2-Oxoglutarate-dependent dioxygenases are sensors of energy metabolism, oxygen availability, and iron homeostasis: potential role in the regulation of aging process

Abstract

Recent studies have revealed that the members of an ancient family of nonheme Fe²⁺/2-oxoglutarate-dependent dioxygenases (2-OGDO) are involved in the functions associated with the aging process. 2-Oxoglutarate and O₂ are the obligatory substrates and Fe²⁺ a cofactor in the activation of 2-OGDO enzymes, which can induce the hydroxylation of distinct proteins and the demethylation of DNA and histones. For instance, ten-eleven translocation 1-3 (TET1-3) are the demethylases of DNA, whereas Jumonji C domain-containing histone lysine demethylases (KDM2-7) are the major epigenetic regulators of chromatin landscape, known to be altered with aging. The functions of hypoxia-inducible factor (HIF) prolyl hydroxylases (PHD1-3) as well as those of collagen hydroxylases are associated with age-related degeneration. Moreover, the ribosomal hydroxylase OGFOD1 controls mRNA translation, which is known to decline with aging. 2-OGDO enzymes are the sensors of energy metabolism, since the Krebs cycle intermediate 2-oxoglutarate is an activator whereas succinate and fumarate are the potent inhibitors of 2-OGDO enzymes. In addition, O₂ availability and iron redox homeostasis control the activities of 2-OGDO enzymes in tissues. We will briefly elucidate the catalytic mechanisms of 2-OGDO enzymes and then review the potential functions of the above-mentioned 2-OGDO enzymes in the control of the aging process.

Introduction

The aging of an organism is a progressive degeneration process, which leads to functional decline and exposes to many age-related diseases. The origin of the age-related deterioration was probably associated with the appearance of multicellular organisms involving mechanisms, which seem to be conserved across the evolution. Genetic studies have revealed that perhaps the aging process is not as programmed as the development but it includes common hallmarks through the metazoa, e.g., epigenetic changes in chromatin, mitochondrial dysfunction, loss of proteostasis, and cellular senescence [1]. Molecular studies on model organisms have revealed that the lifespan can be extended by many experimental treatments, such as dietary restriction [2], hypoxic insults [3], mitochondrial dysfunction [4], and suppression of protein synthesis [5]. Recent studies have emphasized the role of energy metabolism in the regulation of gene expression through the modulation of chromatin landscape [6, 7]. Oxygen availability controls the energy metabolism but it also has a fundamental role in the gene expression adapting cells and organisms to hypoxic conditions. There is substantial evidence that aging impairs mitochondrial metabolism, which can have disturbing effects on the epigenetic regulation of gene expression, possibly enhancing the aging process [8].

Recent studies have revealed that the enzymes removing methyl groups from DNA and histones are included in the superfamily of 2-oxoglutarate-dependent dioxygenases (2-OGDO) [9, 10]. DNA and histone methylation has a fundamental role in the epigenetic regulation of gene expression. Intriguingly, the collagen prolyl and lysyl hydroxylases as well as hypoxia-inducible factor (HIF) prolyl hydroxylases (PHDs) are the members of 2-OGDO family [11]. Moreover, the ribosome hydroxylase, called 2-oxoglutarate and Fe²⁺-dependent oxygenase-1 (OGFOD1), is also included in the same enzyme family [12]. The catalytic activity of 2-OGDO enzymes is controlled by the availability of O₂, energy metabolites (2-oxoglutarate/succinate), and ferrous iron (Fe²⁺) (Fig. 1). It is probable that these basic cellular elements might have been important sensors for ancient multicellular organisms. Our purpose is to elucidate the catalytic mechanisms and the regulation of these fundamental 2-OGDO enzymes and discuss their potential role in the control of the aging process.

Fig. 1.

A diagram illustrating the common catalytic reaction of Fe²⁺/2-oxoglutarate-dependent dioxygenases. 2-Oxoglutarate and O₂ are obligatory substrates in the hydroxylation of target molecules by 2-OGDO enzymes. The reaction also produces succinate and CO₂. The catalytic activity requires ferrous (Fe²⁺) iron, which is generated by ascorbate (vitamin C) and glutathione. The 2-OGDO reaction is inhibited by distinct energy metabolites, i.e., succinate, fumarate and R(-)-2-hydroxyglutarate. Moreover, many metals, ROS, and Fe²⁺ chelators are potent inhibitors. GSH reduced glutathione, KDM2-7 histone lysine demethylases, OH hydroxyl group, 2-OG 2-oxoglutarate, 2-OGDO 2-oxoglutarate-dependent dioxygenases, OGFOD1 2-oxoglutarate and Fe²⁺-dependent oxygenase domain-containing protein 1, P3H and P4H collagen prolyl 3- and 4-hydroxylases, PHD1-3 prolyl hydroxylase domain-containing proteins 1-3, ROS reactive oxygen species, TET1-3 ten-eleven translocation 1-3

2-OGDO enzymes are sensors of energy metabolism, hypoxia and iron homeostasis

The 2-OGDO enzymes comprise a large family of non-heme containing Fe²⁺ and 2-oxoglutarate-dependent dioxygenases [11, 13–15]. The family includes several, evolutionarily conserved subfamilies, which have diverse functions both in prokaryotes, micro-organisms, plants, and metazoans. The common feature of these 2-OGDO subfamilies is their catalytic mechanism, i.e., they all are 2-oxoglutarate-dependent oxidases/hydroxylases, which are also dependent on molecular oxygen as a co-substrate and ferrous (Fe²⁺) iron as a catalyzing cofactor (Fig. 1). Attributable to these characteristics, the 2-OGDO enzymes are cellular sensors for the changes in (i) energy metabolism through the availability of 2-oxoglutarate, a Krebs cycle intermediate, (ii) the oxygen level consequently inducing hypoxic responses, and (iii) the Fe²⁺ redox status indicating alterations in iron metabolism and oxidative stress [7, 11, 16, 17].

