Integrative analysis of methylomic and transcriptomic data in fetal sheep muscle tissues in response to maternal diet during pregnancy

Abstract

Background: Numerous studies have established a link between maternal diet and the physiological and metabolic phenotypes of their offspring. In previous studies in sheep, we demonstrated that different maternal diets altered the transcriptome of fetal tissues. However, the mechanisms underlying transcriptomic changes are poorly understood. DNA methylation is an epigenetic mark regulating transcription and is largely influenced by dietary components of the one-carbon cycle that generate the methyl group donor, SAM. Therefore, in the present study, we tested whether different maternal diets during pregnancy would alter the DNA methylation and gene expression patterns in fetal tissues.

Results: Pregnant ewes were randomly divided into two groups which received either hay or corn diet from mid-gestation (day 67 ± 5) until day 131 ± 1 when fetuses were collected by necropsy. A total of 1516 fetal longissimus dorsi (LD) tissues were used for DNA methylation analysis and gene expression profiling. Whole genome DNA methylation using methyl-binding domain enrichment analysis revealed 60 differentially methylated regions (DMRs) between hay and corn fetuses with 39 DMRs more highly methylated in the hay fetuses vs. 21 DMRs more highly methylated in the corn fetuses. Three DMRs (LPAR3, PLIN5-PLIN4, and the differential methylation of a novel lincRNA) were validated using bisulfite sequencing. These DMRs were associated with differential gene expression. Additionally, significant DNA methylation differences were found at the single CpG level. Integrative methylome and transcriptome analysis revealed an association between gene expression and inter-/intragenic methylated regions. Furthermore, intragenic DMRs were found to be associated with expression of neighboring genes.

Conclusions: The findings of this study imply that maternal diet from mid- to late-gestation can shape the epigenome and the transcriptome of fetal tissues, and putatively affect phenotypes of the lambs.

Keywords: DNA methylation; Differentially-methylated region; Gene expression; Maternal diet.

Fig. 1

Principal Component Analysis based on whole genome DNA methylation in the sheep fetal LD muscle from hay- and corn-fed ewes

Fig. 2

DNA methylation levels in fetal LD muscle from hay- and corn-fed ewes. a) LPAR3 DMR, b) PLIN5-PLIN4 DMR, c) lincRNA DMR. HF: hay females; HM: hay males; CF: corn females; CM: corn males; Hay: male and female pools combined; Corn: male and female pools combined. The percentage of DNA methylation per sequence was calculated using the following formula: number of methylated CpG sites / total number of successfully sequenced CpG sites. The BISMA online tool was used to assess the methylation percentage

Fig. 3

Methylation levels at individual CpG sites. a) LPAR3 DMR, b) PLIN5-PLIN4 DMR, c) lincRNA DMR. HF: hay females; HM: hay males; CF: corn females; CM: corn males; Hay: male and female pools combined; Corn: male and female pools combined. The percentage of DNA methylation for each CpG site was calculated using the following formula: number of methylated cytosines at the CpG site/total number of successfully sequenced cytosines at the CpG site under evaluation

Fig. 4

Gene expression profiles of DMRs in the following genes: a) LPAR3, b) PLIN4, c) PLIN5 and d) PLCB4. The fold difference in expression was calculated using the 2^-∆∆CT method. Corn: male and female pools combined; Hay: male and female pools combined; CF: corn females; HF: hay females; CM: corn males; HM: hay males

Fig. 5

Normalized gene expression values: a) LPAR3, b) PLIN4 and c) PLCB4. CF: corn females; CM: corn males; HF: hay females; HM: hay males. ∆CT (CT_{Target gene} – CT_{reference gene ‘RPL19’)} values for individual samples were averaged. Error bars represent the standard error