An all-to-all approach to the identification of sequence-specific readers for epigenetic DNA modifications on cytosine

Advancing TET Inhibitor Development: From Structural Insights to Biological Evaluation

Ten-eleven translocation (TET) methylcytosine dioxygenases are part of the epigenetic regulatory machinery that erases DNA methylation. Aberrant TET activities are frequently found in hematopoietic malignancies, where loss of TET2 function leads to DNA hypermethylation. A comprehensive understanding of the biological role of TETs is essential to elucidate disease pathogenesis and identify novel therapeutic strategies. We present a robust pipeline integrating protein X-ray crystallography, molecular modeling, and pharmacophore analysis to advance the current TET inhibitor development. In addition, we have synthesized and evaluated a series of 8-hydroxyquinoline (8-HQ) derivatives, demonstrating their potential as chemical tools to explore TET function further. These findings lay the groundwork for a TET-centered therapeutic approach.

Active DNA demethylation is upstream of rod-photoreceptor fate determination and required for retinal development

Retinal cell fate specification from multipotent retinal progenitors is governed by dynamic changes in chromatin structure and gene expression. Methylation at cytosines in DNA (5mC) is actively regulated for proper control of gene expression and chromatin architecture. Numerous genes display active DNA demethylation across retinal development; a process that requires oxidation of 5mC to 5-hydroxymethylcytosine (5hmC) and is controlled by the ten-eleven translocation methylcytosine dioxygenase (TET) enzymes. Using an allelic series of conditional TET enzyme mutants, we determine that DNA demethylation is required upstream of NRL and NR2E3 expression for the establishment of rod-photoreceptor fate. Using histological, behavioral, transcriptomic, and base-pair resolution DNA methylation analyses, we establish that inhibition of active DNA demethylation results in global changes in gene expression and methylation patterns that prevent photoreceptor precursors from adopting a rod-photoreceptor fate, instead producing a retina in which all photoreceptors specify as cones. Our results establish the TET enzymes and DNA demethylation as critical regulators of retinal development and cell fate specification, elucidating a novel mechanism required for the specification of rod-photoreceptors.

Identifying the "stripe" transcription factors and cooperative binding related to DNA methylation

DNA methylation plays a critical role in gene regulation by modulating the DNA binding of transcription factors (TFs). This study integrates TFs' ChIP-seq profiles with WGBS profiles to investigate how DNA methylation affects protein interactions. Statistical methods and a 5-letter DNA motif calling model have been developed to characterize DNA sequences bound by proteins, while considering the effects of DNA modifications. By employing these methods, 79 significant universal "stripe" TFs and cofactors (USFs), 2360 co-binding protein pairs, and distinct protein modules associated with various DNA methylation states have been identified. The USFs hint a regulatory hierarchy within these protein interactions. Proteins preferentially bind to non-CpG sites in methylated regions, indicating binding affinity is not solely CpG-dependent. Proteins involved in methylation-specific USFs and cobinding pairs play essential roles in promoting and sustaining DNA methylation through interacting with DNMTs or inhibiting TET binding. These findings underscore the interplay between protein binding and methylation, offering insights into epigenetic regulation in cellular biology.

Evolved Readers of 5-Carboxylcytosine CpG Dyads Reveal a High Versatility of the Methyl-CpG-Binding Domain for Recognition of Noncanonical Epigenetic Marks

Mammalian genomes are regulated by epigenetic cytosine (C) modifications in palindromic CpG dyads. Including canonical cytosine 5-methylation (mC), a total of four different 5-modifications can theoretically co-exist in the two strands of a CpG, giving rise to a complex array of combinatorial marks with unique regulatory potentials. While tailored readers for individual marks could serve as versatile tools to study their functions, it has been unclear whether a natural protein scaffold would allow selective recognition of marks that vastly differ from canonical, symmetrically methylated CpGs. We conduct directed evolution experiments to generate readers of 5-carboxylcytosine (caC) dyads based on the methyl-CpG-binding domain (MBD), the widely conserved natural reader of mC. Despite the stark steric and chemical differences to mC, we discover highly selective, low nanomolar binders of symmetric and asymmetric caC-dyads. Together with mutational and modelling studies, our findings reveal a striking evolutionary flexibility of the MBD scaffold, allowing it to completely abandon its conserved mC recognition mode in favour of noncanonical dyad recognition, highlighting its potential for epigenetic reader design.

