Diet has been shown to influence epigenetic key players, such as DNA methylation, which can regulate the gene expression potential in both parents and offspring. Diets enriched in omega-6 and deficient in omega-3 PUFAs (low dietary omega-3/omega-6 PUFA ratio), have been associated with the promotion of pathogenesis of diseases in humans and other mammals. In this study, we investigated the impact of increased dietary intake of arachidonic acid (ARA), a physiologically important omega-6 PUFA, on 2 generations of zebrafish. Parental fish were fed either a low or a high ARA diet, while the progeny of both groups were fed the low ARA diet. We screened for DNA methylation on single base-pair resolution using reduced representation bisulfite sequencing (RRBS). The DNA methylation profiling revealed significant differences between the dietary groups in both parents and offspring. The majority of differentially methylated loci associated with high dietary ARA were found in introns and intergenic regions for both generations. Common loci between the identified differentially methylated loci in F0 and F1 livers were reported. We described overlapping gene annotations of identified methylation changes with differential expression, but based on a small number of overlaps. The present study describes the diet-associated methylation profiles across genomic regions, and it demonstrates that parental high dietary ARA modulates DNA methylation patterns in zebrafish liver.
Citation: Adam A-C, Lie KK, Whatmore P, Jakt LM, Moren M, Skjærven KH (2019) Profiling DNA methylation patterns of zebrafish liver associated with parental high dietary arachidonic acid. PLoS ONE 14(8): e0220934. https://doi.org/10.1371/journal.pone.0220934
Editor: Sriharsa Pradhan, New England Biolabs Inc., UNITED STATES
Received: February 20, 2019; Accepted: July 26, 2019; Published: August 9, 2019
Copyright: © 2019 Adam et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript, its Supporting Information files are held in a public database. RNA-sequencing data has been deposited in NCBI’s Gene Expression Omnibus and is accessible through GEO Series accession number GSE104692 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104692). DNA methylation data has been stored in SRA (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA418670) and is accessible through accession number PRJNA418670.
Funding: This research was financed through The Research Council of Norway (www.forskningsradet.no): EpiSip 228877 (KHS) and EpiFeedFish 225250/E40 (KHS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Methylation of the cytosine nucleotide, generally referred to as DNA methylation, is the most widely studied epigenetic mechanism. DNA methylation is crucial for regulating cell differentiation and development [1–3], and can be affected in response to environmental stimuli, such as nutrition [4–7]. Cytosine methylation plays a key role in transcriptional regulation, and can thereby influence physiological processes [8, 9]. Methylation primarily occurs in CpG sites where a cytosine is followed by a guanine. CpG islands are genomic regions highly enriched in CpG sequences, and often associated with promoter regions of a gene . CpG methylation located in promoters [11, 12], first exons , gene bodies [14–16], or enhancer elements  of the genome can influence the transcriptional activity of neighbouring genes, though the mechanisms by which methylation affects transcription in other genomic regions is still largely unknown. Additionally, epigenetic patterns can be transmitted to the next generation through the germline or the developing embryonic environment , but to what degree nutrients in the parental diet influence gene regulation in the next generation is not fully understood.
Nutrition during early life stages can have an impact on health and disease disposition later in life [19–22]. Both maternal and paternal diets have been shown to influence egg production, viability and gene expression of the embryos in different model species [23–25]. As DNA methylation may be mitotically stable and heritable through the germline [26–29], dietary influence or other environmental exposures can potentially be ‘remembered’ in the epigenome and transmitted to the next generation . It is conceivable that parental diet can influence gene expression potential in the progeny, especially under different environmental conditions, such as lifestyle and diet [31–33]. It has been shown in teleost species that dietary micronutrient levels given to the parents affected gene expression in embryo and fry [23, 33] as well as DNA methylation in livers of adult offspring .
High dietary intake of omega-6 (n-6) PUFAs, especially arachidonic acid (ARA, 20:4n-6) has evoked some concern due to their role in inflammatory processes and in relation to prevalence of certain diseases [34–37]. In fish, increased ARA in the diet has an impact on bone development [38, 39], growth and reproduction through improving hatching rate and accumulating ARA in ovary and eggs [40, 41]. PUFAs are capable of modulating DNA methylation patterns as shown in studies in mammals [42, 43], though studies in mice have observed no link between DNA methylation differences in adult progeny livers and the maternal high fat diet . However, the latter study did show a strong effect of the maternal diet on the hepatic expression of genes directing to inflammation, cholesterol synthesis and RXR activation . A study using human THP-1 monocytes revealed general DNA hypermethylation in response to ARA, and it was suggested that the changes in β-oxidation and PPAR-α were important mediators of ARA-induced DNA methylation changes . Another study on human vascular endothelial cells suggested that ARA metabolism was sensitive to changes in DNA methylation, but whether these methylation changes affected other enzymes regulating the ARA metabolism was not known . Although there are findings showing an effect of dietary fatty acids on methylation profiles, more studies are needed to investigate both mechanisms and how changes in DNA methylation can influence gene regulation in adult progeny.
Recently, we showed that parental high dietary ARA levels affected hepatic gene expression in the offspring, and levels of immune-related eicosanoids, lipids and oxidised metabolites in the first generation of zebrafish [47, 48]. In the present study, we used zebrafish as a nutritional model to investigate the effect of a parental high ARA diet on hepatic DNA methylation patterns in both parents and progeny using reduced representation bisulfite sequencing (RRBS). We compared our DNA methylation results with previously described transcriptomic profiles in the livers .
Materials and methods
Animal care and performance of the experimental trial conform to the principles of the Norwegian Animal Research Authority and the study was approved by the Norwegian Food Safety Authority (division no. 54, reference 2012/145126).
Zebrafish feeding trial
Study design, zebrafish husbandry and standardized operating procedures for mating, handling and feeding for both F0 and F1 generation of wildtype AB zebrafish (Danio rerio) has been described previously . In short, fish were kept in 10 gender mixed tanks (containing 60 fish each) per dietary group. F0 and F1 larvae were fed with Gemma micro (Skretting, Norway) and Artemia nauplii (Silver Star Artemia, USA) as start feed from 5 days post fertilization (DPF) and from 7 DPF until 26 DPF, respectively (Fig 1). The experimental diets were given twice a day from 27 DPF onwards. The F0 generation got either the control diet that was lower in ARA (1.87 mg ARA/g diet), or the high ARA diet (20.66 mg ARA/g diet). Diets were based on the requirement levels for carp . Ingredients and nutritional composition of the diets are given in S1 File. F0 fish were mated at 97 DPF to receive F1 progeny. The F1 generation from both parental dietary groups (Control and high ARA group) were fed the control diet from 27 DPF onwards until liver dissection for DNA and RNA extraction.
