Developmental arsenic exposure impairs cognition, directly targets DNMT3A, and reduces DNA methylation

Article4 April 2022free access Source DataTransparent process Developmental arsenic exposure impairs cognition, directly targets DNMT3A, and reduces DNA methylation Ni Yan Ni Yan orcid.org/0000-0001-5775-5488 Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China Contribution: Data curation, Software, Formal analysis, Validation, ​Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review & editing Search for more papers by this author Yuntong Li Yuntong Li orcid.org/0000-0003-2164-0069 Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China Contribution: Software, Formal analysis, ​Investigation, Methodology, Writing - review & editing Search for more papers by this author Yangfei Xing Yangfei Xing Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Software, Formal analysis, ​Investigation, Writing - review & editing Search for more papers by this author Jiale Wu Jiale Wu Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Formal analysis, ​Investigation, Methodology Search for more papers by this author Jiabing Li Jiabing Li Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Software, Methodology Search for more papers by this author Ying Liang Ying Liang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Yigang Tang Yigang Tang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Zhengyuan Wang Zhengyuan Wang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Huaxin Song Huaxin Song Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: ​Investigation Search for more papers by this author Haoyu Wang Haoyu Wang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Shujun Xiao Shujun Xiao Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China Contribution: ​Investigation Search for more papers by this author Min Lu Corresponding Author Min Lu [email protected] orcid.org/0000-0001-6902-3751 Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Visualization, Methodology, Writing - original draft, Project administration, Writing - review & editing Search for more papers by this author Ni Yan Ni Yan orcid.org/0000-0001-5775-5488 Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China Contribution: Data curation, Software, Formal analysis, Validation, ​Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review & editing Search for more papers by this author Yuntong Li Yuntong Li orcid.org/0000-0003-2164-0069 Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China Contribution: Software, Formal analysis, ​Investigation, Methodology, Writing - review & editing Search for more papers by this author Yangfei Xing Yangfei Xing Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Software, Formal analysis, ​Investigation, Writing - review & editing Search for more papers by this author Jiale Wu Jiale Wu Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Formal analysis, ​Investigation, Methodology Search for more papers by this author Jiabing Li Jiabing Li Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Software, Methodology Search for more papers by this author Ying Liang Ying Liang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Yigang Tang Yigang Tang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Zhengyuan Wang Zhengyuan Wang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Huaxin Song Huaxin Song Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: ​Investigation Search for more papers by this author Haoyu Wang Haoyu Wang Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Methodology Search for more papers by this author Shujun Xiao Shujun Xiao Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China Contribution: ​Investigation Search for more papers by this author Min Lu Corresponding Author Min Lu [email protected] orcid.org/0000-0001-6902-3751 Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Contribution: Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Visualization, Methodology, Writing - original draft, Project administration, Writing - review & editing Search for more papers by this author Author Information Ni Yan1,2,†, Yuntong Li1,3,†, Yangfei Xing1, Jiale Wu1, Jiabing Li1, Ying Liang1, Yigang Tang1, Zhengyuan Wang1, Huaxin Song1, Haoyu Wang1, Shujun Xiao1,2 and Min Lu *,1 1Shanghai Institute of Hematology, State Key Laboratory of Medical Genomics, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China 2School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China 3Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China † These authors contributed equally to this work *Corresponding author. Tel: +86 021 64370045 610805; E-mail: [email protected] EMBO Reports (2022)23:e54147https://doi.org/10.15252/embr.202154147 