The redox basis of epigenetic modifications: from mechanisms to functional consequences

Abstract

Epigenetic modifications represent mechanisms by which cells may effectively translate multiple signaling inputs into phenotypic outputs. Recent research is revealing that redox metabolism is an increasingly important determinant of epigenetic control that may have significant ramifications in both human health and disease. Numerous characterized epigenetic marks, including histone methylation, acetylation, and ADP-ribosylation, as well as DNA methylation, have direct linkages to central metabolism through critical redox intermediates such as NAD(+), S-adenosyl methionine, and 2-oxoglutarate. Fluctuations in these intermediates caused by both normal and pathologic stimuli may thus have direct effects on epigenetic signaling that lead to measurable changes in gene expression. In this comprehensive review, we present surveys of both metabolism-sensitive epigenetic enzymes and the metabolic processes that may play a role in their regulation. To close, we provide a series of clinically relevant illustrations of the communication between metabolism and epigenetics in the pathogenesis of cardiovascular disease, Alzheimer disease, cancer, and environmental toxicity. We anticipate that the regulatory mechanisms described herein will play an increasingly large role in our understanding of human health and disease as epigenetics research progresses.

FIG. 1.

Epigenetics as an integrating mechanism. Epigenetic processes can be conceptualized as a common effector for multiple different inputs to affect chromatin structure and gene expression. Epigenetics can thus be considered as an integrating mechanism that takes into account signals from diverse processes to produce a phenotypic outcome.

FIG. 2.

Example modifications of histones H3 and H4. Post-translational modification of histones can produce myriad different combinations and effects on gene expression. Here we provide illustration of H3 and H4 N-terminal tails and the locations that can be methylated or acetylated as examples. Critically, these marks do not necessarily exist independently of one another, leading to remarkable combinatorial diversity and signaling subtleties. Arginine demethylation is specially denoted as it may be symmetric or asymmetric around the guanidino moiety, with different transcriptional outcomes. See Wysocka et al. (330) for further review.

FIG. 3.

Histone lysine and arginine methyltransferase mechanisms. (A) Illustration of lysine methylation reaction utilizing SAM as a methyl donor. Individual enzymes determine the specificity for individual histone lysines. Lysines may be mono-, di-, or tri-methylated by serial iteration of methyltransferase reactions (not pictured). (B) Arginine methyltransferase reaction. These enzymes also use SAM as a methyl donor, and can catalyze the formation of symmetric or asymmetric dimethylarginine (not pictured) through serial addition of methyl groups. For detailed review of these topics, see Spannhoff et al. (287). SAM, S-adenosyl methionine.

FIG. 4.

Histone lysine demethylase mechanisms. (A) JmjC demethylase mechanism. JmjC proteins catalyze the hydroxylation of methyllysine residues in a 2-OG and Fe(II)-dependent manner, releasing succinate and CO₂. In a second, nonenzymatic step, formaldehyde is spontaneously released after the decomposition of the N-hydroxymethyl moiety. (B) LSD1 demethylase mechanism. LSD1 catalyzes the oxidation of the C-N methylamine bond in an FAD-dependent mechanism, producing an unstable intermediate that decomposes to release formaldehyde. As a consequence of this mechanism, LSD1 cannot eliminate trimethyllysine, whereas jmjC demethylases can. 2-OG, 2-oxoglutarate; jmjC, jumonji-domain containing; LSD1, lysine specific demethylase 1; Succ, succinate.

FIG. 5.

HAT mechanism. Broadly, HAT enzymes each use a similar mechanism, with acetyl-CoA serving as the acetyl donor. Notably, HAT enzymes are not generally specific for lysine residues on histones, but rather are responsible for pan-acetylation reactions. HAT, histone acetyltransferase.

FIG. 6.

Histone deacetylase mechanisms. (A) HDAC class I, II, IV mechanism. Classical HDAC enzymes utilize a redox-active metal [Zn(II) or Fe(II)] to coordinate the hydrolysis of acetate from lysine residues. (B) Sirtuin enzyme mechanism. Sirtuins remove acetyl groups at the expense of NAD, releasing nicotinamide and the unique metabolite O-acetyl-ADP-ribose. Both HDAC and sirtuin enzymes may catalyze the deacetylation of substrates other than histones. HDAC, histone deacetylase; OAADPr, 2′-O-acetyl-ADP-ribose; Nam, nicotinamide.

FIG. 7.

PARP activity and histone shuttling. (A) Canonical model for PARP family activity. PARP enzymes consume NAD⁺ to generate the poly ADP-ribose mark, releasing nicotinamide as a byproduct. This can target both glutamate and lysine residues on proteins. (B) Histone shuttling diagram. PARP1 is a nuclear PARP that recognizes DNA damage, promoting autocatalysis that results in the poly (ADP-ribosyl)ation of PARP1 itself. Poly(ADP-ribose) is a strong competitor for DNA binding to histones, which shuttles histones away from DNA lesions and allows repair enzymes access to the damaged region. PARP, poly(ADP-ribose) polymerase.

FIG. 8.

DNA methyltransferase and putative demethylase. Cytosine methylation is catalyzed by the action of the DNMT enzymes in a context-dependent fashion (see text for further information). The recently discovered Tet1 protein is a 2-OG and Fe(II)-dependent dioxygenase that catalyzes the formation of hydroxymethylcytosine, which may represent a critical step in active oxidative DNA demethylation or may itself comprise a novel epigenetic mark. DNMT, DNA methyltransferase; dR, deoxyribose.