Both generations of zebrafish were fed Gemma micro and Artemia nauplii as start feed from 5 and 7 days post fertilization (DPF) until 26 DPF, respectively. The experimental diet (control and high ARA) was given from 27 DPF until sampling for F0 only. Adult F0 were mated at 97 DPF to generate F1. Both F1 groups were fed the control diet from 27 DPF. Body weight and liver sampling for reduced representation bisulfite sequencing were performed at day 98 for F0 while at 140–142 DPF for F1.
Liver tissue sampling
Mature zebrafish were deprived for food 18 h prior to sampling. They were anesthetized with 0.05% Tricaine Methane Sulphonate (PHARMAQ AS, Norway), blotted dry on tissue paper, weighed and euthanized by cutting the cardinal vein prior to liver dissection. Six single male livers (98 DPF (F0) and 140–142 DPF (F1)) from six tank populations (6n) of each dietary group were dissected, rinsed in 1x PBS, snap frozen with liquid nitrogen and stored at -80°C until DNA extraction. For RNA extraction, six pooled male livers from each of the six separate tanks from both groups were randomly sampled over two days between 154–156 DPF (F0) and 140–142 DPF (F1) due to simultaneous sampling for other analyses connected to this feeding trial.
DNA and RNA extraction
Liver tissue was treated with RNase A (50ng/μL, 10 min at room temperature, Wizard SV Genomic DNA Purification System, Promega, USA) and proteinase K (20μg/μL, 1.5h at 55°C, NEB #P8102S, New England Biolabs (NEB), USA) prior to DNA extraction following the manufacturer’s instructions (Wizard SV Genomic DNA Purification System). DNA was eluted in nuclease-free water, quantity was measured using Qubit fluorometric quantitation (Life Technologies, USA) and extracted DNA was stored at -20°C.
Total RNA was extracted using QIAzol Lysis Reagent (Qiagen, Germany) and DNase treated with the Ambion™ DNA-free™ DNA Removal Kit (Invitrogen, USA). RNA quantity was measured using NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, USA). RNA integrity (RIN), which was 9.06 ± 0.39 on average, was determined using an Agilent 2100 Bioanalyzer (RNA 6000 Nano LabChip kit, Agilent Technologies, USA).
RRBS and DNA methylation calling
Reduced representation bisulfite sequencing (RRBS) was used to measure DNA methylation with single-base resolution, and to enrich amplified genomic regions . RRBS and initial processing of the RRBS data was completed by the Biomedical Sequencing Facility (BSF, Vienna, Austria). RRBS library preparation was based on previous RRBS studies [51, 52]. Genomic DNA was digested by MspI (NEB #R0106L, 20 U/μL), followed by single-end preparation of the DNA fragments consisting of adapter ligation and A-tailing. Fragments were size-selected by performing a 0.75× cleanup with AMPure XP beads (A63881, Beckman Coulter, Inc, USA) for bisulfite conversion, enriched by PCR amplification and subsequent sequencing on an Illumina HiSeq 2000 platform in a 50/51bp single read mode . FastQC software  was used for quality control of the sequences. Bisulfite reads were trimmed for low-quality and adapter sequences using a custom pipeline . Bisulfite conversion metrics are given in S2 File. Reads were aligned to the Zebrafish Genome Assembly GRCz10 (danRer10) using BSMAP . DNA methylation calls were performed using BiSeq . Due to low DNA quality for sequencing of two samples and divergent sequencing results of one sample (all three samples were control F0), they were excluded from further downstream analysis. The data discussed has been stored in SRA (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA418670) and is accessible through accession number PRJNA418670.
Differential methylation and functional enrichment analysis
Differential methylation and downstream analysis was completed in R version 3.3.2 . Methylation calls were filtered after a minimum read depth per locus of 10, and replicates were combined into a single table based on genomic loci that were present in all replicates and control samples to compare the same loci for differential methylation in all samples. One outlier replicate was removed after principal component analysis and pairwise sample distance plot inspection (dendrogram, heatmap). Differentially methylated loci (DML, differentially methylated cytosine in CpG context) were detected using the methylKit package, based on logistic regression analysis and Benjamini-Hochberg false discovery correction of p-values (q-values) . Methylation events with a methylation difference +/- ≥25% and q-value ≤0.01 were considered for assigning hypermethylation (positive percentage) or hypomethylation (negative percentage) in the high ARA group compared to control. The ‘genomation’ package  was used to annotate DML by genomic regions such as promoters (±1000bp from the transcription start site), introns, exons, intergenic regions (gene nearest to the DML), CpG islands and CpG shores (±2000bp flanking regions around a CpG island). Definitions for genomic regions were obtained from the UCSC Genome Browser , which were coordinate mapped to GRCz10 Ensembl transcripts. Using a hypergeometric test in R, we also examined if genomic regions (CpG islands, CpG island shores, exons, introns and promoters) were significantly enriched or depleted (p<0.05) for DML. In addition, we calculated enrichment scores based on the ratio of DML to methylated loci within the genomic region (i.e. detected number of DML) divided by the ratio of total DML to total methylated loci within the entire genome (expected number of DML). Positive scores indicate the genomic region is enriched for DML (i.e. more DML detected than expected) and a negative score indicates depletion for DML (S3 File). Annotation of DML to gene identifiers such as Ensembl and Entrez gene identifier, gene symbol and gene description was completed using the genomicRanges  and biomaRt packages . The annotated DML (S4 File) were used for functional annotation for KEGG pathways and GO terms by over-representation testing using the R package ‘clusterProfiler’  (S5 File).
Gene expression data
RNA-sequencing (RNA-seq) was performed by the Norwegian Sequencing Centre (NSC) doing the library preparation using TruSeq™ Stranded mRNA Library Prep Kit (Illumina, Inc, USA), and sequencing on the NextSeq500 platform (Illumina, Inc, USA) to generate single-end 75bp reads as previously reported . Briefly, reads were mapped to the GRCz10 (Genome Reference Consortium Zebrafish Build 10) assembly based on Ensembl annotation data,  using the default parameters of HISAT2 . Read counts per gene were quantified using featureCounts  and pre-filtered to exclude combined mean read counts smaller than 10. Differential gene expression was estimated using DESeq2  and the complete list of expressed genes from F0 and F1 livers is reported in the S6 File. We identified significant differential expressed genes (DEG) using an adjusted p-value cut-off <0.05 for F1 DEG and <0.1 for F0 DEG due to fewer genes in the latter . The gene expression data was obtained from a different cohort of zebrafish than the DNA methylation data, though the experimental treatment was the same as for the zebrafish in this present study. Raw data is accessible at the NCBI’s Gene Expression Omnibus  through GEO Series accession number GSE104692 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104692).
