Report

Early-life nutrition modulates the epigenetic state of specific rDNA genetic variants in mice

See allHide authors and affiliations

Science  29 Jul 2016:
Vol. 353, Issue 6298, pp. 495-498
DOI: 10.1126/science.aaf7040

Mom's diet affects growth

Nutrition is important during development. This appears to be true even in utero, with potential long-lasting effects on adult phenotype and disease. Epigenetic factors are prime suspects in identifying the corresponding molecular mechanism because they can be maintained throughout cell division. Holland et al. show in mice that in utero nutritional deficits influence offspring growth through epigenetic states at multicopy ribosomal DNA elements. This effect is influenced by the genetic variation that naturally exists within the ribosomal RNA.

Science, this issue p. 495

Abstract

A suboptimal early-life environment, due to poor nutrition or stress during pregnancy, can influence lifelong phenotypes in the progeny. Epigenetic factors are thought to be key mediators of these effects. We show that protein restriction in mice from conception until weaning induces a linear correlation between growth restriction and DNA methylation at ribosomal DNA (rDNA). This epigenetic response remains into adulthood and is restricted to rDNA copies associated with a specific genetic variant within the promoter. Related effects are also found in models of maternal high-fat or obesogenic diets. Our work identifies environmentally induced epigenetic dynamics that are dependent on underlying genetic variation and establishes rDNA as a genomic target of nutritional insults.

Exposure to an adverse in utero environment can have a long-lasting influence on adult phenotypes in mammals, a process termed “developmental programming” (1, 2). Consequently, there is great interest in identifying the molecular mechanisms that underlie developmental programming, and, in this regard, modulation of the epigenome has emerged as a potentially key contributing factor (3, 4).

To explore epigenetic mechanisms involved in developmental programming, we employed a maternal protein restriction model (5). Inbred C57BL/6J mice were mated, and G0 females were assigned to either a protein-restricted diet (PR) (8% protein) or a control diet (C) (20% protein) (table S1) until their G1 offspring were weaned. Only male G1s were studied in detail (n = 146). From weaning onward, both G1-PR and G1-C mice were kept on a control diet until they were killed at 16 to 20 weeks. Consistent with previous work, G1-PR males were ~25% lighter than G1-C at weaning (5) (Fig. 1A) (P = 2 × 10−6). PRs also displayed reduced spontaneous locomotor activity (fig. S1) and reduced glucose-stimulated insulin secretion (fig. S2).

Fig. 1 Maternal PR induces a correlation between rDNA methylation and weaning weight.

(A) Weaning weight of G1-PR males (red; 62 individual mice from 17 different litters) was reduced compared with G1-C (black; 84 individual males from 20 different litters) (t test, P = 2 × 10−6 using litter means, and P < 2.2 × 10−6 using individual mice). Small points represent individual mice; larger squares represent the mean of a given G1 litter. (B) RRBS analysis of rDNA in G1 sperm shows that PRs (n = 8) are hypermethylated compared to controls (n = 8). The line represents mean methylation, and points represent individual mice. The rDNA schematic shows the rRNA subunits, transcriptional start site (TSS), external transcribed spacer (ETS), and internal transcribed spacer (ITS). The Rn45S regions identified in the initial RRBS analysis is 98% homologous to the region shaded blue. (C) The correlation coefficient (τ) between weaning weight (ww) and DNA methylation across the rDNA. Highlighted are examples of a positive correlation (green), close to zero (purple), and negative (orange). CpG-133 is circled in blue.

Several studies have shown that developmental programming can perturb DNA methylation profiles (1). We used reduced representation bisulfite sequencing (RRBS) to generate genome-scale, single-base resolution DNA methylomes for eight G1-PR and eight G1-C mice, initially focusing on sperm, because it can be isolated to a high degree of purity. After genome-wide correction, we identified a single 1916–base pair (bp) differentially methylated region (DMR) hypermethylated in G1-PR males that mapped to Rn45s on chromosome 17 (mm10) (table S2). Further analysis revealed that Rn45s displays 98% homology to the 973- to 2883-bp region of the ribosomal DNA (rDNA) consensus (Fig. 1B). rDNA is excluded from genome assemblies because of its multicopy nature. We therefore remapped the RRBS data to the consensus sequence for mouse rDNA (BK000964) and confirmed extensive hypermethylation in PR sperm across the entire promoter and coding regions (~13.5 kb) (Fig. 1B). Directly correlating weaning weight with rDNA methylation levels revealed that G1-PR displayed a significantly greater negative correlation between weaning weight and DNA methylation compared with G1-C (Wilcoxon rank sum test; P < 2.2 × 10−16) (Fig. 1C). This correlation was not confounded by weight or age at death (fig. S3).