FIG. 9.

The citric acid cycle and key metabolic intersections. The citric acid cycle is a critical intersection between catabolic and anabolic processes that has many implications for epigenetic events. To supplement the basic 8-reaction cycle, we have included the utilization of 2-OG as a substrate by the 2-OG and Fe(II)-dependent dioxygenases, as well as the ability of both succinate and fumarate to inhibit these enzymes. Additionally, superoxide produced during conditions of oxidative stress may interfere with both SDH and aconitase, leading to different substrate balances that could affect downstream processes. SDH, succinate dehydrogenase.

FIG. 10.

SDH and regulation of dioxygenases. SDH is an inner-membrane-bound protein in the mitochondrion and its activity may be altered by a number of stressors such as thenoyltrifluoroacetate and superoxide. In this case, succinate buildup would occur, leading to increased succinate levels in the cytoplasm which could negatively impact the activity of multiple 2-OG and Fe(II)-dependent dioxygenases, including the jmjC histone demethylases, the Tet1 methylcytosine hydroxylase (MeCH), and the PHD HIF hydroxylases. Regulation of any of these processes could have potentially large effects on downstream gene expression. HIF, hypoxia-inducible factor; PHD, prolyl hydroxylase domain; TTFA, thenoyltrifluoroacetate.

FIG. 11.

2-Hydroxyglutarate metabolism and potential epigenetic effects. L-2-HG is produced normally in low amounts by malate dehydrogenase in the mitochondrial matrix, and is normally metabolized back to 2-OG by the L-2-HG dehydrogenase. D-2-HG is an oncometabolite produced by the indicated isocitrate dehydrogenase mutation identified from malignant gliomas, and is also produced from various glyoxalate detoxification pathways (not shown). A D-2-HG dehydrogenase metabolizes D-2-HG back to 2-OG. When 2-HG production exceeds the capacity of the dehydrogenases to convert back to 2-OG, we hypothesize that 2-HG may interfere with the 2-OG and Fe(II)-dependent dioxygenases, thereby perturbing epigenetic modifications governed by jmjC and Tet1 proteins. This may also represent a normal signaling axis (for details, see text). L-2-HG, L-2-hydroxyglutarate.

FIG. 12.

Methionine cycle and glutathione biosynthesis. Methionine serves as the starting material for a variety of different biologically important compounds, including SAM, SAH, and cysteine. Under normal conditions, regular dietary intake of methionine offsets losses to the transsulfuration pathway. Under conditions of oxidative stress, when the demand for glutathione may exceed the capacity of the cell to produce it, there is a net reduction of the methionine pool, leading to an insufficiency of SAM and SAH that likely has significant effects on downstream epigenetic events. BHMT, betaine-homocysteine methyltransferase; MTase, generalized methyltransferase. SAH, S-adenosyl homocysteine.

FIG. 13.

NAD⁺ biosynthesis and salvage pathways. NAD⁺ is a critical metabolite for numerous metabolic processes, and both de novo biosynthesis and salvage of NAD⁺ byproducts are important for maintaining overall pools of NAD. Importantly, both the sirtuins and PARPs consume NAD⁺ and produce nicotinamide as a byproduct, which suggests that the salvage pathway is particularly important when considering the function of these enzymes. NaPRT, nicotinic acid phosphoribosyl transferase; NMNAT, nicotinic acid/nicotinamide mononucleotide adenylyltransferase; QPRT, quinolinate phosphoribosyl transferase.

FIG. 14.

Relationship between PARP and sirtuin enzymes in a local environment. As both PARPs and sirtuins consume NAD, the relative activity of either enzyme family affects the other within a local environment such as the nucleus. When PARPs are highly active, sirtuin activity is likely blunted, and vice-versa, suggesting a potential regulatory linkage that was initially proposed by Jie Zhang in 2003 (345).

FIG. 15.

The labile iron pool and iron homeostasis. Before being utilized in intracellular processes, iron that is imported into the cell passes through the labile iron pool. This transient and normally low-concentration pool is a critical player in redox homeostasis, because labile iron may participate in harmful Haber-Weiss and Fenton reactions, leading to ROS production. Consequently, cells have strict regulatory controls on labile iron to ensure that it does not damage local processes. Tf, transferrin; DMT, divalent metal transporter.

FIG. 16.

A global illustration of oxidative stress and epigenetics. Oxidative stress may contribute to epigenetic perturbation through each of the means discussed in this review. The number of ways in which epigenetics may potentially be affected by metabolism strongly suggests that study of the overall epigenetic status will become increasingly important as we move toward better understanding of disease processes. Numbers adjacent to individual sections on the chart indicate the figures that focus more exclusively on each topic. It is also critical to note that some of the arrows indicating the trending directions of individual components may be reversed under certain conditions. For example, some generalized oxidative stress would move the arrow of the NAD⁺/NADH ratio up, not down as we have suggested. However, when one considers specific oxidative stressors such as alcohol, we can observe changes in the indicated directions.

FIG. 17.

A model for metabolic-epigenetic synergism in cancer. (A) In the normal condition, intermediates produced through metabolism may influence and maintain appropriate epigenetic regulation of critical metabolic loci in the genome. (B) In cancer, there is strong evidence for metabolic perturbation and mitochondrial dysfunction. This may promote deregulation of the epigenetic signals in the nucleus, leading to a synergistic feedback mechanism that promotes increasing metabolic and epigenetic derangement.