DEG were converted to human orthologues to identify upstream regulators that can explain the observed gene expression changes using Ingenuity Pathway Analysis software suite (IPA, Ingenuity Systems, USA). Human orthologues were predicted using the OrthoRetriever v1.2 software. An overlap p-value using Fisher’s Exact test and an activation z-score (positive or negative) for upstream regulators is calculated based on prior knowledge stored in the Ingenuity Knowledge Base. Upstream regulators with missing z-scores and p >0.05 were filtered out from our analysis. Upstream regulators, which were only differentially methylated in F0, were also excluded from the final list. Positive and negative z-scores predict an activation and inhibition of an upstream regulator, respectively. Higher positive z-scores or lower negative z-scores indicate stronger directional relationship.
Body weight of 98 DPF (F0, control: 36 fish, high ARA: 36 fish) and 140–142 DPF (F1, control: 48 fish, high ARA: 47 fish) zebrafish was initially tested for tank variances (10 tanks in F0 and 6 tanks in F1) using one-way ANOVA (GraphPad Prism 8, GraphPad Software, Inc, USA). An unpaired, two-tailed t-test (GraphPad Prism 8, GraphPad Software, Inc, USA) was used for significance testing (p-value <0.05) of body weight from all tank populations within a dietary groups given as mean ± standard deviation. For RRBS and RNA-seq, statistical treatment such as filtering, multiple comparison correction and hypergeometric test are described in previous respective sections. Downstream statistical analysis of RRBS and RNA-Seq data was completed using R v3.3.2 (http://cran.rproject.org/).
No significant differences in body weight were observed between the dietary groups of either F0 or F1. Mean body weight of F0 zebrafish was 0.31 ± 0.12 g in the control and 0.35 ± 0.13 g in the high ARA group. Mature F1 progeny body weight was on average 0.35 ± 0.04 g in the control and 0.34 ± 0.05 g in the high ARA group.
Differential DNA methylation in F0 and F1 livers
General DNA methylation patterns.
An average of 84% of reads per sample were mapped to the zebrafish genome using BSMAP (S2 File). Bisulfite conversion rate was on average 99% for all sample groups and consistent between the sample groups (S2 File). Principal component analysis (Fig 2A) and pairwise distance clustering of the samples (dendrogram/heatmap) (Fig 2B) revealed that the overall DNA methylation pattern distinguishes samples by generation and to a lesser extent by dietary group. It must be noted that livers were sampled at a different age of the zebrafish in F0 (98 DPF) and F1 (140–142 DPF). The F0 generation was clearly separated between the dietary groups although there were only three samples left in the control group as the other three samples did not meet DNA quality cut-offs for sequencing. The PCA plot shows a higher degree of overlap for the F1 generation compared to the F0 generation.
Principal component analysis (A), and pairwise distance clustering of the samples (dendrogram and heatmap) (B) show the grouping of dietary groups based on percent methylation per locus in F0 livers from fish fed either the control or high ARA diet, and F1 fed the control diet. The scale for Fig 2B (0:1200) indicates Euclidean distance between samples as calculated by the base R package ‘dist’ (http://cran.rproject.org/).
Among all samples, the number of common methylated loci was less for F0 than for F1 livers (Table 1). After filtering by a minimum read depth per locus of 10, and combining samples based on methylated loci common to all samples within a treatment comparison group (e.g. F1 high ARA vs F1 control), the remaining methylated loci were analysed for differential methylation between the dietary groups. We found that the number of DML between high ARA and control group was marginally higher in F0 livers (2338) than in F1 livers (2142). In addition, the number of hypermethylated (1091) compared to hypomethylated (1051) loci was almost identical in F1 livers, whereas F0 livers showed greater hypermethylation (1411) than hypomethylation (927). We found no difference between high ARA and control livers when comparing total CpG methylation rates, both groups showed on average 85% methylation (S2 File).
Differential methylation across genomic regions.
The distribution of DML was investigated in promoters, exons, introns, and intergenic regions (Fig 3A) to verify if the DML were randomly distributed or specifically enriched in specific locations of the genome. The results show that the overall DML distribution across genomic regions was similar between F0 and F1 livers, with fewer DML in promoters and exons compared to DML in introns and intergenic regions. However, the increased number of hypermethylation within the F0 group is predominantly in the introns (625) and especially in the intergenic regions (658) compared to F1 (introns: 490, intergenic: 483). In addition, exons of F0 have proportionally more hypermethylation (119) than hypomethylation (65) compared to F1 (hypermethylation: 104, hypomethylation: 97).
Differential methylation here shown as hyper- and hypomethylated loci across promoters, exons, introns and intergenic regions (A), CpG islands (CpGi) and shores (B).
Searching for DML in CpG islands and CpG shores, we noticed increased differential methylation in CpG islands (F0: 702, F1: 551) than in the CpG shores (F0: 294, F1: 328) for both generations (Fig 3B). Again, we found that the F0 generation had more hypermethylation especially in the CpG islands (425) than F1 (292). However, the majority of DML fell outside the CpG island and CpG shore annotation categories, into other genomic regions (F0: 1342, F1: 1263), where again F0 high ARA livers show noticeably more hypermethylation (809) and less hypomethylation (533) than F1 (617/646). In terms of enrichment and depletion of DML within genomic regions, for the F0 cohort only CpG islands were significantly depleted with no significant difference found within CpGi shores, exons, introns and promoters. For the F1 cohort, all genomic regions were significantly enriched or depleted for DML, with CpG islands and introns depleted and CpGi shores, exons and promoters enriched for DML (S3 File).
Functional annotation of DML.
We associated DML within promoters and gene bodies with specific biological pathways (S5 File). No KEGG pathways and only one GO term (proteinaceous extracellular matrix) was significantly enriched (q-value cut-off <0.05) in F0, whereas in F1 no significantly enriched terms were identified.
Among the most differentially methylated loci between the dietary groups in F1 livers, crebbpb was dominating with three hyper- and one hypomethylated DML (S4 File). crebbpb codes for the nuclear coactivator cAMP-response element-binding protein (CREB) binding protein that plays a key role in various signaling pathways through interacting with numerous transcription factors.
Common F0 and F1 DML.
DML from high ARA vs. control group analysis in both generations were merged to find common loci between the F0 and F1 methylation differences (S4 File). In total, 190 DML assigned to promoter, gene body and intergenic regions were common between F0 and F1 DML. Among them, 5 DML exclusively assigned to promoters were at identical sites in F0 and F1 following both same (cryabb and nup160) and opposite methylation patterns (rsf1b.1, si:dkey−4c2.11 and zbtb24) between the generations. For cryabb, nup160 and zbtb24 that have one assigned DML in F0 livers, at least two or more DML have been associated to same gene showing same methylation patterns in the F1 livers (S4 File). Among the genes, cryabb (crystallin alpha B) assigned to four F1 DML codes for a small heat shock protein that is involved in the stabilization of stress-related cellular processes such as cell cycle, differentiation, apoptosis and redox homeostasis.
Differential methylation and gene expression
Genes linked to differential methylation and differential expression.
