TDRD3 is an effector molecule for arginine-methylated histone marks.

Specific sites of histone tail methylation are associated with transcriptional activity at gene loci. These methyl marks are interpreted by effector molecules, which harbor protein domains that bind the methylated motifs and facilitate either active or inactive states of transcription. CARM1 and PRMT1 are transcriptional coactivators that deposit H3R17me2a and H4R3me2a marks, respectively. We used a protein domain microarray approach to identify the Tudor domain-containing protein TDRD3 as a "reader" of these marks. Importantly, TDRD3 itself is a transcriptional coactivator. This coactivator activity requires an intact Tudor domain. TDRD3 is recruited to an estrogen-responsive element in a CARM1-dependent manner. Furthermore, ChIP-seq analysis of TDRD3 reveals that it is predominantly localized to transcriptional start sites. Thus, TDRD3 is an effector molecule that promotes transcription by binding methylarginine marks on histone tails.

Local sequence and sequencing depth dependent accuracy of RNA-seq reads.

BACKGROUND: Many biases and spurious effects are inherent in RNA-seq technology, resulting in a non-uniform distribution of sequencing read counts for each base position in a gene. Therefore, a base-level strategy is required to model the non-uniformity. Also, the properties of sequencing read counts can be leveraged to achieve a more precise estimation of the mean and variance of measurement. RESULTS: In this study, we aimed to unveil the effects on RNA-seq accuracy from multiple factors and develop accurate modeling of RNA-seq reads in comparison. We found that the overdispersion rate decreased when sequencing depth increased on the base level. Moreover, the influence of local sequence(s) on the overdispersion rate was notable but no longer significant after adjusting the effect from sequencing depth. Based on these findings, we propose a desirable beta-binomial model with a dynamic overdispersion rate on the base-level proportion of sequencing read counts from two samples. CONCLUSIONS: The current study provides thorough insights into the impact of overdispersion at the position level and especially into its relationship with sequencing depth, local sequence, and preparation protocol. These properties of RNA-seq will aid in improvement of the quality control procedure and development of statistical methods for RNA-seq downstream analyses.

Age-related epigenetic drift in the pathogenesis of MDS and AML.

The myelodysplastic syndrome (MDS) is a clonal hematologic disorder that frequently evolves to acute myeloid leukemia (AML). Its pathogenesis remains unclear, but mutations in epigenetic modifiers are common and the disease often responds to DNA methylation inhibitors. We analyzed DNA methylation in the bone marrow and spleen in two mouse models of MDS/AML, the NUP98-HOXD13 (NHD13) mouse and the RUNX1 mutant mouse model. Methylation array analysis showed an average of 512/3445 (14.9%) genes hypermethylated in NHD13 MDS, and 331 (9.6%) genes hypermethylated in RUNX1 MDS. Thirty-two percent of genes in common between the two models (2/3 NHD13 mice and 2/3 RUNX1 mice) were also hypermethylated in at least two of 19 human MDS samples. Detailed analysis of 41 genes in mice showed progressive drift in DNA methylation from young to old normal bone marrow and spleen; to MDS, where we detected accelerated age-related methylation; and finally to AML, which markedly extends DNA methylation abnormalities. Most of these genes showed similar patterns in human MDS and AML. Repeat element hypomethylation was rare in MDS but marked the transition to AML in some cases. Our data show consistency in patterns of aberrant DNA methylation in human and mouse MDS and suggest that epigenetically, MDS displays an accelerated aging phenotype.

Stabilization of the promoter nucleosomes in nucleosome-free regions by the yeast Cyc8-Tup1 corepressor.

