The Metabolic Baton: Conducting the Dance of m⁶A Writing and Erasing

Abstract

The RNA modification m⁶A plays an important role determining the functional output of gene expression programs. Throughout the transcriptome, levels of m⁶A are tightly regulated by the opposing activities of methyltransferases and demethylases as well as the interaction of modified transcripts with m⁶A-dependent RNA binding proteins that modulate transcript stability, often referred to as writers, erasers, and readers. The enzymatic activities of both writers and erasers are tightly linked to the cellular metabolic environment, as these enzymatic reactions rely on metabolism intermediaries as co-factors. In this review we highlight examples of intersection between metabolism and m⁶A-dependent gene regulation and discuss the different contexts where this interaction plays important roles.

Introduction

RNA post-transcriptional modifications represent a crucial layer of non-coding information embedded in RNA transcripts. These modifications have the capacity to alter the physical properties of the modified transcript and to modulate interactions between the modified RNA and trans regulatory factors, such as RNA binding proteins [1]. N6-methyladenosine (m⁶A), is the most abundant modification in eukaryotic mRNAs [2]. The stoichiometry of m⁶A modification at each site can differ significantly across different cell types, cells exposed to different culture conditions, and even cells in the same population [3–10]. This variability suggests a dynamic mechanism that enables gene regulatory pathways to promptly adapt to changes in the cellular environment. Additionally, the m⁶A modification is also present in the cap adjacent N6,2’-O-dimethyladenosine (m⁶A_m). The levels of m⁶A and m⁶A_m are controlled in part by the opposing activities of methyltransferases, often referred to as writers, and demethylases, known as erasers (Figure 1). The presence of m⁶A on mRNA and lncRNAs is interpreted by readers, RNA binding proteins whose interaction with RNA is influenced by the presence of m⁶A [1]. The interaction between mRNAs and m⁶A readers plays a crucial role in regulating various aspects of RNA metabolism, such as splicing, nuclear export, stability, and translation [1]. The majority of m⁶A modifications added to RNA Polymerase II transcripts is catalyzed by a large nuclear RNA methyltransferase complex (Figure 1). The catalytic core of this complex is composed of a heterodimer of Methyltransferase-like 3 (METTL3) and Methyltransferase-like 14 (METTL14). While METTL3 is responsible for the catalytic activity, METTL14 plays a crucial role as a substrate-binding scaffold, enhancing the methyltransferase activity of the complex [11–13]. The catalytic core interacts with a regulatory complex, [14–17] which ensures proper cellular localization and modulates interactions between target RNAs and the catalytic core. Additionally, m⁶A can also be deposited on mRNAs in internal positions by Methyltransferase-like 16 (METTL16), which modifies the small nuclear RNA U6 and the mRNA coding for the Methionine Adenosyltransferase 2A (MAT2A) protein [18–20] (Figure 1). Phosphorylated CTD-interacting factor 1 (PCIF1) is responsible for the methylation at the N6 position of 2’-O-methylated adenosines adjacent to the cap to form m⁶A_m [21–23] (Figure 1). The methyl group in m⁶A can be removed by the enzymes fat mass and obesity-associated protein (FTO) and ALKB homolog 5 (ALKBH5) (Figure 1). FTO can remove the methyl group from both m⁶A and m⁶A_m [24,25] while ALKBH5 catalyzes the removal of internal m⁶A only [26]. The enzymatic activities of both ‘writers’ and ‘erasers’ requires cellular metabolites as co-factors, establishing a significant connection between cellular metabolism and m⁶A-dependent regulation of gene expression [27]. In this review, we discuss examples that highlight the intricate interplay between the cellular metabolic environment and the opposing activities of m⁶A writers and erasers. We aim to elucidate how changes in the cellular environment can drive alterations in gene expression, shedding light on the intricate regulatory mechanisms governing cellular processes.

Figure 1:

Schematic representation illustrating the regulation of N6-methyladenosine (m6A) levels on mRNA by methyltransferases (writers) and demethylases (erasers). METTL3/METTL14 from the core of the methyltransferase complex for the majority of m⁶A sites on mRNA. METTL3/METTL14 interact with other cellular proteins, including RNA binding motif protein 15 (RBM15), Hakai, Virilizer, zinc finger CCCH domain-containing protein 13 (Zc3h13), and Wilms Tumor 1-associated protein (WTAP). Phosphorylated CTD-interacting factor 1 (PCIF1) methylates at the N6 position of 2’-O-methylated adenosines adjacent to the cap to form m⁶A_m. METTL16 deposits m⁶A on the small nuclear RNA U6 an internal position of some mRNAs, such as the mRNA coding for the MAT2A protein. Conversely, the demethylase FTO can remove methyl groups from both m⁶A and m⁶A_m, while demethylase ALKBH5 can only remove methyl groups from internal m⁶A residues. This image was created with BioRender (https://biorender.com/).