We searched for a connection between DML (S4 File) and DEG (S6 File) for F0 and F1 generation. We found 5 concordant genes between DML and DEG in F0 and 37 concordant genes in F1 (Fig 4A and Table 2). All overlapping genes for both generations are listed in S7 File.
A: Plot showing genomic locus-specific methylation differences and gene expression of concordant genes in F0 (5) and F1 (37) livers. Differentially expressed genes (adjusted p <0.05 for F1 and p <0.1 for F0) and DML (methylation difference +/- ≥25% and q-value ≤0.01) were obtained from F0 and F1 livers comparing high ARA and control group. Some of the genes in F1 were annotated to more than one DML, which led to 46 comparisons in total (S8 File). A best fit line with equation and goodness of fit (R2) quantified by linear regression (S8 File) is shown for F1 DML located in gene bodies and intergenic regions. B: Bar graphs showing individual read counts and individual CpG methylation levels of single genes tagged in Fig 4A for each replicate of the high ARA and control group. Gene expression is shown as mean normalized read counts (MNRC) in high ARA versus control livers. CpG methylation was calculated as the percentage of total methylated CpGs over the total number of CpGs assessed.
Plotting methylation differences against gene expression in F1, we found 29 overlaps (corresponding to 24 genes due to several DML annotated to the same gene) between DML assigned to a gene body and DEG showing no significant correlation between methylation (hyper/hypo) and gene expression (up/down). 16 overlaps (corresponding to 14 genes) involving F1 DML assigned to intergenic regions were significantly and positively correlated as shown in Fig 4A (p = 0.0011). Correlation and linear regression analysis is reported in S8 File. For single overlapping F1 genes flagged in Fig 4A, individual read counts (RNA-seq) and individual CpG methylation (RRBS) for each of the replicates in high ARA and control group are shown in Fig 4B. Excluding uncharacterized genes, DML associated with an F1 locus showing the greatest hyper- or hypomethylation were mgat4b (mannosyl (alpha-1,3-)-glycoprotein beta-1,4-N-acetylglucosaminyltransferase, isozyme B) and roraa (RAR-related orphan receptor A, paralog a), respectively.
Three hypomethylated loci in F1 were associated with the closest gene, roraa, which encodes a nuclear hormone receptor involved in lipid regulation. roraa was significantly downregulated in F1 high ARA livers. esr2a (estrogen receptor 2a), involved in retinoid signalling, was found as the closest gene to a hypomethylated locus in F1. esr2a showed significant downregulation in F1 high ARA livers. A hypomethylated locus in F1 overlapped with sult1st3 (sulfotransferase family 1, cytosolic sulfotransferase 3), which functions in lipid homeostasis. sult1st3 showed significant upregulation in F1 high ARA livers. One of the key genes in the methionine cycle, called mat1a (methionine adenosyltransferase I, alpha), was overlapping with a hypermethylated locus in both F0 (28% increased methylation) and F1 (29%). mat1a was significantly upregulated in F1 high ARA livers compared to the control, but not differentially expressed in F0. Three hypomethylated loci were overlapping with ccnf (cyclin F) that was upregulated in F1 livers.
One of the overlapping F0 genes was mboat2a (membrane bound O-acyltransferase domain containing 2a), which is involved in lysophospholipid metabolism. A hypomethylated locus was associated with mboat2a as the closest gene and mboat2a was upregulated in F0 livers.
Upstream regulators affecting downstream biological functions were predicted using IPA based on F1 DEG. Among the 399 predicted upstream regulators, 44 showed an overlap with the F0 and F1 list of DML linked to a gene annotation (S9 File). After filtering, the cannabinoid receptor 1 (CNR1) had the strongest positive z-score of the upstream regulators predicted by IPA (S9 File). cnr1 was overlapping with a hypomethylated locus in F1. However, cnr1 was not among the DEG as it was filtered out prior to differential expression calling due to low read count in the sequenced livers. The majority of the upstream regulators were transcriptional regulators (PPARGC1A, NCOA2) and nuclear receptors (PPARD, RORA, PPARA, ESR2, NR0B2) involved in the regulation of fatty acids, lipids, estrogen and energy metabolism.
In the present study, we have shown that feeding the parents a diet high in ARA, an n-6 PUFA, can alter the DNA methylation patterns in livers of both the parents and their adult progenies. This demonstrates that the epigenetic DNA methylation pattern is sensitive to dietary ARA levels over generations. We reported previously, that a diet high in ARA did not alter the growth of adult zebrafish in either parents or progeny [47, 48]. Earlier studies have shown that nutrition can influence DNA methylation, which affects phenotypes [5, 66, 67]. To our knowledge, this is the first study showing that changing the fatty acid profile in the parental diet can affect DNA methylation at locus-specific sites in zebrafish liver.
DNA methylation is the most studied epigenetic mechanism, which is known to regulate gene expression potential. When comparing the DNA methylation pattern and gene expression profiles in mature zebrafish livers in this study, it was hard to find a consistent pattern between direction of methylation and changes in gene expression. In general, we observed a higher degree of overlap between methylation and transcription profiles in the F1 generation than in the F0 probably due to a greater number of differentially expressed genes detected in F1. In response to the ARA fortified diet, we observed a larger distribution of DML in introns and intergenic regions than in exons and promoters for both generations. However, the latter was a significantly enriched region in F1 livers, meaning more DML detected than expected. A latest study  examined DNA methylation in livers of zebrafish fed a micronutrient-modified diet, also reported similar findings, with promoters enriched for DML. Methylation in promoters is widely believed to alter gene expression  though methylation outside promoters has also been shown to play a role in gene regulation . Although not differentially expressed, crebbpb is one gene that was assigned to four among the most differentially methylated loci present in either an exon or an intron in the F1 livers. CREBBP is playing critical roles in embryonic development, growth control, and homeostasis by coupling chromatin remodeling to transcription factor recognition [70, 71]. A previous zebrafish study suggested that methylation associated with promoters was not the major determinant of transcription when comparing methylation and transcriptional data . This is in line with our results, which showed relatively few promoter region associations between DML and DEG. Regardless, we reported interesting patterns among promoter DML that are identical between F1 and F0 DML (5 DML), where e.g. two identical DML show the same methylation pattern in both generations. Interestingly, those genes were assigned to one F0 DML and at least two or more F1 DML showing same methylation patterns. One of them we reported was cryabb that functions in skeletal muscle tissue development, repair and stabilization of stress-related cellular processes such as cell cycle, differentiation, apoptosis and redox homeostasis [73, 74]. Whether some of the reported methylation differences across different genomic regions are due to random chance and others due to the dietary treatment in this study, remains unanswered. Although the number of total DML were similar in both generations of the present study, a slightly higher number of hypermethylated DML were observed in the parental F0 compared to the F1. A general increase in genome methylation following ARA exposure has previously been shown in mammalian cells . As noted by us and by others using rodents , DNA methylation patterns are sensitive to both parental (previous) and within generation environmental changes. This highlights nutrition as important modulator of DNA methylation pattern changes, as earlier implied for vitamins by others .
