In the earlier post titled ' Role of polycomb and trithorax in developmental regulation' we have seen that the main function of the polycomb group proteins (PcGs) is to repress target gene expression, their major targets being the homeobox (Hox) genes. These proteins bind to Polycomb Response Elements (PREs), their target sequences on the DNA. The recruitment of the PcGs to these sites occurs due to covalent histone modifications in the DNA. Here we look into this aspect in some detail.
Till date three distinct PcG complexes have been discovered in Drosophila Polycomb repressor complex 1 and 2 (PRC1 and 2) and Pho repressive complex (Pho RC). All three contain multiple subunits encoded by PcG genes that are crucial for Hox gene silencing. PRC2 is known to have histone methyl transferase activity. It has H3K27 methylase activity. The PRC1 subunit inhibits nucleosome remodelling and transcription and brings about chromatin compaction. The PRC1 and 2 have various catalytic and non catalytic subunits which play a role in silencing target sites. The Pho repressive complex contains Pho and dSfmbt. Pho subunit has a distinct sequence specific DNA binding activity which is essential for targeting. dSfmbt binds to H3 or H4 tail peptides which are mono or di methylated at H3K9 or H4K20 which is crucial for repression.
The above figure depicts a speculative model for long range interactions between the PRE tethered PcG complexes and the methylated nucleosomes in the flanking chromatin. PRC1, PRC2 and PhoRC are locally bound to the PREs whereas trimethylation at H3K27, H3K9 and H4K20 is observed over an extended domain emcompassing the promoter and the coding regions.
Left : In case of PRC1 the Pc chromadomain (red triangle) would dock onto the nucleosomes that are tri methylated at H3K27 and through such bridging interactions will bring them in proximity to the other PRE bound PcG proteins. This may permit the other PRC1 subunits to block remodelling of the target nucleosomes or facilitate methylation of the neighbouring hypomethylated (pink) nucleosomes by PRC2.
Right : In case of PhoRC, the MBT (Malignant Brain Repeats) of dSfmbt (pink triangle) would interact with mono or dimethylated nucleosomes at H3K9 or H4K20.This bridging interaction would permit PRC2 to efficiently tri methylate H3K27 in hypomethylated nucleosomes.
Its is also possible that as yet unindentified HMTases which tri methylate these histones are also localized at the PREs.Similarly proteins with specificity for binding to other mrthyl-lysine modifications in histones might also be localized to the PREs.Various proteins have been purified which are associated to PcG complexes that are strictly essential for silencing. Pcl (Polycomb like) has been shown to be associated with PRC2 and mutations in this protein affect PRC2 functioning. Another protein zeste is a component of PRC1 and it has DNA binding activity.
As described above biochemical purification of PRC1, PRC2 and PhoRC suggests that these three complexes are separate entities. However since they co localize at the PREs various studies are aimed at finding out the physical interactions between subunits from the different complexes which may explain how PRC1 and PRC2 are localized to the PREs. Different models have been put forward based on various studies. In particular Wang et al (2004), reported that E(z) and Esc (subunits of PRC2) directly interact with Pho, leading to the proposal that Pho directly tethers PRC2 to the PREs. the PRE tethered PRC2 then locally trimethylates H3K27 thereby creating binding sites for the chromodomain of the PRC1 subinut Pc. More recent studies however challenge this model. First quantitative ChIp studies by two different labs suggest that PREs at the Ubx gene are infact devoid of nucleosomes.(V Pirrotta). Moreover PREs constitue of hypersensitive sites providing additional evidence that PRE DNA is not packaged into nucleosomes. Second studies by Mohd Sarip et al (2002), suggested that Pho can directly interact with PRC1 subunits, and that in in vitro studies Pho and PRC1 can co assemble on a naked PRE DNA template in absence of nucleosomes. This assembly assumes a conformation that is difficult to reconcile with the conformation of a nucleosome core particle. Thus available evidence suggests that not only PhoRC, but also PRC1 and PRC2 localize onto the PRE through interactions with Pho and other DNA binding proteins and not through covalent histone modifications or interactions with nucleosomes.
This conclusion then raises a major question with regards to the role of histone modification (trimethylation) in PcG silencing. Quantitative ChIp analysis that compared the Ubx gene in its off and on states in Drosophila larvae suggest that trimethylation in the promoter and coding region is essential for PcG silencing. In the 'off' state extensive trimethylation is present throughout the upstream control, promoter and coding regions but in the 'on' state this methylation is restricted only to the upstream control region and not seen in the promoter and coding regions.This led to the proposal that trimethylation at H3K9, H3K27 and H4K20 in the promoter and coding regions is required to demarcate the chromatin interval that is targeted for repression by the PRE-tethered PcG protein complexes. In this context, an important aspect about the relationship between the PREs and histone modifications of the target genes comes from studies with PRE reporter genes in Drososphila, in which the PRE DNA was flanked with FRT ( Flip Recombinase Target) sites and could thus be deleted from the reporter gene. This study indicated that excision of the PRE DNA from the silenced reporter gene, caused loss of PcG silencing, even when the excision was done late in development. Thus although the PcG proteins appear to repress transcription by methylation in promoter and coding regions, silencing depends on the continuous presence of the PREs and the PcGs tethered to them.
Although much progress has been made towards understanding the mechanism of PcG silencing, much remains to be learned. It is still not understood how the PcG proteins are targeted to the PREs beacuse Pho binding sites alone do not make a PRE. ChIp studies may also help elucidate additional DNA binding PcG proteins.Also there may be many additional PcG complexes and PRC1, PRC2 and PhoRC may represent only the most stable or most abundant PcG complexes. Also it seems likely that these findings of PcG silencing in Drosophila will prove important in inderstanding the mechanism of PcG silencing in mammals, given the recent discovery that mammalian PcG proteins seem to be acting as global repressors of developmental control genes in embryonic stem cells.
Reference:MULLER, J., & KASSIS, J. (2006). Polycomb response elements and targeting of Polycomb group proteins in Drosophila Current Opinion in Genetics & Development, 16 (5), 476-484 DOI: 10.1016/j.gde.2006.08.005