Structural studies on the 2-OGDO enzymes have revealed that they contain a modified double-stranded β-helix structure, which involves conserved Fe²⁺ and 2-oxoglutarate binding sites [11, 18]. This catalytic domain has been called the jelly-roll fold comprising eight antiparallel β-strands and separate secondary structures, being diverse in different enzymes [18]. The crystal structures of 2-OGDO enzymes, e.g., those of collagen P4H [19], PHD2 [20], KDM4A [21] and TET2 [22], have revealed specific binding modes for distinct substrates, as well as sites for Fe²⁺ and 2-oxoglutarate in the catalytic domains of these 2-OGDO enzymes. The Jumonji domain present in histone KDM2-7 is the most thoroughly studied jelly-roll fold structure of the 2-OGDO enzymes [9, 23, 24]. The members of Jumonji subfamily can remove the methyl group from the distinctly methylated lysine residues of histones. Moreover, certain 2-OGDO enzymes can target proline residues, e.g., P4H and PHD1-3, as well as the asparagine residues, such as FIH-1. On the other hand, the members of TET subfamily target the 5-methylated cytosine (5mC) of DNA and convert it to 5-hydroxymethylcytosine (5hmC) and thus trigger DNA demethylation [10]. In addition, there is a large range of miscellaneous compounds, which can be the targets of 2-OGDO enzymes, e.g., in plants during flavonoid synthesis [15]. The most typical reactions in mammals are oxidation/hydroxylation and demethylation, whereas in plants and micro-organisms, some subfamilies can mediate, e.g., molecular cyclization, desaturation, and chlorination reactions [14].

Activation of 2-OGDO enzymes starts with the binding of 2-oxoglutarate and Fe²⁺ into their specific sites in the catalytic domain of enzyme. Consequently, dioxygen (O2) binds to ferrous (Fe²⁺) iron, which stimulates the oxidative decarboxylation of 2-oxoglutarate to succinate and CO2, and provokes the formation of oxidized ferryl (^Fe3+/4+) intermediates (Fig. 1.). Subsequently, these high-valent iron oxidant species hydroxylate the substrate molecules in the enzyme–substrate complex [25–27]. The catalytic activity is dependent not only on substrate but also on the availability of 2-oxoglutarate and O2 as well as on the maintenance of Fe²⁺ redox status. 2-Oxoglutarate is an obligatory co-substrate for the 2-OGDO enzymes (Fig. 1). Given that 2-oxoglutarate is the key metabolite in the function of Krebs cycle, it means that energy metabolism has a central role in the control of 2-OGDO enzymes. There is a 2-oxoglutarate carrier (OGC) which mediates the efflux of 2-oxoglutarate from the mitochondria to the cytoplasm [28]. In the cytoplasm, 2-oxoglutarate can be used for amino acid synthesis, i.e., amino acid aminotransferases convert 2-oxoglutarate to glutamate and consequently glutamate can be metabolized to some other amino acids [29]. Moreover, 2-oxoglutarate can be metabolized through the reductive carboxylation pathway to acetyl-CoA and subsequently used for fatty acid synthesis [30–32]. Reductive carboxylation is an active metabolic pathway in the condition where mitochondrial metabolism is compromised, such as hypoxia and tumor growth.

Moreover, several studies have indicated that succinate and fumarate can bind to the 2-oxoglutarate site but they are unsuitable substrates for the decarboxylation reaction and thus they are potent competitive inhibitors of 2-OGDO enzymes, e.g., those of PHDs [33–35], KDMs [36], and TETs [37]. Recent studies have revealed that distinct mutations in the SDH enzyme can provoke the accumulation of succinate and subsequently lead to carcinogenesis, probably due to the inhibition of TET enzymes since SDH mutations are associated with a robust hypermethylation of DNA [38, 39]. Mutations in isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) can convert isocitrate to 2-hydroxyglutarate, instead of 2-oxoglutarate [40, 41]. Xu et al. [42] revealed that 2-hydroxyglutarate was a competitive inhibitor of 2-OGDO enzymes, which inhibited TET and KDM enzymes and consequently increased DNA and histone methylation in human gliomas. Currently, succinate, fumarate, and 2-hydroxyglutarate are called oncometabolites, since they can provoke carcinogenesis by epigenetically modifying DNA and histone methylation [43].

Molecular O₂ has a fundamental role in the catalytic process of 2-OGDO enzymes, since one oxygen atom is involved in the oxidative decarboxylation of 2-oxoglutarate and another in the hydroxylation of the prime substrate [44, 45]. Especially PHDs are sensitive sensors of oxygen availability in tissues, i.e., PHDs are active in normoxia and they hydroxylate HIF-1 factors inducing their degradation. On the contrary, hypoxia inhibits PHD activity, and consequently HIF-1 factors will be stabilized activating the hypoxia-related signaling pathways [17]. Many studies have confirmed that the oxygen level also has a crucial role in the activation of other 2-OGDO enzymes, e.g., those of Jumonji type of KDMs [46, 47]. Sanchez-Fernandez et al. [46] reported that the activity of recombinant KDM4E enzyme showed a linear increase in the response to increasing O₂ level. They also observed that the reaction kinetics was dependent on the methylation degree of H3K9 site; a higher methylation level enhanced the activity. Zhou et al. [48] demonstrated that hypoxia increased the amount of H3K4me3 in cultured cells by inhibiting the demethylation activity of H3K4 site, an activating site for gene expression. They also reported that the hypoxia-induced increase in histone methylation was attributed to the inhibition of KDM5A demethylase activity. Interestingly, although hypoxia decreased the enzymatic activity of KDMs, the expression levels of some KDMs were significantly increased during hypoxia [49–51]. HIF-1α was involved in this signaling response to hypoxia, reflecting probably the feedback regulation to maintain chromatin status during hypoxia. Moreover, HIF-1α can also induce the expression of PHD2 and PHD3 in cultured cells, which is a negative feedback mechanism in oxygen sensing [52].

The 2-OGDO enzymes contain a mononuclear non-heme iron center, which is activated by the binding of molecular O₂. The reaction involves the oxidation of ferrous iron to the high-valent, Fe³⁺ and Fe⁴⁺-superoxo species, which subsequently can decarboxylate 2-oxoglutarate and hydroxylate substrate [45, 53]. These ferryl-oxo intermediates are highly reactive and can hydroxylate a number of different substrates. Ascorbate and glutathione are capable of reducing these oxidized Fe^3+/4+ species back to Fe²⁺, which restores the activity of 2-OGDO enzymes [54, 55]. Several studies have revealed that ascorbate (vitamin C) can control oxygen sensing and DNA methylation by stimulating PHDs and TETs [54, 56, 57]. There are contradictory results whether ROS and other oxidants can regulate the function of 2-OGDO enzymes. It is known that nitric oxide (NO) inhibits histone KDMs by generating histidyl-iron-nitrosyl complexes in the catalytic Jumonji domain [58]. The inhibition of KDM3A activity by NO increased the histone H3K9 methylation in cultured cells. However, the expression of KDM3A increased to the same extent as with the exposure to hypoxia, probably indicating the homeostatic feedback response. Metzen et al. [59] demonstrated that the NO treatment inhibited the activity of PHDs in a dose-dependent manner and provoked the accumulation of HIF-1α into the nuclei. It seems that NO can stimulate pseudohypoxic state at the normoxic conditions. Recent studies have also indicated that mitochondrial ROS production did not directly inhibit PHD activity or stabilize HIF-1α [60, 61].