Several genes associated with F0 and F1 DML overlap between the generations, however, no significantly enriched pathways were observed for either F0 or F1 DML except from one GO term in the F0 generation. Despite few overlaps between DML and DEG in F0, we associated DML with the metabolic profiles reported previously . The F0 metabolic profiles of juvenile zebrafish from the present study showed that high dietary ARA induced both inflammatory and anti-oxidative responses affecting lipids and amino acids . Interestingly, mboat2a was one of the overlapping genes between DML and DEG in the F0 generation. This gene functions in lysophospholipid metabolism, which is consistent with the reported shift in lysophospholipid profiles for F0 fish fed high ARA levels . However, surprisingly few concordant genes to F0 DML and DEG were detected. Many factors might have masked some differences in the F0 generation leading to few overlaps for instance general low differential gene expression, plasticity of DNA methylation [77, 78], age  and variation introduced by the cell type specific nature of DNA methylation .
An intriguing finding was the significant positive correlation between intergenic DML and DEG in offspring livers (and no correlation between gene body DML and DEG). Genes annotated to intergenic DML were based on the nearest gene TSS to the DML, which varied between a few thousand to hundreds of thousands of bases distant. Regulatory elements, such as promoters, enhancers, silencers, noncoding RNA, etc, can be many thousands of bases distant from the genes. They often have a considerable and complex combinatorial effect on gene expression and phenotype . We need to stress that though we found a significant correlation between DML and DEG in intergenic regions, this is still based on a relatively small number of overlaps (n = 16) and we have not identified any causal links. That we found a positive correlation, i.e. hypermethylation correlating with upregulation (e.g. solute carriers 38a3b and 26a2) and hypomethylation correlating with downregulation (e.g. roraa and esr2a), was surprising, given that the focus of functional methylation has been primarily on promoter regions, where hypermethylation has been shown to functionally repress gene expression . However, the functional associations between gene expression and methylation in different genomic regions (such as gene bodies) is still poorly understood [83, 84]. We have reported the neighboring genes and associated intergenic DML as a reference for further functional exploration. It is possible that future studies will identify functionally associated regulatory elements using the DEG associated DML we have reported here.
DNA methylation changes might control gene expression at early developmental stages and in specific tissues that can prime gene expression changes later in life controlling specific pathways [30, 85, 86]. In the present study, we found several genes with parental diet-associated methylation changes that also showed differential gene expression in the mature progeny livers. We found no significant pathway enrichment for F1 DML, but interestingly some of the F1 DML genes related to pathways previously reported to be affected by ARA at both a transcriptional and a metabolic level [47, 48]. Of the genes overlapping between DML and DEG in F1, some function in the methionine cycle (mat1a), lipid (roraa, sult1st3, lpar2a, abca12) and estrogen signalling (esr2a). In the present study, we looked for possible upstream regulators based on the F1 DEG and if some of them could explain the observed differences in both DNA methylation and gene expression. Several of the upstream regulators predicted by IPA were linked to DML, of which most were transcriptional regulators and nuclear receptors regulating fatty acid, lipid, estrogen and energy metabolism. Another gene, cannabinoid receptor 1 (CNR1), with an associated DML was suggested as the most activated upstream regulator in the progeny livers. Cannabinoid receptor 1 gets activated by endocannabinoids in liver, affecting de novo lipogenesis and fatty acid catabolism . Remarkably, we observed increased endocannabinoid levels in F0 fish fed the high ARA diet , but no causal relationship has been found.
We showed that the overall changes in methylation were bigger in between generations and to a lesser extent between dietary groups. This can be attributed to natural occurring variation across generations representing a covariate and differences in the age of the samples analyzed for DNA methylation . Based on the few associations we made between the methylation and transcription changes in grown offspring, it is also difficult to conclude on whether parental (F0) diet influenced the environment of the developing embryo while priming hepatic gene expression of progeny [18, 33, 88]. Two studies in rodents could not link parental diet-associated DNA methylation changes with gene expression in the progeny [44, 89]. This underlines the need for more detailed knowledge on the complex link between nutrients and epigenetic modifications, for example, in the germline and early embryo resulting in altered gene expression and metabolic phenotypes in grown offspring . Nevertheless, we reported several DNA methylation differences present in the livers of grown offspring. Although we could not link overall methylation differences to DEG in the liver, it is conceivable that the reported methylation changes rather regulate gene expression during early development or in other tissues. This indicates that epigenetic mechanisms of gene regulation may act in a spatio-temporal matter, meaning tissue-specific and specific for developmental stages. More studies are needed to further illuminate and validate the mechanisms.
This study has shown that parental dietary ARA influences DNA methylation in zebrafish liver. Hepatic methylation patterns across different genomic regions have been reported in two generations. We identified 2338 loci in the parental livers and 2142 loci in the livers of their progeny showing differential methylation between the dietary groups. We compared the DNA methylation changes in progeny livers to existing gene expression changes, but only based on few gene annotations. Although we reported several genes possibly regulated by diet-associated methylation changes, our results are limited to liver tissue, and causal or functional associations remain undiscovered. Thus, it is possible that the relation between DNA methylation and gene expression changes is stronger during early developmental stages or in tissues other than liver. The effects of nutritional induced DNA methylation changes at specific CpG loci in a transgenerational context and the extent of epigenetic gene regulation need to be verified by further studies.
S1 File. Ingredients, nutritional and selected fatty acid composition of control and high ARA diet.
S2 File. Global methylation means, bisulfite conversion and mapping rates from reduced representation bisulfite sequencing of F0 and F1 livers.
S3 File. Log2 transformed enrichment ratios for differentially methylated (DM) loci within genomic regions (CpG islands, CpG island shores, exons, introns and promoters) in F0 (A) and F1 (B) zebrafish liver.
S4 File. F0 and F1 differentially methylated loci annotation (Ensembl) by genomic regions and identical differentially methylated loci between generations.
S5 File. Functional annotation of F0 and F1 differentially methylated loci for KEGG pathways and GO terms.
S6 File. Differentially expressed genes from RNA-sequencing of F0 and F1 livers comparing control and high ARA group (Ensembl).
S7 File. Common gene annotations to differentially methylated loci (DML) and differentially expressed genes (DEG) in F0 and F1 zebrafish livers.
S8 File. Differential methylation and differential gene expression of overlapping genes from F1 and F0 generation comparing high ARA and control group.
