ESC originally appeared to be unique among PcG

ESC originally appeared to be unique among PcG proteins in being required predominantly during early embryogenesis. Temperature shift experiments with a temperature-sensitive esc allele suggested that esc is required only during early embryogenesis (Struhl and Brower, 1982) and similar experiments with a heat-inducible hsp70-esc transgene also suggested that early esc LY3009120 is sufficient to promote normal development (Simon et al., 1995). This early requirement for esc is reflected in its temporal expression profile: esc mRNA is most abundant in early embryos, peaking at 8 h (Gutjahr et al., 1995, Sathe and Harte, 1995) and subsequently declines to almost undetectable levels by the end of embryogenesis. Similarly, the ESC protein is present at high levels during the first half of embryogenesis, peaking at mid-embryogenesis and declining to barely detectable levels by first instar (Simon et al., 1995). In contrast, the sole mammalian ESC ortholog, EED, appears to be expressed and required continuously (Schumacher et al., 1996). The temporal profile of esc expression suggested that ESC might be specifically required only for the establishment but not the subsequent maintenance of Polycomb silencing. E(Z), however, is required continuously throughout development (Beuchle et al., 2001), and recent biochemical studies demonstrate that ESC is required for E(Z) HMTase activity both in vitro (Czermin et al., 2002, Nekrasov et al., 2005) and in vivo, at least during embryogenesis (Ketel et al., 2005). This suggested that this essential function of ESC may be carried out by another protein after ESC levels drop. One obvious possibility was that such a protein would be similar to ESC itself. When the complete sequence of the Drosophila melanogaster genome became available, we conducted a BLASTP search using the ESC protein sequence as a query and identified a single predicted protein with a high degree of sequence similarity to ESC. This protein is encoded by the CG5202 gene, which we have renamed extra sex combs-like (escl). While this work was in progress, Wang et al. (2006) demonstrated that recombinant ESCL can substitute for ESC in reconstituted PRC2 complexes in an in vitro histone H3K27 methylation assay, indicating that their biochemical functions are similar if not identical. We report here that ESCL protein exhibits a temporal expression profile that is complementary to that of ESC, including substantially higher levels of expression during larval and adult stages than embryogenesis. We show that ESCL, like ESC, binds directly to E(Z) via its WD repeats and is physically associated with E(Z), SU(Z)12, PCL and p55 in vivo. While RNAi-mediated knockdown of either ESC or ESCL in S2 cells has no appreciable effect on histone H3K27 methylation, simultaneous knockdown of both strongly reduces tri-methyl and di-methyl (3m and 2m) H3K27 levels. Some ESCL is bound to the Ubx PRE in Drosophila Kc cells and this binding is increased when ESC is depleted by RNAi. A strong escl mutation is viable and fertile but enhances the phenotypes of PcG mutants, consistent with a role in Polycomb silencing. Genetic analysis reveals that the well-known “maternal rescue” of esc− embryos to viable adults requires ESCL. GAL4-driven constitutive maternal and zygotic ESCL expression can fully substitute for ESC in vivo, indicating that their functions are qualitatively indistinguishable if not identical. We discuss why two apparently functionally equivalent proteins have been retained during Drosophila evolution.