Transcription factor acetylation

Much has been written about histones and regulation of gene expression and the roles that environmental factors, including diet and specific nutrients, play as modifiers of regulation⁽Reference Jaenisch and Bird^7–Reference Fukuda, Sano and Muto⁹⁾ and which histone modifications are truly epigenetic⁽Reference Kouzarides¹⁰⁾. While much attention has focused on methylation as a post-translational modifier of histones, recently major strides have been made in understanding how histone acetylation, also a post-translational modification, influences gene expression⁽Reference Fukuda, Sano and Muto⁹⁾. The concept of histone acetylation as a factor in gene regulation was first advanced 35 years ago⁽Reference Allfrey, Faulkner and Mirsky¹¹⁾. Acetylation of the lysine residues of histone tails removes positive charges, thus decreasing histone–DNA affinity. This process yields easier access for RNA polymerase II and transcription factors to promoter regions. Targeted regions of chromatin are acted on by histone acetyltransferases and histone deacetylases to regulate transcription of specific genes.

Recently, an elegant example of histone acetylation has been shown to link acetyl-CoA production from glycolysis to the expression of genes that influence glucose metabolism⁽Reference Wellen, Hatzivassiliou and Sachdeva¹²⁾. Focusing specifically on ATP-citrate lyase (ACL), an enzyme that generates acetyl-CoA from citrate, it was found that, surprisingly, ACL is present in both cytoplasm and the nucleus. This finding suggests that citrate may diffuse into the nucleus, where it may facilitate acetyl-CoA production. Silencing of ACL with small interfering RNA (siRNA) decreases the acetylation of numerous histones. The effect on histone acetylation is prevented by the addition of acetate, which is an acetylation substrate used by histone acetyltransferases. ACL siRNA suppression reduces the expression of the glucose transporter GLUT4 as well as genes needed for glucose metabolism, i.e. hexokinase, phosphofructokinase-1 and lactate dehydrogenase A. Other genes not involved in glucose metabolism are not influenced by ACL siRNA. Acetate also reverses effects on those genes that are inhibited by ACL siRNA. Finally, immunoprecepitation studies with the GLUT4 promoter using antibodies for two histone acetylases (Ac-H3 and Ac-H4) have shown that histone acetylation at the GLUT4 promoter is reduced when ACL production is inhibited with siRNA. GLUT4 promoter activity is rescued with acetate. It was not possible to fully rule out that the observed effects on the GLUT4 gene and those for the glucose-metabolizing enzymes (hexakinase 2, phosphofructokinase and lactate dehydrogenase A) are produced by global changes in acetylation. Nevertheless, histone acetylation by ACL at promoters relevant to a particular metabolic pathway opens a new avenue of inquiry.

The experiments on ACL acetylation have all been carried out with cells in culture (see Rathmell & Newgard⁽Reference Rathmell and Newgard¹³⁾). However, experiments in animals are lending support for the role of acetylation in nutrient-responsive transcription factor activation. A notable example is the transcriptional activation of the fatty acid synthase gene on feeding⁽Reference Wong, Chang and Hudak¹⁴⁾. In this model a transcription factor upstream, transcription factor 1, associates with histone deacetylase 9, which acts as a transcriptional repressor through upstream transcription factor 1 deacetylation. On feeding, upstream transcription factor 1 is acetylated and activated, thus allowing recruitment of accessory factors and increased fatty acid synthase expression. The results in relation to fatty acid synthase gene expression suggest that histone deacetylases may influence transcription factor access to specific promoters of other genes that respond to nutrients. In this context histone deacetylase 9 has been shown to associate with numerous transcription factors⁽Reference Mejat, Ramond and Bassel-Duby¹⁵⁾. This mode of nutrient–transcription factor responsiveness is illustrated further in Fig. 1(B). It needs to be further explored through further research.

Nutrient transport-influenced transcription factors

Homeostatically-influenced transport of nutrients control cell function through provision of needed substrates. Recent evidence from experiments with various cell types and integrative systems suggest targeted nutrient transport has physiological consequences. Of particular interest for the authors' laboratory has been the linkage of Zn transport with cellular signalling. The idea that Zn could act as a specific second messenger was initially presented in 1984⁽Reference Williams¹⁶⁾. Cellular compartmentalization and transport are central to the signalling role of Zn. Zn transporters are from either the ZnT family (ten members) or ZIP family (fourteen members)⁽Reference Lichten and Cousins¹⁷⁾. The most-clearly-defined evidence for transporter-mediated signalling roles for Zn have been derived for the ZIP family. Specifically, ZIP6, ZIP8 and ZIP10 have each been identified with signalling processes⁽Reference Yamashita, Miyagi and Fukada^18–Reference Murakami and Hirano²⁰⁾. Of the three transporters, nutrient–transcription factor communication has been shown for ZIP8 in transcriptional activation of the interferon-γ gene⁽Reference Aydemir, Liuzzi and McClellan¹⁹⁾.

