Polycomb-mediated Gene Silencing: research reveals new clues
A re-examination of an experimental result obtained several years ago has led to a potentially new way of thinking about the mechanism of polycomb-mediated gene silencing.
Polycomb proteins play a key role in developmental gene regulation in most multicellular organisms, from plants to mammals. They were first defined in Drosophila as factors which allow cells to ‘remember’ patterns of gene expression through successive cell divisions, providing a memory of the cell fate or identity established through differentiation and development.
Drosophila and mouse Polycomb mutants. Each set of images shows the wild type (wt) alongside Polycomb mutants which show distorted body structure
In mammals, polycomb proteins have important roles throughout an organism’s life. In embryonic stem (ES) cells, they play a crucial role in restraining the cells from going down differentiation pathways, targeting around 1500 genes. They are vital in the establishment and maintenance of different cell types. Polycomb proteins are also found on the inactive X chromosome.
Polycomb proteins function in multiprotein complexes which are recruited to and silence selected genes. They accomplish silencing by chemically modifying the histone proteins around which the DNA is entwined.
Now Professor Neil Brockdorff in the department, in collaboration with Dr Sarah Elderkin at the Babraham Institute, have found evidence that points to a new way of looking at the problem of how polycomb complexes find their way onto the right genes. Their findings are published in Cell (1).
Polycomb complexes exist in two forms – PRC1 and PRC2. PRC2 loads onto the chromatin first, modifying a lysine on one of the resident histone proteins by adding methylation groups – H3K27me3. This methylated lysine is recognised by a protein called CBX in the PRC1 complex.
PRC1 has a different chromatin modifying activity known as ubiquitylation. Bringing PRC1 onto the site leads to ubiquitylation of another histone protein, H2A. It is this modification that inhibits transcription on the genes located at that site, possibly by blocking the enzyme which transcribes the DNA into RNA.
‘Perhaps there is some underlying mechanism that is common to all polycomb recruitment.’
Key to selective gene silencing by polycomb is the recruitment of complexes onto the chromatin. But the primary signal for recruitment remains unknown - transcription factors appear to be important, but there may also be a role for non-coding RNAs.
Professor Brockdorff and Dr Elderkin were intrigued by a puzzling finding some years ago related to the PRC1 complex, and decided to investigate further.
‘We had just made the connection between H2A ubiquitylation and the PRC1 complex,’ explains Professor Brockdorff. ‘We took mouse ES cells which lacked PRC2 which should also be losing H2A ubiquitylation because all the H2A ubiquitylation in the nucleus is from PRC1. But we didn’t find this was the case.’
They found instead that H2A ubiquitylation looked almost unchanged in the PRC2-deficient ES cells, although they detected no H3K27 methylation.
More detailed scrutinisation of the polycomb complexes recruited across the genome showed that the levels of PRC1 dropped whilst H2A ubiquitylation was retained. It also revealed that the same set of genes was being targeted in the absence of H3K27 methylation as in the presence.
RYBP-PRC1 complex (green) and associated H2A ubiquitylation (H2AK119u1, red) localise to bright foci in the nucleus (stained blue in merged image) of cells with an inactive X chromosome, both in the presence (top panels) and absence (bottom panels) of PRC2 and associated H3K27 methylation (Click to enlarge)
The results suggested that there was a novel pathway to H2A ubiquitylation that did not require the H3K27 methylation pathway. The challenge was to understand how this was happening.
Clues started to emerge when the group used mass spectrometry to tease apart the different components of the PRC1 complexes purified from ES cells, explains Professor Brockdorff. ‘We tagged a protein called MEL-18, one of the core components of PRC1, and pulled down associated proteins, and then carried out mass spectrometry to identify these proteins.’
The experiments provided evidence for the existence of a PRC1 complex different from any previously described. MEL-18 had pulled down a protein called RYBP which is known to interact with RING1B, one of the catalytic subunits of the complex.
‘A paper from 2010 had shown that RYBP interacts with RING1B on exactly the same surface as does CBX protein, so the interactions have to be mutually exclusive,’ explains Professor Brockdorff. ‘So this couldn’t be part of the conventional complex with a CBX protein in it.’
A further series of experiments confirmed that there are two distinct flavours of PRC1. These complexes contain core PRC1 proteins, and either RYBP or a CBX protein. When reconstituted in vitro, they have similar ubiquitylation activity.
RYBP could also be detected specifically on the target genes themselves – in the absence or presence of H3K27 methylation – providing evidence for how H2A ubiquitylation is maintained in PRC2-deficient ES cells. This was also the case on the inactive X chromosome.
Schematic showing the different polycomb complexes and their constituent proteins, and the ways in which the complexes may be recruited onto chromatin (Click to Enlarge)
Given the widespread occurrence of polycomb silencing, the researchers decided to look at the distribution of the two forms of the complex across different cell types.
‘We saw the same phenomenon in neuronal stem cells and fibroblasts,’ says Professor Brockdorff. ‘So it seems that the co-existence of these two flavours of PRC1 complex is broadly conserved, at least in mammalian systems.’
The findings throw up some interesting ideas about the recruitment and function of polycomb repressors, says Professor Brockdorff. ‘Our paper describes a second complex that is somehow getting to the right places without the CBX protein. This suggests that for the two complexes there is some shared aspect of the signal that recruits polycomb to target sites.’
‘This has led us to think that perhaps the current ideas about how polycomb is sent to its different sites are not the whole story – that people are missing the main point. Perhaps there is some underlying mechanism that is common to all polycomb recruitment.’
Professor Brockdorff and his group have begun the search for such a unifying mechanism.