In the C57BL/6J genome, rDNA is composed of hundreds of copies in large arrays on chromosomes 12, 15, 18, and 19, but only a subset are actively transcribed (6). Silenced copies are methylated at a CpG site located 133-bp upstream of the 45S-rRNA transcriptional start site (Fig. 1C), and this prevents binding of the transcription factor UBF (upstream binding factor) and assembly of RNA polymerase I (7). We therefore focused on CpG-133 in the rest of the study using high-throughput sequencing (>1000X coverage) of bisulfite polymerase chain reaction (PCR) amplicons (bisPCR-seq). BisPCR-seq analysis of the same samples profiled by RRBS revealed strong concordance between the two methods (fig. S4) (τ = 0.77, P = 1 × 10−5).

As rDNA copies within a single genome are genetically polymorphic (8), we designed the bisPCR-seq amplicon targeting CpG-133 to simultaneously assay previously documented genetic variation at position –104 (C or A, Fig. 2A). (Note that this variant does not overlap a CpG site) (9). CpG-133 methylation levels were substantially lower for the C-variant relative to the A-variant (Fig. 2A), and there was no interaction between C-variant–associated CpG-133 methylation and weaning weight in G1-PR or G1-C sperm (fig. S5). On the other hand, CpG-133 methylation levels of A-variant rDNA (which we denote as CpG-133A) were negatively correlated with weaning weight (Fig. 2B) (τ = –0.43, P = 0.017). Figure 2B incorporates additional males (nine G1-PR and seven G1-C from litters not represented in the RRBS data), reinforcing the negative correlation between weaning weight and total CpG-133 methylation observed in the RRBS data set. BisPCR-seq analysis of in vitro methylated samples confirmed that there was no amplification bias associated with either variant (fig. S6). We also confirmed sperm purity by analysis of several parentally imprinted regions (fig. S7). Analysis of liver using BisPCR-seq revealed a strong correlation with sperm within individual G1-C (fig. S8) (τ = 0.72, P = 0.00028) or G1-PR animals (fig. S8) (τ = 0.54, P = 0.0041). Liver CpG-133A methylation was negatively correlated with weaning weight in G1-PR (τ = –0.46, n = 24, P = 0.0016) but not in G1-C (n = 26) (Fig. 2C). Collectively, these data demonstrate that PR exposure induces not just rDNA hypermethylation but also a linear relationship between a phenotypic outcome (weaning weight) and CpG-133A methylation in sperm and liver, which is maintained into adulthood.

Fig. 2 Diet-induced methylation dynamics are restricted to a specific genetic variant of rDNA.

(A) BisPCR-seq amplicons were generated to simultaneously analyze methylation at CpG-133 (methylation indicated by black circle) and genetic variation at position –104 (A or C) (left panel). CpG-133 methylation levels in sperm for each genetic variant is shown for G1-C (black; n = 15), and G1-PR (red; n = 17). (B) In sperm, methylation levels at A-variant–associated CpG-133 sites (CpG-133A) and weaning weight are not correlated in G1-C (black; n = 15, τ = 0.20, P = 0.30) but are negatively correlated in G1-PR (red; n = 17, τ = –0.43, P = 0.017). (C) In liver, CpG-133A methylation levels and weaning weight are not correlated in G1-C (black; n = 26, τ = –0.14, P = 0.32) but are negatively correlated in G1-PR (red; n = 24, τ = –0.46, P = 0.0016). (D) In sperm, CpG-133A methylation levels are uncorrelated with the percentage of total rDNA copies with an A-variant (%A) in G1-C (black; n = 15, τ = –0.07, P = 0.77) but are positively correlated in G1-PR (red; n = 17, τ = 0.71, P = 1.9 × 10−5).

