Abstract DNA modifications, represented by 5‐methylcytosine (5mC), 5‐hydroxymethylcytosine (5hmC), 5‐formylcytosine (5fC), and 5‐carboxylcytosine (5caC), play important roles in epigenetic regulation of biological processes. The specific recognition of DNA modifications by the transcriptional protein machinery is thought to be a potential mechanism for epigenetic‐driven gene regulation, and many modified DNA‐specific binding proteins have been uncovered. However, the panoramic view of the roles of DNA modification readers at the proteome level remains largely unclear. Here, a recently developed concatenated tandem array of consensus transcription factor (TF) response elements (catTFREs) approach is employed to profile the binding activity of TFs at DNA modifications. Modified DNA‐binding activity is quantified for 1039 TFs, representing 70% of the TFs in the human genome. Additionally, the modified DNA‐binding activity of 600 TFs is monitored during the mouse brain development from the embryo to the adult stages. Readers of these DNA modifications are predicted, and the hierarchical networks between the transcriptional protein machinery and modified DNA are described. It is further demonstrated that ZNF24 and ZSCAN21 are potential readers of 5fC‐modified DNA. This study provides a landscape of TF–DNA modification interactions that can be used to elucidate the epigenetic‐related transcriptional regulation mechanisms under physiological conditions. Keywords: DNA modification, epigenetic regulation, mass spectrometry, proteomics __________________________________________________________________ Modified‐transcription factor response elements (TFREs) are developed, which profile the binding activity of TFs at DNA modifications (5mC, 5hmC, and 5fC), covering 70% of the TFs in the genome. A landscape of TF–DNA modification interaction is provided and the epigenetic‐related transcriptional regulation mechanisms are elucidated under physiological conditions. Potential readers of 5fC‐modified DNA are demonstrated and validated. graphic file with name ADVS-8-2101426-g002.jpg 1. Introduction Epigenetic modification histones and DNA play critical roles in regulating gene expression during development, differentiation, diseases, and other physiological and pathological processes.^[ [50]^1 ^] Methylation at the fifth position of cytosine (5mC) is the predominant epigenetic DNA modification and is highly related to the embryogenesis, development, aging, and carcinogenesis.^[ [51]^2 ^] The 5mC is an epigenetic marker linked to gene silencing. It is reported that the DNA methylation‐related gene downregulation results from specific interactions between transcription factors (TFs) and methylated DNA.^[ [52]^1a ^] Only proteins with a methyl‐CpG binding domain (MBD) can interact with methylated DNA while the majority of TFs cannot.^[ [53]^3 ^] Recently, evidence of interactions between methylated DNA and TFs has emerged, such as CEBP,^[ [54]^4 ^] ZFHX3,^[ [55]^5 ^] RFX1,^[ [56]^6 ^] SIX4, AKSCAN3, and FOXK2,^[ [57]^7 ^] however, the information of the MBDs still remained unclear. DNA methylation patterns are established and maintained by DNA methyltransferases (DNMTs).^[ [58]^8 ^] In 2009, the ten‐eleven translocation (TET) family dioxygenase was discovered to recognize and oxidize 5mC into 5‐hydroxymethylcytosine (5hmC).^[ [59]^9 ^] TET can successively oxidize 5hmC into 5‐formylcytosine (5fC) and 5‐carboxylcytosine (5caC).^[ [60]^10 ^] The oxidized derivatives 5fC and 5caC are recognized and excised by the mammalian thymine DNA glycosylase (TDG) and are subsequently converted to cytosine, contributing to DNA demethylation and gene regulation.^[ [61]^11 ^] Recent technological advances have made it possible to decode DNA methylomes and even hydroxymethylomes at single‐base pair resolution under various physiological conditions.^[ [62]^12 ^] DNA modifications can be screened at the genome level to assess gene expression patterns and their potential mechanisms.^[ [63]^13 ^] For example, while 5mC is a gene silencing epigenetic marker, hydroxylation of 5mC to 5hmC might retrieve transcription through dissociation of 5mC‐binding proteins and/or recruitment of effector proteins.^[ [64]^14 ^] Affinity enrichment‐based methods and modified bisulfite sequencing (BS‐seq) studies^[ [65]^15 ^] on pluripotent stem cells and differentiated tissues have indicated that 5hmC is obviously abundant in embryonic stem (ES) cells and neuronal Purkinje cells, and is enriched in highly transcribed gene bodies.