BIOLOGICAL OVERVIEW

Pleiohomeotic (Pho) is the first Polycomb group (PcG) member to be identified as a DNA binding protein (Brown, 1998). Genes of the Drosophila Polycomb group encode proteins necessary for the maintenance of transcriptional repression of homeotic genes. PcG proteins are thought to act by binding as multiprotein complexes to DNA through Polycomb group response elements (PREs). Although recruitment of PcG proteins to PREs is central to PcG action, it has not been clear whether PcG proteins recognize PREs through interactions with sequence-specific DNA binding proteins or by recognizing a particular chromatin structure. Pleiohomoetic has been identfied as a PRE binding protein in a search for proteins that interact with a PRE from the segment polarity gene engrailed. Sequencing of pho reveals that it codes for a homolog of vertebrate Ying-Yang (YY1), a ubiquitously expressed zinc finger DNA binding protein that is able to act as either a transcriptional repressor or activator in different regulatory contexts (Brown, 1998).

PREs have been shown to have an unusual silencing activity. When a PRE is included in a P-transformation vector with the white gene, it silences white expression in transgenic Drosophila. This silencing is much stronger in flies homozygous for the PRE-white vector; thus, this type of repression has been called pairing-sensitive silencing: it requires two copies of the PRE-white vector. This silencing activity was first reported for a 2.4 kb fragment of DNA from the Drosophila engrailed gene (Kassis, 1991). This fragment of DNA was further dissected and shown to contain two strong and one weak pairing-sensitive silencing fragments (Kassis, 1994). One of the fragments (a 176 bp fragment located at -576 to -400 bp of the engrailed locus) that acts as a strong pairing-sensitive site, has been the focus of an attempt to discover PRE binding proteins (Brown, 1998).

A homologous DNA fragment from the distantly related Drosophila species Drosophila virilis also acts as a pairing-sensitive silencer in D. melanogaster (Kassis, 1994). Sequence conservation between these two fragments reveals five blocks of near sequence identity. The first block is a 17 bp sequence located -547 to -531 bp upstream of the transcription start site of the Drosophila engrailed gene. When this 17 bp is deleted, the 176 bp fragment no longer acts as a pairing-sensitive silencer in transgenic flies. This result shows that the 17 bp sequence is essential for pairing-sensitive silencing. This 17 bp sequence was therefore used to search for DNA binding proteins that might be required for silencing. The 17 bp sequence was radioactively labeled and used as a gel shift probe with nuclear extracts from 0-22 hr Drosophila embryos. Three shifted complexes were detected, all of which competed specifically with the unlabeled probe but not with nonspecific double-stranded probes or the single-stranded 17 base oligonucleotides. A multimer (nine copies) of the 17 bp sequence was prepared and used to screen a cDNA expression library made from 20-24 hr Drosophila embryos. Of three clones coding for a protein that showed specific binding to a labeled fragment, one contained a full-length cDNA for a protein specifically binding the 17 bp oligonucleotide. Sequencing of the cDNA revealed a protein homologous to YY1 (Brown, 1998).

Previous analyses have uncovered GAGA sites as the only obvious sequence similarity between diverse PREs. The sequences of three different PREs were therefore examined for the presence of the YY1 core consensus site CCATNTT or the other reported YY1 consensus binding site 5'(C/g/a)(G/t)(C/t/a)CATN(T/a)(T/g/c)-3'. YY1 consensus binding sites were found within a PRE from Antennapedia (Zink, 1991), a PRE from polyhomeotic (Fauvarque, 1993), and a PRE from Ubx (Chang, 1995 ), suggesting that Pho binds to all of these PREs. Thus, Pho may be important for the function of many different PREs. It is unlikely that Pho acts alone in recruiting PcG proteins to the DNA for three reasons: (1) another fragment of engrailed DNA (from -2407 to -1944) contains potential Pho binding sites but does not exhibit pairing-sensitive repression of white (Kassis, 1994 ); (2) deletion of another block of conserved DNA (from -481 to -470) of the 176 bp silencer also leads to a loss of pairing-sensitive silencing, and (3) a multimerized version of the 17 bp oligonucleotide was not sufficient for pairing-sensitive silencing of mini-white. Thus, it is likely that more than one DNA binding protein is required for pairing-sensitive silencing. It has been suggested that Pho binds to PREs and, in conjunction with other unidentified DNA binding proteins, recruits PcG proteins to the DNA. At different promoters and in different developmental contexts, Pho may interact with different protein complexes. At the engrailed locus, different subsets of PcG complexes may act together to silence engrailed expression (Brown, 1998 and references).

