatin prepared from primary epididymis cells and skin fibroblasts. The results shown in Figures 3(B) and 3(C) demonstrate that the three-dimensional structure of the CFTR locus shows a dramatic difference between CFTR-expressing cells (epididymis) and non-expressing cells (fibroblasts). In epididymis cells, the 3′-flanking region (encompassing the +6.8 kb DHS) is closely associated with the promoter region, as shown by its enrichment after 3C using a probe in the CFTR promoter. We have also demonstrated strong interactions between the promoter, the intron 1 DHS and other intronic DHS in other cell types  (C.J. Ott, N.P. Blackledge, J.E. Kerschner, G.E. Crawford, C.U. Cotton and A. Harris, unpublished work). In contrast, there are no significant interactions between the CFTR promoter and distal parts of the gene in skin fibroblasts.
3C analysis of the CFTR locus
We predict that looping of CFTR (Figure 3D), possibly induced by CTCF, enables key intronic cis-acting regulatory elements and others flanking the gene to move into close proximity to the CFTR promoter, so activating cell-type-specific expression.
Biochemical Basis of Respiratory Disease: Biochemical Society Focused Meeting held at AstraZeneca, Loughborough, U.K., 5–6 March 2009. Organized and Edited by Colin Bingle (Sheffield, U.K.) and Alan Wallace (AstraZeneca, U.K.).
ankyrin repeat, SAM (sterile α-motif) and basic leucine zipper
chromosome conformation capture
capping protein (actin filament) muscle Z-line α2
cystic fibrosis transmembrane conductance regulator
cortactin-binding protein 2
DNase I-hypersensitive site(s)
hepatocyte nuclear factor 1
Wingless-type MMTV (murine-mammary-tumour virus) integration site 2
yeast artificial chromosome
This work was supported by the Cystic Fibrosis Foundation USA [grant number Harris07PO], the National Institutes of Health [grant number NIH R01 HL094585], the Cystic Fibrosis Trust UK and the Children's Memorial Research Center. N.P.B. was the recipient of a scholarship from the Medical Research Council for part of this work.
Present address: Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, U.K.