X-ray crystallography reveals a new link between metabolism and epigenetics

De novo lipogenesis (DNL) is the metabolic pathway used primarily by the liver and adipose (fat-storing) tissue to synthesise fatty acids from excess carbohydrates or other precursors [1]. An increased rate of DNL is observed in cancer and in many characteristically "Western" diseases of affluence, such as metabolic disorder, type 2 diabetes, chronic liver disease and cardiovascular disease [2]. In response to metabolic changes, the liver and adipose tissue regulate the expression of genes encoding proteins important for de novo lipogenesis (DNL) through a variety of mechanisms. One of these processes involves the introduction of acetyl groups into histones, the proteins that package DNA, without altering their DNA sequence, thereby affecting gene expression. However, the molecular details that underlie this level of regulation are poorly understood.

In this work, an unexpected breakthrough has come from an initially unrelated investigation into the hyper-acetylation of chromatin during mammalian sperm development. Among the most highly abundant factors in male germ cells that bind acetyl-coenzyme A (acetyl-CoA), a key metabolite for histone acetylation, the protein NME1 and its closely related variant NME2 were identified. These ubiquitous metabolic enzymes, collectively called NME1/2, are best known for balancing proportions of nucleotides in the cell. They perform this role by transferring a phosphate group from a nucleoside triphosphate, typically ATP, to a nucleoside diphosphate. The surprising discovery of NME1/2 as a factor that also binds to acetyl-CoA prompted a more detailed investigation via a combination of structural, biophysical, genetic and epigenetic approaches.

Click image to enlarge

Fig. 1: Ligand recognition by NME1. a) Crystal structure of NME1, a hexameric enzyme with D3 symmetry. b) Details of the recognition of the nucleotide moiety of acetyl-CoA (AcCoA). Direct and water-mediated hydrogen bonds between NME1 and acetyl-CoA are shown as red and grey dashed lines, respectively. c) Superimposition of the acetyl-CoA and ADP ligands showing the distinct binding mode of acetyl-CoA, achieved through a rotation of the adenine base and a repositioning of the α- and β-phosphate groups.

Crystallographic data collected at beamline ID30A-1 revealed that NME1 recognises acetyl-CoA through the same binding site that recognises ATP and other nucleotides. Despite the similarity with ATP, NME1 recognises the nucleotide moiety of acetyl-CoA by a distinct binding mode that is critically dependent on the ligand's 3' phosphate group (Figure 1). Meanwhile, native mass spectrometry combined with biochemical assays showed that acetyl-CoA binding by NME1 is inhibited by ATP-mediated phosphorylation and is hence sensitive to the cellular energy status. Strikingly, when fed a high-fat diet, a mouse knockout model with greatly reduced NME1/2 levels exhibited excessive triglyceride synthesis and liver steatosis (fat buildup). In liver cells, NME1/2 was found to mediate a gene transcriptional response to a high-fat diet via the targeted acetylation of histones that inhibits fatty acid accumulation by repressing genes encoding key transcription factors involved in DNL and fatty acid metabolism.

These findings identify NME1/2 as a key regulator of the competing processes of histone acetylation and fatty acid synthesis. The results shed important light on how DNL is regulated and reveal a previously unsuspected level of cross-talk between metabolic and epigenetic pathways.

Principal publication and authors
Nucleoside Diphosphate Kinases 1 and 2 regulate a liver protective response to a high-fat diet, D. Iuso (a), I. Garcia-Saez (b), Y. Couté (c), Y. Yamaryo-Botté (a), E. Boeri Erba (b), A. Adrait (c), N. Zeaiter (d), M. Tokarska-Schlattner (d), Z. Macek Jilkova (a,e), F. Boussouar (a), S. Barral (a), L. Signor (b), K. Couturier (d), A. Hajmirza (a), F. Chuffart (a), E. Bourova-Flin (a), A.-L. Vitte (a), L. Bargier (f), D. Puthier (f), T. Decaens (a,e), S. Rousseaux (a), C. Botté (a), U. Schlattner (g), C. Petosa (b), S. Khochbin (a), Sci. Adv. 9, eadh0140 (2023); https://doi.org/10.1126/sciadv.adh0140
(a) Univ. Grenoble Alpes, CNRS, INSERM, Institute for Advanced Biosciences, La Tronche (France)
(b) Univ. Grenoble Alpes, CNRS, CEA, Institut de Biologie Structurale, Grenoble (France)
(c) Univ. Grenoble Alpes, INSERM, CEA, BioSanté, Grenoble (France)
(d) Univ. Grenoble Alpes, INSERM, Laboratory of Fundamental and Applied Bioenergetics, Grenoble (France)
(e) CHU Grenoble Alpes, Service d’hépato‐gastroentérologie, La Tronche (France)
(f) Aix Marseille Université, INSERM, TAGC, TGML, Marseille (France)
(g) Univ. Grenoble Alpes, INSERM, Institut Universitaire de France, Laboratory of Fundamental and Applied Bioenergetics, Grenoble (France)

References
[1] M. Wallace, C.M. Metallo, Semin. Cell Dev. Biol. 108, 65-71 (2020).
[2] F. Ameer et al., Metabolism 63, 895-902 (2014).