Given that iron has an obligatory function in the catalytic activation of 2-OGDO enzymes; it means that the maintenance of cellular iron redox status has an important role in the control of 2-OGDO activities [16]. Iron chelators, e.g., desferoxamine and 1,10-phenanthrolin, inhibited PHD2 and induced the expression of HIF-1α [62]. Quercetin, a flavonoid and an effective iron chelator, was also a potent inducer of HIF-1α in cultured cells [63]. On the other hand, many studies have indicated that the divalent transition metals, e.g., cobalt, nickel, and zinc, inhibited different 2-OGDO enzymes by substituting iron in the active center [64, 65] and thus they could disturb the function of epigenome provoking toxic effects and cancers [66]. The availability of free iron is indispensable for the function of many different enzyme systems and thus the cellular iron homeostasis is tightly controlled [67]. Many studies have demonstrated that HIF-1 factors increase the expression of several key components of iron regulation, e.g., transferrin and its receptor, and heme oxygenase [16]. This can have deleterious effects leading to the iron overload, e.g., in cerebral ischemia. Ding et al. [68] reported that the enhanced expression of transferrin receptor 1 (TfR1) increased iron uptake into the ischemic cerebral cortex. Iron chaperones (PCBP1 and 2) have a crucial role in the regulation of balance between cytosolic iron pool and ferritin-bound storage [69]. Nandal et al. [70] reported that the depletion of PCBP1 and PCBP2 reduced the activity of PHD2 and increased the expression of HIF-1α in cultured cells. They also revealed that PHD2 and PCBP1 interacted in vivo and the lack of PCBPs reduced the incorporation of Fe²⁺ into the catalytic domain of PHD2. These observations indicated that PCBPs are involved in the targeting of Fe²⁺ into the 2-OGDO enzymes. It seems that the 2-OGDO enzymes are at the nexus of interdependent control between energy metabolism and the availability of oxygen and iron in the regulation of cellular homeostasis.

2-OGDO enzymes are regulators of epigenetic and non-epigenetic functions

Given that 2-OGDO enzymes are dependent on the presence of oxygen, the increase in atmospheric oxygen concentration during the Cambrian period provided circumstances required for the evolution of oxygenase/hydroxylase systems [71, 72]. Oxygen has also been a prerequisite for the evolution of epigenetic regulation, since most of the epigenetic marks can be removed by 2-OGDO enzymes, i.e., DNA and histone methylation by TETs and KDMs, respectively. Simultaneously, there evolved multicellular organisms, which are dependent on the synthesis of collagen, a structural protein of connective tissues. Thus, it is not a surprise that collagen synthesis is regulated by 2-OGDO enzymes [73]. Moreover, protein synthesis also is under the regulation of 2-OGDO enzymes, i.e., OGFOD1 ribosomal hydroxylase [12] which is conserved during evolution from prokaryotes to humans [74]. The evolution of aging process is still elusive although the expansion of multicellular organisms was most probably associated with the appearance of aging process and the death of parental organisms. There seems to be differences between cellular senescence and the aging of multicellular organisms, although cellular senescence can also appear in mammalian tissues [75].

TET1-3 control DNA demethylation

DNA methylation is regulated by four DNA methyltransferases (DNMT1, 3A, 3B and 3L) and three demethylases, i.e., Ten-Eleven Translocations (TET1-3) [10, 76, 77]. The methylation of cytosine sites (CpG) and CpG-rich islands is a well-characterized process, whereas the removal of the methyl group from 5-methylcytosine (5mC) has remained a more elusive mechanism. In 2009, Tahiliani et al. [78] demonstrated that TET1 enzyme oxidized the 5mC residue to 5-hydroxymethylcytosine (5hmC), which was associated with the demethylation process of DNA. Interestingly, they revealed that TET1 was included in the 2-OGDO enzyme family being dependent on the presence of oxygen, 2-oxoglutarate, and Fe²⁺. The 5hmC is an intermediate in the removal of epigenetic 5mC mark, the process which can be further continued by TET enzymes oxidizing 5hmC to the 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC). These intermediates activate thymine DNA glycosylase and modified residues are finally substituted with unmodified residues by the base excision repair system [79]. Thus, the TET enzymes initiate the DNA demethylation process, although the hydroxylated intermediates could also have their own signaling functions. To confirm the typical 2-OGDO properties of TET enzymes in cultured cells, Xiao et al. [37] knocked down SDH and FH enzymes inducing the accumulation of succinate and fumarate, the inhibitors of 2-OGDO enzymes. They reported that these treatments reduced the TET-catalyzed 5hmC formation and increased global DNA methylation in cultured cells, which indicated that succinate and fumarate inhibited TET enzymes, also in vivo [37]. On the other hand, Minor et al. [80] observed that ascorbate, retaining a reduced level of Fe²⁺, enhanced the generation of 5hmC in cultured embryonic fibroblasts. Ascorbate did not affect the expression of TET genes. Recently, Zhao et al. [81] reported that several redox-active quinones increased the conversion of 5mC to 5hmC in cultured cells. They also demonstrated that the quinones stimulated the activity of TET enzymes by increasing the availability of labile iron pool, whereas an iron chelator reduced the level of 5hmC. These observations indicated that TETs are 2-OGDO enzymes in vivo and that energy metabolism and iron homeostasis can control epigenetic responses.

