Position-effect protection and enhancer blocking by the chicken beta-globin insulator are separable activities

Abstract

The 1.2-kb DNA sequence element (5'HS4) at the 5' end of the chicken beta-globin locus has the two defining properties of an insulator: it prevents an "external" enhancer from acting on a promoter when placed between them ("enhancer blocking") and acts as a barrier to chromosomal position effect (CPE) when it surrounds a stably integrated reporter. We previously reported that a single CTCF-binding site in 5'HS4 is necessary and sufficient for enhancer blocking. We show here that a 250-bp "core" element from within 5'HS4 is sufficient to confer protection against silencing of transgenes caused by CPE. Further dissection of the core reveals that 5'HS4 is a compound element in which it is possible to separate enhancer blocking and barrier activities. We demonstrate that full protection against CPE is conferred by mutant 5'HS4 sequences from which the CTCF-binding site has been deleted. In contrast, mutations of four other protein binding sites within 5'HS4 result in varying reductions in the ability to protect against CPE. We find that binding sites for CTCF are neither necessary nor sufficient for protection against CPE. Comparison of the properties of 5'HS4 with those of other CTCF-binding enhancer-blocking elements suggests that CPE protection is associated with maintenance of a high level of histone acetylation near the insulator, conferred by insulator binding-proteins other than CTCF.

Figure 1

(a) The chicken β-globin domain, the 5′ insulator region, and the neighboring domains. The DNase I-hypersensitive site 5′HS4 (with both enhancer-blocking and barrier properties) marks the 5′ end of the globin locus. Upstream is a 16-kb-long condensed chromatin region, and beyond that is a folate receptor gene (22). At the 3′ end of the β-globin locus is another constitutive hypersensitive site, 3′HS, with enhancer-blocking properties, and beyond that is a gene for an odorant receptor (12). (b) Reporter used in assays for position effects (the pGI′ vector). Plasmids were stably transformed into 6C2 cells, which were analyzed for IL2R expression on the cell surface by flow cytometry (Materials and Methods).

Figure 2

CTCF-binding sites are neither necessary nor sufficient for barrier activity. Two copies of test fragments were placed on each side of the IL2R reporter and stably integrated into 6C2 cells. Results of FACS analyses for IL2R expression at 0, 40, and 80 days after removal of hygromycin selection are shown; number of cells (y) versus intensity of antibody staining (x). A representative number of single-copy lines (numbered to the left of each set of panels) are shown unless otherwise stated. Data from all lines can be found online in Figs. 6–21. (a) A 250-bp core of the 5′HS4 insulator has full barrier activity. All lines express high levels of reporter IL2R at day 0; expression levels are maintained for at least 80 days in culture. The data are representative of the nine single-copy and eight multicopy lines tested (Figs. 6 and 7). (b) A 730-bp fragment of the 3′HS enhancer-blocking element does not have barrier activity. All lines (Fig. 8) lose recombinant IL2R expression by 20–40 days in culture. The lines shown are representative of the five single-copy and five multicopy lines tested (Fig. 8). (c and d) Deletion of the single binding site (FII) for CTCF has no effect on barrier activity of the 250-bp core or full 1.2-kb 5′HS4 insulators, respectively. The data are representative of the 10 single-copy and 12 multicopy core ΔII (Figs. 9 and 10) and the six single-copy and six multicopy 1.2ΔII lines tested (Figs. 11 and 12).

Figure 3

A comparison of enrichment for CTCF and histone H3 acetylated on Lys-9 and Lys-14 (K9 and K14) at locations across the chicken β-globin locus in 10-day chicken erythrocytes after chromatin immunoprecipitation and quantitative PCR analysis (see Materials and Methods). Locations of PCR primer sets are shown below on a map of the chicken β-globin neighborhood drawn to approximate scale. The data from chromatin preparations using anti-CTCF (solid line) or anti-acetyl K9 and K14 histone H3 (dotted line) antibodies are shown. The differences in DNA site enrichment represent the ratio of the immunoprecipitated DNA divided by the input DNA with all points for anti-acetyl K9 and K14 histone H3 divided by two for better comparison

Figure 4

Deletion of binding sites for proteins other than CTCF disrupt barrier activity of the core insulator. The results of FACS analyses for recombinant IL2R expression at 0, 20, 40, and 80 days after removal of hygromycin selection are shown as described for Fig. 2. (a) Deletion of the FI protein-binding site disrupts barrier activity. The lines show variable expression at day 0. Most lines lose IL2R expression by day 40. The data are representative of three single-copy and nine multicopy lines tested (Fig. 15). (b) Deletion of the FIII protein-binding site disrupts barrier activity. All lines show variable expression at day 0 but maintain levels of recombinant IL2R expression at least to day 80. The data are representative of the 12 single-copy and eight multicopy lines tested (Figs. 16 and 17). (c) Deletion of the FIV protein-binding site disrupts barrier activity. All multicopy lines display high levels of recombinant IL2R expression at day 0 but become silenced with time in culture. The data are representative of the 10 multicopy lines tested (Fig. 18). Deletion of the FV protein-binding site had a similar affect on the activity of the core (10 multicopy lines, Fig. 19). (d) The FI and FII protein-binding sites together are dispensable for barrier activity. A minimal fragment containing only FIII, IV, and V has near full barrier activity. All lines maintain high levels of expression for at least 40 days. Some extinction is observed in multicopy lines after 80 days. The data shown are representative of the nine single-copy and 10 multicopy lines tested (Figs. 20 and 21).

Figure 5

Summary of the barrier activities of wild-type and mutant insulator fragments. (Upper) A selected line that exhibits a biphasic expression pattern. Cells not expressing recombinant IL2R (shaded gray) are gated based on a control where no secondary antibody was used. (Lower) Percentages of cells expressing IL2R from all lines were averaged at each time point and plotted. Raw FACS plots used for this quantitation are shown in Figs. 6–21.