The yeast Cyc8 (also known as Ssn6)-Tup1 complex regulates gene expression through a variety of mechanisms, including positioning of nucleosomes over promoters of some target genes to limit accessibility to the transcription machinery. To further define the functions of Cyc8-Tup1 in gene regulation and chromatin remodeling, we performed genome-wide profiling of changes in nucleosome organization and gene expression that occur upon loss of CYC8 or TUP1 and observed extensive nucleosome alterations in both promoters and gene bodies of derepressed genes. Our improved nucleosome profiling and analysis approaches revealed low-occupancy promoter nucleosomes (P nucleosomes) at locations previously defined as nucleosome-free regions. In the absence of CYC8 or TUP1, this P nucleosome is frequently lost, whereas nucleosomes are gained at -1 and +1 positions, accompanying up-regulation of downstream genes. Our analysis of public ChIP-seq data revealed that Cyc8 and Tup1 preferentially bind TATA-containing promoters, which are also enriched in genes derepressed upon loss of CYC8 or TUP1. These results suggest that stabilization of the P nucleosome on TATA-containing promoters may be a central feature of the repressive chromatin architecture created by the Cyc8-Tup1 corepressor, and that releasing the P nucleosome contributes to gene activation.

TET1 is a maintenance DNA demethylase that prevents methylation spreading in differentiated cells.

TET1 is a 5-methylcytosine dioxygenase and its DNA demethylating activity has been implicated in pluripotency and reprogramming. However, the precise role of TET1 in DNA methylation regulation outside of developmental reprogramming is still unclear. Here, we show that overexpression of the TET1 catalytic domain but not full length TET1 (TET1-FL) induces massive global DNA demethylation in differentiated cells. Genome-wide mapping reveals that 5-hydroxymethylcytosine production by TET1-FL is inhibited as DNA methylation increases, which can be explained by the preferential binding of TET1-FL to unmethylated CpG islands (CGIs) through its CXXC domain. TET1-FL specifically accumulates 5-hydroxymethylcytosine at the edges of hypomethylated CGIs, while knockdown of endogenous TET1 induces methylation spreading from methylated edges into hypomethylated CGIs. We also found that gene expression changes after TET1-FL overexpression are relatively small and independent of its dioxygenase function. Thus, our results identify TET1 as a maintenance DNA demethylase that does not purposely decrease methylation levels, but specifically prevents aberrant methylation spreading into CGIs in differentiated cells.

Molecular mechanism by which the nucleoid occlusion factor, SlmA, keeps cytokinesis in check.

In Escherichia coli, cytokinesis is orchestrated by FtsZ, which forms a Z-ring to drive septation. Spatial and temporal control of Z-ring formation is achieved by the Min and nucleoid occlusion (NO) systems. Unlike the well-studied Min system, less is known about the anti-DNA guillotining NO process. Here, we describe studies addressing the molecular mechanism of SlmA (synthetic lethal with a defective Min system)-mediated NO. SlmA contains a TetR-like DNA-binding fold, and chromatin immunoprecipitation analyses show that SlmA-binding sites are dispersed on the chromosome except the Ter region, which segregates immediately before septation. SlmA binds DNA and FtsZ simultaneously, and the SlmA-FtsZ structure reveals that two FtsZ molecules sandwich a SlmA dimer. In this complex, FtsZ can still bind GTP and form protofilaments, but the separated protofilaments are forced into an anti-parallel arrangement. This suggests that SlmA may alter FtsZ polymer assembly. Indeed, electron microscopy data, showing that SlmA-DNA disrupts the formation of normal FtsZ polymers and induces distinct spiral structures, supports this. Thus, the combined data reveal how SlmA derails Z-ring formation at the correct place and time to effect NO.

Access to the nucleus and functional association with c-Myc is required for the full oncogenic potential of ΔEGFR/EGFRvIII.