Modeling methyl-sensitive transcription factor motifs with an expanded epigenetic alphabet

BACKGROUND

Transcription factors bind DNA in specific sequence contexts. In addition to distinguishing one nucleobase from another, some transcription factors can distinguish between unmodified and modified bases. Current models of transcription factor binding tend not to take DNA modifications into account, while the recent few that do often have limitations. This makes a comprehensive and accurate profiling of transcription factor affinities difficult.

RESULTS

Here, we develop methods to identify transcription factor binding sites in modified DNA. Our models expand the standard A/C/G/T DNA alphabet to include cytosine modifications. We develop Cytomod to create modified genomic sequences and we also enhance the MEME Suite, adding the capacity to handle custom alphabets. We adapt the well-established position weight matrix (PWM) model of transcription factor binding affinity to this expanded DNA alphabet. Using these methods, we identify modification-sensitive transcription factor binding motifs. We confirm established binding preferences, such as the preference of ZFP57 and C/EBPβ for methylated motifs and the preference of c-Myc for unmethylated E-box motifs.

CONCLUSIONS

Using known binding preferences to tune model parameters, we discover novel modified motifs for a wide array of transcription factors. Finally, we validate our binding preference predictions for OCT4 using cleavage under targets and release using nuclease (CUT&RUN) experiments across conventional, methylation-, and hydroxymethylation-enriched sequences. Our approach readily extends to other DNA modifications. As more genome-wide single-base resolution modification data becomes available, we expect that our method will yield insights into altered transcription factor binding affinities across many different modifications.

StableDNAm: towards a stable and efficient model for predicting DNA methylation based on adaptive feature correction learning

BACKGROUND

DNA methylation, instrumental in numerous life processes, underscores the paramount importance of its accurate prediction. Recent studies suggest that deep learning, due to its capacity to extract profound insights, provides a more precise DNA methylation prediction. However, issues related to the stability and generalization performance of these models persist.

RESULTS

In this study, we introduce an efficient and stable DNA methylation prediction model. This model incorporates a feature fusion approach, adaptive feature correction technology, and a contrastive learning strategy. The proposed model presents several advantages. First, DNA sequences are encoded at four levels to comprehensively capture intricate information across multi-scale and low-span features. Second, we design a sequence-specific feature correction module that adaptively adjusts the weights of sequence features. This improvement enhances the model's stability and scalability, or its generality. Third, our contrastive learning strategy mitigates the instability issues resulting from sparse data. To validate our model, we conducted multiple sets of experiments on commonly used datasets, demonstrating the model's robustness and stability. Simultaneously, we amalgamate various datasets into a single, unified dataset. The experimental outcomes from this combined dataset substantiate the model's robust adaptability.

CONCLUSIONS

Our research findings affirm that the StableDNAm model is a general, stable, and effective instrument for DNA methylation prediction. It holds substantial promise for providing invaluable assistance in future methylation-related research and analyses.

The mechanisms and factors that induce trained immunity in arthropods and mollusks

Besides dividing the organism's immune system into adaptive and innate immunity, it has long been thought that only adaptive immunity can establish immune memory. However, many studies have shown that innate immunity can also build immunological memory through epigenetic reprogramming and modifications to resist pathogens' reinfection, known as trained immunity. This paper reviews the role of mitochondrial metabolism and epigenetic modifications and describes the molecular foundation in the trained immunity of arthropods and mollusks. Mitochondrial metabolism and epigenetic modifications complement each other and play a key role in trained immunity.

