The Drosophila esc and E(z) proteins are direct partners in polycomb group-mediated repression

Abstract

The extra sex combs (esc) and Enhancer of zeste [E(z)] proteins are members of the Drosophila Polycomb group (Pc-G) of transcriptional repressors. Here we present evidence for direct physical interaction between the esc and E(z) proteins using yeast two-hybrid and in vitro binding assays. In addition, coimmunoprecipitation from embryo extracts demonstrates association of esc and E(z) in vivo. We have delimited the esc-binding domain of E(z) to an N-terminal 33-amino-acid region. Furthermore, we demonstrate that site-directed mutations in the esc protein previously shown to impair esc function in vivo disrupt esc-E(z) interactions in vitro. We also show an in vitro interaction between the heed and EZH1 proteins, which are human homologs of esc and E(z), respectively. These results suggest that the esc-E(z) molecular partnership has been conserved in evolution. Previous studies suggested that esc is primarily involved in the early stages of Pc-G-mediated silencing during embryogenesis. However, E(z) is continuously required in order to maintain chromosome binding by other Pc-G proteins. In light of these earlier observations and the molecular data presented here, we discuss how esc-E(z) protein complexes may contribute to transcriptional silencing by the Pc-G.

FIG. 1

Tests for esc-E(z) interaction in the yeast two-hybrid system. (A) Growth tests with EGY48 cells containing the following combinations of plasmids are shown: left, pEG202-esc and pJG4-5-E(z), which express LexA-esc and AD-E(z), respectively; middle, pRFHMø, which does not produce a LexA protein, and pJG4-5-E(z); right, pEG202-esc and pJG4-5 (no insert). Yeast strains were streaked in parallel on media lacking leucine and containing galactose (top) or glucose (bottom). Expression of AD fusion proteins from pJG4-5 and its derivatives was induced on galactose medium but not on glucose medium. Yeast strains were grown for 4 days at 30°C. (B) X-Gal assays of yeast strains containing the following combinations of plasmids: left, pEG202-E(z) and pJG4-5-esc; middle, pEG202-E(z) and pJG4-5; right, pRFHMø and pJG4-5-esc. Yeast strains were grown for 48 h on X-Gal indicator medium containing galactose (top) or glucose (bottom).

FIG. 2

Coimmunoprecipitation of HA-esc and E(z) from Drosophila embryo extracts. Proteins were immunoprecipitated from embryo extracts with either anti-HA or anti-E(z) antibodies as indicated. Equal amounts of immunoprecipitates were separated on SDS gels, transferred to nitrocellulose filters, and incubated with anti-HA (A) or anti-E(z) (B) antibodies. Lanes: 1, 30 μg of extract from HA-esc embryos; 2, 30 μg of extract from y w embryos; 3, mock immunoprecipitation from HA-esc extract by protein (Prot.) A-Sepharose without antibody; 4, immunoprecipitation from HA-esc extract with anti-E(z) antibody; 5, immunoprecipitation from y w extract with anti-E(z) antibody; 6, mock immunoprecipitation from HA-esc extract by protein G-Sepharose without antibody; 7, immunoprecipitation from HA-esc extract with anti-HA antibody; 8, immunoprecipitation from y w extract with anti-HA antibody. Bands corresponding to HA-esc and E(z) are indicated by arrowheads. In panel B, lanes 1 and 2, the smaller species are E(z) degradation products. Signals indicated by asterisks in panel A, lanes 7 and 8, and panel B, lanes 4 and 5, are due to cross-reactivity between the secondary antibodies and the heavy chains of the antibodies used in immunoprecipitations. Numbers at left of panels are kilodaltons.

