Tet proteins oxidize 5-methylcytosine (mC) to generate 5-hydroxymethyl (hmC), 5-formyl (fC), and 5-carboxylcytosine (caC). The exact function of these oxidative cytosine bases remains elusive. We applied quantitative mass-spectrometry-based proteomics to identify readers for mC and hmC in mouse embryonic stem cells (mESC), neuronal progenitor cells (NPC), and adult mouse brain tissue. Readers for these modifications are only partially overlapping, and some readers, such as Rfx proteins, display strong specificity. Interactions are dynamic during differentiation, as for example evidenced by the mESC-specific binding of Klf4 to mC and the NPC-specific binding of Uhrf2 to hmC, suggesting specific biological roles for mC and hmC. Oxidized derivatives of mC recruit distinct transcription regulators as well as a large number of DNA repair proteins in mouse ES cells, implicating the DNA damage response as a major player in active DNA demethylation.

Ten eleven translocation (Tet) enzymes oxidize the epigenetically important DNA base 5-methylcytosine (mC) stepwise to 5-hydroxymethylcytosine (hmC), 5-formylcytosine and 5-carboxycytosine. It is currently unknown whether Tet-induced oxidation is limited to cytosine-derived nucleobases or whether other nucleobases are oxidized as well. We synthesized isotopologs of all major oxidized pyrimidine and purine bases and performed quantitative MS to show that Tet-induced oxidation is not limited to mC but that thymine is also a substrate that gives 5-hydroxymethyluracil (hmU) in mouse embryonic stem cells (mESCs). Using MS-based isotope tracing, we show that deamination of hmC does not contribute to the steady-state levels of hmU in mESCs. Protein pull-down experiments in combination with peptide tracing identifies hmU as a base that influences binding of chromatin remodeling proteins and transcription factors, suggesting that hmU has a specific function in stem cells besides triggering DNA repair.

Posttranslational modifications on core histones can serve as binding scaffolds for chromatin-associated proteins. Proteins that specifically bind to or "read" these modifications were previously identified in mass spectrometry-based proteomics screens based on stable isotope-labeling in cell lines. Here we describe a sensitive, label-free histone peptide pull-down technology with extracts of different mouse tissues. Applying this workflow to the classical activating and repressive epigenetic marks on histone H3, H3K4me3, and H3K9me3, we identified known and putative readers in extracts from brain, liver, kidney, and testis. A large class of proteins were specifically repelled by H3K4me3. Our screen reached near-saturation of direct interactors, most of which are ubiquitously expressed. In addition, it revealed a number of specialized readers in tissues such as testis. Apart from defining the chromatin interaction landscape in mouse tissues, our workflow can be used for peptides with different modifications and cell types of any organism.

NuRD (nucleosome remodeling and histone deacetylase) is a versatile multi-protein complex with roles in transcription regulation and the DNA damage response. Here, we show that ZMYND8 bridges NuRD to a number of putative DNA-binding zinc finger proteins. The MYND domain of ZMYND8 directly interacts with PPPLΦ motifs in the NuRD subunit GATAD2A. Both GATAD2A and GATAD2B exclusively form homodimers and define mutually exclusive NuRD subcomplexes. ZMYND8 and NuRD share a large number of genome-wide binding sites, mostly active promoters and enhancers. Depletion of ZMYND8 does not affect NuRD occupancy genome-wide and only slightly affects expression of NuRD/ZMYND8 target genes. In contrast, the MYND domain in ZMYND8 facilitates the rapid, poly(ADP-ribose)-dependent recruitment of GATAD2A/NuRD to sites of DNA damage to promote repair by homologous recombination. Thus, these results show that a specific substoichiometric interaction with a NuRD subunit paralogue provides unique functionality to distinct NuRD subcomplexes.