Role of TET1-mediated epigenetic modulation in Alzheimer's disease

Alzheimer's disease (AD) is a neurodegenerative disorder influenced by a complex interplay of environmental, epigenetic, and genetic factors. DNA methylation (5mC) and hydroxymethylation (5hmC) are DNA modifications that serve as tissue-specific and temporal regulators of gene expression. TET family enzymes dynamically regulate these epigenetic modifications in response to environmental conditions, connecting environmental factors with gene expression. Previous epigenetic studies have identified 5mC and 5hmC changes associated with AD. In this study, we performed targeted resequencing of TET1 on a cohort of early-onset AD (EOAD) and control samples. Through gene-wise burden analysis, we observed significant enrichment of rare TET1 variants associated with AD (p = 0.04). We also profiled 5hmC in human postmortem brain tissues from AD and control groups. Our analysis identified differentially hydroxymethylated regions (DhMRs) in key genes responsible for regulating the methylome: TET3, DNMT3L, DNMT3A, and MECP2. To further investigate the role of Tet1 in AD pathogenesis, we used the 5xFAD mouse model with a Tet1 KO allele to examine how Tet1 loss influences AD pathogenesis. We observed significant changes in neuropathology, 5hmC, and RNA expression associated with Tet1 loss, while the behavioral alterations were not significant. The loss of Tet1 significantly increased amyloid plaque burden in the 5xFAD mouse (p = 0.044) and lead to a non-significant trend towards exacerbated AD-associated stress response in 5xFAD mice. At the molecular level, we found significant DhMRs enriched in genes involved in pathways responsible for neuronal projection organization, dendritic spine development and organization, and myelin assembly. RNA-Seq analysis revealed a significant increase in the expression of AD-associated genes such as Mpeg1, Ctsd, and Trem2. In conclusion, our results suggest that TET enzymes, particularly TET1, which regulate the methylome, may contribute to AD pathogenesis, as the loss of TET function increases AD-associated pathology.

Chemical Biology Approaches to Identify and Profile Interactors of Chromatin Modifications

In eukaryotes, DNA is packaged with histone proteins in a complex known as chromatin. Both the DNA and histone components of chromatin can be chemically modified in a wide variety of ways, resulting in a complex landscape often referred to as the "epigenetic code". These modifications are recognized by effector proteins that remodel chromatin and modulate transcription, translation, and repair of the underlying DNA. In this Review, we examine the development of methods for characterizing proteins that interact with these histone and DNA modifications. "Mark first" approaches utilize chemical, peptide, nucleosome, or oligonucleotide probes to discover interactors of a specific modification. "Reader first" approaches employ arrays of peptides, nucleosomes, or oligonucleotides to profile the binding preferences of interactors. These complementary strategies have greatly enhanced our understanding of how chromatin modifications effect changes in genomic regulation, bringing us ever closer to deciphering this complex language.

Aberrant DNA Methylation Patterns of Deleted in Liver Cancer 1 Isoforms in Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC), a common primary liver cancer, is the third leading cause of death worldwide. DNA methylation changes are common in HCC and have been studied to be associated with hepatocarcinogenesis. In our study, we used the MassARRAY^® EpiTYPER technology to investigate the methylation differences of deleted in liver cancer 1 (DLC1) (isoform 1 and 3) promoter between HCC tissues and corresponding adjacent noncancerous tissues and the association between methylation levels and clinicopathological features. In addition, the modified CRISPR-Cas9 system and the DNA methyltransferase inhibitor (DNMTi) were utilized to explore the functional correlation of epigenetic modifications and DLC1 gene regulation. The methylation levels of the DLC1 isoforms in HCC samples were found significantly lower than those in the adjacent noncancerous tissues (all p < 0.0001). Also, we found that the expression of DLC1 could be bidirectionally regulated by the modified CRISPR-Cas9 system and the DNMTi. Moreover, the hypomethylation of DLC1 in HCC samples was connected with the presence of satellite lesions (p = 0.0305) and incomplete tumor capsule (p = 0.0204). Receiver operator characteristic curve analysis demonstrated that the methylation levels of DLC1 could be applied to discriminate HCC patients (area under the curve = 0.728, p < 0.0001). The hypomethylation status was a key regulatory mechanism of DLC1 expression and might serve as a potential biomarker for HCC.

Autism Spectrum Disorder: Neurodevelopmental Risk Factors, Biological Mechanism, and Precision Therapy

Autism spectrum disorder (ASD) is a heterogeneous, behaviorally defined neurodevelopmental disorder. Over the past two decades, the prevalence of autism spectrum disorders has progressively increased, however, no clear diagnostic markers and specifically targeted medications for autism have emerged. As a result, neurobehavioral abnormalities, neurobiological alterations in ASD, and the development of novel ASD pharmacological therapy necessitate multidisciplinary collaboration. In this review, we discuss the development of multiple animal models of ASD to contribute to the disease mechanisms of ASD, as well as new studies from multiple disciplines to assess the behavioral pathology of ASD. In addition, we summarize and highlight the mechanistic advances regarding gene transcription, RNA and non-coding RNA translation, abnormal synaptic signaling pathways, epigenetic post-translational modifications, brain-gut axis, immune inflammation and neural loop abnormalities in autism to provide a theoretical basis for the next step of precision therapy. Furthermore, we review existing autism therapy tactics and limits and present challenges and opportunities for translating multidisciplinary knowledge of ASD into clinical practice.

