Island-hopping protein found to recruit polycomb complex in embryonic stem cells
A recent paper from Dr Rob Klose’s group and colleagues sheds new light on how mammals use chromatin based processes to lay down the very earliest transcription patterns for development.
Dr Klose’s group teamed up with a number of Oxford researchers including Dr Chris Ponting and Dr Benedikt Kessler to carry out the work which is published in one of the first issues of the new journal eLIFE (1).
ChIP-seq profiles showing the distribution of DNA-binding proteins KDM2A and KDM2B at the Lhx4 gene in mouse ES cells. KDM2B is specifically enriched and KDM2A depleted at this polycomb repressed CGI (Click to Enlarge)
Their research focuses on the Polycomb Repressive Complex which plays a key role in regulating gene expression at the earliest stages of development. The complex alters the chromatin structure, contributing to gene silencing in the region and preventing cells from inappropriate expression of differentiation-specific genes. Polycomb-group proteins are found throughout nature and are crucial for insect, plant and mammalian development.
In mammals, non-methylated regions of DNA in the genome known as CpG islands (CGIs) are areas of intense gene regulatory activity. Most CGIs are marked with a permissive chromatin architecture which allows gene expression. But a minority exists in an alternative more repressed chromatin state – and it is Polycomb Repressive Complexes 1 and 2 (PRC1 and PRC2) which are responsible for setting up this state.
Although the non-methylated state of CGIs was always assumed to have a passive role in gene expression, CGIs are now considered to play a proactive role by recruiting proteins which create unique chromatin environments. Several recent findings in particular suggest that the PRC may be guided to CGIs by their specific chromatin features.
Whilst researchers know a lot about the genes regulated by polycomb and how the regulation is achieved, an outstanding question is how polycomb selects regions of the genome to act upon. Dr Klose’s long-standing interest in CGIs unexpectedly provided him with an opportunity to address this question.
He had been studying proteins bound to CGIs in mouse embryonic stem cells and had identified a Zn-finger domain containing protein called KDM2B. It was when Dr Klose used a genome-wide approach to study KDM2B’s binding pattern that a connection between KDM2B and polycomb – specifically PRC1 - was revealed.
‘Studying polycomb and CGIs is difficult because polycomb functions at so many sites that we have to use genomic approaches to understand,’ Dr Klose explains. ‘If you look at a few genes, then you just get a small snapshot – you might be seeing part of the story - but it’s not until you look across the whole genome that you can make generalisations.’
With assistance from Dr Chris Ponting’s computational biology group (Department of Physiology, Anatomy and Genetics), Dr Klose and colleagues determined that KDM2B associates with CGIs throughout the genome but that it is enriched at polycomb occupied CGIs.
Through collaboration with Dr Benedikt Kessler (Nuffield Department of Clinical Medicine, Centre for Cellular and Molecular Physiology), the researchers were able to couple the genome-wide profiling with a proteomics approach and gain an understanding of the protein components at these sites.
A schematic representation of the specific KDM2B/PRC1 complex identified by Dr Klose and colleagues in mouse ES cells (Click to Enlarge)
‘This is really where we made the advance,’ comments Dr Klose. ‘Through purifying the DNA-binding protein KDM2B, we found that it was physically associated with polycomb. That gave us the first link.’
‘Using genomics and genetic approaches to manipulate the system, we were able to look genome-wide at how these proteins co-localise and what happens when you take away different components.’
In this way, the group could build up an increasingly detailed picture of the proteins associated with polycomb repressed sites. Their findings demonstrate the first real link between recognition of CGI DNA and polycomb factors in mammalian cells, and provide evidence that KDM2B, through recognition of CGIs, can recruit PRC1.
Dr Klose believes that there is a division of labour in setting up polycomb repression at CGIs. ‘The KDM2B-polycomb complex may be important for the initiation phase,’ he explains, ‘but once initiation has taken place, there is a ‘photocopying phase’ that goes on – it already recognises the fact that the site has been selected, and this can be copied.’
The observation that there is only a subset of CGIs on which the complex accumulates even though it is found at all CGIs at a low level, suggests that KDM2B may sample CGIs. ‘The complex probably comes on and off, sampling whether the region is receptive to polycomb repression. Where there is transcription, it may be pushed off – unable to be retained.’
As Dr Klose says, their data provide evidence for how PRC 1 is targeted to regions of the genome. But to prove that it is indeed recruited in this way, they must probe the mechanism by disrupting it in vivo and investigating the consequences. Polycomb plays such a major role in regulating gene expression through development that understanding this step is at the heart of understanding polycomb function.
The paper’s cross-Oxford authorship demonstrates the benefits to tapping into complementary expertise at the University. This includes the contribution of researchers in Professor Neil Brockdorff’s group in the same department, whose own work focuses on polycomb. Much of the experimental work was carried out by Dr Klose’s graduate student Anca Farcas working alongside the other contributors.
Having successfully secured funding for the next five years in the form of a Wellcome Trust Senior Fellowship in the Basic Sciences, Dr Klose is in a strong position to probe CGI function even deeper. The overall aim of this new programme of work will be to understand better how the systems for targeting chromatin modifications to CGIs contribute to gene regulatory processes.
‘There has been an explosion in interest in this area,’ comments Dr Klose, who adds that with the development of new techniques the area has become more tractable. ‘It’s been known for a very long time that CGIs can and often do become hypermethylated in cancer. This probably contributes to the transformation process but nobody really understands why that is. To know how things go wrong, you first have to know how things are normally set up.’
With the publishing of this paper, Dr Klose and colleagues become amongst the first researchers in the University to use eLIFE. Dr Klose says he found the journal’s review process constructive and efficient and is very positive about its future. ‘It’s destined to be a success no matter what, because the right people are backing it and the right people are submitting papers to it,’ he adds.