Recent genome-wide studies have indicated that the TET-induced hydroxylation of 5mC represents a priming step of transcription, locating in, e.g., the enhancers and gene bodies of actively transcribed genes [82, 83]. Hon et al. [84] observed with the genomic base-resolution technique that 5hmC was enriched into enhancers and the deletion of TET2 induced an extensive loss of 5hmC at enhancer sites. Subsequently, DNA methylation was increased and the target genes were silenced. However, there are clear functional differences between TET proteins [85], although their exact functions need to be clarified. Moreover, the activities of TETs and the level of 5hmC fluctuate during mammalian development [77]. The total amounts of 5hmC are also specific for different tissues and not dependent on the level of TET gene expression [86]. In addition to the demethylation activity of TETs, Xu et al. [87] demonstrated that TET1 was bound to CpG-rich regions to prevent the binding of DNA methyltransferases to these active promoter sites in mouse embryonic stem cells (ESC). The binding was mediated via the CXXC domain of TET1 protein. Moreover, Wu et al. [88] reported that TET1 could even recruit polycomb repressive complex 2 (PRC2) to the distinct CpG-rich promoters to repress the expression of certain developmental regulators in ESC. Neri et al. [89] observed that this interaction was especially present in undifferentiated ESC but not in differentiated cells. These studies indicated that TET enzymes have an important dual role in the development of ESC controlling the balance between pluripotency and lineage commitment [90, 91].

KDM2-7 control histone demethylation

Histone methylation has a fundamental role in the epigenetic regulation of chromatin landscape, which not only controls gene expression but also affects the genome reprogramming, e.g., during cellular differentiation and tissue development, cancer formation, and many diseases [92, 93]. It seems that histone methylation can also have a crucial role in the transgenerational longevity in C. elegans [94]. The methylation status of histone amino termini contains the Histone Code, which controls the function of chromatin template [95]. The methylation of distinct lysine and arginine residues of histone tails either activates or represses gene transcription, e.g., H3K4 and H3K36 are potent activating sites, whereas H3K9 and H3K27 are repressing positions. Histone methylation represents the dynamic balance between the function of histone methyltransferases and histone demethylases [92, 96, 97]. Most of the lysine demethylase groups (KDM2-7) are 2-OGDO enzymes containing the Jumonji C domain. The FAD⁺-dependent lysine demethylases, KDM1A/LSD1 and KDM1B/LSD2, are amine oxidases, which can demethylate H3K4me1,2, along with the enzymes of KDM2 and KDM5 [92, 96]. The enzymes of KDM2 and KDM5 groups demethylate H3K4me2,3 and H3K36me1,2, which are the activating sites and thus they are the inhibitors of gene expression. On the other hand, the demethylases of KDM3, KDM6, and KDM7 remove methyl groups from the repressive sites H3K9me1,2 and H3K27me1-3, thus activating gene expression [9, 96, 98]. There are many studies indicating that the enzymes of KDM2-7 subfamilies are 2-OGDO enzymes also in vivo and thus oxygen level, energy metabolites, and iron redox status can affect their activities and the cellular methylation status (see above). The specific regulation of distinct KDM enzymes still needs to be clarified, given that they have, e.g., opposite functional effects on gene expression.

The functional differences between KDMs can be attributed to the fact that the proteins of distinct KDM subgroups have unique binding domains, which target enzymes to specific complexes [9, 47, 96, 98]. The plant homeobox finger domain (PHD) is present in the most of KDMs (2, 4, 5, and 7). It is known that the PHD domains can recognize methylated lysine positions, e.g., H3K4 sites, and thus they are important epigenetic readers [99]. The mutations or dysregulation of PHD fingers has been detected in many diseases, such as immunodeficiency syndromes and neurological disorders [100]. More specific domains are, e.g., F-box (KDM2), Tudor (KDM4), ARID (KDM5) and tetratricopeptide repeats (TPR) (KDM6) [9, 98]. Tudor domain is another histone methylation reader present in the KDM4 type of histone demethylases. KDM4 specifically demethylates H3K9me3 and H3K36me3 sites but it can also regulate DNA damage response via targeting methylated H4K20 site [101]. The ARID (AT-rich interaction domain) segment is a DNA-binding domain, which is present in the KDM5 type of histone demethylases, also called JARID family. KDM5B demethylases can bind to distinct genes, e.g., E2F-target genes, and enhance cellular senescence [102]. KDM5A/B (JARID1) demethylates H3K4me2/3 sites, thus inhibiting gene expression. In addition to the differences in the complex assembly, there are also great variations in the Km values between KDM enzymes, such as commonly observed for 2-OGDO enzymes. Given that the chromatin platform functions as a complex entity [103], it is a challenging task to evaluate the functional role of distinct components, especially since the chromatin has a great capacity to be reprogrammed in the changing environmental conditions.

PHDs and FIH-1 regulate hypoxic responses

Hypoxia is a common environmental stress, e.g., in many pathological conditions such as ischemia/stroke and inflammatory disorders, which stimulates a robust gene expression via the activation of hypoxia-inducible transcription factor (HIF) [104–106]. The oxygen sensing mechanisms were discovered the last decade and it turned out that the 2-OGDO enzymes have a fundamental role in the response to tissue oxygen deficiency. In 2001, Bruick and McKnight [107] and independently Epstein et al. [108] identified three genes encoding novel prolyl 4-hydroxylases (prolyl hydroxylase domain-containing proteins 1-3, PHD1-3), which hydroxylated two proline residues in HIF-1α proteins provoking their degradation. The PHDs are also called HIF prolyl hydroxylases 1-3 and EGLN1-3. The hydroxylation of proline residues (Pro402 and Pro564 in HIF-1α) triggers the binding of von Hippel-Lindau (pVHL) E3 ubiquitin ligase complex to HIF proteins, inducing their ubiquitination and subsequently proteasomal degradation [44]. In hypoxia, the activity of PHDs decreases, which inhibits the hydroxylation of HIF proteins and thus prevents their degradation. This increases the protein level of HIF factors and thus enhances the transactivation capacity. There are considerable differences between PHD enzymes in their capacities to hydroxylate different HIF factors, e.g., in respect to their catalytic and inhibitory constants [109–111]. For instance, PHD2 is highly reactive to the deficiency of oxygen and it is the major sensor for hypoxia, whereas PHD3 has a preference for targeting the HIF-2α protein rather than HIF-1α as a substrate. The HIF factors are the major substrates for PHDs, although recently many other targets have been identified, mostly specific for PHD3 isoenzyme [112]. For instance, PHD3 hydroxylated pyruvate kinase muscle isoform 2 (PKM2), which enhanced the binding of PKM2 to HIF-1α and consequently potentiated the transcriptional activity of HIF-1α [113]. There are also observations that PHD hydroxylases could inhibit NF-κB signaling and thus prevent inflammatory responses [112].

