Mapping polycomb response elements at the Drosophilla melanogaster giant locus

Abstract

Polycomb-group (PcG) proteins are highly conserved epigenetic transcriptional regulators. They are capable of either maintaining the transcriptional silence of target genes through many cell cycles or enabling a dynamic regulation of gene expression in stem cells. In Drosophila melanogaster, recruitment of PcG proteins to targets requires the presence of at least one polycomb response element (PRE). Although the sequence requirements for PREs are not well-defined, the presence of Pho, a PRE-binding PcG protein, is a very good PRE indicator. In this study, we identify two PRE-containing regions at the PcG target gene, giant, one at the promoter, and another approximately 6 kb upstream. PRE-containing fragments, which coincide with localized presence of Pho in chromatin immunoprecipitations, were shown to maintain restricted expression of a lacZ reporter gene in embryos and to cause pairing-sensitive silencing of the mini-white gene in eyes. Our results also reinforce previous observations that although PRE maintenance and pairing-sensitive silencing activities are closely linked, the sequence requirements for these functions are not identical.

Keywords: epigenetic; gene silencing; pairing-sensitive silencing; polycomb; polycomb response element.

Figure 1

Pho binds to two regions within the giant cis-regulatory region that colocalizes with peaks of E(z) and Pc distribution. (A) Schematic representation showing the giant genomic region and flanking genes. Regions amplified by PCR in ChIP assays (1–13) are shown. Region 1 and 13 are, respectively, within the CG32797 and tko genes and serve as negative controls. (B) ChIP analysis of Oregon-R embryos 2 h 50 min to 3 h 20 min after egg lay shows two Pho peaks, one close to the gt promoter (region 4) and the other ∼6 kb upstream (region 9). E(z) and Pc, subunits of PRC2 and PRC1, respectively, colocalize with Pho but are more broadly distributed. Preliminary ChIP assays using 33 primer sets spanning this 19-kb region did not show the presence of Pho at additional sites (data not shown). ChIP was performed as indicated with anti-Pho antibody (upper right panel), anti-E(z) antibody (lower left panel), anti-Pc antibody (lower right panel), and rabbit preimmune antiserum for mock (upper left panel). ChIP signals are presented as a percent of input chromatin. Error bars represent SD.

Figure 2

The giant inserts tested for their abilities to maintain en-like expression pattern of β-galactosidase. (A) A schematic of gt upstream regulatory region showing fragments gt1–gt5 that were cloned into the SD10 vector and tested for PRE activity The locations of previously mapped gt enhancers gt_(-1), gt_(-3), and gt_(-6) are indicated in brackets. Pho-positive regions are indicated by arrows and correspond to PCR amplified regions 4 and 9 (Figure 1). (B) A schematic representation of the en-lacZ reporter construct SD10. Inserts are flanked by FRT sites. (C) Stage 14 embryos from transgenic lines stained for β-galactosidase expression. Lateral views of embryos are shown, anterior to the left, dorsal up. Transgenic lines tested are indicated to the left of the embryos. Expression patterns are representative of those produced by multiple lines of each construct. However, lines that failed to maintain the en-like pattern exhibited varying degrees of ectopic expression. Embryos containing intact transgenes are on the left. ΔInsert lines (right) are FLP recombinase–mediated deletion derivatives of the same lines shown on the left.

Figure 3

Maintenance of SD10-gt reporter repression is PcG-dependent. Stage 14 embryos from transgenic lines stained for β-galactosidase expression. Embryos with a wild-type PcG background are on the left. Embryos from crosses of transgenic lines to ph-d⁴⁰¹ ph-p⁶⁰² double mutants are on the right. Orientation and identification of embryos are the same as in Figure 2C.

Figure 4

The gt fragments exhibit pairing-sensitive silencing. (A) The number of transgenic lines exhibiting pairing-sensitive silencing (PSS) relative to total number of lines tested. The middle column lists the results for all lines tested for each construct. The right column lists the results for the subset of lines for which deletion-derivatives were generated. These are the subset of lines that were tested for β-galactosidase expression (Table 1) and that were homozygous-viable. (B) Eye pigmentation of SD10-gt transgenic flies. Specific transgenic lines are indicated to the left. Flies were either heterozygous (P[w+]/+) or homozygous (P[w+]/P[w+]) for the transgene. SD10-gt1 and SD10-gt5 homozygotes showed reduced eye pigmentation compared to heterozygotes (PSS). None of the SD10-gt2 or SD10-gt3 lines exhibited pairing-sensitive silencing. Eight of nine SD10-gt4 lines did not exhibit PSS. (C) Deletion lines corresponding to the same lines in (B) after excision of the gt fragment by FLP recombinase. PSS seen in the SD10-gt1.2 and SD10-gt5.4 homozygotes was lost upon deletion of the gt fragment. (D) Quantitative assays of fly eye pigments of the same lines shown in (B) and (C). yw is the y w^67c23 stock that contains the transgenes. The absorbance values for heterozygotes and homozygotes of each transgene are, respectively, illustrated as orange and blue bars. Error bars represent SD. There was a statistically significant difference between absorbance values for heterozygotes and homozygotes of each line. gt4.1 and Δgt4.1, P < 0.05. All other lines, P < 0.001.

Figure 5

Locations of Pho consensus binding sites within gt fragments. Schematic of gt fragments with Pho core consensus sites (GCCAT) in red. Lines above each gt fragment represent regions amplified by PCR in ChIP assays. Pho-positive regions (see Figure 1B) are indicated with asterisks.