A 1-megadalton ESC/E(Z) complex from Drosophila that contains polycomblike and RPD3

Abstract

Polycomb group (PcG) proteins are required to maintain stable repression of the homeotic genes and others throughout development. The PcG proteins ESC and E(Z) are present in a prominent 600-kDa complex as well as in a number of higher-molecular-mass complexes. Here we identify and characterize a 1-MDa ESC/E(Z) complex that is distinguished from the 600-kDa complex by the presence of the PcG protein Polycomblike (PCL) and the histone deacetylase RPD3. In addition, the 1-MDa complex shares with the 600-kDa complex the histone binding protein p55 and the PcG protein SU(Z)12. Coimmunoprecipitation assays performed on embryo extracts and gel filtration column fractions indicate that, during embryogenesis E(Z), SU(Z)12, and p55 are present in all ESC complexes, while PCL and RPD3 are associated with ESC, E(Z), SU(Z)12, and p55 only in the 1-MDa complex. Glutathione transferase pulldown assays demonstrate that RPD3 binds directly to PCL via the conserved PHD fingers of PCL and the N terminus of RPD3. PCL and E(Z) colocalize virtually completely on polytene chromosomes and are associated with a subset of RPD3 sites. As previously shown for E(Z) and RPD3, PCL and SU(Z)12 are also recruited to the insertion site of a minimal Ubx Polycomb response element transgene in vivo. Consistent with these biochemical and cytological results, Rpd3 mutations enhance the phenotypes of Pcl mutants, further indicating that RPD3 is required for PcG silencing and possibly for PCL function. These results suggest that there may be multiple ESC/E(Z) complexes with distinct functions in vivo.

FIG. 1.

PCL is physically associated with RPD3, SU(Z)12, E(Z), ESC and p55 in vivo. (A) Zero- to 16-h embryo nuclear extracts (NE) were separated by SDS-11% PAGE, and PCL was detected by Western with affinity purified anti-PCL antibodies. No signal was detected with the rabbit preimmune serum. (B) Co-IP of PCL, ESC, and E(Z) with SU(Z)12 and RPD3 from 0- to 16-h whole-embryo extracts by use of protein G Sepharose and purified rabbit anti-PCL (lane 3), anti-ESC (lane 4), and anti-E(Z) antibodies (lane 5). Preimmune rabbit serum was used in lane 2 for negative control. SU(Z)12 degradation product (lower band in the second panel) was detected in whole-embryo extracts but not in nuclear extracts (Fig. 4A, panel 5). (C) Purified anti-PCL antibodies were coupled to CNBr-activated Sepharose to generate an affinity gel. Embryo extracts (lane 1) were mixed with the affinity gel. After extensive washing, bound proteins were eluted by 0.1 M glycine (pH 2.7) and detected in lane 3. Lane 2 shows a parallel control with preimmune serum-coupled gel. (D) GST pulldown assay with use of GST-RPD3 (lane 3) and 0- to 16-h nuclear extracts (NE). Lanes 1 show 10% (B) and 20% (C and D) of the total sample used. Proteins were detected by Western blotting and were indicated at the right side of each panel. The bottom gels (B, C, and D) show that p105, which is associated with p55 in the abundant ∼600-kDa CAF-1 complex, is not present in immunoprecipitates, indicating that immunoprecipitates are not contaminated with CAF-1.

FIG. 2.

Identification of an ESC/E(Z) complex that contains PCL and RPD3. (A) Freshly made wild-type 0- to 16-h embryo nuclear extracts (NE) (0.2 ml) were fractionated on a Superose 6 HR 10/30 column. The protein in each fraction (fraction number indicated at top of each lane) was analyzed by Western blotting and is indicated at the right side of each panel. The column was calibrated with a series of standard proteins, and mass is indicated at the top with arrows. Fraction 6 is void volume. It is estimated that fraction 10 (or 9) is approximately 2 MDa, based on the presence of PRC1 (PSC) in fractions 8 to 11. (B) EtBr-treated fraction samples from panel A were used for co-IP with anti-ESC (panels 1 to 5), anti-SU(Z)12 (panels 7 to 11), and anti-PCL (panels 13 to 16). Preimmune serum was used for negative controls for IP from fractions and showed no signal (see lanes marked “−” in panels 7 to 11). Forty microliters of NE was diluted four times in the column elution buffer and was used for co-IP in the absence of anti-PCL antibodies, which was also served as negative control (lane C) in panels 13 to 16. Panel 12 shows PCL in each fraction used for IP in panels 13 to 16. Zero- to 16-h nuclear extracts from embryos expressing FLAG-ESC were fractionated as for Fig. 2A. And then EtBr-treated fraction samples were used for co-IP with anti-FLAG M2 gel. p55 eluted from the M2 gel by the FLAG peptide was analyzed by Western blotting for panel 6. Wild-type nuclear extract control on M2 gel was performed as previously described (47), showing no p55 or p105.