Peripheral blood mononuclear cells provide a resource to evaluate the effects of Zn supplementation on expression of specific genes in human subjects. Using T-cells purified by negative magnetic selection and a procedure that simulates antigen presentation, Zip8 mRNA has been identified as the most responsive Zn transporter transcript on cell activation⁽Reference Aydemir, Liuzzi and McClellan¹⁹⁾. The concurrent activation-induced increase in interferon-γ expression is further stimulated by supplemental Zn. Suppression of ZIP8 by siRNA silencing markedly reduces interferon-γ. ZIP8 has been found by confocal microscopy to be localized to the lysosome. On activation Zn is transported from the acidic environment of the lysosome, as shown using a Zn-responsive fluorophore (FluoZin 3). Analysis of the T-cell activation pathway involving calcineurin reveals that activation of the transcription factor cAMP-response element-binding protein, i.e. phosphorylated cAMP-response element-binding protein, is responsive to Zn. In addition, activation of cAMP-response element-binding protein is inversely correlated with calcineurin phosphatase activity. This activity is very sensitive to inhibition by Zn. It is concluded that ZIP8 produces a release of lysosomal Zn into the cytoplasm thereby inhibiting dephosphorylation by calcineurin and maintaining the cAMP-response element-binding protein transcription factor in its active phosphorylated form⁽Reference Aydemir, Liuzzi and McClellan¹⁹⁾.

It is proposed that this experimental T-cell model is but one example of how, through mediated transport, Zn may communicate with a transcription factor. In this particular example, the influence could be through a classical enzyme inhibition mechanism involving a transcription factor that is active in its phosphorylated form. Similar roles for Zn and Zn transport influencing receptor-initiated events has been proposed⁽Reference Beyersmann and Haase²¹⁾ as shown in Fig. 1(C).

Nutrient-influenced gene repression via response element placement

Homeostatic balance in nutrient utilization implies that positive and negative control points are needed. Tight control of cellular Zn flux is of paramount importance because, unlike some other nutrients, Zn is not stored. The identification of the first mammalian Zn transporter, ZnT1, was through its ability to rescue cells in conditions of high extracellular Zn⁽Reference Palmiter and Findley²²⁾. The positive response to Zn was traced to MTF-1. Indeed, nutritional transitions to an elevated Zn supply increase ZnT1 expression in the intestine and pancreas⁽Reference McMahon and Cousins^23, Reference Liuzzi, Bobo and Lichten²⁴⁾. Most of the twenty-four genes in the ZnT and Zip families are refractory to changes in dietary Zn intake in these same tissues⁽Reference Liuzzi, Bobo and Lichten²⁴⁾.

An expanding base of information has led to an analysis of most of these transporter genes in many tissues and their response to conditions of both dietary Zn restriction and Zn supplementation. Of particular note is the up-regulation of Zip10 expression with Zn restriction in mice⁽Reference Lichten, Liuzzi and Cousins^25, Reference Ryu, Lichten and Liuzzi²⁶⁾.The experiments have shown that an increase in Zip10 expression occurs in liver and brain and is also detected in the erythrocyte membrane at the protein level on Zn restriction. Those observations merge well with those made with the Zip10 orthologue in zebrafish (Danio rerio) gill⁽Reference Zheng, Feeney and Kille²⁷⁾. Sequence analysis of the Zip10 promoter has revealed no KLF4 binding site, which has been implicated in the similar up-regulation of Zip4⁽Reference Liuzzi, Guo and Chang²⁸⁾. However, there is a conserved metal response element for mouse, human and zebrafish Zip10. In each species this response element is placed downstream from the transcription start site of Zip10.

It has been found that MTF-1 is associated with the Zip10 promoter in a manner that is proportional to cellular Zn status⁽Reference Lichten, Liuzzi and Cousins²⁵⁾. In this context MTF-1 is responding to Zn occupancy in the classical fashion, i.e. the transcription factor translocates to the nucleus on binding Zn in the cytoplasm. Silencing of MTF-1 with siRNA increases Zip10 expression. The most likely explanation for this outcome is that nuclear MTF-1 acts as a repressor of Zip10 under conditions of normal Zn status. Chromatin immunoprecipitation analysis of RNA polymerase II binding to the Zip10 gene has been conducted with antibodies for Ser2 and Ser5 phosphorylated polymerase II forms⁽Reference Lichten, Liuzzi and Cousins²⁵⁾. The results show that Zn-restricted conditions allow active transcription with clear elongation activity of RNA polymerase II. In Zn-supplemented conditions elongation does not occur. Placement of the metal response element at +17 of the Zip10 gene allows MTF-1 binding, in response to Zn-stimulated nuclear translocation, but then prevents RNA polymerase II movement from the transcription start sites of the Zip10 gene. In this fashion MTF-1 can act as an activator, e.g. for metallothionein and ZnT1, but because of metal response element placement also acts as a repressor of Zip10. A similar conclusion has been reached to explain the inverse relationship between Zn and zebrafish Zip10⁽Reference Zheng, Feeney and Kille²⁷⁾. This mode of nutrient regulation is shown in Fig. 1(D).

Summary

The examples provided here have been derived from recent literature and point to a new level of complexity in nutrient–gene communication. Nutrition and gene regulation in this context has been commented on previously⁽Reference Cousins²⁹⁾. However, the application of new technologies such as siRNA silencing, quantitative PCR, global and targeted microarrays, cell transfection and reporter assays, plus the emergence of the understanding of histone acetyltransferases and deacetylases in transcriptional activation and repression, have given a far greater breadth to this area of nutritional science research than was possible a decade ago. Clearly, gene-regulation studies have a place in understanding the challenges and approaches directed at over-and undernutrition, the theme of the Nutrition Society meeting.