One of the other proteins involved in recruiting PcG proteins to the DNA may be GAGA factor, a protein involved in nucleosome remodeling. GAGA factor has been found to colocalize with PC protein on regulatory elements of the bithorax complex. GAGA binding sites are found in some PREs, and mutations in GAGA factor have been found to influence pairing-sensitive silencing mediated by the Fab-7 PRE. Three potential GAGA binding sites are present within the 176 bp silencer. Within the engrailed pairing-sensitive silencer, located from -1944 to -1503, there are two YY1-consensus binding sites and three GAGA binding sites. Interestingly, the engrailed fragment at -2407 to -1944 contains a YY1-consensus site but no GAGA sites and does not act as a pairing-sensitive silencer (Brown, 1998 and references).

What is the relationship between YY1 and the Polycomb group genes? Mammalian YY1 is a ubiquitous multifunctional protein involved in the expression of many different genes (reviewed in Shi, 1997). It can act as an activator, repressor, or initiator of transcription depending on cellular and promoter context. Transcriptional activation by YY1 involves a bipartite transactivation domain (comprising two acidic regions at the N terminus), a spacer region (with modulatory activity), and the DNA binding domain. Transcriptional repression is largely mediated by the zinc finger domain that is required both for protein-protein and protein-DNA interactions. No known transactivation domains were detected in the Pho sequence, and only the zinc finger region and a small region of the spacer of mammalian YY1 were found to be conserved. Thus, the function of Pho may be limited to repression. This would explain why no Drosophila YY1 activity was detected in a previous study (Seto, 1991), since the assays employed required YY1-driven transcriptional initiation of reporter constructs (Brown, 1998 and references).

How does mammalian YY1 repress transcription, and how is this related to the mechanism of repression by the PcG genes? YY1 has been shown to repress transcription by multiple mechanisms, including competition with an activator for overlapping binding sites, direct protein-protein interactions with transcriptional activators, and interference with activator/transcriptional machinery interactions. Mammalian YY1 has also been shown to interact with RPD3 (see Drosophila Rpd3), a histone deacetylase, through a domain not present in Pho. In HIV-1, YY1 seems to interact with another transcription factor, LSF, to repress transcription through an unknown mechanism. Thus, the mechanism of transcriptional repression by YY1 seems to be promoter-specific. It is suggested that Pho may play multiple roles in the regulation of gene expression in Drosophila: that is, Pho may be considered a general transcription factor that plays a role in stable silencing by binding to the DNA and recruiting chromatin-associated silencing proteins. The severe pleiotropic effects of removing pho function from Drosophila eggs may reflect this dual role (Brown, 1998).

PROTEIN STRUCTURE

Amino Acids - 520

Structural Domains

Comparison of the Pleiohomeotic sequence with GenBank shows that it encodes a Drosophila homolog of Yin Yang-1 (YY1). YY1 is a DNA binding zinc finger transcription factor known to be conserved from Xenopus laevis to mammals. Pho shows a remarkable 112/118 amino acid identity with mammalian YY1 over the region encoding the four zinc fingers and is 100% identical over zinc fingers 2 and 3. Pho contains all of the amino acids identified by X-ray crystallography as being involved in contacting the DNA (Houbaviy, 1996 ). In addition to the zinc finger domain, mammalian YY1 contains three other types of domains: two acidic domains, a glycine-alanine-rich domain, and a region called the spacer (Austen, 1997). A small region of the spacer is present in the Drosophila protein, but other regions of similarity between Pho and mammalian YY1 were found (Brown, 1998).

date revised: 30 September 98

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.