We are grateful to Øyvind Reinshol and Synnøve Wintertun for assistance at the in-house zebrafish facilities, Hui-Shan Tung for DNA extraction and Eva Mykkeltvedt for RNA extraction at the Institute of Marine Research (Bergen, Norway). We thank the Biomedical Sequencing Facility at CeMM Research Centre for Molecular Medicine of the Austrian Academy of Sciences (Vienna, Austria) for RRBS library preparation, sequencing and bisulfite read pre-processing. RNA library preparation and sequencing was performed by the Norwegian High-Throughput Sequencing Centre (NSC) in Oslo, Norway. This research was financed through The Research Council of Norway (EpiSip 228877 and EpiFeedFish 225250/E40).
- 1. Huang K, Fan G. DNA methylation in cell differentiation and reprogramming: an emerging systematic view. Regen Med. 2010 Jul;5(4):531–44. pmid:20632857
- 2. Jaenisch R, Bird A. Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat Genet. 2003 Mar;33 Suppl:245–54.
- 3. Bogdanovic O, Lister R. DNA methylation and the preservation of cell identity. Curr Opin Genet Dev. 2017 Oct;46:9–14. pmid:28651214
- 4. Anderson OS, Sant KE, Dolinoy DC. Nutrition and epigenetics: an interplay of dietary methyl donors, one-carbon metabolism and DNA methylation. J Nutr Biochem. 2012 Aug;23(8):853–9. pmid:22749138
- 5. Skjaerven KH, Jakt LM, Fernandes JMO, Dahl JA, Adam AC, Klughammer J, et al. Parental micronutrient deficiency distorts liver DNA methylation and expression of lipid genes associated with a fatty-liver-like phenotype in offspring. Sci Rep. 2018 Feb 14;8(1):3055. pmid:29445184
- 6. Veron V, Marandel L, Liu J, Velez EJ, Lepais O, Panserat S, et al. DNA methylation of the promoter region of bnip3 and bnip3l genes induced by metabolic programming. BMC Genomics. 2018 Sep 17;19(1):677. pmid:30223788
- 7. Zhang N. Role of methionine on epigenetic modification of DNA methylation and gene expression in animals. Anim Nutr. 2018 Mar;4(1):11–6. pmid:30167479
- 8. Park LK, Friso S, Choi SW. Nutritional influences on epigenetics and age-related disease. Proc Nutr Soc. 2012 Feb;71(1):75–83. pmid:22051144
- 9. Kaelin WG Jr., McKnight SL. Influence of metabolism on epigenetics and disease. Cell. 2013 Mar 28;153(1):56–69. pmid:23540690
- 10. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet. 2012 Jul;13(7):484–92. pmid:22641018
- 11. Li E, Zhang Y. DNA methylation in mammals. Cold Spring Harb Perspect Biol. 2014 May 1;6(5):a019133. pmid:24789823
- 12. Zhang C, Hoshida Y, Sadler KC. Comparative Epigenomic Profiling of the DNA Methylome in Mouse and Zebrafish Uncovers High Interspecies Divergence. Front Genet. 2016;7:110. pmid:27379160
- 13. Brenet F, Moh M, Funk P, Feierstein E, Viale AJ, Socci ND, et al. DNA Methylation of the First Exon Is Tightly Linked to Transcriptional Silencing. PLoS One. 2011 Jan 18;6(1).
- 14. Yang X, Han H, De Carvalho DD, Lay FD, Jones PA, Liang G. Gene body methylation can alter gene expression and is a therapeutic target in cancer. Cancer Cell. 2014 Oct 13;26(4):577–90. pmid:25263941
- 15. Jjingo D, Conley AB, Yi SV, Lunyak VV, Jordan IK. On the presence and role of human gene-body DNA methylation. Oncotarget. 2012 Apr;3(4):462–74. pmid:22577155
- 16. Zemach A, McDaniel IE, Silva P, Zilberman D. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science. 2010 May 14;328(5980):916–9. pmid:20395474
- 17. Clermont PL, Parolia A, Liu HH, Helgason CD. DNA methylation at enhancer regions: Novel avenues for epigenetic biomarker development. Front Biosci (Landmark Ed). 2016 Jan 1;21:430–46.
- 18. Zenk F, Loeser E, Schiavo R, Kilpert F, Bogdanovic O, Iovino N. Germ line-inherited H3K27me3 restricts enhancer function during maternal-to-zygotic transition. Science. 2017 Jul 14;357(6347):212–6. pmid:28706074
- 19. Waterland RA, Jirtle RL. Early nutrition, epigenetic changes at transposons and imprinted genes, and enhanced susceptibility to adult chronic diseases. Nutrition. 2004 Jan;20(1):63–8. pmid:14698016
- 20. Burdge GC, Lillycrop KA. Nutrition, epigenetics, and developmental plasticity: implications for understanding human disease. Annu Rev Nutr. 2010 Aug 21;30:315–39. pmid:20415585
- 21. Waterland RA. Is epigenetics an important link between early life events and adult disease? Horm Res. 2009 Jan;71 Suppl 1:13–6.
- 22. Lillycrop KA, Phillips ES, Jackson AA, Hanson MA, Burdge GC. Dietary protein restriction of pregnant rats induces and folic acid supplementation prevents epigenetic modification of hepatic gene expression in the offspring. J Nutr. 2005 Jun;135(6):1382–6. pmid:15930441
- 23. Skjaerven KH, Jakt LM, Dahl JA, Espe M, Aanes H, Hamre K, et al. Parental vitamin deficiency affects the embryonic gene expression of immune-, lipid transport- and apolipoprotein genes. Sci Rep. 2016 Oct 12;6:34535. pmid:27731423
- 24. Newman T, Jhinku N, Meier M, Horsfield J. Dietary Intake Influences Adult Fertility and Offspring Fitness in Zebrafish. PLoS One. 2016;11(11):e0166394. pmid:27870856
- 25. Polak M, Simmons LW, Benoit JB, Ruohonen K, Simpson SJ, Solon-Biet SM. Nutritional geometry of paternal effects on embryo mortality. Proc Biol Sci. 2017 Oct 11;284(1864).
- 26. McRae AF, Powell JE, Henders AK, Bowdler L, Hemani G, Shah S, et al. Contribution of genetic variation to transgenerational inheritance of DNA methylation. Genome Biol. 2014 May 29;15(5):R73. pmid:24887635
- 27. Schubeler D. Function and information content of DNA methylation. Nature. 2015 Jan 15;517(7534):321–6. pmid:25592537
- 28. Lee HJ, Hore TA, Reik W. Reprogramming the methylome: erasing memory and creating diversity. Cell Stem Cell. 2014 Jun 5;14(6):710–9. pmid:24905162
- 29. Heard E, Martienssen RA. Transgenerational epigenetic inheritance: myths and mechanisms. Cell. 2014 Mar 27;157(1):95–109. pmid:24679529
- 30. Burdge GC, Slater-Jefferies J, Torrens C, Phillips ES, Hanson MA, Lillycrop KA. Dietary protein restriction of pregnant rats in the F0 generation induces altered methylation of hepatic gene promoters in the adult male offspring in the F1 and F2 generations. Br J Nutr. 2007 Mar;97(3):435–9. pmid:17313703
- 31. Hoile SP, Lillycrop KA, Thomas NA, Hanson MA, Burdge GC. Dietary protein restriction during F0 pregnancy in rats induces transgenerational changes in the hepatic transcriptome in female offspring. PLoS One. 2011;6(7):e21668. pmid:21750721
- 32. Triggs AM, Knell RJ. Parental diet has strong transgenerational effects on offspring immunity. Funct Ecol. 2012 Dec;26(6):1409–17.
