Lukas Burger

Transcriptome-wide identification of RNA-binding protein and microRNA target sites by PAR-CLIP

RNA transcripts are subject to posttranscriptional gene regulation involving hundreds of RNA-binding proteins (RBPs) and microRNA-containing ribonucleoprotein complexes (miRNPs) expressed in a cell-type dependent fashion. We developed a cell-based crosslinking approach to determine at high resolution and transcriptome-wide the binding sites of cellular RBPs and miRNPs. The crosslinked sites are revealed by thymidine to cytidine transitions in the cDNAs prepared from immunopurified RNPs of 4-thiouridine-treated cells. We determined the binding sites and regulatory consequences for several intensely studied RBPs and miRNPs, including PUM2, QKI, IGF2BP1-3, AGO/EIF2C1-4 and TNRC6A-C. Our study revealed that these factors bind thousands of sites containing defined sequence motifs and have distinct preferences for exonic versus intronic or coding versus untranslated transcript regions. The precise mapping of binding sites across the transcriptome will be critical to the interpretation of the rapidly emerging data on genetic variation between individuals and how these variations contribute to complex genetic diseases.

DNA-binding factors shape the mouse methylome at distal regulatory regions

Methylation of cytosines is an essential epigenetic modification in mammalian genomes, yet the rules that govern methylation patterns remain largely elusive. To gain insights into this process, we generated base-pair-resolution mouse methylomes in stem cells and neuronal progenitors. Advanced quantitative analysis identified low-methylated regions (LMRs) with an average methylation of 30%. These represent CpG-poor distal regulatory regions as evidenced by location, DNase I hypersensitivity, presence of enhancer chromatin marks and enhancer activity in reporter assays. LMRs are occupied by DNA-binding factors and their binding is necessary and sufficient to create LMRs. A comparison of neuronal and stem-cell methylomes confirms this dependency, as cell-type-specific LMRs are occupied by cell-type-specific transcription factors. This study provides methylome references for the mouse and shows that DNA-binding factors locally influence DNA methylation, enabling the identification of active regulatory regions.

A quantitative analysis of CLIP methods for identifying binding sites of RNA-binding proteins

Cross-linking and immunoprecipitation (CLIP) is increasingly used to map transcriptome-wide binding sites of RNA-binding proteins. We developed a method for CLIP data analysis, and applied it to compare CLIP with photoactivatable ribonucleoside–enhanced CLIP (PAR-CLIP) and to uncover how differences in cross-linking and ribonuclease digestion affect the identified sites. We found only small differences in accuracies of these methods in identifying binding sites of HuR, which binds low-complexity sequences, and Argonaute 2, which has a complex binding specificity. We found that cross-link–induced mutations led to single-nucleotide resolution for both PAR-CLIP and CLIP. Our results confirm the expectation from original CLIP publications that RNA-binding proteins do not protect their binding sites sufficiently under the denaturing conditions used during the CLIP procedure, and we show that extensive digestion with sequence-specific RNases strongly biases the recovered binding sites. This bias can be substantially reduced by milder nuclease digestion conditions.

Genomic profiling of DNA methyltransferases reveals a role for DNMT3B in genic methylation.

DNA methylation is an epigenetic modification associated with transcriptional repression of promoters and is essential for mammalian development. Establishment of DNA methylation is mediated by the de novo DNA methyltransferases DNMT3A and DNMT3B, whereas DNMT1 ensures maintenance of methylation through replication. Absence of these enzymes is lethal, and somatic mutations in these genes have been associated with several human diseases. How genomic DNA methylation patterns are regulated remains poorly understood, as the mechanisms that guide recruitment and activity of DNMTs in vivo are largely unknown. To gain insights into this matter we determined genomic binding and site-specific activity of the mammalian de novo DNA methyltransferases DNMT3A and DNMT3B. We show that both enzymes localize to methylated, CpG-dense regions in mouse stem cells, yet are excluded from active promoters and enhancers. By specifically measuring sites of de novo methylation, we observe that enzymatic activity reflects binding. De novo methylation increases with CpG density, yet is excluded from nucleosomes. Notably, we observed selective binding of DNMT3B to the bodies of transcribed genes, which leads to their preferential methylation. This targeting to transcribed sequences requires SETD2-mediated methylation of lysine 36 on histone H3 and a functional PWWP domain of DNMT3B. Together these findings reveal how sequence and chromatin cues guide de novo methyltransferase activity to ensure methylome integrity.