FIG. 3.

In vitro binding of PCL and RPD3. In vitro translated ³⁵S-labeled PCL (full-length and truncated forms) and RPD3 proteins (input lanes in panels A to D), as well as GST fusion proteins (indicated at top of appropriate lanes), were used to test direct interaction of PCL and RPD3 and to map regions responsible for interactions. Lanes with GST alone in panels A through D serve as negative controls and show background binding. Signals that are ≥3-fold higher than background were considered significant specific binding. Proteins were separated by SDS-PAGE and visualized by autoradiography (A to D). Purified GST fusion proteins used for panel D were shown in panel E with Coomassie blue staining. (F and G) Schematic summary of results of in vitro binding experiments and the PCL and RPD3 constructs tested. For PCL (F), the highly conserved PHD fingers are indicated by a white box, and the serine-threonine-rich region (S/T) is indicated by a gray box. For RPD3 (G), the highly conserved N terminus is indicated by a gray box. The in vitro binding result (+ or −) is indicated to the right of each construct. Note that the PCL constructs containing both PHD fingers (PCL423-567) and the RPD3 constructs containing the N terminus (24-129) retain binding activity. a.a., amino acids.

FIG. 4.

The association of PCL with ESC, E(Z), and SU(Z)12 is no longer detectable by late embryogenesis. (A) EtBr-treated 0- to 16-h and 18- to 24-h embryo nuclear extracts (type of extract indicated at the left side of each panel) were used for co-IP assays with rabbit anti-PCL (lane 3), anti-ESC (lane 4), and anti-E(Z) antibodies (lane 5) and preimmune rabbit serum (lane 2). Other proteins present in immunoprecipitates were detected by Western blotting with rabbit anti-PCL and anti-RPD3 antibodies, mouse anti-E(Z) antibodies, and chicken anti-SU(Z)12 antibodies (antibody type indicated at the right side of each panel). (B) Fractionation of 18- to 24-h nuclear extracts (NE) on a Superose 6 column, same as for Fig. 3A.

FIG. 5.

Colocalization of PCL with RPD3 and E(Z) on salivary gland polytene chromosomes. PCL binding sites on chromosomes were detected with affinity-purified anti-PCL primary antibodies and fluorescein isothiocyanate-labeled secondary antibodies (green [A and D]). RPD3 and E(Z) were detected by sequential immunofluorescence staining by using these antibodies as second primary and Texas red-labeled second secondary antibodies on the same chromosomes. Panels C and F show merged images of panels A and B and panels D and E, respectively. Yellow bands (C and F) indicate sites of colocalization. Control experiments performed in parallel, where the second primary antibodies were omitted, showed no signal due to Texas red-labeled secondary antibody binding to the first primary antibody.

FIG. 6.

(A to F) Binding of PCL, SU(Z)12, and RPD3 to the Ubx PRE in vivo. The chromosomal binding sites of PCL (red [A and B]), SU(Z)12 (C and D), and RPD3 (E and F) on the distal portion of chromosome arm 3L are shown in a wild-type larva (wt) (A, C, and E) and in a transformant (PRE) (B, D, and F) containing the 670-bp PstI-NdeI fragment of the Ubx PRE inserted at 65B. A landmark site at 65D, which is bound by PCL and other PcG proteins, is indicated by a white line. The arrows indicate new PCL, SU(Z)12, and RPD3 binding sites created at the insertion site of the PRE construct in the transformant (B, D, and F); no band is present at this site in wild-type controls (A, C, and E). The RPD3 binding site at the insertion site of the PRE construct (F) was further confirmed by double staining of PCL and RPD3 (data not shown). (G) Typical transformation of T2 to T1 leg identity seen in over 90% of Pcl¹¹/+; Rpd3³⁰³/+ double heterozygotes. (H) Typical A4-to-A5 transformation seen in 35% Pcl⁷/+; Rpd3³⁰³/+ double heterozygotes.