In 2001, Mahon et al. [114] identified a factor inhibiting HIF-1 (FIH-1), which was bound to HIF-1α and in that manner inhibited its transactivation capacity. Next year, Lando et al. [115] demonstrated that FIH-1 also was a 2-OGDO enzyme hydroxylating the asparagine residue (Asn803) at the C-terminal transactivation domain of HIF-1α protein. They also reported that the FIH-1-induced hydroxylation prevented the interaction between p300 transcription coactivator and HIF-1α protein. The catalytic properties of FIH-1 differ from those of PHDs, e.g., the Km values of FIH-1 are lower for both O₂ and 2-oxoglutarate than those of PHDs, but instead higher for ascorbate [116]. It seems that FIH-1 is not as responsive to hypoxia as PHDs but more responsive to oxidant stress than PHDs [117]. Moreover, in contrast to PHDs, FIH-1 is not inhibited by succinate and fumarate [34]. These observations also indicated that FIH-1 is not as responsive as PHDs to the changes in energy metabolism. Moreover, there are differences in the substrate specificity between the hydroxylases of PHDs and FIH-1. Several studies have revealed that the ankyrin repeat domains (ARD) are sensitive targets for asparagine hydroxylation by FIH-1 [118, 119]. For instance, the ARD domains of Notch [119] and the members of IκB family [118] are hydroxylated even at the higher affinity than the prolines of HIF-1 factors. Zheng et al. [119] revealed that FIH-1 hydroxylated the Notch intracellular domain (ICD) inhibiting the transactivation mediated by Notch ICD. The suppression of Notch activity enhanced neuronal and myogenic differentiation. Many observations have indicated that the ARD binding of FIH-1 prevents its capacity to hydroxylate HIF-1α, which seems to potentiate the transactivation of HIF-1α [119]. In conclusion, it seems that in collaboration HIF-1 and FIH-1 can adapt cells not only to hypoxia but more generally to cellular stress.

There are three HIF factors, i.e., HIF-1α, HIF-2α, and HIF-3α, which are basic-helix-loop-helix transcription factors [104, 105, 120]. In hypoxia, these factors form heterodimers with HIF-1β (also called aryl hydrocarbon receptor nuclear translocator, ARNT) and subsequently bind to the hypoxia response elements (HRE) of distinct gene promoters. There are studies indicating that HIF factors control the expression of 200–1000 genes, being context-specific responses and depending on the tissue type. The typical genes are those enhancing tissue oxygenation, angiogenesis, glucose uptake, and glycolytic metabolism, i.e., the responses which adapt cells to hypoxia [106, 121]. Moreover, HIF factors can regulate autophagy/apoptosis, cell cycle, stem cell differentiation, wound healing, and inflammatory responses [106, 112]. There are significant functional differences between HIF factors, sometimes even opposing activities. For instance, the HIF-3α variant can inhibit the HIF-1α-mediated transcription [122] and, moreover, it seems that HIF-1α is driving the acute responses to hypoxia, whereas HIF-2α controls the chronic responses [123]. These differences can partially be attributed to the specific post-translational modifications of HIF factors, e.g., acetylation, hydroxylation, phosphorylation, sumoylation, and ubiquitination and, on the other hand, to the presence of context-dependent transcription cofactors. Lim et al. [124] demonstrated that SIRT1 interacted with HIF-1α which was acetylated by PCAF at Lys674, deacetylated it, and thus repressed the HIF-1α-induced transactivation by blocking the recruitment of p300 protein. In fact, this mechanism enhanced the HIF-1α transactivation capacity in hypoxia, since the level of NAD⁺ decreased and the activity of SIRT1 was attenuated during hypoxia. In conclusion, the 2-OGDO enzymes, PHDs and FIH-1, are the major regulators of hypoxic and pseudohypoxic responses, both in physiological processes and many diseases such as coronary artery diseases [125].

Collagen hydroxylases regulate collagen synthesis

In 1967, Kivirikko and Prockop [126] demonstrated that procollagen was enzymatically hydroxylated at proline and lysine residues during the collagen synthesis. Moreover, they revealed that the enzymatic reaction was dependent on 2-oxoglutarate, Fe²⁺, and ascorbate concentrations. The kinetic studies on prolyl hydroxylation revealed for the first time that these enzymes were included in the family of 2-OGDO enzymes [73]. Oxygen was an obligatory substrate for the decarboxylation of 2-oxoglutarate, whereas succinate was a potent competitive inhibitor of the collagen hydroxylases. Later molecular studies revealed that there are two different types of collagen prolyl hydroxylases, i.e., prolyl 3-hydroxylases (P3H) and prolyl 4-hydroxylases (P4H) [127, 128]. The P3H enzymes include two (P3HA1-2) and the P4H three (P4HA1-3) α-isoforms. In addition, collagen lysyl hydroxylases involve a family of three procollagen-lysine 2-oxoglutarate 5-dioxygenases (PLOD1-3) [129]. The proline hydroxylases act as α2β2 heterotetramers hydroxylating collagen peptides in -X-Pro-Gly sequences forming the collagen triple-helix structure within the endoplasmic reticulum. The β-subunits include a domain of protein disulfide isomerase (PDI), which assembles disulfide bonds between peptides stabilizing the tertiary fibrillary structure of collagen [130]. On the other hand, lysine hydroxylase 3, i.e., PLOD3, contains domains for glucosyltransferase (GGT) and galactosyltransferase (GT) enzymes and thus it can control the glycosylation of collagen fibrils [129]. Defect in the glycosylation capacity of PLOD3 impaired the formation of basement membranes [131]. Moreover, PLOD3 can be secreted from the endoplasmic reticulum to the extracellular space, where it affects protein glycosylation and cellular matrix remodeling [132].

OGFOD1 ribosomal hydroxylase

In addition to collagen hydroxylases, the specific ribosomal protein hydroxylases, included in the 2-OGDO family, have been identified in both prokaryotes and humans [74]. These dioxygenases contain a Jumonji domain, which is capable of hydroxylation but not demethylation reactions. Recently, Singleton et al. [12] demonstrated that the human 2-oxoglutarate and Fe²⁺-dependent oxygenase domain-containing protein 1 (OGFOD1) is a prolyl hydroxylase, which targets the Pro62 residue of the small ribosomal protein component 23 (RPS23). They also revealed that the knockdown of OGFOD1 in cell lines arrested the mRNA translation and provoked the formation of stress granules. There are OGFOD1 homologs, e.g., in yeast (TPA1) [133] and Drosophila (Sudestada1) [134]. Katz et al. [134] reported that the knockdown of Sudestada1 impaired the translation of mRNA and consequently reduced the normal growth of Drosophila. They also revealed that the translational stress triggered the unfolded protein response of endoplasmic reticulum and the appearance of stress granules in the flies. Wehner et al. [135] observed that human OGFOD1 is mainly a nuclear protein but it is translocated to stress granules in arsenite and thapsigargin-induced stresses. Particularly, OGFOD1 interacted with the G3BP1 protein in stress granules provoking apoptotic cell death. Given that OGFOD1 is present in the nuclei, it can mediate nuclear signals to the cytoplasm and also control other functions than stress granule formation.