Competition between DNA methylation and transcription factors determines binding of NRF1

Eukaryotic transcription factors (TFs) are key determinants of gene activity, yet they bind only a fraction of their corresponding DNA sequence motifs in any given cell type. Chromatin has the potential to restrict accessibility of binding sites; however, in which context chromatin states are instructive for TF binding remains mainly unknown. To explore the contribution of DNA methylation to constrained TF binding, we mapped DNase-I-hypersensitive sites in murine stem cells in the presence and absence of DNA methylation. Methylation-restricted sites are enriched for TF motifs containing CpGs, especially for those of NRF1. In fact, the TF NRF1 occupies several thousand additional sites in the unmethylated genome, resulting in increased transcription. Restoring de novo methyltransferase activity initiates remethylation at these sites and outcompetes NRF1 binding. This suggests that binding of DNA-methylation-sensitive TFs relies on additional determinants to induce local hypomethylation. In support of this model, removal of neighbouring motifs in cis or of a TF in trans causes local hypermethylation and subsequent loss of NRF1 binding. This competition between DNA methylation and TFs in vivo reveals a case of cooperativity between TFs that acts indirectly via DNA methylation. Methylation removal by methylation-insensitive factors enables occupancy of methylation-sensitive factors, a principle that rationalizes hypomethylation of regulatory regions.

PAR-CliP-a method to identify transcriptome-wide the binding sites of RNA binding proteins

RNA transcripts are subjected to post-transcriptional gene regulation by interacting with hundreds of RNA-binding proteins (RBPs) and microRNA-containing ribonucleoprotein complexes (miRNPs) that are often expressed in a cell-type dependently. To understand how the interplay of these RNA-binding factors affects the regulation of individual transcripts, high resolution maps of in vivo protein-RNA interactions are necessary1. A combination of genetic, biochemical and computational approaches are typically applied to identify RNA-RBP or RNA-RNP interactions. Microarray profiling of RNAs associated with immunopurified RBPs (RIP-Chip)2 defines targets at a transcriptome level, but its application is limited to the characterization of kinetically stable interactions and only in rare cases3,4 allows to identify the RBP recognition element (RRE) within the long target RNA. More direct RBP target site information is obtained by combining in vivo UV crosslinking5,6 with immunoprecipitation7-9 followed by the isolation of crosslinked RNA segments and cDNA sequencing (CLIP)10. CLIP was used to identify targets of a number of RBPs11-17. However, CLIP is limited by the low efficiency of UV 254 nm RNA-protein crosslinking, and the location of the crosslink is not readily identifiable within the sequenced crosslinked fragments, making it difficult to separate UV-crosslinked target RNA segments from background non-crosslinked RNA fragments also present in the sample. We developed a powerful cell-based crosslinking approach to determine at high resolution and transcriptome-wide the binding sites of cellular RBPs and miRNPs that we term PAR-CliP (Photoactivatable-Ribonucleoside-Enhanced Crosslinking and Immunoprecipitation) (see Fig. 1A for an outline of the method). The method relies on the incorporation of photoreactive ribonucleoside analogs, such as 4-thiouridine (4-SU) and 6-thioguanosine (6-SG) into nascent RNA transcripts by living cells. Irradiation of the cells by UV light of 365 nm induces efficient crosslinking of photoreactive nucleoside-labeled cellular RNAs to interacting RBPs. Immunoprecipitation of the RBP of interest is followed by isolation of the crosslinked and coimmunoprecipitated RNA. The isolated RNA is converted into a cDNA library and deep sequenced using Solexa technology. One characteristic feature of cDNA libraries prepared by PAR-CliP is that the precise position of crosslinking can be identified by mutations residing in the sequenced cDNA. When using 4-SU, crosslinked sequences thymidine to cytidine transition, whereas using 6-SG results in guanosine to adenosine mutations. The presence of the mutations in crosslinked sequences makes it possible to separate them from the background of sequences derived from abundant cellular RNAs. Application of the method to a number of diverse RNA binding proteins was reported in Hafner et al.18

Transcription Factor Occupancy Can Mediate Active Turnover of DNA Methylation at Regulatory Regions