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract Developmental arsenic exposure has been associated with cognitive deficits in epidemiological studies, but the underlying mechanisms remain poorly understood. Here, we establish a mouse model of developmental arsenic exposure exhibiting deficits of recognition and spatial memory in the offspring. These deficits are associated with genome-wide DNA hypomethylation and abnormal expression of cognition-related genes in the hippocampus. Arsenic atoms directly bind to the cysteine-rich ADD domain of DNA methyltransferase 3A (DNMT3A), triggering ubiquitin- and proteasome-mediated degradation of DNMT3A in different cellular contexts. DNMT3A degradation leads to genome-wide DNA hypomethylation in mouse embryonic fibroblasts but not in non-embryonic cell lines. Treatment with metformin, a first-line antidiabetic agent reported to increase DNA methylation, ameliorates the behavioral deficits and normalizes the aberrant expression of cognition-related genes and DNA methylation in the hippocampus of arsenic-exposed offspring. Our study establishes a DNA hypomethylation effect of developmental arsenic exposure and proposes a potential treatment against cognitive deficits in the offspring of pregnant women in arsenic-contaminated areas. SYNOPSIS Arsenic pollution has been linked to cognitive deficits. This study shows that arsenic binds the DNMT3A protein and disturbs cognition-related gene expression in mouse brains, which can be rescued by metformin treatment. Developmental arsenic exposure induces genome-wide DNA hypomethylation in the mouse hippocampus. Arsenic directly binds to DNMT3A and induces DNMT3A degradation through the ubiquitin-proteasome pathway. Arsenic treatment leads to genome-wide DNA hypomethylation in mouse embryonic fibroblasts but not in non-embryonic cell lines. Metformin treatment ameliorates the arsenic-induced cognitive deficits and normalizes the arsenic-induced hippocampal DNA hypomethylation. Introduction Arsenic is a widespread environmental contaminant that threatens over 200 million people worldwide due to exposure through the drinking water (Nordstrom, 2002; Naujokas et al, 2013; Podgorski & Berg, 2020). Numerous epidemiological studies have shown that arsenic readily passes the placenta of pregnant women and potentially leads to cognitive impairment in the offspring (Grandjean & Landrigan, 2014; Freire et al, 2018; Vahter et al, 2020). For instance, a prospective cohort study of 1,523 children in Bangladesh reported that arsenic exposure during early pregnancy was inversely associated with children's cognitive functions (Vahter et al, 2020). How developmental arsenic exposure leads to cognitive deficits in the offspring is a long-standing question, and one possible mechanism is the dysregulation of DNA methylation. As an epigenetic process catalyzed by DNA methyltransferases (DNMTs) (Jaenisch & Bird, 2003), DNA methylation has been well established to play an essential role in nervous system development and cognitive functions (Miller & Sweatt, 2007; Day & Sweatt, 2010; Moore et al, 2013). For example, inhibition of DNMT activity in the hippocampus, a brain region associated primarily with cognitive processes (Eichenbaum, 2004), was reported to disturb the expression of memory-related genes and block memory formation (Miller & Sweatt, 2007). There is epidemiological evidence for a relationship between arsenic exposure and dysregulation of DNA methylation, even though the role of arsenic in regulating DNA methylation remains inconclusive (Brocato & Costa, 2013). Some studies have linked arsenic exposure to DNA hypomethylation (Intarasunanont et al, 2012; Green et al, 2016; Guo et al, 2018), such as a cohort study of 343 mother–infant pairs, which related placenta arsenic levels to gene-specific hypomethylation (Green et al, 2016). However, it was also demonstrated in other studies that arsenic exposure induces DNA hypermethylation (Pilsner et al, 2007; Niedzwiecki et al, 2013; Ameer et al, 2017). To date, the precise mechanisms by which arsenic causes DNA hypo- or hypermethylation remain elusive, although several mechanisms have been proposed, including the arsenic-induced transcriptional dysregulation of DNMTs (Zhao et al, 1997; Benbrahim-Tallaa et al, 2005; Reichard et al, 2007; Majumdar et al, 2010; Ma et al, 2020; Zhou et al, 2020). These inconsistent reports leave open the question whether and how arsenic exposure affects DNA methylation, leading to cognitive deficits. Here, we established a mouse model of developmental arsenic exposure with cognitive deficits in the offspring. We further identified the roles of arsenic-mediated DNMT3A degradation and hippocampal hypomethylation in this model. Finally, we found that metformin could effectively