The links between metabolism and writer activity

The addition of m⁶A to RNA transcripts by m⁶A writers requires S-Adenosylmethionine (SAM), the cell’s principal methyl donor and the second most used enzyme substrate after adenosine triphosphate (ATP) [28]. The byproduct of the reaction, S-Adenosyl-L-homocysteine (SAH), acts as an allosteric inhibitor of METTL3 [29]. SAM levels are tightly regulated within the cell, and disruption of SAM homeostasis is known to influence methyltransferase activity toward multiple substrates involved in gene expression regulation [30]. This includes methylation of DNA and histones, influencing transcriptional processes, as well as methylation of RNA, affecting post-transcriptional processing. Moreover, dysregulation of SAM levels has been implicated as a potential driver of cancer progression [31]. SAM is synthetized by Methionine Adenosyltransferase using methionine and adenosine triphosphate as substrates. RNA methylation plays an important role in preserving SAM homeostasis by regulating the levels of MAT2A. Depending on the availability of SAM in the cell, the interaction between METTL16 and the mRNA encoding the MAT2A yields distinct outcomes, impacting mRNA stability or splicing and, consequently, protein production [18–20] (Figure 2). This establishes a feedback loop where METTL16, in response to low SAM availability, drives expression of MAT2A expression, and thereby enhances production of SAM. In essence, METTL16 effectively functions as a SAM sensor. MAT2A activity is also linked to cell growth and metabolism through Mechanistic Target of Rapamycin Complex 1 (mTOR complex 1). This important regulator of cell growth and metabolism activates MYC, which promotes expression of MAT2A. Additionally, mTORC1 impacts m⁶A methyltransferase activity by increasing expression of Wilms tumor 1-associated protein (WTAP), one of the components of the methyltransferase complex [32,33]. In the methylation reaction, a methyl group is transferred to the substrate and SAM is converted to SAH, which in then converted to adenosine and homocysteine. 5-methyl-tetrahydrofolate, generated in the folate cycle, is used to convert homocysteine to methionine, [34] (Figure 2). There are several examples that demonstrate how one carbon metabolism, the use of and regeneration of SAM, affect m⁶A levels. The interplay between SAM levels and amino acid availability is well exemplified in cancer cells, where serine availability impacts methyl group transfer to nucleic acids. Under methionine starvation, serine contributes one-carbon units to support the methionine cycle, while in methionine fed cells, serine availability impacts the SAM cycle by allowing for de novo ATP synthesis to support conversion of methionine to SAM [35]. In clear cell renal cell carcinoma, the methionine required for the maintenance of cancer stem cells (CSCs) is secreted by a specific population of pericytes in response to accumulation of succinate in the tumor microenvironment [36]. In the CSCs, methionine supports m⁶A modification of transcripts such as the mRNA for ATPase-family-AAA-domain-containing 2 (ATAD2), which plays a pivotal role in maintaining a stemness. This example illustrates how metabolism across different cell types can impact m⁶A dependent regulation of gene expression. Another connection between m⁶A modification and metabolism in clear cell renal cell carcinoma, is the elevated expression of Methylenetetrahydrofolate Dehydrogenase (NADP+ Dependent) 2 (MTHFD2), an enzyme in one-carbon metabolism. Elevated levels of MTHFD2 support increased nucleic acid methylation, including RNA m⁶A modification. Among hypermethylated RNAs, increased levels of m⁶A on the mRNA coding for Hypoxia-Inducible Factor 2 Alpha (HIF-2α), enhance translation and promote aerobic glycolysis [37]. The balance between SAM and SAH can also be locally controlled to enhance m⁶A modification of RNA. The enzyme Adenosylhomocysteinase (AHCY), catalyzes the conversion of SAH to adenosine and L-homocysteine, reducing the levels of SAH. Upon Zika virus (ZIKV) infection, AHCY is recruited to ZIKV-remodeled endoplasmic reticulum membranes, facilitating m⁶A modification of ZIKV RNA (Figure 2). Modification of ZIKV RNA with m⁶A dampens the interferon response and allows for accumulation of viral RNA [38]. In contrast to METTL16, which responds to SAM levels, the K_m of METTL3 is below cellular SAM levels [39] suggesting METTL3 is active even in the presence of low levels of SAM. On the other hand, cellular levels of SAH are above the IC50 for METTL3, suggesting METTL3 might instead respond to cellular SAH levels [39]. These examples highlight the reciprocal modulation between the ‘writer’ activity and the metabolic milieu within the cell.

