Mdm2 as a chromatin modifier

Abstract

Mdm2 is the key negative regulator of the tumour suppressor p53, making it an attractive target for anti-cancer drug design. We recently identified a new role of Mdm2 in gene repression through its direct interaction with several proteins of the polycomb group (PcG) family. PcG proteins form polycomb repressive complexes PRC1 and PRC2. PRC2 (via EZH2) mediates histone 3 lysine 27 (H3K27) trimethylation, and PRC1 (via RING1B) mediates histone 2A lysine 119 (H2AK119) monoubiquitination. Both PRCs mostly support a compact and transcriptionally silent chromatin structure. We found that Mdm2 regulates a gene expression profile similar to that of PRC2 independent of p53. Moreover, Mdm2 promotes the stemness of murine induced pluripotent stem cells and human mesenchymal stem cells, and supports the survival of tumour cells. Mdm2 is recruited to target gene promoters by the PRC2 member and histone methyltransferase EZH2, and enhances PRC-dependent repressive chromatin modifications, specifically H3K27me3 and H2AK119ub1. Mdm2 also cooperates in gene repression with the PRC1 protein RING1B, a H2AK119 ubiquitin ligase. Here we discuss the possible implications of these p53-independent functions of Mdm2 in chromatin dynamics and in the stem cell phenotype. We propose that the p53-independent functions of Mdm2 should be taken into account for cancer drug design. So far, the majority of clinically tested Mdm2 inhibitors target its binding to p53 but do not affect the new functions of Mdm2 described here. However, when targeting the E3 ligase activity of Mdm2, a broader spectrum of its oncogenic activities might become druggable.

Keywords: EZH2; Mdm2; histone methylation; histone ubiquitination; polycomb repressor complex.

Figure 1

Hypothetical, non-mutually exclusive models of the Mdm2−PcG interactive network. Mdm2 is interacting with several members of the PcG protein family and enhances both H3K27 trimethylation and H2AK119 monoubiquitination. H3K27me3 may be enhanced by Mdm2 through activity changes of EZH2, or indirectly because PRC2 activity is enhanced by adjacent H2AK119ub1. H2AK119 could be directly ubiquitinated by Mdm2, but additional or alternative working models are also conceivable. (A) Mdm2 could provide E3 ubiquitin ligase function on distinct PRC2 target gene loci, making the maintenance of gene repression by H2AK119ub1 independent of the presence of PRC1. What argues against this model as the sole scenario is our observation that the global H2AK119ub1 levels in cells severely drop upon RING1B depletion, even when Mdm2 is still present. (B) The second and third working models are based on quite recent observations summarized by Comet and Helin (2014). According to their studies, PRC1 is characterized by a canonical or variant form containing CBX or RYBP, respectively. Mdm2 might facilitate the interaction between vPRC1 and PRC2 through simultaneous interaction with RING1B/RYBP and EZH2/SUZ12. Although we did not observe a decrease in RING1B on PRC2 target promoters upon removing Mdm2, it does not exclude that certain PRC1 species (like canonical and variant) might still associate with Mdm2-bound chromatin. Mdm2 would have two functions according to this model: bridging PRC1 and PRC2, as well as H2AK119 ubiquitination. (C) Similar to B, Mdm2 might bind to specific PRC1 subspecies, either canonical or variant. Here, Mdm2 could facilitate the recruitment of cPRC1 to PRC2 or H3K27me3. Instead of ubiquitinating H2AK119, this complex was reported to act by compacting chromatin through unknown mechanisms (Eskeland et al., 2010). We speculate that these hypothetical mechanisms may act together and may depend on each other, as is often the case when several repressive modifications trigger chromatin compaction.