Genomic approaches uncover increasing complexities in the regulatory landscape at the human SCL (TAL1) locus

Abstract

The SCL (TAL1) transcription factor is a critical regulator of haematopoiesis and its expression is tightly controlled by multiple cis-acting regulatory elements. To elaborate further the DNA elements which control its regulation, we used genomic tiling microarrays covering 256 kb of the human SCL locus to perform a concerted analysis of chromatin structure and binding of regulatory proteins in human haematopoietic cell lines. This approach allowed us to characterise further or redefine known human SCL regulatory elements and led to the identification of six novel elements with putative regulatory function both up and downstream of the SCL gene. They bind a number of haematopoietic transcription factors (GATA1, E2A LMO2, SCL, LDB1), CTCF or components of the transcriptional machinery and are associated with relevant histone modifications, accessible chromatin and low nucleosomal density. Functional characterisation shows that these novel elements are able to enhance or repress SCL promoter activity, have endogenous promoter function or enhancer-blocking insulator function. Our analysis opens up several areas for further investigation and adds new layers of complexity to our understanding of the regulation of SCL expression.

Figure 1. Histone H3 modification profiles across the human SCL locus in the K562 cell line.

The modifications studied are named at the left of each panel. Dots on the solid joined-up lines represent the data obtained for each genomic tiling array element. In each panel, the x-axis is the genomic sequence co-ordinate (NCBI build 35) and the y-axis is the enrichment obtained in ChIP-chip assays expressed in log₂ scale. Schematic diagram at the bottom of the figure shows the genomic organisation of SCL and its neighbouring genes. Exons are shown as vertical blocks with gene names and direction of transcription shown above. Transcripts denoted by a,b and c refer to transcripts of unknown function (see also text). Vertical lines at the bottom (with dotted lines through all the panels) show the location of known and novel regulatory regions at the SCL locus. Promoters are denoted by P. Other nomenclature refers to the distance in kb from SCL pro1a (P^1a). Novel regions are highlighted in bold at the bottom and with arrows at the top of the figure.

Figure 2. CTCF, PolII and TAFII 250 binding, chromatin accessibility and nucleosome density profiles across the human SCL locus in the K562 cell line.

Assays used are named at the left of each panel. Dots on the solid joined-up lines represent the data obtained for each genomic tiling array element. In panel E, the profiles for histone H2B and histone H3 are shown as dotted joined-up line with triangles and solid joined-up lines with dots for H2B and H3 respectively. All other aspects of the figure are as in Figure 1.

Figure 3. Treeview diagram of hierarchical clustering of genomic tiles across the human SCL locus for histone modifications in the K562 cell line.

Each tile is represented by a horizontal bar and is shaded in red scale or green scale to denote level of enrichment or depletion respectively. The tree diagram is shown to the left of the figure. Two branches of the tree and the histone modifications studied are shown in the enlarged panel to the right of the figure. Each genomic tile in the enlarged panel is named according to its distance in kb relative to SCL pro1a or to the promoter of its closest known gene. Tiles representing SCL known and novel regulatory elements are denoted by an arrow and highlighted in bold.

Figure 4. Transient reporter assays of novel regulatory elements at the human SCL locus.

A. Enhancer/repressor activity of six novel regulatory elements in K562 (erythroid) and HPB-ALL (lymphoid) cell lines. The y-axis shows the fold increase/decrease in luciferase activity relative to the pGL3 basic negative control construct (not shown). The x axis shows constructs under the control of the SV40 promoter (SV40 pro) or the promoter 1a of SCL (SCL pro1a). All constructs under the control of these two promoters are named according to the region being tested based on their distance in kb from the SCL pro1a (−  =  upstream; +  =  downstream). The SV40pro and SCLpro1a control lanes shown were derived from transfections with constructs containing a sequence having no regulatory function cloned into the BamHI site. Luciferase expression for the SV40 promoter in combination with the SV40 enhancer (SV40 pro/en) is also shown as a positive control. Standard error bars are shown. B. Promoter assays of the +53 region in K562 (erythroid) and HPB-ALL (lymphoid) cell lines. The y-axis shows the fold increase/decrease in luciferase activity relative to the pGL3 basic negative control construct. The x axis shows luciferase expression driven by pGL3 basic, SCL pro1a only (SCL pro1a positive control), SV40 pro (SV40 promoter positive control), and the two +53 constructs (SCL+53u and SCL+53d). The +53 region was cloned in one (u) or the other (d) orientation. Standard error bars are shown. C. Enhancer blocking assays of regions at the SCL locus which bind CTCF in the K562 cell line. The y-axis shows enhancer blocking activity (G418-resistant colony counts) for the pNI and pNI-CD (negative and positive controls) and four putative insulator regions from the human SCL locus. Insulator constructs are named according to the region being tested based on their distance in kb from the SCL pro1a (−  =  upstream; +  =  downstream). Standard error bars are shown. Note: Enhancer blocking assays were performed only in the K562 cell line as expression of the pNI-based vector is regulated by the murine HS2 β-globin enhancer which is active only in erythroid cells.

Figure 5. ChIP-chip analysis of the human SCL erythroid enhancer region in SCL-expressing and non-expressing haematopoietic cell lines.

A. Binding of the SCL-containing erythroid complex (SEC) of transcription factors in the K562 cell line. E47 and E12 are two isoforms of E2A. B. CTCF binding in the K562 and U937 cell lines. C. TAFII 250 binding in the K562 and U937 cell lines. D. PolII binding in K562 and U937 cell lines. E. Histone H3 K4me3 in the K562, U937, HPB-ALL and HL-60 cell lines. F. Histone H3 K4me2 in the K562, U937, HPB-ALL and HL-60 cell lines. G. Histone H3 K4me1 in the K562, U937, HPB-ALL and HL-60 cell lines. H. Histone H3 K9/K14ac in the K562, U937, HPB-ALL and HL-60 cell lines. All enrichments obtained in ChIP-chip assays are expressed in log₂ scale in the figure. Schematic at the bottom shows the 4 kb of genomic DNA surrounding the SCL erythroid enhancer. Genomic sequence co-ordinates are from NCBI build 35. Numbering of the blocks of DNA (with dotted lines through all the panels) refers to the distance in kb from SCL pro1a. The vertical dotted lines are positioned at the mid-point of each of the 1kb blocks and do not define positions of regulatory regions per se.