TAP of Pcl protein complexes from Drosophila embryonic nuclear extracts. Protein complexes purified from wild-type (wt) and TAP-Pcl; Pcl^+/+ embryos. Purified material was separated on a 4–12% polyacrylamide gel and visualized by silver staining; ‘M' indicates molecular weight marker. Input material for mock purification from wild-type embryos, and for purification from transgenic embryos was normalized by protein concentration and equivalent amounts of material eluted from calmodulin affinity resin was loaded. Complexes were eluted with EGTA under non-denaturing conditions. Approximately 0.67 pmol of Pcl–PRC2 complex is loaded, this amount was determined by comparing with mass spectrometry analysis of defined amounts of purified recombinant Pcl, E(z) and Su(z)12 (not shown, see Figure 3). Indicated proteins consistently co-purified with CBP-Pcl in several independent experiments and were identified by microsequencing; CBP in fusion proteins refers to the calmodulin-binding moiety of the TAP-tag. Note presence of PRC2 subunits Su(z)12, E(z), NURF55 and Esc in TAP-Pcl lane; Esc stains poorly and runs as a diffuse band, but is identified by multiple peptides like the other proteins (Supplementary Figure S1). See Supplementary Figure S1 for information on additional proteins identified by mass spectrometry.

HMTase activity of Pcl-PRC2 and PRC2 in vitro. (A) Molecular weight marker (lanes 1 and 14) and HMTase reactions (lanes 2–13) performed with 2 pmol recombinant PRC2 (lanes 2–5), no complex (‘mock', lanes 6–9) or 0.67 pmol Pcl-PRC2 (lanes 10–13), [¹⁴C]-SAM (lanes 2–13) and 2.8 pmol wild-type (wt) or mutant (H3-K9A, H3K27A and H3-K9A/K27A) mononucleosomes as indicated, were resolved on a 20% SDS–polyacrylamide gel. The gel was stained with Coomassie (top) and exposed for autoradiography (below). Note that Pcl-PRC2 and PRC2 both specifically methylate H3-K27 (compare lanes 2–5 with 10–13), no methylation is observed on nucleosomes with H3 containing the K27A mutation. (B) Mass spectrometry analysis of H3 following HMTase reactions with Pcl-PRC2 (top) or PRC2 (bottom). HMTase reactions were performed as in lanes 2 and 10 in panel A except that non-radioactive SAM was used and reactions were allowed to proceed for 12 h; excised H3 bands were digested and analyzed by reverse-phase chromatography and quantitative mass spectrometry. Chromatography separated modified and unmodified H3_KSAPATGGVK peptides by only ∼95 s. Therefore a single mass spectrum can show ions of unmodified and all methylation states of H3_KSAPATGGVK peptides during their elution. The right lane shows the ion volumes (thompson^*sec) of unmodified and modified H3_KSAPATGGVK peptides after HMTase reaction. The ion volume is the summed ion intensity of a peptide over its elution time (Fraterman et al, 2007). Note that Pcl-PRC2 and PRC2 generate comparable ion volumes of mono-, di- and trimethylated H3-K27 in this assay. The error bars indicate the standard deviation of a duplicate analysis.

Pcl is required for repression and H3-K27 trimethylation at PcG target genes. (Top) Wild-type (wt, top row) and Pcl^22M21 homozygous (Pcl^−/−, bottom row) embryos stained with antibodies against Ubx, Abd-B, Engrailed (En), Wingless (Wg), Caudal (Cad) and Distal-less (Dll) proteins. In Pcl^−/− embryos, Ubx and Abd-B are misexpressed outside of their normal expression domains; arrowheads mark anterior margin of ps5 and ps10 that correspond to the anterior boundary of the Ubx and Abd-B expression domain in wt animals, respectively. Expression of En is essentially normal in Pcl^−/− animals and only very few En-positive cells that are not visible here are present in addition to the wild-type En expression pattern in the posterior compartment of every segment (cf. Moazed and O'Farrell, 1992). The complex expression pattern of Wg, expression of Cad at the posterior end of the embryo (arrowhead) and Dll expression in the head and thorax (arrowheads) are all normal in Pcl^−/− animals and no misexpression is detected. The slightly reduced levels of Dll expression in the imaginal disc primordia in thoracic segments (arrowheads) in Pcl^−/− animals is likely due to downregulation by misexpressed BX-C proteins (Vachon et al, 1992). (Bottom). X-ChIP analysis in wild-type (wt, purple bars) and Pcl mutant (Pcl^−/−, yellow bars) embryos, respectively. Each bar shows the result from at least three independent immunoprecipitation reactions on independently prepared batches of chromatin, performed with the indicated antibodies against H3-K27me3, H3-K27me2, H3-K27me1, unmodified H3, Su(z)12, Pho or Ph; X-ChIP signal levels are presented as percentage of input chromatin precipitated for each region. The location of PREs (blue boxes) and other regions with respect to the transcription start sites are indicated in kilobases, see Supplementary Figure S4 for information on exact location of PCR primer pairs. Note that at each region, H3 X-ChIP signals are comparably high in wt and Pcl mutant embryos. The comparably lower H3-X-ChIP signal at each PREs in both wt and Pcl mutant embryos suggests inefficient detection of nucleosomes or, more likely, reduced nucleosome occupancy at PREs, as previously observed at Ubx (Kahn et al, 2006; Mohd-Sarip et al, 2006; Papp and Müller, 2006). In wt animals, H3-K27me3 X-ChIP signals are enriched at all target genes compared to control regions 1 and 2. Note that in Pcl mutant embryos, H3-K27me3 signals are at least two-fold lower in each region; only at Dll, H3-K27me3 signals are unaltered. In wild-type embryos, H3-K27me1 and H3-K27me2 signals at target genes are overall lower than in control regions 1 and 2. Note that in Pcl mutants H3-K27me1 and H3-K27me2 signals are strongly increased at target genes and that they are also increased at control regions 1 and 2 (see text for details). Note the specific enrichment of Su(z)12, Pho and Ph at PREs. In Pcl mutant embryos, binding of Pho and Ph is undiminished compared to wild-type embryos but Su(z)12 binding is reduced. Note that Su(z)12 X-ChIP signals at the bxd PRE in Pcl mutants are still higher than in control regions 1 and 2. Compared to X-ChIP assays in imaginal discs (Papp and Müller, 2006), we find that in embryos a smaller fraction of input material is precipitated with Pho, Ph and, in particular, with Su(z)12 antibodies. It is possible that the intrinsically different fixation procedure is responsible for a lower crosslinking efficiency in embryos compared with discs. Asterisks (^*) indicate regions where X-ChIP signals were not measured.