Further exploration of the bisPCR-seq data revealed interindividual variation in the relative copy number of rDNA harboring the A-variant at position –104, even in an inbred genetic background. This underlying copy number variation (which we denote as %A, i.e., the percentage of A-variant reads relative to total coverage for this amplicon) was positively correlated between sperm and liver of both G1-C (fig. S9) (τ = 0.77, P = 7 × 10−5) and G1-PR animals (fig. S9) (τ = 0.73, P = 3.7 × 10−5). The accuracy of the bisPCR-seq–derived estimates of %A were confirmed by whole-genome resequencing of six mice (fig. S10) (τ = 1, P = 0.0028). Furthermore, CpG-133A methylation correlated positively with %A in G1-PR sperm (Fig. 2D) (τ = 0.71, P = 1.9 × 10−5) and liver (fig. S11) (τ = 0.31, P = 0.034) but not in G1-C sperm (Fig. 2D) or liver (fig. S11). Therefore, early-life PR induces an interdependence between underlying variation in the relative abundance of a specific genetic variant of rDNA and methylation state of this variant at a functionally relevant CpG site.

rDNA copies that lack methylation at CpG-133 have the potential to be transcriptionally active (7). As most methylation is localized to A-variant rDNA, both the level of methylation at CpG-133A and the relative abundance of this variant (i.e., %A) will contribute toward transcriptional competency. This interaction can be represented as the percentage of total rDNA copies that are both A-variant and unmethylated at CpG-133 (which we denote as %AUN). (Note that %AUN is different from simply considering the percentage of CpG-133A that is unmethylated.) As expected, %AUN correlates between the sperm and liver of G1-C and G1-PR mice (fig. S12). To confirm the functional importance of %AUN, we analyzed a regulatory noncoding RNA [promoter-associated RNA (pRNA)] that spans the rDNA promoter (Fig. 3A). pRNA is transcribed from early replicating and unmethylated rDNA copies (10). It functions in trans to recruit nucleolar chromatin remodeling complex and DNA methyltransferase to silenced rDNA copies (11). Using reverse transcription quantitative PCR (RT-qPCR), we generated a pRNA-derived amplicon spanning the genetic polymorphism at position –104 and determined the percentage of A-variant reads after high-throughput sequencing [pRNA(%A)]. The pRNA(%A) reads in liver were consistently and positively correlated with %AUN (Fig. 3B) but not %A (fig. S13). Therefore, %AUN is indicative of transcriptional competency at rDNA.

Fig. 3 Functional consequences of altered rDNA dynamics.

(A) pRNA is transcribed from early replicating rDNA copies (assumed to be unmethylated at CpG-133). Therefore, the percentage of pRNA reads that encode an A at position –104 [pRNA(%A), indicated in blue, right] should reflect the proportion of A-variant rDNA copies that are unmethylated at CpG-133 (%AUN) (B) pRNA(%A) positively correlates with %AUN in both G1-C (black) and G1-PR (red) liver (total, n = 23, τ = 0.61, P = 1.4 × 10−5). (C) %AUN is not correlated with the abundance of 45S-rRNA in liver of G1-C (black; n = 14, τ = 0.03, P = 0.91), but is positively correlated in liver of G1-PR (red; n = 12, τ = 0.52, P = 0.021).

The 45S-rRNA is cotranscriptionally cleaved at position +650 within the 5′ external transcribed spacer, and the first 650 nucleotides (nt) are then rapidly degraded (12). We assessed the abundance of the nascent, uncleaved 45S-rRNA precursor via RT-qPCR targeting the first 650 nt. In the liver of G1-C, 45S-rRNA abundance did not correlate with CpG-133A methylation, %A, or %AUN (Fig. 3C and fig. S14). In PR males, 45S-rRNA levels did not correlate with CpG-133A methylation or %A but correlated positively with %AUN (Fig. 3C) (τ = 0.52, P = 0.021) (fig. S14). Therefore, PR exposure induces a correlation between transcriptional competency and 45S-rRNA levels.