Utilize a few features to classify presynaptic and postsynaptic neurotoxins

Neurotoxins are a class of proteins that have a significant damaging effect on nerve tissue. Neurotoxins are classified into presynaptic neurotoxins and postsynaptic neurotoxins, and accurate identification of neurotoxins plays a key role in drug development. In this study, 90 presynaptic neurotoxins and 165 postsynaptic neurotoxins were classified. The features of the presynaptic and postsynaptic neurotoxin sequences were extracted using the AutoProp feature extraction method and feature selection was performed using the maximum relevance maximum distance (MRMD) program, Finally, only two features were retained to achieve 84.7% classification accuracy. Moreover, it was found that the two retained features were present in the conserved sites and motifs of presynaptic neurotoxins and could represent the critical structures of presynaptic neurotoxins. This method demonstrates that using a few key features to classify proteins can effectively identify critical protein structures.

A high-throughput effector screen identifies a novel small molecule scaffold for inhibition of ten-eleven translocation dioxygenase 2

Ten-eleven translocation dioxygenases (TETs) are the erasers of 5-methylcytosine (mC), the central epigenetic regulator of mammalian DNA. TETs convert mC to three oxidized derivatives with unique physicochemical properties and inherent regulatory potential, and it initializes active demethylation by the base excision repair pathway. Potent small molecule inhibitors would be useful tools to study TET functions by conditional control. To facilitate the discovery of such tools, we here report a high-throughput screening pipeline and its application to screen and validate 31.5k compounds for inhibition of TET2. Using a homogenous fluorescence assay, we discover a novel quinoline-based scaffold that we further validate with an orthogonal semi-high throughput MALDI-MS assay for direct monitoring of substrate turnover. Structure-activity relationship (SAR) studies involving >20 derivatives of this scaffold led to the identification of optimized inhibitors, and together with computational studies suggested a plausible model for its mode of action.

Evolved DNA Duplex Readers for Strand-Asymmetrically Modified 5-Hydroxymethylcytosine/5-Methylcytosine CpG Dyads

5-Methylcytosine (mC) and 5-hydroxymethylcytosine (hmC), the two main epigenetic modifications of mammalian DNA, exist in symmetric and asymmetric combinations in the two strands of CpG dyads. However, revealing such combinations in single DNA duplexes is a significant challenge. Here, we evolve methyl-CpG-binding domains (MBDs) derived from MeCP2 by bacterial cell surface display, resulting in the first affinity probes for hmC/mC CpGs. One mutant has low nanomolar affinity for a single hmC/mC CpG, discriminates against all 14 other modified CpG dyads, and rivals the selectivity of wild-type MeCP2. Structural studies indicate that this protein has a conserved scaffold and recognizes hmC and mC with two dedicated sets of residues. The mutant allows us to selectively address and enrich hmC/mC-containing DNA fragments from genomic DNA backgrounds. We anticipate that this novel probe will be a versatile tool to unravel the function of hmC/mC marks in diverse aspects of chromatin biology.

Immunoglobulin Classification Based on FC* and GC* Features

Immunoglobulins have a pivotal role in disease regulation. Therefore, it is vital to accurately identify immunoglobulins to develop new drugs and research related diseases. Compared with utilizing high-dimension features to identify immunoglobulins, this research aimed to examine a method to classify immunoglobulins and non-immunoglobulins using two features, FC* and GC*. Classification of 228 samples (109 immunoglobulin samples and 119 non-immunoglobulin samples) revealed that the overall accuracy was 80.7% in 10-fold cross-validation using the J48 classifier implemented in Weka software. The FC* feature identified in this study was found in the immunoglobulin subtype domain, which demonstrated that this extracted feature could represent functional and structural properties of immunoglobulins for forecasting.