prevent the arsenic-induced cognitive deficits in the mouse model, offering a potential preventive strategy for pregnant women in arsenic-contaminated areas. Results Developmental arsenic exposure leads to cognitive deficits and hippocampal DNA hypomethylation in mice To investigate the effects of developmental arsenic exposure to the cognition of offspring and elucidate its possible mechanism(s), we first established a mouse model of developmental arsenic exposure with cognitive deficits in the offspring as previously described (Martínez et al, 2011). Specifically, parental C57BL/6 mice were continually exposed to 100 μg/l (ppb) or 5,000 ppb arsenic trioxide (ATO) in drinking water from the pre-mating period to the lactation period, and the offspring mice continued receiving the ATO treatments in drinking water after weaning until the end of the experiment. It is worth noting that 100 or 5,000 ppb ATO in drinking water for a C57BL/6 mouse is equivalent to 7.6 or 379 ppb arsenic for a 60-kg human being, respectively (Reagan-Shaw et al, 2008). These two arsenic concentrations are comparable to the permissible limit suggested by the World Health Organization (10 ppb) (World Health Organization, 1993), and the concentration found in many arsenic-contaminated areas worldwide (Smedley & Kinniburgh, 2002; Rodríguez-Lado et al, 2013), respectively. Arsenic exposure at these concentrations did not significantly change the water intake or body weight gain in maternal or offspring mice during the treatment period (Fig EV1A). When the pups grew to 9–10 weeks of age, an open-field test (OFT) was performed to assess their spontaneous locomotor activity and anxiety-like behavior. Arsenic-exposed mice displayed normal locomotor activity, as demonstrated by the total distance traveled during the 10-min test (Fig EV1B, left panel). Based on the total distance moved and total time spent in the center zone, the arsenic-exposed mice also showed normal anxiety-like behavior (Fig EV1B, middle and right panels). Next, we examined potential cognitive deficits in each group using the novel object recognition test (NORT; Fig 1A). At the training stage, all groups explored the two identical objects (F) with similar preferences. In the testing stage 1 h later, both the 100- and 5,000-ppb ATO groups spent less time exploring the novel object (N) than the control group (Fig 1A; F(2,39) = 4.848; Ctr vs. 100 ppb ATO: P = 0.016; Ctr vs. 5,000 ppb ATO: P = 0.006), indicating the successful establishment of the mouse model of developmental arsenic exposure with cognitive deficits in the offspring. Click here to expand this figure. Figure EV1. Developmental arsenic exposure leads to cognitive deficits and hippocampal genome-wide DNA hypomethylation in mice Maternal daily water intake, maternal growth curve (n = 6 mice per group), offspring's daily water intake, and offspring's growth curve (Ctr: n = 6; 100 ppb ATO: n = 12; 5,000 ppb ATO: n = 14 mice) were recorded during experiment and are shown. Drinking behavior was measured by the average water intake per mouse in a group from one cage. Data are presented as mean for water intake curve and as mean ± SEM for growth curve, and further analyzed by one-way ANOVA followed by Bonferroni's post hoc test. OFT. Total distance traveled in 10-min test, distance moved in the center zone, and time spent in the center zone of each group. Data are presented as mean ± SEM. ns, not significant by one-way ANOVA followed by Bonferroni's post hoc test. MeDIP assay using 5meC or IgG antibody. The signals were determined by qPCR using IAP (5meC-positive) and CSa (5meC-negative) sequences. Data are presented as mean ± SD of n = 3 technical replicates. ND, not detected; ****P < 0.0001 by Student's t-test. Primer targets for MeDIP-qPCR assay. Annotated lines below genes illustrate loci for CpG island and primer-amplified regions. Relative CpG methylation levels [log2(fold change)] at CGI, CGI shore, promoter [transcription start site (TSS) ± 1,000 bp], exon, intron, and intergenic region in the hippocampal DNA samples (ATO vs. Ctr), as determined by MethylC-seq. Genomic distribution of DMRs with respect to the gene model. GO enrichment analysis for the DMRs based on molecular function and cellular compartment GO categories. Green dot indicates gene number in each term. Profiles showing the top three negatively enriched terms in GSEA. NES, normalized enrichment score. GO enrichment analysis for the 153 DEGs in the ATO group based on molecular function and cellular compartment GO categories. Green dot indicates DEG number in each term. Download figure Download PowerPoint Figure 1. Developmental arsenic exposure leads to cognitive deficits and hippocampal genome-wide DNA hypomethylation in mice A. Test procedure