Because rDNA expression is sensitive to nutrient availability (13), the types of effects we describe could be a conserved feature of other nutritional developmental programming models. We identified a recent study in which the authors fed C57BL/6J G0 females a low-fat (LF) or high-fat (HF) diet from 3 weeks before pregnancy until the male G1 offspring were weaned at 3 weeks onto a LF diet until they were killed at 9 weeks (14). Their RRBS analysis of G1 liver did not identify any maternal diet–induced DNA methylation differences. We mapped their raw sequencing reads to rDNA and found that early-life exposure to HF induces CpG-133A hypermethylation in the G1s (Fig. 4A) (P = 0.0098); again, CpG-133C showed lower methylation levels that were not affected by diet. Unfortunately, there were insufficient mice to examine correlations between %A and methylation or weaning weight. Next, we generated bisPCR-seq data for G1 male C57BL/6J mice from a model of maternal obesogenic diet (15) (elevated fat and sugar content). G0 females were fed either control or obesogenic diet 6 weeks before mating until the G1 offspring were weaned at 3 weeks onto a control diet and killed at 6 months. G1 males exposed in utero to obesogenic diet showed hypermethylation at CpG-133A (Fig. 4B) (P = 0.017).

Fig. 4 Maternal high-fat or obesogenic diet induces hypermethylation at CpG-133A.

(A) RRBS raw sequencing reads [obtained from (14)] were mapped to the rDNA consensus. G0 dams that were fed either a low- (ML) or high-fat (MH) diet before conception and up until the G1s were weaned. Data shown here are from the livers of 9-week-old G1 males that were placed on a low-fat diet from weaning up until they were killed.n = 10 mice per group. We only reanalyzed data from (14) for the dietary groups analogous to the design of our PR model (B) G0 dams were fed either a control(C) or obesogenic (O) diet6 weeks before conception and up until the G1s were weaned. bsPCR-seq data shown here are from the livers of 6-month-old G1 males that were placed on a control diet from weaning until they were killed (CC, n = 7; OC, n = 8).

Recently, Shea et al. reported a study in which they exposed male C57BL/6J mice to one of three different diets (PR, HF, or caloric restriction) postweaning (16). They identified substantial interindividual genetic and methylomic variability at rDNA but no consistent diet-induced effects. Although part of the reason for the discrepant conclusions could be that they did not discriminate between the A or C genetic variants, the more likely explanation is differences in developmental timing of the dietary insults because we analyzed exposures spanning only the period between conception and weaning. Previous human epidemiological and animal studies suggest that early life is a critical time when exposures can have long-term phenotypic effects on the offspring (17).

In summary, we have described an example of a mammalian “epiallele” whose epigenetic state is influenced by an interaction between the underlying genotype and early-life environment, and this correlates with transcriptional and phenotypic outcomes. A schematic model of the effects we describe is presented in fig. S15. Our work, in combination with previous demonstrations in flies and yeast (18, 19), identifies rDNA as a genomic target of various nutritional insults that is conserved among nonmammalian and mammalian models. Exploration of such interactions at rDNA in humans could provide novel insights into the molecular basis of some complex phenotypes and diseases.

Supplementary Materials

www.sciencemag.org/content/353/6298/495/suppl/DC1

Materials and Methods

Figs. S1 to S15

Tables S1 to S4

References (2023)

References and Notes

  1. Acknowledgments: This work was supported by the following grants and fellowships: Biotechnology and Biological Sciences Research Council, UK(BB/M012494/1) to V.K.R. and (BB/G00711/X/1) to V.K.R. andC.G.; and a Research Council UK Academic Fellowship to M.L.H.R.L. is supported by EU-FP7 BLUEPRINT. S.E.O. is supported by the British Heart Foundation (FS/12/64/30001) and the Medical Research Council (MC_UU_12012/4). This research used Queen Mary’s MidPlus computational facilities, supported by Queen Mary University of London Research-IT and funded by Engineering and Physical Sciences Research Council grant EP/K000128/1. We thank King’s College London FWB Genomics Centre and Barts and The London Genome Centre for performing high-throughput sequencing. Author contributions: M.L.H., C.G., P.W.C., V.K.R., A.A.M.C., and E.L. performed all experiments. R.L. and G.C. conducted the bioinformatic analyses. V.K.R. and S.E.O. provided reagents and contributed to experimental design. M.L.H., R.L., and V.K.R. conceived the study and prepared the manuscript. All authors discussed the results and interpretation and approved the final manuscript. Data deposition statement: Whole-genome sequencing data and pRNA sequencing were submitted to BioProject under ID PRJNA293403. RRBS and BisPCR-seq were all submitted to the National Center for Biotechnology Information’s Gene Expression Omnibus under accession GSE72610. All data will be linked under the BioProject ID. Competing interest declaration: The authors declare no conflict of interest.
View Abstract

Stay Connected to Science

Navigate This Article