TAP of E(z) protein complexes from Drosophila embryonic nuclear extracts. (A) Protein complexes purified from wild-type (wt) and TAP-E(z); E(z)^+/+ embryos. Purified material was separated on a 4–12% polyacrylamide gel and visualized by silver staining. Input material for mock purification from wild-type embryos and for purification from transgenic embryos was normalized by protein concentration and equivalent amounts of material eluted from calmodulin affinity resin was loaded; complexes were eluted by boiling of calmodulin resin in SDS buffer. Indicated proteins consistently co-purified with CBP-E(z) in several independent experiments and were identified by microsequencing; CBP in fusion proteins refers to the calmodulin-binding moiety of the TAP-tag. PRC2 subunits Su(z)12, Nurf55 and the weak Pcl band in TAP-E(z) lane were all identified by multiple peptides (Supplementary Figure S1) in excised gel bands; Esc was not detected on the gel. See Supplementary Figure S1 for information on additional proteins identified by mass spectrometry. (B) Western blot analysis of total embryonic nuclear extract input material (IN, lanes 1 and 2) from wild-type (wt) and TAP-E(z) transgenic embryos, and material eluted from calmodulin affinity resin (E, lanes 3 and 4) after purification. All panels come from the same batch of input material, and eluates were all from the same batch of material purified from wild-type and TAP-E(z) embryos, respectively; the same ratio of input versus eluate material was loaded in all cases. CBP-E(z) (red asterisk), TAP-E(z) (blue asterisk) and endogenous E(z) (black asterisk) are indicated; in lane 2, TAP-E(z) is also detected by other antibodies due to protein A tag. Compare the relative enrichment of E(z), Su(z)12 and Pcl in lane 4. For unknown reasons, Su(z)12 and Pcl in lane 4 migrate with slightly higher mobility. Note the lack of signals for HDAC, Pc and Pho in lane 4.

Pcl is needed for high levels of H3-K27 trimethylation in Drosophila. (A) Western blot analysis of extracts from 16- to 18-h-old wild-type (wt) and Pcl^22M21 homozygous (Pcl^−/−) embryos, probed with antibodies against the indicated proteins; the anti-tubulin (a-Tub) Western blot provides control for loading of equal amounts of extract. Note that in Pcl^−/− embryos, no Pcl protein is detected, but that Su(z)12 and E(z) protein levels are comparable to wt embryos. (B) Wing imaginal discs with clones of cells that are homozygous for E(z)⁷³¹ (E(z)), Su(z)12⁴ (Su(z)12), or Pcl^21M22 (Pcl), stained with antibodies against H3-K27me3 (red signal, left column) or H3-K27me1 (red signal, right column). In each case, clones of mutant cells are marked by the absence of GFP signal (green) and discs were analyzed 96 h after clone induction. Su(z)12 and E(z) mutant clones show complete loss of H3-K27me3 and H3-K27me1 signals (arrowheads). Pcl mutant clones also show a clear reduction of H3-K27me3 levels (arrowheads), but H3-K27me1 levels are unaffected (empty arrowheads) and are comparable to those in neighboring wild-type cells. Representative clones (white frame) are shown at higher magnification (C) Western blot analysis of extracts from imaginal disc and CNS tissues from second instar larvae (larval extract, left column) or from 16- to 18-h-old embryos (embryo extract, right column) of the following genotypes: wild-type (wt), Su(z)12⁴/Su(z)12⁴ (Su(z)12^−/−), E(z)⁷³¹/E(z)⁷³¹ (E(z)^−/−), Pcl^22M21/Pcl^22M21 (Pcl^−/−). Recombinant histone octamer (rec. oct.) reconstituted from Xenopus histones expressed in Escherichia coli served as additional control for H3-K27me antibody specificity. In each case, the membrane was simultaneously probed with the antibody against the indicated H3-K27 methylation state and with an antibody against unmodified histone H4, to control for equal extract loading and Western blot processing. Note that in E(z)^−/− and Su(z)12^−/− larvae, H3-K27me3 Western blot signals are comparable to those observed in wt larvae and only H3-K27me1 and H3-K27me2 signals are detectably reduced. In Pcl^−/− embryos, H3-K27me1, H3-K27me2 and H3-K27me3 Western blot signals are comparable to those observed in wt embryos.