**Polycomb, or How Flies Have Memory**

Polycomb is a protein first identified in 1947 by Pamela Lewis in *Drosophila melanogaster*, the small fruit fly beloved by geneticists. It has given its name to a group of proteins essential to cellular identity memory, highly conserved throughout evolution. For a long time, these proteins were considered loyal but simple guardians of specific chromatin states. However, recent studies reveal that their role is more complex than previously thought.

During early development, an organism's cells differentiate, each one assigned a specific destiny. Once established, this identity must be preserved throughout life. In *Drosophila*, a segmented insect, each cell’s identity within a segment is determined by the expression profile of specific genes along the animal's anterior-posterior axis. These genes, known as homeotic genes, are grouped into two clusters on a single chromosome, called the *Antennapedia* and *Bithorax* complexes. The establishment of the expression profile of these homeotic genes, which grants each cell its identity, results from a cascade of maternal and zygotic transcription factors that are only temporarily present in the embryo. Most transcription factors disappear midway through embryogenesis, but the transcriptional state of homeotic genes is maintained throughout life by Polycomb group proteins (PcG) and trithorax group proteins (trxG). Their roles are antagonistic: PcG proteins sustain the repression of homeotic genes that were inactivated during early embryogenesis, while trxG proteins maintain the active state of genes transcribed at the onset of development.

**PRC1 and PRC2/3**

PcG proteins function in multimeric complexes known as PRC1 and PRC2/3, short for Polycomb Repressive Complex. These complexes bind to DNA to repress gene activity. PRC1 includes the Polycomb (PC) and Polyhomeotic (PH) proteins, among others. PRC2/3 complexes sustain gene silencing initiated during early embryogenesis by methylating specific amino acids on histones H3 and H1. PRC1 is recruited by methylated histone H3 and can repress transcription by preventing chromatin remodeling by the Swi/Snf complex or by forming direct contacts with the transcription machinery. This way, PRC complexes condense chromatin, rendering it inaccessible to transcription factors.

**A More Dynamic Role Than Expected**

“We studied the binding of PC and PH proteins on part of the X chromosome and a portion of chromosome 2 in *Drosophila* at different developmental stages (Nègre et al., 2006). The method used is called ChIP on chip. Chromatin is immunoprecipitated with anti-PC and anti-PH antibodies. The DNA sequences bound to PC and PH are retrieved, labeled with a fluorochrome, and used to hybridize DNA microarrays containing much of the *Drosophila* genome. The result is unexpected and surprising! Until now, we believed that once PC and PH proteins bind to target sequences, they remain there, stably maintaining repression at these loci. This is not what we observed. PC and PH are bound to certain sites at the beginning of embryonic development but are no longer detected there at later stages,” explains Giacomo Cavalli, Research Director at CNRS, UPR 1142, “Chromatin and Cell Biology” at the Human Genetics Institute in Montpellier. Conversely, on other target sequences, PC and PH are not detectable in the embryo but are bound to them once the fly reaches adulthood. These results suggest that the role of PcG proteins is not as univocal as anticipated, at some loci. In addition to their function of cellular identity memory, they appear to regulate certain genes. “This interpretation corroborates previous findings. We had shown that PcG proteins regulate the *Cyclin A* gene, involved in cell cycle control (Martinez et al., 2006),” adds Cavalli. Another study from the same institute revealed that the *hedgehog* gene is also regulated by PcG proteins (Chanas et al., 2005). *Cyclin A* and *hedgehog* are direct PcG targets. However, in certain tissues and specific developmental stages, they are not associated with PcG proteins and are regulated independently of them.

**A Difference Between the Sexes**

The embryo and adult male show the most distinct distribution of PC and PH proteins. A previous study indicated that PcG proteins are involved in the specific regulation of male germline differentiation in *Drosophila* (Chen et al., 2005). These new data suggest that PcG protein-mediated gene regulation also varies between males and females in somatic cells.

**Regulation of Transcription Factors**

In this part of the X chromosome and chromosome 2, the binding sites of PC and PH proteins encompass 37 potential target genes. Among these, several function categories are represented. Most strikingly, 38% of these genes are transcription factors, even though they represent only 6.4% of the genes in the studied sequences. These factors regulate a broad range of developmental processes, though most are involved in embryonic patterning and neurogenesis. “We conclude that PcG proteins could act as pleiotropic transcriptional regulators, coordinating various developmental processes in addition to serving as stable repressors of other genes. Now that we know this, such analyses should be expanded to the entire genome and further explored in different cell types and developmental stages,” says Cavalli.

**Polycomb Foci**

In all cells across all organisms, DNA molecules are compacted into three dimensions to fit within a nucleus that measures 5 to 10 µm on average. Chromosomal architecture is not monotonous; it is utilized to regulate gene function. “Polycomb protein regulates its target genes by directly binding to chromatin. When attempting to locate it in the nucleus, we observe areas of high concentration while it is scarcely present in the rest of the nucleus. There are only thirty to fifty Polycomb foci per nucleus, even though ChIP on chip studies show that this protein binds to around two hundred different genomic loci. We think that PcG target genes cluster in the nucleus to better organize their regulation, segregating into compartments that are more or less favorable for repression,” adds Cavalli.

**Importance of Nuclear Gene Compartmentalization**

In *Drosophila*, PRC1 and PRC2/3 complexes repress genes by directly binding to DNA sequences known as PREs (PcG Response Elements). Conversely, trxG proteins act antagonistically by binding to different DNA sequences called TREs (TrxG Response Elements).

*Fab-7* is a regulatory element of the *Abdominal-B* homeotic gene that contains a PRE. In flies homozygous for the *Fab-7* element, repression of the adjacent gene is enhanced when homologous chromosomes pair up. This phenomenon was demonstrated in transgenic *Drosophila* carrying a gene responsible for red eye pigmentation under the control of *Fab-7*. In homozygous flies, the gene is strongly repressed, and the eyes are depigmented. This pairing-sensitive silencing suggests that proximity between two homologous PREs enhances their individual repression function. In other transgenic flies with several *Fab-7* copies inserted at loci on different chromosomes (the trans-heterozygous configuration), *Fab-7* can again intensify neighboring gene repression.

**Involvement of RNAi Machinery Elements**

Recent studies show that components of the RNAi machinery are involved in PcG-dependent repression at *Fab-7* loci (Grimaud et al., 2006). These components are also required for the production of small interfering RNAs (siRNAs), as well as for long-distance interaction between *Fab-7* elements. “We observed colocalization of Dicer-2, PIWI, and Argonaute1, three proteins involved in RNAi, with Polycomb foci. A mutation in one of these complexes significantly reduces the frequency of *Fab-7* loci pairing without preventing PcG proteins from binding to these target elements,” explains Cavalli. These results indicate that RNAi machinery may play a role in regulating nuclear organization of PREs.

**In Mammals**

PcG proteins are essential for early embryogenesis in all metazoans, but their exact role in mammals remains unclear. To date, no PRE has been identified in these organisms. However, PcG target genes have recently been identified in human and murine stem cells using a ChIP on chip approach (Boyer et al., 2006; Lee et al., 2006). Most of these genes are involved in embryonic development. PcG proteins thus appear to be vital for maintaining these stem cells' undifferentiated state.