Trimethylation of histone H3 lysine 27 (H3K27me3) by Polycomb repressive complex 2 (PRC2) is essential for transcriptional silencing of Polycomb target genes, whereas acetylation of H3K27 (H3K27ac) has recently been shown to be associated with many active mammalian genes. The Trithorax protein (TRX),which associates with the histone acetyltransferase CBP, is required for maintenance of transcriptionally active states and antagonizes Polycomb silencing, although the mechanism underlying this antagonism is unknown. Here we show that H3K27 is specifically acetylated by Drosophila CBP and its deacetylation involves RPD3. H3K27ac is present at high levels in early embryos and declines after 4 hours as H3K27me3 increases. Knockdown of E(Z)decreases H3K27me3 and increases H3K27ac in bulk histones and at the promoter of the repressed Polycomb target gene abd-A, suggesting that these indeed constitute alternative modifications at some H3K27 sites. Moderate overexpression of CBP in vivo causes a global increase in H3K27ac and a decrease in H3K27me3, and strongly enhances Polycomb mutant phenotypes. We also show that TRX is required for H3K27 acetylation. TRX overexpression also causes an increase in H3K27ac and a concomitant decrease in H3K27me3 and leads to defects in Polycomb silencing. Chromatin immunoprecipitation coupled with DNA microarray (ChIP-chip) analysis reveals that H3K27ac and H3K27me3 are mutually exclusive and that H3K27ac and H3K4me3 signals coincide at most sites. We propose that TRX-dependent acetylation of H3K27 by CBP prevents H3K27me3 at Polycomb target genes and constitutes a key part of the molecular mechanism by which TRX antagonizes or prevents Polycomb silencing.

The cotranslational incorporation of the unusual amino acid selenocysteine (Sec) into both prokaryotic and eukaryotic proteins requires the recoding of a UGA stop codon as one specific for Sec. The recognition of UGA as Sec in mammalian selenoproteins requires a Sec insertion sequence (SECIS) element in the 3 untranslated region as well as the SECIS binding protein SBP2. Here we report a detailed analysis of SBP2 structure and function using truncation and site-directed mutagenesis. We have localized the RNA binding domain to a conserved region shared with several ribosomal proteins and eukaryotic translation termination release factor 1. We also identified a separate and novel functional domain N-terminal to the RNA binding domain which was required for Sec insertion but not for SECIS binding. Conversely, we showed that the RNA binding domain was necessary but not sufficient for Sec insertion and that the conserved glycine residue within this domain was required for SECIS binding. Using glycerol gradient sedimentation, we found that SBP2 was stably associated with the ribosomal fraction of cell lysates and that this interaction was not dependent on its SECIS binding activity. This interaction also occurred with purified components in vitro, and we present data which suggest that the SBP2-ribosome interaction occurs via 28S rRNA. SBP2 may, therefore, have a distinct function in selecting the ribosomes to be used for Sec insertion.

SUMMARY Rapid mitotic divisions and a fixed transcription rate limit the maximal length of transcripts in early Drosophila embryos. Previous studies suggested that transcription of long genes is initiated but aborted, as early nuclear divisions have short interphases. Here, we identify long genes that are expressed during short nuclear cycles as truncated transcripts. The RNA binding protein Sex-lethal physically associates with transcripts for these genes and is required to support early termination to specify shorter transcript isoforms in early embryos of both sexes. In addition, one truncated transcript for the gene short-gastrulation encodes a product in embryos that functionally relates to a previously characterized dominant-negative form, which maintains TGF-β signaling in the off-state. In summary, our results reveal a developmental program of short transcripts functioning to help temporally regulate Drosophila embryonic development, keeping cell signaling at early stages to a minimum in order to support its proper initiation at cellularization.

Caudal visceral mesoderm (CVM) cells migrate from posterior to anterior of the Drosophila embryo as two bilateral streams of cells to support the specification of longitudinal muscles along the midgut. To accomplish this long-distance migration, CVM cells receive input from their environment, but little is known about how this collective cell migration is regulated. In a screen we found that wunen mutants exhibit CVM cell migration defects. Wunens are lipid phosphate phosphatases known to regulate the directional migration of primordial germ cells (PGCs). PGC and CVM cell types interact while PGCs are en route to the somatic gonadal mesoderm, and previous studies have shown that CVM impacts PGC migration. In turn, we found here that CVM cells exhibit an affinity for PGCs, localizing to the position of PGCs whether mislocalized or trapped in the endoderm. In the absence of PGCs, CVM cells exhibit subtle changes, including more cohesive movement of the migrating collective, and an increased number of longitudinal muscles is found at anterior sections of the larval midgut. These data demonstrate that PGC and CVM cell migrations are interdependent and suggest that distinct migrating cell types can coordinately influence each other to promote effective cell migration during development.

Drosophila Polycomb Repressive Complex 2 (PRC2) is a lysine methyltransferase that trimethylates histone H3 lysine 27 (H3K27me3), a modification essential for Polycomb silencing. Mutations in its catalytic subunit, E(Z), that abolish its methyltransferase activity disrupt Polycomb silencing, causing derepression of Polycomb target genes in cells where they are normally silenced. In contrast, the unusual E(z) mutant allele Trithorax mimic (E(z)(Trm)) causes dominant homeotic phenotypes similar to those caused by mutations in trithorax (trx), an antagonist of Polycomb silencing. This suggests that E(z)(Trm) causes inappropriate silencing of Polycomb target genes in cells where they are normally active. Here we show that E(z)(Trm) mutants have an elevated level of H3K27me3 and reduced levels of H3K27me1 and H3K27me2, modifications also carried out by E(Z). This suggests that the E(z)(Trm) mutation increases the H3K27 trimethylation efficiency of E(Z). Acetylated H3K27 (H3K27ac), a mark of transcriptionally active genes that directly antagonizes H3K27 methylation by E(Z), is also reduced in E(z)(Trm) mutants, consistent with their elevated H3K27me3 level causing inappropriate silencing. In 0-4h E(z)(Trm) embryos, H3K27me3 accumulates prematurely and to high levels and does so at the expense of H3K27ac, which is normally present at high levels in early embryos. Despite their high level of H3K27me3, expression of Abd-B initiates normally in homozygous E(z)(Trm) embryos, but is substantially lower than in wild type embryos by completion of germ band retraction. These results suggest that increased H3K27 trimethylation activity of E(Z)(Trm) causes the premature accumulation of H3K27me3 in early embryogenesis, "predestining" initially active Polycomb target genes to silencing once Polycomb silencing is initiated.