DNA methylation is an epigenetic modification that is generally associated with repression of transcription initiation at CpG-island promoters. Here we argue that, on the basis of recent high-throughput genomic and proteomic screenings, DNA methylation can also have different outcomes, including activation of transcription. This is evidenced by the fact that transcription factors can interact with methylated DNA sequences. Furthermore, in certain cellular contexts, genes containing methylated promoters are highly transcribed. Interestingly, this uncoupling between methylated DNA and repression of transcription seems to be particularly evident in germ cells and pluripotent cells. Thus, contrary to previous assumptions, DNA methylation is not exclusively associated with repression of transcription initiation.

dX chromosome inactivation is a remarkable example of chromosome-wide gene silencing and facultative heterochromatin formation. Numerous histone posttranslational modifications, including H3K9me2 and H3K27me3, accompany this process, although our understanding of the enzymes that lay down these marks and the factors that bind to them is still incomplete. Here we identify Cdyl, a chromodomain-containing transcriptional corepressor, as a new chromatin-associated protein partner of the inactive X chromosome (Xi). Using mouse embryonic stem cell lines with mutated histone methyltransferase activities, we show that Cdyl relies on H3K9me2 for its general association with chromatin in vivo. For its association with Xi, Cdyl requires the process of differentiation and the presence of H3K9me2 and H3K27me3, which both become chromosomally enriched following Xist RNA coating. We further show that the removal of the PRC2 component Eed and subsequent loss of H3K27me3 lead to a reduction of both Cdyl and H3K9me2 enrichment on inactive Xi. Finally, we show that Cdyl associates with the H3K9 histone methyltransferase G9a and the MGA protein, both of which are also found on Xi. We propose that the combination of H3K9me2 and H3K27me3 recruits Cdyl to Xi, and this, in turn, may facilitate propagation of the H3K9me2 mark by anchoring G9a.

BackgroundDNA methylation is an epigenetic modification that plays a crucial role in a variety of biological processes. Methylated DNA is specifically bound by Methyl-CpG Binding Proteins (MBPs). Three different types of MBPs have been identified so far: the Methyl-CpG Binding Domain (MBD) family proteins, three BTB/POZ-Zn-finger proteins, and UHRF1. Most of the known MBPs have been identified via homology with the MBD and Zn-finger domains as present in MeCP2 and Kaiso, respectively. It is conceivable that other proteins are capable of recognizing methylated DNA.Methodology/Principal FindingsFor the purpose of identifying novel ‘readers’ we set up a methyl-CpG pull-down assay combined with stable-isotope labeling by amino acids in cell culture (SILAC). In a methyl-CpG pull-down with U937 nuclear extracts, we recovered several known MBPs and almost all subunits of the MBD2/NuRD complex as methylation specific binders, providing proof-of-principle. Interestingly, RBP-J, the transcription factor downstream of Notch receptors, also bound the DNA in a methylation dependent manner. Follow-up pull-downs and electrophoretic mobility shift assays (EMSAs) showed that RBP-J binds methylated DNA in the context of a mutated RBP-J consensus motif.Conclusions/SignificanceThe here described SILAC/methyl-CpG pull-down constitutes a new approach to identify potential novel DNAme readers and will advance unraveling of the complete methyl-DNA interactome.

Summary Mammalian genomes contain long domains with distinct average compositions of A/T versus G/C base pairs. In a screen for proteins that might interpret base composition by binding to AT-rich motifs, we identified the stem cell factor SALL4, which contains multiple zinc fingers. Mutation of the domain responsible for AT binding drastically reduced SALL4 genome occupancy and prematurely upregulated genes in proportion to their AT content. Inactivation of this single AT-binding zinc-finger cluster mimicked defects seen in Sall4 null cells, including precocious differentiation of embryonic stem cells (ESCs) and embryonic lethality in mice. In contrast, deletion of two other zinc-finger clusters was phenotypically neutral. Our data indicate that loss of pluripotency is triggered by downregulation of SALL4, leading to de-repression of a set of AT-rich genes that promotes neuronal differentiation. We conclude that base composition is not merely a passive byproduct of genome evolution and constitutes a signal that aids control of cell fate.