ΔEGFR is a potent glioblastoma oncogene which has been studied primarily as a plasma membrane kinase. Using intracranial xenograft studies in mice, we show that blocking ΔEGFR access to the nucleus attenuates its tumorigenicity and, conversely, that promoting nuclear accumulation enhances this, providing the first in vivo evidence that the nuclear actions of ΔEGFR contribute strongly to its oncogenic function. Nuclear actions of ΔEGFR include regulation of gene expression by participation in chromatin-bound complexes, and genome-wide mapping of these sequences by chromatin immunoprecipitation and massively parallel sequencing identified 2294 peaks. Bioinformatic analysis showed enrichment of the E-box motif in the dataset, and c-Myc and ΔEGFR were corecruited to the promoters of and transcriptionally activated a subset of nuclear ΔEGFR chromatin targets. Knockdown of c-Myc decreased the expression of these targets and diminished ΔEGFR-stimulated anchorage-independent colony formation. We conclude that transcriptional regulation of target genes by association with gene regulatory chromatin in cooperation with c-Myc by nuclear ΔEGFR makes a unique contribution to its oncogenicity and propose that this venue provides new targets for therapeutic intervention.

Widespread and tissue specific age-related DNA methylation changes in mice.

Aberrant methylation of promoter CpG islands in cancer is associated with silencing of tumor-suppressor genes, and age-dependent hypermethylation in normal appearing mucosa may be a risk factor for human colon cancer. It is not known whether this age-related DNA methylation phenomenon is specific to human tissues. We performed comprehensive DNA methylation profiling of promoter regions in aging mouse intestine using methylated CpG island amplification in combination with microarray analysis. By comparing C57BL/6 mice at 3-mo-old versus 35-mo-old for 3627 detectable autosomal genes, we found 774 (21%) that showed increased methylation and 466 (13%) that showed decreased methylation. We used pyrosequencing to quantitatively validate the microarray data and confirmed linear age-related methylation changes for all 12 genomic regions examined. We then examined 11 changed genomic loci for age-related methylation in other tissues. Of these, three of 11 showed similar changes in lung, seven of 11 changed in liver, and six of 11 changed in spleen, though to a lower degree than the changes seen in colon. There was partial conservation between age-related hypermethylation in human and mouse intestines, and Polycomb targets in embryonic stem cells were enriched among the hypermethylated genes. Our findings demonstrate a surprisingly high rate of hyper- and hypomethylation as a function of age in normal mouse small intestine tissues and a strong tissue-specificity to the process. We conclude that epigenetic deregulation is a common feature of aging in mammals.

Genome architecture marked by retrotransposons modulates predisposition to DNA methylation in cancer.

Epigenetic silencing plays an important role in cancer development. An attractive hypothesis is that local DNA features may participate in differential predisposition to gene hypermethylation. We found that, compared with methylation-resistant genes, methylation-prone genes have a lower frequency of SINE and LINE retrotransposons near their transcription start site. In several large testing sets, this distribution was highly predictive of promoter methylation. Genome-wide analysis showed that 22% of human genes were predicted to be methylation-prone in cancer; these tended to be genes that are down-regulated in cancer and that function in developmental processes. Moreover, retrotransposon distribution marks a larger fraction of methylation-prone genes compared to Polycomb group protein (PcG) marking in embryonic stem cells; indeed, PcG marking and our predictive model based on retrotransposon frequency appear to be correlated but also complementary. In summary, our data indicate that retrotransposon elements, which are widespread in our genome, are strongly associated with gene promoter DNA methylation in cancer and may in fact play a role in influencing epigenetic regulation in normal and abnormal physiological states.

REST regulates oncogenic properties of glioblastoma stem cells.

Glioblastoma multiforme (GBM) tumors are the most common malignant primary brain tumors in adults. Although many GBM tumors are believed to be caused by self-renewing, glioblastoma-derived stem-like cells (GSCs), the mechanisms that regulate self-renewal and other oncogenic properties of GSCs are only now being unraveled. Here we showed that GSCs derived from GBM patient specimens express varying levels of the transcriptional repressor repressor element 1 silencing transcription factor (REST), suggesting heterogeneity across different GSC lines. Loss- and gain-of-function experiments indicated that REST maintains self-renewal of GSCs. High REST-expressing GSCs (HR-GSCs) produced tumors histopathologically distinct from those generated by low REST-expressing GSCs (LR-GSCs) in orthotopic mouse brain tumor models. Knockdown of REST in HR-GSCs resulted in increased survival in GSC-transplanted mice and produced tumors with higher apoptotic and lower invasive properties. Conversely, forced expression of exogenous REST in LR-GSCs produced decreased survival in mice and produced tumors with lower apoptotic and higher invasive properties, similar to HR-GSCs. Thus, based on our results, we propose that a novel function of REST is to maintain self-renewal and other oncogenic properties of GSCs and that REST can play a major role in mediating tumorigenicity in GBM.

dsPIG: a tool to predict imprinted genes from the deep sequencing of whole transcriptomes.