Figure 2:

Schematic representation of metabolic pathways that can influence m⁶A methylation processes. Activity of methyltransferases converts SAM, which is generated in the one-carbon cycle, into SAH. SAH is an allosteric inhibitor of methyltransferase activity. Levels of MAT2A, the enzyme responsible for SAM synthesis, are controlled by METTL16, which acts as a SAM sensor. mTORC1 influences m⁶A deposition by controlling expression of MAT2A and WTAP, a component of the methyltransferase complex. WTAP enhances c-Myc activity by inhibiting MXD2. During Zika virus (ZIKV) infection, AHCY is recruited to remodeled endoplasmic reticulum membranes, reducing levels of SAH to facilitate m⁶A modification of ZIKV RNA. Black arrows represent the metabolic flow in the SAM cycle, while blue arrows denote regulation of the pathway at the protein level. This image was created with BioRender (https://biorender.com/).

The link between metabolism and eraser activity

The methyl group in m⁶A can be removed by FTO and ALKBH5, enzymes that belong to the 2-Oxoglutarate-dependent dioxygenase (2OGDD) family. All enzymes in this family require oxygen, reduced iron and α-ketoglutarate for activity [40]. Each enzyme in this family has a different affinity to each one of the necessary co-factors, allowing enzymes in this family to act as sensors of the cell’s environment. In addition, activity of enzymes in this family can be inhibited by metabolites that are structurally related to α-ketoglutarate, such as fumarate, succinate and 2-hydroxyglutarate (2-HG) (Figure 3) [40]. These metabolites are generated in the citric acid (TCA) cycle and are known to accumulate in response to mutations in enzymes in this important metabolic pathway. Both FTO and ALKBH5 have been shown to interact in vitro with metabolites structurally related to α-ketoglutarate and the activity of these enzymes is inhibited by changes in the ratio between α-ketoglutarate and structurally related metabolites in cells [41–43]. In cells derived from Hereditary leiomyomatosis and renal cell cancer (HLRCC) patients, a disease driven by loss of fumarate hydratase (FH) activity, accumulation of high levels of fumarate results in m⁶A hypermethylation of mRNAs. Interestingly, activity of ALKBH1, a different enzyme in this family was not impacted by high levels of fumarate [41]. In Acute myeloid leukemia (AML) models, accumulation of R-2-hydroxyglutarate (R-2HG) inhibits FTO activity, increasing levels of m⁶A on mRNA. Hypermethylation of the transcripts coding for MYC reduces mRNA stability, suppressing pathways required for proliferation and viability of leukemic cells [43]. Changes in the pool TCA metabolites don’t always result in inhibition of enzymatic activity. In a sub type of Clear Cell Renal Cell Carcinoma, an increase in the α-ketoglutarate–to-succinate ratio results in FTO activation, lowering m⁶A levels and stabilizing the mRNA coding for Bromodomain Containing 9 (BRD9). Activity of FTO has also been shown to be modulated by metabolites outside of the TCA cycle (Figure 3). Ascorbic acid (vitamin C) is a water-soluble antioxidant and a cofactor for numerous enzymes including 2OGDDs, the family to which FTO and ALKBH5 belong [44]. Ascorbic acid binds to FTO and activates its demethylase activity in vitro [45,46]. In line with an activation of the demethylase activity, treating cells with ascorbic acid reduces global levels of m⁶A. The loss of m⁶A is more prominent at sites modified at high stoichiometry [46]. Importantly, while ascorbic acid stimulates 2OGDD activity, high levels of ascorbic acid can have an inhibitory effect. This is particularly significant considering that Vitamin C deficiency remains a common health issue, and has been reported in cancer patients [44,47,48]. The activity of FTO can be also modulated by Nicotinamide adenine dinucleotide phosphate (NADP), despite NADP not being a substrate or product of the demethylation reaction. NADP directly binds to FTO, increasing FTO-mediated m⁶A demethylation [45]. Furthermore, signaling molecules can also impact demethylase activity (Figure 3). Nitric oxide (NO) binds to the catalytic iron center of FTO and inhibits its activity. Disruptions in NO production have been observed in various cancer types and correlated with unfavorable patient outcomes [49]. These examples underscore how differences in metabolite levels between cells can lead to differences in m⁶A-dependent gene regulation.