Studies on Saccharomyces cerevisiae have revealed that TPA1 is involved in different cellular functions, e.g., translation termination [133] but also in the repair of DNA alkylation damages [136]. However, Loenarz et al. [137] demonstrated that OGFOD1 increased the translational accuracy rather than affected the termination of translation in animal cells. The hydroxylation of RPS23 augmented the binding of OGFOD1 to the RPS23 component, which subsequently influenced the function of the ribosomal decoding center. There are also observations that the activation of OGFOD1 can induce a transient phosphorylation of eIF2α, which is a survival response to stress [12, 135]. It seems that OGFOD1 and its homologs have a crucial role in the ribosomal function, e.g., in the control of protein synthesis, which is clearly declined during the aging process (see below).

Potential impact of 2-OGDO enzymes on the aging process

Given that the 2-OGDO enzymes is an ancient protein family and involved in the regulation of fundamental cellular processes, it is plausible that 2-OGDO enzymes have a role in the regulation of the aging process. Next, we will review the literature indicating the potential connections of aforementioned 2-OGDO enzymes to the age-related degenerative processes (Fig. 2).

Fig. 2.

A schematic presentation on the age-related pathological functions of distinct 2-oxoglutarate-dependent dioxygenases (2-OGDO). The functions of 2-OGDO enzymes are controlled by disturbances in energy metabolism, O₂ availability, and iron homeostasis. The 2-OGDO enzymes provoke many pathological disturbances associated with the aging process. The pathological responses induced by distinct 2-OGDO enzymes are detailed in the text

Epigenetic control of chromatin landscape with aging

Currently, modern techniques, e.g., epigenome-wide association studies (EWAS), have been used to screen epigenetic changes induced by the aging process in different tissues. Several studies have revealed significant changes in the DNA methylation pattern of many tissues with aging [138–141]. However, it is still difficult to conclude whether they are directing the aging process or only coincidences with frailty and many diseases. The genome-wide screening studies have revealed a global decrease in DNA methylation status, although concurrently the methylation level was clearly increased in distinct DNA loci, especially in CpG islands [138–140]. This mode is called an epigenetic drift [142]. This kind of tendency has been observed in many diseases, e.g., in the late-onset Alzheimer’s disease [143]. Clear alterations in DNA methylation profiles have been observed in progeroid syndromes, which could have a causative role in the premature aging process [141]. However, it seems to be difficult to repeat the results on the age-related changes in DNA methylation profiles, which might indicate that changes are stochastic rather than induced through a programmed mechanism. The age-related changes in DNA methylation also seem to be tissue-specific responses [138]. Histone methylation is a more flexible epigenetic mark than DNA methylation, although the H3K9me3, H3K27me3 and H4K20me3 modifications, associated with, e.g., heterochromatin formation, can be rather stable structures. Recent studies have revealed significant alterations in the histone methylome during the aging process, both at the global level and more specific loci [144, 145]. For instance, the levels of repressive histone marks, H3K9me3 and H3K27me3, decrease with aging and cellular senescence, probably linked to a decline in heterochromatin [144, 146]. Interestingly, the decrease in the level of H3K27me3 is associated with the aging process of C. elegans (see below).

Emerging studies have revealed that distinct KDM2-7 subgroups, in particular KDM2 and KDM6, are potentially involved in the regulation of the aging process and many age-related diseases [8, 98, 147]. Next, we will shortly review the KDM2-mediated epigenetic control of ribosomal DNA locus, which is associated with cellular senescence and the aging process, as described earlier [146, 148, 149]. Recent studies have also revealed that Cockayne syndrome proteins A (CSA) and B (CSB), the mutations of which induce a premature senescence, control the ribosomal biogenesis [150, 151]. There are many studies indicating that KDM2A and 2B are potent regulators of ribosomal DNA (rDNA) locus by demethylating H3K4me3 and H3K36me2 sites, which are activating epigenetic marks. Frescas et al. [152] demonstrated that KDM2B was localized in the nucleoli of several cell types and its activation decreased the level of H3K4me3 in the rDNA locus, which repressed the transcription of ribosomal RNA genes. Impaired ribosomal biogenesis decreased the cell growth, which appeared as reduced cell size and proliferation rate. Later, Tanaka et al. [153] observed that KDM2A also repressed the transcription of ribosomal RNA genes by demethylating H3K36me2 mark. Interestingly, the starvation of MCF-7 cells activated KDM2A and reduced the level of H3K36me2 on the rDNA promoter region repressing the transcription of RNA genes. They also demonstrated that the treatment of cells with cell-permeable succinate, an inhibitor of 2-OGDO enzymes, inhibited the activity of KDM2A, prevented the demethylation of H3K36me2 sites, and subsequently restored the transcription of ribosomal RNA during cellular starvation.

The members of KDM6 subfamily, KDM6A (UTX) and KDM6B (JMJD3), have been linked to the aging process and inflammatory diseases. KDM6 enzymes activate gene expression by demethylating H3K27me2,3 sites, which are repressive epigenetic marks inhibiting gene expression [154]. Moreover, methylated H3K27 sites also have a crucial role in the maintenance of chromatin organization through the binding of polycomb complexes [155]. In 2011, both Jin et al. [156] and Maures et al. [157] demonstrated that the inhibition of UTX-1 expression in C. elegans extended the lifespan of worms by 30 %, which was associated with a significant delay in the age-related decrease in the global level of H3K27me3 mark. Both research groups also revealed that the extension of lifespan was dependent on insulin/IGF-signaling pathway, which is commonly linked to the aging process [158]. Jin et al. [156] also observed that the deficiency of UTX-1 (KDM6A) in C. elegans reduced the expression of genes of DAF-2 pathway (insulin/IGF in mammals), e.g., daf-2, akt-1 and akt-2. A defect in the phosphorylation of DAF16 (FOXO proteins in mammals) enhanced its nuclear translocation. DAF16/FOXO proteins are well-known longevity and stress resistance factors [158]. In fact, Jin et al. [156] reported that the UTX-1 knockout mice were resistant to heat, UV and oxidative stress. They also confirmed that the depletion of UTX-1 increased the level of H3K27me3 on the promoters of daf-2, akt-1 and akt-2, which indicated that UTX-1 (KDM6A) regulated the insulin/IGF-1 pathway in a demethylase-dependent manner.