BACKGROUND: Dysregulation of imprinted genes, which are expressed in a parent-of-origin-specific manner, plays an important role in various human diseases, such as cancer and behavioral disorder. To date, however, fewer than 100 imprinted genes have been identified in the human genome. The recent availability of high-throughput technology makes it possible to have large-scale prediction of imprinted genes. Here we propose a Bayesian model (dsPIG) to predict imprinted genes on the basis of allelic expression observed in mRNA-Seq data of independent human tissues. RESULTS: Our model (dsPIG) was capable of identifying imprinted genes with high sensitivity and specificity and a low false discovery rate when the number of sequenced tissue samples was fairly large, according to simulations. By applying dsPIG to the mRNA-Seq data, we predicted 94 imprinted genes in 20 cerebellum samples and 57 imprinted genes in 9 diverse tissue samples with expected low false discovery rates. We also assessed dsPIG using previously validated imprinted and non-imprinted genes. With simulations, we further analyzed how imbalanced allelic expression of non-imprinted genes or different minor allele frequencies affected the predictions of dsPIG. Interestingly, we found that, among biallelically expressed genes, at least 18 genes expressed significantly more transcripts from one allele than the other among different individuals and tissues. CONCLUSION: With the prevalence of the mRNA-Seq technology, dsPIG has become a useful tool for analysis of allelic expression and large-scale prediction of imprinted genes. For ease of use, we have set up a web service and also provided an R package for dsPIG at http://www.shoudanliang.com/dsPIG/.

Accuracy of RNA-Seq and its dependence on sequencing depth.

BACKGROUND: The cost of DNA sequencing has undergone a dramatical reduction in the past decade. As a result, sequencing technologies have been increasingly applied to genomic research. RNA-Seq is becoming a common technique for surveying gene expression based on DNA sequencing. As it is not clear how increased sequencing capacity has affected measurement accuracy of mRNA, we sought to investigate that relationship. RESULT: We empirically evaluate the accuracy of repeated gene expression measurements using RNA-Seq. We identify library preparation steps prior to DNA sequencing as the main source of error in this process. Studying three datasets, we show that the accuracy indeed improves with the sequencing depth. However, the rate of improvement as a function of sequence reads is generally slower than predicted by the binomial distribution. We therefore used the beta-binomial distribution to model the overdispersion. The overdispersion parameters we introduced depend explicitly on the number of reads so that the resulting statistical uncertainty is consistent with the empirical data that measurement accuracy increases with the sequencing depth. The overdispersion parameters were determined by maximizing the likelihood. We shown that our modified beta-binomial model had lower false discovery rate than the binomial or the pure beta-binomial models. CONCLUSION: We proposed a novel form of overdispersion guaranteeing that the accuracy improves with sequencing depth. We demonstrated that the new form provides a better fit to the data.

BM-BC: a Bayesian method of base calling for Solexa sequence data.

Base calling is a critical step in the Solexa next-generation sequencing procedure. It compares the position-specific intensity measurements that reflect the signal strength of four possible bases (A, C, G, T) at each genomic position, and outputs estimates of the true sequences for short reads of DNA or RNA. We present a Bayesian method of base calling, BM-BC, for Solexa-GA sequencing data. The Bayesian method builds on a hierarchical model that accounts for three sources of noise in the data, which are known to affect the accuracy of the base calls: fading, phasing, and cross-talk between channels. We show that the new method improves the precision of base calling compared with currently leading methods. Furthermore, the proposed method provides a probability score that measures the confidence of each base call. This probability score can be used to estimate the false discovery rate of the base calling or to rank the precision of the estimated DNA sequences, which in turn can be useful for downstream analysis such as sequence alignment.

