A number of disease states in humans are the result of subtle changes in gene expression caused by alterations in chromatin structure (epigenetic modifications; Robertson & Wolffe 2000
). Epigenetic marks such as the covalent methylation of cytosine residues in DNA are set in the chromatin during development and determine the accessibility of a particular gene to the transcriptional machinery (Spiegelman & Heinrich 2004
). By regulating gene expression through changes in the promoter region, these epigenetic modifications represent another mechanism for the nutritional regulation of gene expression.
Epigenetic modifications may result from the direct effects of changes in metabolism as the process of chromatin remodelling depends on a number of products derived from intermediary metabolism such as S-adenosyl methionine (SAM), acetyl CoA and nicotinamide adenine dinucleotide (NAD(+)). The modification of cytosine residues with methyl groups derived from SAM serves two purposes, transcriptional repression and genome defence (Grace Goll & Bestor 2005). There is a growing body of evidence showing that transcriptional repression and genome defence is perturbed by changes in metabolism. Embryo culture techniques in farm animals frequently expose the preimplantation embryo to an inappropriate nutritional environment in vitro, leading to defective epigenetic programming and a number of developmental abnormalities collectively known as the large offspring syndrome (Young et al. 2001). Studies have revealed that the expression of a number of key genes is altered in these large offspring (Young & Beaujean 2004). Defects in epigenetic programming can also affect metabolism, the surviving offspring of mice produced by nuclear transfer (where the transferred nucleus undergoes extensive epigenetic reprogramming) exhibit an obese phenotype (Tamashiro et al. 2002). Many of the cytosine residues modified by methylation reside within parasitic DNA elements or retrotransposons, such as endogenous retroviruses. One of the best-studied examples is the Agouti mouse (Wolff et al. 1998). In this animal an endogenous retrovirus-like transposon sequence is inserted close to the gene coding for the Agouti protein. Normally a cryptic promoter within the retrotransposon is silenced by methylation allowing normal tissue-specific and regulated agouti expression. However, if this site is undermethylated the promoter is active and drives constitutive ectopic expression of the agouti gene, leading to yellow coat colour and obesity. It has been shown that the methylation status of these inserted viral DNA sequences can be modified by the methionine, folic acid and choline content of the maternal diet, illustrating that the extent of methylation is determined by events occurring in utero (Cooney et al. 2002). Adding additional methyl donors to the maternal diet increases methylation of the retrotransposon, suppresses ectopic gene expression and improves the outcome for the offspring. However, the wider relevance of this model still unclear and it is not known whether nutritional regulation of similar retrotransposon sequences is responsible for disease in humans.
The methylation of cytosine bases in DNA is tightly associated with modification of histone tails and other changes in chromosome structure (reviewed by Santos & Dean 2004). DNA is assembled on core histones; the tails of which are exposed on the nucleosome surface, where they are subject to a variety of enzyme-catalysed, post-translational modifications. The acetylation and methylation of histones, mediated by histone acetyl transferases and histone methyl transferases, changes the structure of the nucleosome and allows other proteins access to the DNA (Hermanson et al. 2002). A group of histone deacetylases reverses the acetylation of histones in a reaction that requires NAD(+) (Blander & Guarente 2004). These processes are essential for the differentiation of cells such as hepatocytes and adipocytes. It has been suggested that methylation and acetylation of histones introduces a ‘histone code’ which acts as a form of cellular memory influencing the subsequent binding and therefore function of transcriptional activators responsible for the differentiated phenotype (Turner 2002).
Chromatin remodelling is particularly important in the oocyte and early embryo. In these cells the pattern of co-activator expression is different from that observed in somatic cells, reflecting the specialised events occurring during early development (Zheng et al. 2005). In particular the modification of chromatin with poly ADP-ribose (derived from NAD(+)) is essential for normal preimplantation development (Imamura et al. 2004). It is interesting to speculate that this requirement for NAD(+) during the early stages of development may be associated with the dependence of the pre-implantation embryo on pyruvate metabolism (Houghton & Leese 2004). The NAD(+) dependent deacetylases are also essential during the later stages of fetal development where loss of activity leads to reduced fetal growth and defects in the lung and pancreas (McBurney et al. 2003). In the liver of rat fetuses, the nutritional stresses produced by uteroplacental insufficiency produce specific changes in the association of acetylated histones with the promoters of genes dependent on PPAR (Fu et al. 2004). Nutrient sensitive transcriptional activators, such as the C/EBPs and PPARs, are able to determine local chromatin structure through interactions with co-activator proteins. This provides a second indirect route for nutritionally mediated epigenetic programming of gene expression (Jia et al. 2004).