of NORT (left panel) and the bar graph showing the exploratory preference for novel object of offspring mice (right panel). F, familiar object; N, novel object. B, C. 5meC enrichment analysis of the IAP (B), and Xist and H19 loci (C) in the hippocampus of offspring mice from each group, as measured by MeDIP-qPCR. Each of the biological replicates was run in three technical replicates, and means were shown as cycles in the graph. D. Relative CpG methylation level [log2(fold change)] at each chromosome in the hippocampal DNA samples (ATO vs. Ctr), as determined by MethylC-seq. E. Number of hypo- (lower methylation in the ATO group) and hyper- (higher methylation level in the ATO group) differentially methylated tiles (1-kb), as determined by MethylC-seq. F. GO enrichment analysis for the DMRs based on the biological process GO categories. Green dot indicates gene number in each term. G. Volcano plot showing DEGs in the offspring's hippocampus upon 5,000-ppb ATO exposure (n = 2 mice for both control and ATO groups). Dash lines indicate fold change ≥ 2 or ≤ 0.5 and adjusted P value < 0.05. Genes to be further validated for expression levels are labeled. H. GO enrichment analysis for the 153 DEGs in the ATO group based on the biological process GO category. Green dot indicates DEG number in each term. I. PPI network for the 153 DEGs based on STRING database. Node colors indicate fold changes; node sizes indicate link numbers. J. Heatmap showing expression levels (Z-score) of the 12 cognition-related DEGs with node size ≥ 6 in the PPI network. K. RT-qPCR validation of the 12 selected cognition-related DEGs. Data information: Data are presented as mean ± SEM. ns, not significant; *P < 0.05; **P < 0.01; and ***P < 0.001. For (A), data are analyzed by two-way repeated-measures ANOVA followed by Bonferroni's post hoc test. For (B), (C), and (K), data are analyzed by one-way ANOVA followed by Bonferroni's post hoc test. Download figure Download PowerPoint Given that DNA methylation plays an indispensable role in cognitive functions (Miller & Sweatt, 2007; Day & Sweatt, 2010; Moore et al, 2013), we next detected whether developmental arsenic exposure alters DNA methylation in the offspring's hippocampus. The DNA methylation level was measured by methylated DNA immunoprecipitation (MeDIP), a technique that isolates methylated DNA fragments using an antibody specific for 5′-methylcytosine (5meC; Fig EV1C). Upon exposure to 100 ppb ATO, the enrichment of methylated DNA revealed a mild, but not significant, decrease of the methylation level of intracisternal A-particle (IAP) repetitive sequences (Fig EV1D), which are representative of global DNA methylation (Waterston et al, 2002; Borowczyk et al, 2009; Wang et al, 2009) (Fig 1B; P = 0.452). Notably, exposure to 5,000 ppb ATO induced significant global DNA hypomethylation (Fig 1B; P = 0.024). We further tested the methylation levels of two specific genes, Xist and H19 (Fig EV1D), which are respectively related to X chromosome inactivation and genomic imprinting, and are both used as model genes in epigenetic studies (Fedoriw et al, 2012). Upon exposure to 100 ppb ATO, a significant reduction in DNA methylation was observed on Xist (Fig 1C; P = 0.029), but not on H19 (Fig 1C; P = 0.106). Consistently, more dramatic decreases were observed in the methylation levels of both genes following exposure to 5,000 ppb ATO (Fig 1C; Xist: P = 0.008, H19: P < 0.001). We further assessed the hippocampal DNA methylation profiles of the Ctr and 5,000-ppb ATO groups by target-captured MethylC-seq. The methylation status of 2.3 million CpG sites with ≥ 10× coverage in both samples was used for the initial comparative analysis. As shown in Fig 1D, the average CpG site methylation on all chromosomes was lower in the ATO group. Extending the analysis to various genomic features, we found that arsenic exposure decreased DNA methylation levels on all elements, including CGI (CpG island), CGI shore, promoter, exon, intron, and intergenic region (Fig EV1E). When dividing the genome into 1-kb tiles, we identified 9,173 differentially methylated tiles with significant differences between the Ctr and ATO groups (Dataset EV1; methylation difference < −10 or > 10 and adjusted P-value < 0.05), with more than 70% (6,452) of the tiles being hypomethylated in the ATO group (Fig 1E). The majority (62.55%) of the differentially methylated regions (DMRs) was associated with gene regions, with 8.47% localizing to promoters and 54.08% to gene bodies (Fig EV1F and Dataset EV2). In addition, gene ontology (GO) analyses suggested that the DMR-related genes were significantly associated with cognition-related terms such as axon development (Figs 1F and EV1G). After determining that developmental arsenic