Neuroendocrine prostate cancer xenografts with large-cell and small-cell features derived from a single patient's tumor: morphological, immunohistochemical, and gene expression profiles.

BACKGROUND: Small-cell carcinoma (SCC) of the prostate is an AR-negative variant of prostate cancer found at progression in 10-20% of castrate-resistant disease. Its finding predicts a distinct clinical course and a poor prognosis. Large-cell neuroendocrine carcinoma (LCNEC) is a much rarer variant that behaves similarly to SCC. The biological mechanisms that drive these disease variants are poorly understood. METHODS: Eight tumor fragments from the salvage pelvic exenteration specimen of a patient with castrate-resistant prostate carcinoma were subcutaneously implanted into 6- to 8-week-old male CB17 SCID mice. Serial tissue sections and tissue microarrays of the resulting MDA PCa 144 xenograft lines were used for histopathologic and immunohistochemical characterization of the xenografts and their tissue of origin. RNA from two representative xenograft sublines was used for gene-expression profiling. RESULTS: All eight fragments formed tumors: four of the MDA PCa 144 xenograft sublines had morphologic characteristics of SCC and four, of LCNEC. All retained high fidelity to their parent tumor tissue, which remained stable through serial passages. Morphological transitions in the specimen of origin suggested LCNEC represents an intermediate step between adenocarcinoma and SCC. Over 2,500 genes were differentially expressed between the SCC (MDA PCa 144-13) and the LCNEC (MDA PCa 144-4) sublines and enriched in "Nervous System Development" Gene Ontology subtree. CONCLUSION: The eight xenograft models described represent the spectrum of neuroendocrine carcinomas in prostate cancer and will be valuable preclinical tools to study the pathogenesis of and therapy targets for this increasingly recognized subset of lethal prostate cancer.

BM-map: Bayesian mapping of multireads for next-generation sequencing data.

Next-generation sequencing (NGS) technology generates millions of short reads, which provide valuable information for various aspects of cellular activities and biological functions. A key step in NGS applications (e.g., RNA-Seq) is to map short reads to correct genomic locations within the source genome. While most reads are mapped to a unique location, a significant proportion of reads align to multiple genomic locations with equal or similar numbers of mismatches; these are called multireads. The ambiguity in mapping the multireads may lead to bias in downstream analyses. Currently, most practitioners discard the multireads in their analysis, resulting in a loss of valuable information, especially for the genes with similar sequences. To refine the read mapping, we develop a Bayesian model that computes the posterior probability of mapping a multiread to each competing location. The probabilities are used for downstream analyses, such as the quantification of gene expression. We show through simulation studies and RNA-Seq analysis of real life data that the Bayesian method yields better mapping than the current leading methods. We provide a C++ program for downloading that is being packaged into a user-friendly software.

Cell cycle-dependent nucleosome occupancy at cohesin binding sites in yeast chromosomes.

In the budding yeast, cohesin is loaded onto the chromosome during the late G1 phase, establishes sister chromatid cohesion concomitant with DNA replication, and dissociates by the telophase. Here, using oligonucleotide tiling arrays, we show that, at the anaphase, nearly all of the cohesin binding sites contain nucleosome-free regions. The majority of these sites remain nucleosome-free throughout the cell cycle, consistent with the suggestion of a DNA-binding anchoring protein present at these sites, although such a region could also serve as part of a marker for the binding of cohesin in the next cell cycle. However, a third of these sites are remodeled in the G1 phase, being reoccupied by nucleosomes by the G1/S boundary, though their subsequent removal in the S phase appears to be independent of DNA replication. Whether this difference is a result of other functions of cohesin or of the chromatin remains to be elucidated.

Digital Restriction Enzyme Analysis of Methylation (DREAM).