- 33. Seiliez I, Velez EJ, Lutfi E, Dias K, Plagnes-Juan E, Marandel L, et al. Eating for two: Consequences of parental methionine nutrition on offspring metabolism in rainbow trout (Oncorhynchus mykiss). Aquaculture. 2017 Mar 20;471:80–91.
- 34. Simopoulos AP. Evolutionary aspects of diet, the omega-6/omega-3 ratio and genetic variation: nutritional implications for chronic diseases. Biomed Pharmacother. 2006 Nov;60(9):502–7. pmid:17045449
- 35. Patterson E, Wall R, Fitzgerald GF, Ross RP, Stanton C. Health implications of high dietary omega-6 polyunsaturated Fatty acids. J Nutr Metab. 2012;2012:539426. pmid:22570770
- 36. Thomas MH, Pelleieux S, Vitale N, Olivier JL. Dietary arachidonic acid as a risk factor for age-associated neurodegenerative diseases: Potential mechanisms. Biochimie. 2016 Jul 26.
- 37. Lands B. Consequences of Essential Fatty Acids. Nutrients. 2012 Sep;4(9):1338–57. pmid:23112921
- 38. de Vrieze E, Moren M, Metz JR, Flik G, Lie KK. Arachidonic acid enhances turnover of the dermal skeleton: studies on zebrafish scales. PLoS One. 2014;9(2):e89347. pmid:24586706
- 39. Lie KK, Kvalheim K, Rasinger JD, Harboe T, Nordgreen A, Moren M. Vitamin A and arachidonic acid altered the skeletal mineralization in Atlantic cod (Gadus morhua) larvae without any interactions on the transcriptional level. Comp Biochem Phys A. 2016 Jan;191:80–8.
- 40. Norberg B, Kleppe L, Andersson E, Thorsen A, Rosenlund G, Hamre K. Effects of dietary arachidonic acid on the reproductive physiology of female Atlantic cod (Gadus morhua L.). Gen Comp Endocrinol. 2017 May 30;250:21–35. pmid:28576420
- 41. Asil SM, Kenari AA, Miyanji GR, Van Der Kraak G. The influence of dietary arachidonic acid on growth, reproductive performance, and fatty acid composition of ovary, egg and larvae in an anabantid model fish, Blue gourami (Trichopodus trichopterus; Pallas, 1770). Aquaculture. 2017 Jul 1;476:8–18.
- 42. Yu HL, Dong S, Gao LF, Li L, Xi YD, Ma WW, et al. Global DNA methylation was changed by a maternal high-lipid, high-energy diet during gestation and lactation in male adult mice liver. Br J Nutr. 2015 Apr 14;113(7):1032–9. pmid:25778733
- 43. Voisin S, Almen MS, Moschonis G, Chrousos GP, Manios Y, Schioth HB. Dietary fat quality impacts genome-wide DNA methylation patterns in a cross-sectional study of Greek preadolescents. Eur J Hum Genet. 2015 May;23(5):654–62. pmid:25074463
- 44. Cannon MV, Buchner DA, Hester J, Miller H, Sehayek E, Nadeau JH, et al. Maternal nutrition induces pervasive gene expression changes but no detectable DNA methylation differences in the liver of adult offspring. PLoS One. 2014;9(3):e90335. pmid:24594983
- 45. Silva-Martinez GA, Rodriguez-Rios D, Alvarado-Caudillo Y, Vaquero A, Esteller M, Carmona FJ, et al. Arachidonic and oleic acid exert distinct effects on the DNA methylome. Epigenetics. 2016 May 3;11(5):321–34. pmid:27088456
- 46. Xue SS, He JL, Zhang X, Liu YJ, Xue FX, Wang CJ, et al. Metabolomic analysis revealed the role of DNA methylation in the balance of arachidonic acid metabolism and endothelial activation. Biochim Biophys Acta. 2015 Oct;1851(10):1317–26. pmid:26170200
- 47. Adam AC, Lie KK, Moren M, Skjaerven KH. High dietary arachidonic acid levels induce changes in complex lipids and immune-related eicosanoids and increase levels of oxidised metabolites in zebrafish (Danio rerio). Br J Nutr. 2017 May 09:1–11.
- 48. Adam AC, Skjaerven KH, Whatmore P, Moren M, Lie KK. Parental high dietary arachidonic acid levels modulated the hepatic transcriptome of adult zebrafish (Danio rerio) progeny. PLoS One. 2018;13(8):e0201278. pmid:30070994
- 49. NRC. Nutrient requirements of fish and shrimp. Washington, DC: The National Academies Press; 2011.
- 50. Bock C, Tomazou EM, Brinkman AB, Muller F, Simmer F, Gu H, et al. Quantitative comparison of genome-wide DNA methylation mapping technologies. Nat Biotechnol. 2010 Oct;28(10):1106–14. pmid:20852634
- 51. Klughammer J, Datlinger P, Printz D, Sheffield NC, Farlik M, Hadler J, et al. Differential DNA Methylation Analysis without a Reference Genome. Cell Rep. 2015 Dec 22;13(11):2621–33. pmid:26673328
- 52. Sheffield NC, Pierron G, Klughammer J, Datlinger P, Schonegger A, Schuster M, et al. DNA methylation heterogeneity defines a disease spectrum in Ewing sarcoma. Nat Med. 2017 Mar;23(3):386–95. pmid:28134926
- 53. Andrews S. FastQC: a quality control tool for high throughput sequence data. 2017; http://www.bioinformatics.babraham.ac.uk/projects/fastqc.
- 54. Xi Y, Li W. BSMAP: whole genome bisulfite sequence MAPping program. BMC Bioinformatics. 2009 Jul 27;10:232. pmid:19635165
- 55. RCoreTeam. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013.