Intragenomic Decarboxylation of 5-Carboxy-2'-deoxycytidine

Cellular DNA is composed of four canonical nucleosides (dA, dC, dG and T), which form two Watson-Crick base pairs. In addition, 5-methylcytosine (mdC) may be present. The methylation of dC to mdC is known to regulate transcriptional activity. Next to these five nucleosides, the genome, particularly of stem cells, contains three additional dC derivatives, which are formed by stepwise oxidation of the methyl group of mdC with the help of Tet enzymes. These are 5-hydroxymethyl-dC (hmdC), 5-formyl-dC (fdC), and 5-carboxy-dC (cadC). It is believed that fdC and cadC are converted back into dC, which establishes an epigenetic control cycle that starts with methylation of dC to mdC, followed by oxidation and removal of fdC and cadC. While fdC was shown to undergo intragenomic deformylation to give dC directly, a similar decarboxylation of cadC was postulated but not yet observed on the genomic level. By using metabolic labelling, we show here that cadC decarboxylates in several cell types, which confirms that both fdC and cadC are nucleosides that are directly converted back to dC within the genome by C-C bond cleavage.

iDNA-ABT : advanced deep learning model for detecting DNA methylation with adaptive features and transductive information maximization

MOTIVATION

DNA methylation plays an important role in epigenetic modification, the occurrence, and the development of diseases. Therefore, the identification of DNA methylation sites is critical for better understanding and revealing their functional mechanisms. To date, several machine learning and deep learning methods have been developed for the prediction of different methylation types. However, they still highly rely on manual features, which can largely limit the high-latent information extraction. Moreover, most of them are designed for one specific methylation type, and therefore cannot predict multiple methylation sites in multiple species simultaneously. In this study, we propose iDNA-ABT, an advanced deep learning model that utilizes adaptive embedding based on bidirectional transformers for language understanding together with a novel transductive information maximization (TIM) loss.

RESULTS

Benchmark results show that our proposed iDNA-ABT can automatically and adaptively learn the distinguishing features of biological sequences from multiple species, and thus perform significantly better than the state-of-the-art methods in predicting three different DNA methylation. In addition, TIM loss is proven to be effective in dichotomous tasks via the comparison experiment. Furthermore, we verify that our features have strong adaptability and robustness to different species through comparison of adaptive embedding and six handcrafted feature encodings. Importantly, our model shows great generalization ability in different species, demonstrating that our model can adaptively capture the cross-species differences and improve the predictive performance. For the convenient use of our method, we further established an online webserver as the implementation of the proposed iDNA-ABT.

AVAILABILITY

our proposed iDNA-ABT, which is now freely accessible via http://server.wei-group.net/iDNA_ABT and our source codes are available in the GitHub repository (https://github.com/YUYING07/iDNA_ABT).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Human MettL3-MettL14 RNA adenine methyltransferase complex is active on double-stranded DNA containing lesions

MettL3-MettL14 methyltransferase complex has been studied widely for its role in RNA adenine methylation. This complex is also recruited to UV- and X-ray exposed DNA damaged sites, and its methyltransfer activity is required for subsequent DNA repair, though in theory this could result from RNA methylation of short transcripts made at the site of damage. We report here that MettL3-MettL14 is active in vitro on double-stranded DNA containing a cyclopyrimidine dimer – a major lesion of UV radiation-induced products – or an abasic site or mismatches. Furthermore, N6-methyladenine (N6mA) decreases misincorporation of 8-oxo-guanine (8-oxoG) opposite to N6mA by repair DNA polymerases. When 8-oxoG is nevertheless incorporated opposite N6mA, the methylation inhibits N6mA excision from the template (correct) strand by the adenine DNA glycosylase (MYH), implying that the methylation decreases inappropriate misrepair. Finally, we observed that the N6mA reader domain of YTHDC1, which is also recruited to sites of DNA damage, binds N6mA that is located across from a single-base gap between two canonical DNA helices. This YTHDC1 complex with a gapped duplex is structurally similar to DNA complexes with FEN1 and GEN1 – two members of the nuclease family that act in nucleotide excision repair, mismatch repair and homologous recombination, and which incise distinct non-B DNA structures. Together, the parts of our study provide a plausible mechanism for N6mA writer and reader proteins acting directly on lesion-containing DNA, and suggest in vivo experiments to test the mechanisms involving methylation of adenine.