^[ [66]^16 ^] With regard to the oxidized derivatives 5fC and 5caC, due to their low percentages in the genome, rare evidence of their functions on gene expression regulations has been reported. Whether 5fC and 5caC have additional DNA demethylation‐independent functions in gene regulation is not very well studied at present. TFs, which form the second largest gene family in the genome, play critical roles in controlling gene expression patterns in almost all biological processes, including differentiation, development, cell cycle, and cell death. Approximately 1500 TF‐coding genes are annotated in the human genome.^[ [67]^17 ^] TFs recognized the genome sequences of the downstream genes, and the subsequently constructed the TF/transcriptional coregulator (TC)–DNA machinery. These progresses are considered the major mechanisms of gene expression regulations, in which the specific recognition of target genes (TGs) by TFs and the interactions of TF readers with these target genes are the initial steps.^[ [68]^18 ^] Identification of DNA modification readers, which translate modification signals into biological actions, will be crucial for deciphering the epigenetic codes of DNA modification‐mediated biological processes. In addition to the specific binding of MBD to 5mC, the recent reports have revealed that MBD3^[ [69]^19 ^] and MECP2^[ [70]^20 ^] are able to bind to 5hmC. Furthermore, Spruijt et al. have demonstrated the specific binding of Klf4 to 5mC, and Uhrf2 to 5hmC.^[ [71]^14 ^] Iurlaro et al. have indicated the specific binding of RPL26 and PRP8 to 5mC and some members of the FOX family to 5fC.^[ [72]^7 ^] These researchers aimed to screen the potential TF “readers” of DNA modifications, especially for the 5hmC and 5fC, in order to reveal the mechanisms of the DNA modification‐related gene regulation. Although a few studies have provided many TF candidates that can interact with modified DNA in humans and mouse,^[ [73]^4 , [74]^21 ^] however, a panoramic view of the modified DNA binding activity of thousands of TFs and TCs remains elusive. We previously developed an approach that enables the identification and quantification of the DNA‐binding activity of endogenous TFs at the proteome scale.^[ [75]^22 ^] With synthetic DNA containing a concatenated tandem‐array of the consensus TF response elements (catTFRE) as an affinity reagent, the TF–DNA interactions of almost all expressed TFs can be surveyed in cell lines and tissues. In this study, based on a catTFRE sequence, we prepared the 5mC‐TFRE, 5hmC‐TFRE, and 5fC‐TFRE by replacing the regular cytosine with the 5mC, 5hmC, and 5fC during the PCR amplification. The cell lines of HeLa, HepG2, A549, and MCF‐7 were used to profile the binding activity of the TFs on different DNA modifications. The binding activity on four types of DNA (5C, 5mC, 5hmC, and 5fC) was assessed for 1039 TFs, covering 70% of gene‐coding TFs, which produced the most comprehensive dataset on TF‐modified DNA interactions. Based on this dataset, the specific modified‐DNA binding TFs were identified, and the protein machineries that recognized to different DNA modifications were constructed. The different types of DNA‐binding domains (DBDs) of TFs showed diverse priorities in interacting with modified DNA. We further employed the strategy of TF‐modified DNA screening to dissect the landscape of the TF–DNA modification interactome during mouse brain development from the embryo stage to the adult stage. Finally, we validated the candidates of 5fC‐DNA‐binding TF readers, ZNF24 and ZSCAN21, and examined the relationship between epigenetics and transcriptional regulation. This study provides a rich resource that will facilitate scientists in comprehensively accessing DNA modification “readers” and elucidating the mechanisms of the gene regulation connected to the epigenetics.^[ [76]^23 ^] 2. Results 2.1. Workflow for Preparing Different DNA‐Modification Types and the Proteome‐Wide Screening of TF‐Modified DNA Binding Activity Based on our previously developed catTFRE strategy,^[ [77]^22 ^] we further reformed the TFRE sequences to the 5mC‐TFRE, 5hmC‐TFRE, and 5fC‐TFRE by replacing the 5C with the 5mC, 5hmC, and 5fC, respectively, during PCR amplification (Figure [78] 1A). Consistent with 5C‐TFRE, 5mC‐TFRE, 5hmC‐TFRE, and 5fC‐TFRE have DNA lengths of 2.8 kb. To profile the modified DNA‐binding activity of TFs in humans, we made nuclear extracts (NEs) of four cell lines: HeLa, HepG2, A549, and MCF‐7. For mouse brains, the NEs were prepared at different time points during development, from embryonic day (E) E14.5 to the six weeks of age (Figure [79]1B). For each cell line and organ, the NE aliquots were incubated with the four types of modified TFREs (Figure [80]1C), and the DNA‐binding proteins were submitted to the mass spectrometry (MS) for proteome detection (Figure [81]1D). The proteome profiling of the four cell types were also performed for the references of the input.