Figure 3:

Schematic representation of metabolic pathways that can influence m⁶A demethylation processes. The demethylation reaction requires α-ketoglutarate, oxygen (O2), and iron [Fe(II)]. In addition to the co-factor α-ketoglutarate, the citric acid (TCA) cycle also generates α-ketoglutarate related metabolites that inhibit demethylase activity (fumarate, succinate, and 2-hydroxyglutarate). Eraser activity can also be modulated by metabolites outside the TCA cycle. Nitric oxide (NO) can inhibit activity of erasers while nicotinamide adenine dinucleotide phosphate (NADP) or ascorbic acid (AC) activate eraser activity. Dashed arrows represent output from a metabolic pathway. This image was created with BioRender (https://biorender.com/).

Metabolism sets the boundaries for m⁶A regulation

The RNA modification m⁶A holds a pivotal role in regulating the fate of mRNA. The orchestrated activity of writers, erasers, and readers is indispensable during developmental processes and dysregulation in this pathway has been implicated in various diseases [50], underscoring the significance of m⁶A in gene regulation. Notably, the enzymatic activities of both m⁶A writers and eraser can be impacted by changes in the cellular metabolic environment. Consequently, m⁶A-dependent pathways can drive changes in gene expression is response to alterations in the cell’s environment, enabling cells to adjust to changing conditions. However, this link between metabolism and the enzymatic activities of m⁶A writers and erasers implies that alterations in metabolism can constraint programed gene expression programs that rely on m⁶A. For instance, activity of FTO can be modulated by changes in the ratio between α-ketoglutarate and related metabolites, and the availability of ascorbic acid, NADP or nitric oxide. In a cellular context where FTO activity plays an important role in regulating gene expression, fluctuations in the level of these cofactors can disrupt the balance between the writer and eraser activity. For example, mutations in FH or Isocitrate Dehydrogenase (IDH), present in specific type of cancer [51], result in the buildup of α-ketoglutarate related metabolites, consequently inhibiting m⁶A demethylase activity [41,43]. Other factors known to modulate FTO activity, such as ascorbic acid, nitric oxide and NADP, are also known to change in the context of disease [44,47,48,52,53]. These observations suggest that m⁶A-dependent regulation of gene expression can be disrupted even in the absence of mutations in writers, erasers, or readers. As alterations in metabolite levels are not exclusive to disease states [54], it will be interesting to explore whether developmentally programed changes in metabolism modulate m⁶A-depent pathways to fine tune expression of genes important for development. Given that both writers and erasers of m⁶A have been implicated in disease progression, these enzymes are attractive targets for drug development [55]. The link between metabolism and the enzymatic activities of m⁶A writers and erasers is likely to hold significant relevance in the development of therapies targeting components in this pathway. We predict that some contexts, variances in drug efficiency will correlate with alteration in the cellular environment. As an illustration we highlight the example of the use of xanthine derivates as inhibitors of FTO. Although the inhibition mechanism of xanthine derivates has not yet been elucidated, it has been shown to be modulated by L-ascorbic acid, with the inhibitory effect decreasing with higher concentrations of L-ascorbic acid [56]. This example underscores the importance of understanding how different classes of drugs interact with the cellular environment. Another critical aspect where the interplay between metabolism and m⁶A writer and eraser activity is likely to play a role is in heterogeneity of gene expression. Recently developed methods to map m⁶A have unveiled significant differences in methylation level at individual m⁶A sites across different cells in a population. This heterogeneity in methylation at individual m⁶A sites in single cells implies the presence of cell-autonomous mechanisms that modulate the balance between writer and eraser activity [9,10]. We propose that differences in metabolism and access to important co-factors among cells within a heterogenous tissue translate into a distinct balance between m⁶A writer and eraser activity in each cell, contributing to divergent patterns of gene expression. In this review we have highlighted instances illustrating how metabolism interacts with m⁶A-dependent regulation of gene expression by modulating writer and eraser activity. We propose that this interaction plays a significant role in m⁶A-dependent gene regulation across multiple contexts.