exposure induced hippocampal DNA hypomethylation, we next explored the hippocampal transcriptome of mice exposed to 5,000 ppb ATO by RNA-seq. Gene set enrichment analysis (GSEA) revealed that “Transmitter gated channel activity”, “Neurotransmitter receptor activity involved in regulation of postsynaptic membrane potential” and “Positive regulation of synapse assembly” were the most significantly enriched gene sets that were downregulated in the 5,000-ppb ATO group (Fig EV1H; all P = 0.000), suggesting that developmental arsenic exposure could have deleterious effects on neuronal synapses and subsequent cognitive functions. We next focused on the 153 differentially expressed genes (DEGs; Dataset EV3; fold change ≥ 2 or ≤ 0.5 and adjusted P value < 0.05) between the control group and 5,000-ppb ATO group (Fig 1G). GO enrichment analysis of the 153 DEGs revealed that GO terms related to the extracellular matrix (ECM) were the most enriched terms in the biological process, molecular function, and cellular component GO categories (Figs 1H and EV1I). The protein–protein interaction (PPI) network of the 153 DEGs revealed a protein network consisting of abundant ECM molecules, such as collagens (Fig 1I; Col9a3, Col8a1, Col8a2, Col18a1, Col4a4, and Col4a3). ECM molecules are closely related to normal cognition by regulating neuronal and synapse development and function (Lau et al, 2013; Végh et al, 2014; Song & Dityatev, 2018). Thus, DNA hypomethylation induced by developmental arsenic exposure might cause cognitive deficits by dysregulating these ECM molecules. To confirm the RNA-seq results, we selected 12 DEGs with node size ≥ 6 in the PPI network (Fig 1J) for validation by RT–qPCR. Based on the GO term annotations, most of the 12 selected genes (Ttr, Col9a3, Col8a1, Col8a2, Serpinh1, Col18a1, Col4a4, and Col4a3) were related to ECM. In the validation experiments, 10 of the 12 selected genes showed significant changes at mRNA levels with the same trend as in the RNA-seq analysis (Fig 1K). The mRNA levels of Col18a1 and Igf2 were also upregulated, but the difference did not reach statistical significance. In summary, developmental arsenic exposure leads to cognitive deficits in the offspring, which is associated with genome-wide DNA hypomethylation and dysregulated expression of cognition-related genes in the hippocampus of the offspring. Arsenic directly binds to DNMT3A To identify the cellular target through which arsenic induces DNA hypomethylation, we determined the arsenic interactome in mouse embryonic fibroblast cell line NIH3T3. The cells were treated with biotinylated arsenic (Biotin-As) (Zhang et al, 2010) for 2 h, and the arsenic-interacting proteins in the cell lysate were then pulled down with streptavidin agarose beads, followed by determination using liquid chromatography–electrospray ionization–tandem mass spectrometry (LC-ESI-MS/MS). We identified a total of 1,798 arsenic-binding candidate proteins with scores ≥ 10.0 (Dataset EV4). Interestingly, DNA methyltransferase 3a (Dnmt3a, the mouse homolog of human DNMT3A), an enzyme that catalyzes the transfer of methyl groups to cytosine, was among the 1,798 candidates (Fig 2A). We did not identify the other two key DNA methyltransferases, Dnmt1 and Dnmt3b, nor the ten-eleven translocation (Tet) enzymes that are well known to participate in DNA methylation among the candidate proteins in the interactome. To confirm the suggested interaction between arsenic and Dnmt3a, NIH3T3 cells were incubated with 5 μg/ml Biotin-As for 2 or 16 h, followed by a streptavidin agarose pull-down assay and immunoblotting for Dnmt3a. As shown in Fig 2B, arsenic detectably binds to Dnmt3a after 2-h Biotin-As incubation. Interestingly, longer Biotin-As incubation caused a significant decrease of the Dnmt3a level in the input sample. Similar results were observed in C57BL/6 mouse embryonic fibroblasts (MEFs; Fig 2C) and human CCRF-CEM cancer cells (Fig 2D). Figure 2. Arsenic directly binds to DNMT3A A. LC-ESI-MS/MS spectrum of Dnmt3a pulled down by Biotin-As in NIH3T3 cells. NIH3T3 cells were treated with 10 μg/ml Biotin-As for 2 h, followed by pull-down assay and determination using LC-ESI-MS/MS. B–D. Validation of the binding between DNMT3A and Biotin-As at 5 μg/ml in NIH3T3 (B), MEFs (C) and CCRF-CEM (D) cells in the Biotin-As pull-down assay. E. Schematic diagram of the structure of human DNMT3A. Upper panel: All cysteines in DNMT3A are shown. Lower panel: Structure of DNMT3A-ADD (PDB: 3A1A) and the four interested cysteine triads. F. Measurement of DNMT3A-ADD thermodynamic stability upon ATO treatment. Purified DNMT3A-ADD was mixed with ATO at the indicated ratios in 20 mM HEPES (pH 7.5) for 1 h. Melting curves were measured by DSF (left panel), and the calculate