FIG. 3

In vitro binding of the esc and E(z) proteins. (A) Autoradiographs of SDS gels. Radiolabeled esc protein (lanes 1 to 3) and radiolabeled E(z) protein (lanes 4 to 6) were tested for binding to GST or GST fusion proteins as indicated. The input lanes (1 and 4) contain 4% the amount of radiolabeled protein used in the binding assays. (B) Western blots. Proteins from Drosophila HA-esc embryo extracts were tested for binding to GST or GST fusion proteins. Proteins that bound to the indicated GST or GST fusion proteins were separated by SDS-PAGE, transferred to nitrocellulose, and incubated with anti-HA (lanes 1 to 3) or anti-E(z) (lanes 4 to 6) antibodies. Lanes labeled “Sup.” contain 4% the unbound material recovered after incubation of the HA-esc extracts with the GST-E(z) (lane 1) or GST-esc (lane 4) fusion proteins. Numbers at left of panels are kilodaltons.

FIG. 4

Localization of the E(z) domain that mediates binding to esc protein. GST, GST-E(z) (full length, amino acids 1 to 760), and GST-E(z) deletion derivatives were tested for binding to radiolabeled esc protein produced by in vitro translation (A) or HA-esc protein from embryo extracts (B). The GST fusion proteins used are indicated above the lanes; the numbers refer to E(z) amino acids included in each fusion protein. (A) Autoradiograph of SDS gel. (B) Western blot probed with anti-HA antibodies. Lanes labeled “Input” and “Sup.” are as described in the legend to Fig. 3. Numbers at left of panels are kilodaltons.

FIG. 5

Effects of E(z) point mutations on binding to esc protein. (A) Diagram of E(z) protein and amino acid sequence of the esc-interacting domain. (Top) Evolutionarily conserved E(z) regions are depicted as shaded or hatched boxes. Names of these domains and/or percent identities between fly E(z) and mammalian homologs are indicated. (Bottom) Alignment of the esc-interacting domain (amino acids 34 to 66) with homologous regions of the human EZH1 and EZH2 proteins and mouse Ezh1 and Ezh2 proteins (amino acids 39 to 71). Residues that are identical in all five proteins are boxed. Asterisks indicate E(z) residues that were substituted by alanine in this study. (B) Effects of E(z) alanine substitutions on binding to in vitro-translated esc protein. GST fusion proteins containing E(z) amino acids 1 to 218 and the indicated substitution mutations were incubated with radiolabeled esc protein. Equal amounts of bound protein samples were electrophoresed on an SDS gel. Relative band intensities on the autoradiograph were measured with the spot densitometry feature of the AlphaImager 2000 gel documentation system (Alpha Innotech). (C) Effects of E(z) alanine substitutions on binding to HA-esc from embryo extracts. Mutant GST fusion proteins were incubated with HA-esc embryo extracts. Bound proteins were separated by SDS-PAGE and transferred to nitrocellulose, and HA-esc protein was detected with anti-HA antibodies. Lanes labeled “Input” and “Sup.” are as described in the legend to Fig. 3. Numbers at left of panels B and C are kilodaltons.

FIG. 6

Effects of clustered alanine mutations in esc on binding to E(z) protein. Wild-type (esc) or mutant (RDE216, DFST278, GG210, and RD282) esc proteins were tested for binding to GST-E(z) fusion protein in vitro. The RDE216 mutation changes RDE to AAA, DFST278 changes DFST to AFAA, GG210 changes GG to AA, and RD282 changes RD to AA (54). Wild-type and mutant esc proteins were tested as radiolabeled, full-length proteins. Lanes labeled “s” contain 4% of the first supernatant removed after the binding reaction, and lanes labeled “b” contain bound radiolabeled protein after extensive washing. Lanes 1 and 2 and lanes 7 and 8 show two independent binding assays with wild-type esc.

FIG. 7

Tests for in vitro binding of human EZH1 to human eed (heed) and Drosophila esc proteins. Autoradiographs of SDS gels are shown. (A) Radiolabeled, full-length heed was tested for binding to GST alone, full-length human EZH1 (GST-EZH1), or an N-terminal portion of EZH1 [GST-EZH1(1-166)]. (B) Radiolabeled, full-length Drosophila esc protein was tested for binding to the same GST fusion proteins as in panel A. Lanes labeled “s” and “b” contain unbound supernatant and bound protein samples, respectively, as described in the legend to Fig. 6. Numbers at left of panels are kilodaltons.