- 56. Akalin A, Kormaksson M, Li S, Garrett-Bakelman FE, Figueroa ME, Melnick A, et al. methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles. Genome Biol. 2012 Oct 03;13(10):R87. pmid:23034086
- 57. Akalin A, Franke V, Vlahovicek K, Mason CE, Schubeler D. Genomation: a toolkit to summarize, annotate and visualize genomic intervals. Bioinformatics. 2015 Apr 01;31(7):1127–9. pmid:25417204
- 58. Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, et al. The human genome browser at UCSC. Genome Research. 2002 Jun;12(6):996–1006. pmid:12045153
- 59. Lawrence M, Huber W, Pages H, Aboyoun P, Carlson M, Gentleman R, et al. Software for computing and annotating genomic ranges. PLoS Comput Biol. 2013;9(8):e1003118. pmid:23950696
- 60. Durinck S, Spellman PT, Birney E, Huber W. Mapping identifiers for the integration of genomic datasets with the R/Bioconductor package biomaRt. Nat Protoc. 2009;4(8):1184–91. pmid:19617889
- 61. Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012 May;16(5):284–7. pmid:22455463
- 62. Yates A, Akanni W, Amode MR, Barrell D, Billis K, Carvalho-Silva D, et al. Ensembl 2016. Nucleic Acids Res. 2016 Jan 04;44(D1):D710–6. pmid:26687719
- 63. Kim D, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with low memory requirements. Nat Methods. 2015 Apr;12(4):357–60. pmid:25751142
- 64. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014 Apr 01;30(7):923–30. pmid:24227677
- 65. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. pmid:25516281
- 66. Waterland RA, Jirtle RL. Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol Cell Biol. 2003 Aug;23(15):5293–300. pmid:12861015
- 67. Amarasinghe HE, Toghill BJ, Nathanael D, Mallon EB. Allele specific expression in worker reproduction genes in the bumblebee Bombus terrestris. PeerJ. 2015;3:e1079. pmid:26213649
- 68. Curradi M, Izzo A, Badaracco G, Landsberger N. Molecular mechanisms of gene silencing mediated by DNA methylation. Mol Cell Biol. 2002 May;22(9):3157–73. pmid:11940673
- 69. Lou S, Lee HM, Qin H, Li JW, Gao Z, Liu X, et al. Whole-genome bisulfite sequencing of multiple individuals reveals complementary roles of promoter and gene body methylation in transcriptional regulation. Genome Biol. 2014 Jul 30;15(7):408. pmid:25074712
- 70. Goodman RH, Smolik S. CBP/p300 in cell growth, transformation, and development. Genes Dev. 2000 Jul 1;14(13):1553–77. pmid:10887150
- 71. Liu X, Wang L, Zhao K, Thompson PR, Hwang Y, Marmorstein R, et al. The structural basis of protein acetylation by the p300/CBP transcriptional coactivator. Nature. 2008 Feb 14;451(7180):846–50. pmid:18273021
- 72. McGaughey DM, Abaan HO, Miller RM, Kropp PA, Brody LC. Genomics of CpG methylation in developing and developed zebrafish. G3 (Bethesda). 2014 Mar 21;4(5):861–9.
- 73. Ciano M, Allocca S, Ciardulli MC, Della Volpe L, Bonatti S, D’Agostino M. Differential phosphorylation-based regulation of alphaB-crystallin chaperone activity for multipass transmembrane proteins. Biochem Biophys Res Commun. 2016 Oct 14;479(2):325–30. pmid:27641668
- 74. Zhang Y, Li C, Meng H, Guo D, Zhang Q, Lu W, et al. BYD Ameliorates Oxidative Stress-Induced Myocardial Apoptosis in Heart Failure Post-Acute Myocardial Infarction via the P38 MAPK-CRYAB Signaling Pathway. Front Physiol. 2018;9:505. pmid:29867551
- 75. Moody L, Chen H, Pan YX. Postnatal diet remodels hepatic DNA methylation in metabolic pathways established by a maternal high-fat diet. Epigenomics. 2017 Nov;9(11):1387–402. pmid:28885036
- 76. Hore TA. Modulating epigenetic memory through vitamins and TET: implications for regenerative medicine and cancer treatment. Epigenomics. 2017 Jun;9(6):863–71. pmid:28554227
- 77. Szyf M. The dynamic epigenome and its implications in toxicology. Toxicol Sci. 2007 Nov;100(1):7–23. pmid:17675334
- 78. Schneider E, Pliushch G, El Hajj N, Galetzka D, Puhl A, Schorsch M, et al. Spatial, temporal and interindividual epigenetic variation of functionally important DNA methylation patterns. Nucleic Acids Res. 2010 Jul;38(12):3880–90. pmid:20194112
- 79. Yuan T, Jiao Y, de Jong S, Ophoff RA, Beck S, Teschendorff AE. An integrative multi-scale analysis of the dynamic DNA methylation landscape in aging. PLoS Genet. 2015 Feb;11(2):e1004996. pmid:25692570
- 80. McGregor K, Bernatsky S, Colmegna I, Hudson M, Pastinen T, Labbe A, et al. An evaluation of methods correcting for cell-type heterogeneity in DNA methylation studies. Genome Biol. 2016 May 3;17:84. pmid:27142380
- 81. Levo M, Segal E. In pursuit of design principles of regulatory sequences. Nat Rev Genet. 2014 Jul;15(7):453–68. pmid:24913666
- 82. Liu XS, Wu H, Ji X, Stelzer Y, Wu X, Czauderna S, et al. Editing DNA Methylation in the Mammalian Genome. Cell. 2016 Sep 22;167(1):233–47 e17. pmid:27662091
- 83. Zilberman D. An evolutionary case for functional gene body methylation in plants and animals. Genome Biol. 2017 May 9;18(1):87. pmid:28486944
- 84. Hon GC, Hawkins RD, Caballero OL, Lo C, Lister R, Pelizzola M, et al. Global DNA hypomethylation coupled to repressive chromatin domain formation and gene silencing in breast cancer. Genome Res. 2012 Feb;22(2):246–58. pmid:22156296
- 85. Godfrey KM, Sheppard A, Gluckman PD, Lillycrop KA, Burdge GC, McLean C, et al. Epigenetic gene promoter methylation at birth is associated with child’s later adiposity. Diabetes. 2011 May;60(5):1528–34. pmid:21471513
- 86. Gluckman PD, Hanson MA, Low FM. The role of developmental plasticity and epigenetics in human health. Birth Defects Res C Embryo Today. 2011 Mar;93(1):12–8. pmid:21425438
- 87. Alswat KA. The role of endocannabinoids system in fatty liver disease and therapeutic potentials. Saudi J Gastroenterol. 2013 Jul-Aug;19(4):144–51. pmid:23828743
- 88. Sharma U, Rando OJ. Metabolic Inputs into the Epigenome. Cell Metabolism. 2017 Mar 7; 25(3):544–558. pmid:28273477
- 89. Sabet JA, Park LK, Iyer LK, Tai AK, Koh GY, Pfalzer AC, et al. Paternal B Vitamin Intake Is a Determinant of Growth, Hepatic Lipid Metabolism and Intestinal Tumor Volume in Female Apc1638N Mouse Offspring. PLoS One. 2016;11(4):e0154979. pmid:27124183