The KDM6B (JMJD3) controls the activity of many inflammatory mediators, such as NF-κB, IRF4, SMAD3, and STATs, as we have recently reviewed in detail [147]. KDM6B interacts with transcription factors and enhances the initiation of transcription by demethylating repressive H3K27me3 sites and subsequently removing inhibitory polycomb complexes. Moreover, KDM6B can enhance the transcription elongation by demethylating the H3K27me3-enriched gene bodies [159]. Several studies have also revealed that KDM6B is an inducible histone demethylase, stimulated, e.g., by many inflammatory factors as well as metabolic and hypoxic stresses [160, 161]. This is an interesting observation since a low-grade inflammation is associated with the aging process and many age-related diseases. KDM6B (JMJD3) can also control the INK4 box. Barradas et al. [162] demonstrated that the RAS-oncogene, known to induce cellular senescence, activated the expression of KDM6B in human embryonal fibroblasts. Consequently, KDM6B reduced the level of H3K27me3 at the Ink4a/Arf locus and stimulated the expression of INK4a, which arrested the proliferation and provoked senescence. Agherbi et al. [163] observed that the onset of cellular senescence was associated with a loss of H3K27me3 as well as the deficit of polycomb complexes PRC1 and PRC2 at the regulatory domain (RD) of Ink4a/Arf locus, concurrently with the accumulation of KDM6B enzymes to the RD domain. This indicates that KDM6B has a fundamental role in the regulation of cellular senescence and longevity, although there will be other epigenetic players, such as long noncoding RNA ANRIL [164], in the regulation of INK4-dependent cellular senescence.

In conclusion, it seems that the excessive activation of KDM2 and KDM6 demethylases enhance the cellular senescence and could propagate the processes leading to premature aging. Given that these enzymes are 2-OGDO enzymes, their inhibition, e.g., by reduced oxygen availability or increased succinate accumulation, as observed in hypoxia [165], might attenuate the aging process. However, the hypoxia-induced increase in HIF-1α expression also stimulates the expression of other KDMs, e.g., KDM6B [161], which seems to be a feedback response to enhance gene expression during hypoxia although it might jeopardize chromatin stability.

Hypoxia/pseudohypoxia regulation

There is an extensive literature indicating that the hypoxia (or anoxia) resistance (tolerance) is associated with an increased lifespan in many animal species [166, 167]. Generally, the hypoxia resistance seems to be a part of common stress resistance, as e.g., in naked mole rats [168] or ocean quahog Arctica islandica [169]. The blind naked mole rats (Spalax species) are a unique model to study the role of hypoxia resistance in longevity regulation since the rats live in hypoxic, subterranean conditions for about 30 years. Currently, the mechanism behind the multi-resistance, even against cancer formation, has remained elusive, although the transcriptome of mole rats was recently screened [170]. Shams et al. [171] observed that in normoxia, the expression level of HIF-1α was significantly higher in the kidney of Spalax compared with that of Rattus norvegicus. Correspondingly, the hypoxia-induced response was remarkably stronger and faster in Spalax than in Rattus. Moreover, there are noticeable differences in the expression of energy metabolic enzymes and protein synthesis [170], which could affect the longevity by decreasing oxidative metabolism and enhancing anaerobic energy production. Given that the oxygen sensors, i.e., PHD1-3 and FIH-1, have a crucial role in the stabilization of HIF-1α and thus the induction of hypoxia resistance, there are therapeutic approaches to inhibit these hydroxylases, which could prevent ischemic and inflammatory diseases [172]. Recent studies have indicated that PHD1 and HIF-2α axis has an important role in the prevention of ischemia–reperfusion injury [173]. Many genetic screening studies on the mechanism of hypoxia tolerance in Tibetan highlanders have revealed a positive selection of PHD2 and HIF-2α (EPAS) gene variants, which are associated with high-altitude adaptations [174]. Song et al. [175] demonstrated that the Tibetan amino acid substitutions disturbed the binding of PHD2 to the HSP90 co-chaperone p23, an interaction which facilitates HIF-2α hydroxylation. This defect seems to induce the Tibetan high-altitude resistance. Whether or not this PHD2 mutation is linked to more general stress resistance or longevity needs to be clarified.

Different types of experiments have indicated that the tissue responsiveness to hypoxia clearly declines with aging [176, 177]. Generally, the hypoxia-induced increase in the expression levels of HIF-1α and its target genes reduce through the aging process. It seems that there are several causes for this downregulation with aging. One important mechanism could be a common decline of the protein synthesis capacity, which also affects many other stress responses and aggravates age-related diseases (see below). Ndubuizu et al. [176] demonstrated that the expression levels of PHD1 and PHD3 were significantly increased with aging in normoxic rat brain. They also reported that the hypoxia-induced accumulation of HIF-1α and its targets, e.g., GLUT-1 and VEGF, were significantly reduced with aging. Moreover, Rohrbach et al. [178] observed that caloric restriction counteracted the age-related increase in the expression of PHD3 in mouse heart, liver and skeletal muscle that should enhance the hypoxic responses and improve cell survival. However, an increased expression of PHDs does not mean that their activity would be enhanced since they are under the metabolic (2-oxoglutarate vs. succinate) and iron regulation, too. In addition, there are many transcriptional and post-translational mechanisms, which control the function of HIF factors and thus could attenuate the hypoxia response during aging [179].

Hypoxia responses are commonly short-term events and subsequently HIF-1α factor stimulates the expression of PHD2 and PHD3 as a negative feedback response [52]. This attenuates the excessive expression of HIF factors, thus protecting tissues in chronic hypoxia. Oxygen deficiency is not the only way to induce the expression of HIF factors and their induction in normoxia is called pseudohypoxia. The accumulation of succinate and fumarate, e.g., in the case of succinate dehydrogenase and fumarase mutations, inhibits PHD activities and consequently induces the expression of HIF-1α [46, 180]. Given that Fe²⁺ is required for the activation of PHDs, it is possible that dysfunctions in the iron homeostasis can also trigger pseudohypoxic state. There are numerous signaling pathways, which directly control the function of HIF factors and thus these factors can context dependently either protect tissues or provoke detrimental responses. For instance, the NF-κB signaling can enhance HIF-1α transcription [181] or synergistically with HIF-1α stimulate the expression of HIF-1α target genes [182]. In addition, there are many observations indicating that PHDs can inhibit NF-κB signaling by blocking the activity of IKK-complex [183]. This means that pseudohypoxia, e.g., via succinate accumulation, stimulates NF-κB signaling and subsequently enhances hypoxia response along with HIF-1α. Interestingly, several signaling pathways known to be associated with the regulation of longevity are linked to the control of HIF-1α, e.g., the insulin/PI-3K/mTOR signaling activates HIF-1α expression [184], whereas FoxO3a represses the transcription of HIF-1α [185]. Currently, it seems that the PHD/HIF-1 axis can have both positive and negative responses to the control of longevity in a context-dependent manner [186]. The activation of angiogenesis is a crucial outcome of the stimulation of HIF system (Fig. 2.), which can be beneficial in cardiovascular diseases [187] but detrimental in the age-related macular degeneration [188] or in cancer progression [189].