Distal regulatory elements, including enhancers, play a critical role in regulating gene activity. Transcription factor binding to these elements correlates with Low Methylated Regions (LMRs) in a process that is poorly understood. Here we ask whether and how actual occupancy of DNA-binding factors is linked to DNA methylation at the level of individual molecules. Using CTCF as an example, we observe that frequency of binding correlates with the likelihood of a demethylated state and sites of low occupancy display heterogeneous DNA methylation within the CTCF motif. In line with a dynamic model of binding and DNA methylation turnover, we find that 5-hydroxymethylcytosine (5hmC), formed as an intermediate state of active demethylation, is enriched at LMRs in stem and somatic cells. Moreover, a significant fraction of changes in 5hmC during differentiation occurs at these regions, suggesting that transcription factor activity could be a key driver for active demethylation. Since deletion of CTCF is lethal for embryonic stem cells, we used genetic deletion of REST as another DNA-binding factor implicated in LMR formation to test this hypothesis. The absence of REST leads to a decrease of hydroxymethylation and a concomitant increase of DNA methylation at its binding sites. These data support a model where DNA-binding factors can mediate turnover of DNA methylation as an integral part of maintenance and reprogramming of regulatory regions.

Accurate prediction of protein-protein interactions from sequence alignments using a Bayesian method.

Accurate and large‐scale prediction of protein–protein interactions directly from amino‐acid sequences is one of the great challenges in computational biology. Here we present a new Bayesian network method that predicts interaction partners using only multiple alignments of amino‐acid sequences of interacting protein domains, without tunable parameters, and without the need for any training examples. We first apply the method to bacterial two‐component systems and comprehensively reconstruct two‐component signaling networks across all sequenced bacteria. Comparisons of our predictions with known interactions show that our method infers interaction partners genome‐wide with high accuracy. To demonstrate the general applicability of our method we show that it also accurately predicts interaction partners in a recent dataset of polyketide synthases. Analysis of the predicted genome‐wide two‐component signaling networks shows that cognates (interacting kinase/regulator pairs, which lie adjacent on the genome) and orphans (which lie isolated) form two relatively independent components of the signaling network in each genome. In addition, while most genes are predicted to have only a small number of interaction partners, we find that 10% of orphans form a separate class of ‘hub’ nodes that distribute and integrate signals to and from up to tens of different interaction partners.

Disentangling direct from indirect co-evolution of residues in protein alignments.

Predicting protein structure from primary sequence is one of the ultimate challenges in computational biology. Given the large amount of available sequence data, the analysis of co-evolution, i.e., statistical dependency, between columns in multiple alignments of protein domain sequences remains one of the most promising avenues for predicting residues that are contacting in the structure. A key impediment to this approach is that strong statistical dependencies are also observed for many residue pairs that are distal in the structure. Using a comprehensive analysis of protein domains with available three-dimensional structures we show that co-evolving contacts very commonly form chains that percolate through the protein structure, inducing indirect statistical dependencies between many distal pairs of residues. We characterize the distributions of length and spatial distance traveled by these co-evolving contact chains and show that they explain a large fraction of observed statistical dependencies between structurally distal pairs. We adapt a recently developed Bayesian network model into a rigorous procedure for disentangling direct from indirect statistical dependencies, and we demonstrate that this method not only successfully accomplishes this task, but also allows contacts with weak statistical dependency to be detected. To illustrate how additional information can be incorporated into our method, we incorporate a phylogenetic correction, and we develop an informative prior that takes into account that the probability for a pair of residues to contact depends strongly on their primary-sequence distance and the amount of conservation that the corresponding columns in the multiple alignment exhibit. We show that our model including these extensions dramatically improves the accuracy of contact prediction from multiple sequence alignments.

Dynamic DNA methylation orchestrates cardiomyocyte development, maturation and disease

The heart is a highly specialized organ with essential function for the organism throughout life. The significance of DNA methylation in shaping the phenotype of the heart remains only partially known. Here we generate and analyse DNA methylomes from highly purified cardiomyocytes of neonatal, adult healthy and adult failing hearts. We identify large genomic regions that are differentially methylated during cardiomyocyte development and maturation. Demethylation of cardiomyocyte gene bodies correlates strongly with increased gene expression. Silencing of demethylated genes is characterized by the polycomb mark H3K27me3 or by DNA methylation. De novo methylation by DNA methyltransferases 3A/B causes repression of fetal cardiac genes, including essential components of the cardiac sarcomere. Failing cardiomyocytes partially resemble neonatal methylation patterns. This study establishes DNA methylation as a highly dynamic process during postnatal growth of cardiomyocytes and their adaptation to pathological stress in a process tightly linked to gene regulation and activity.