Digital Restriction Enzyme Analysis of Methylation (DREAM) is a method for quantitative mapping of DNA methylation across genomes using next-generation sequencing (NGS) technology. The method is based on sequential cuts of genomic DNA with a pair of restriction enzymes (SmaI and XmaI) at CCCGGG target sites. Unmethylated sites are first digested with SmaI. This enzyme cuts the sites in the middle at CCC^GGG, leaving behind blunt ended fragments. CpG methylation completely blocks SmaI; therefore, only unmethylated sites are cleaved. The remaining methylated sites are digested with XmaI in the next step. This enzyme is not blocked by CpG methylation. It cuts the recognition site sideways at C^CCGGG forming 5'-CCGG overhangs. The sequential cuts thus create distinct methylation-specific signatures at the ends of restriction fragments: 5'-GGG for unmethylated CpG sites and 5'-CCGGG for methylated sites. The DNA fragments resulting from the digestions are ligated to NGS adapters. Sequencing libraries are prepared using hexanucleotide barcodes for sample identification. Individual libraries with distinct barcodes are pooled and sequenced using a paired ends protocol. The sequencing reads are aligned to the genome and mapped to unique CCCGGG target sites. Methylation at individual CpG sites is calculated as the ratio of sequencing reads with the methylated signature to the total number of reads mapping to the site. Sequencing of 25 million reads per sample typically yields 50,000 unique CpG sites covered with hundreds of reads enabling accurate determination of DNA methylation levels. DREAM does not require bisulfite conversion, has a very low background, and has high sensitivity to low levels of methylation. The method is simple, cost-effective, quantitative, highly reproducible, and can be applied to any species.

fREDUCE: detection of degenerate regulatory elements using correlation with expression.

BACKGROUND: The precision of transcriptional regulation is made possible by the specificity of physical interactions between transcription factors and their cognate binding sites on DNA. A major challenge is to decipher transcription factor binding sites from sequence and functional genomic data using computational means. While current methods can detect strong binding sites, they are less sensitive to degenerate motifs. RESULTS: We present fREDUCE, a computational method specialized for the detection of weak or degenerate binding motifs from gene expression or ChIP-chip data. fREDUCE is built upon the widely applied program REDUCE, which elicits motifs by global statistical correlation of motif counts with expression data. fREDUCE introduces several algorithmic refinements that allow efficient exhaustive searches of oligonucleotides with a specified number of degenerate IUPAC symbols. On yeast ChIP-chip benchmarks, fREDUCE correctly identified motifs and their degeneracies with accuracies greater than its predecessor REDUCE as well as other known motif-finding programs. We have also used fREDUCE to make novel motif predictions for transcription factors with poorly characterized binding sites. CONCLUSION: We demonstrate that fREDUCE is a valuable tool for the prediction of degenerate transcription factor binding sites, especially from array datasets with weak signals that may elude other motif detection methods.

Caloric restriction delays age-related methylation drift.

In mammals, caloric restriction consistently results in extended lifespan. Epigenetic information encoded by DNA methylation is tightly regulated, but shows a striking drift associated with age that includes both gains and losses of DNA methylation at various sites. Here, we report that epigenetic drift is conserved across species and the rate of drift correlates with lifespan when comparing mice, rhesus monkeys, and humans. Twenty-two to 30-year-old rhesus monkeys exposed to 30% caloric restriction since 7-14 years of age showed attenuation of age-related methylation drift compared to ad libitum-fed controls such that their blood methylation age appeared 7 years younger than their chronologic age. Even more pronounced effects were seen in 2.7-3.2-year-old mice exposed to 40% caloric restriction starting at 0.3 years of age. The effects of caloric restriction on DNA methylation were detectable across different tissues and correlated with gene expression. We propose that epigenetic drift is a determinant of lifespan in mammals.Caloric restriction has been shown to increase lifespan in mammals. Here, the authors provide evidence that age-related methylation drift correlates with lifespan and that caloric restriction in mice and rhesus monkeys results in attenuation of age-related methylation drift.