Defects in collagen metabolism

Collagen is the main component of connective tissue and thus a critical composition in the development of multicellular organisms. The collagen superfamily has enlarged during the evolution of metazoans and the triple-helix conformation was evolved during the radiation of animal kingdom [190]. Given that the amounts of hydroxylated proline and lysine residues are high in mature collagen, it seems that the collagen hydroxylases have a fundamental role in the stabilization of fibril-forming collagen species in the tissues of animals. There is a substantial literature indicating that the aging process affects the collagen composition and organization in tissues, e.g., in skin [191], bone [192], and vascular wall [193] (Fig. 2). Briefly, there seems to be a clear decrease in the rate of collagen synthesis and structural remodeling in several tissues. The enzymatic assays in vitro have revealed a significant reduction in the activity of prolyl and lysyl hydroxylases of collagen with aging in many rat tissues [194, 195]. The knockout studies of prolyl and lysyl hydroxylases have revealed marked, tissue-specific abnormalities in collagen-rich tissues. For instance, Ruotsalainen et al. [131] demonstrated that the deficiency of PLOD3 activity in mouse mutants disturbed the collagen fibril organization, especially that of type IV collagen in the basement membranes. On the other hand, the knockout of P3H1 resulted depletion in the hydroxylation of type 1 collagen and serious defects in collagen fiber architecture, e.g., in skin, bone, and tendon [196]. It seems that collagen hydroxylases have a crucial role in the maintenance of collagen fibril organization, e.g., in the extracellular matrix, which is remodeled with aging in many tissues.

There is an abundant literature indicating that ascorbate (vitamin C) stimulates collagen biosynthesis, improves wound healing, and prevents the appearance of scurvy [197, 198]. Given that ascorbate is a cofactor for the 2-OGDO enzymes, several reports have linked increased collagen synthesis to enhanced collagen hydroxylation and thus to improved quality control of collagen. However, it seems that ascorbate also stimulates collagen synthesis at the transcriptional level, which could be associated with the ascorbate-induced epigenetic regulation of collagen transcription controlled by DNA and histone demethylases. This is a possible regulation pathway in fibrotic diseases [199]. Considering that ascorbate replenishes Fe²⁺ and activates 2-OGDO enzymes, there are reports indicating that free iron content stimulates collagen synthesis in cultured cells [200]. Gilkes et al. [201] demonstrated that HIF-1α promotes the synthesis of collagen by inducing the expression of collagen prolyl (P4HA1 and P4HA2) and lysyl (PLOD2) hydroxylases in human skin fibroblasts. They also reported that HIF-1α provoked a significant remodeling of extracellular matrix by increasing collagen alignment and stiffness which affected the cell adhesion. However, ascorbate also stimulates the activities of PHDs and thus suppresses the expression of HIF-1α (see above) which reduces the role of HIF-1α in collagen metabolism. Recently, Ewald et al. [202] revealed that the reduction of the insulin/IGF-1 signaling extended the lifespan of C. elegans by activating SKN-1 (human ortholog of NRF2). Surprisingly, they observed that SKN-1 robustly increased the expression of collagen genes and many other genes controlling extracellular matrix composition. Moreover, they reported that different genetic mutants and pharmacological treatments, known to increase lifespan, stimulated the expression of collagen genes in C. elegans. These studies imply that the age-related decrease in the collagen expression and extracellular matrix remodeling have detrimental effects on the maintenance of cellular homeostasis, which might enhance the degenerative aging process.

Decline in protein synthesis with aging

Protein synthesis is one of the fundamental cellular processes, which is tightly controlled through different mechanisms. In 1963, Orgel [203] proposed that a decrease in the translational accuracy could be an important source for age-related degenerative changes. However, this “error catastrophe theory of aging” has not received experimental verification [204]. On the contrary, there is an extensive literature indicating that the rate of protein synthesis and turnover significantly declines with aging in many tissues [205–207]. Especially, it seems that the initiation phase of mRNA translation is delayed with aging, although its significance to the aging process needs to be clarified. Unexpectedly, experimental studies have revealed that the reduction of mRNA translation in ribosomes is associated with the extension of lifespan in C. elegans [5, 208]. It appears that a decrease in the translation rate improves the fidelity of protein synthesis and the folding of peptides [209]. There seems to be a signaling interplay between the mechanisms controlling translation rate and longevity, e.g., insulin/IGF1 pathway [207]. Moreover, the reduction of mRNA translation seems to be a more common survival mechanism, since the depletion of translation factors provided the hypoxic resistance in C. elegans and mouse neurons [210].

Considering that the ribosomal component RPS23 has a central role in the mRNA translation, it is interesting that OGFOD1 controls translation through the hydroxylation of RPS23 [12]. The knockdown of Sudestada1 (OGFOD1 homolog) in Drosophila impaired the translational efficiency and impeded the normal growth of flies [134]. Moreover, Saito et al. [211] reported that the silencing of OGFOD1 prevented ischemic injuries in cultured cells, which could be related to a decrease in mRNA translation rate. It is not known whether OGFOD1 might regulate the longevity. Given that OGFOD1 is a 2-OGDO enzyme, its activity is most probably regulated by energy metabolites, i.e., 2-oxoglutarate and succinate. In this respect, it is a paradox that disturbances in mitochondrial respiration extended the longevity of C. elegans [4, 212]. The mechanism of lifespan extension via mitochondrial disturbances is still elusive. Future studies will reveal whether the longevity regulation related to mRNA translation rate is controlled by OGFOD1 being under the control of energy metabolism, oxygen availability, and iron homeostasis.

Conclusions

The 2-OGDO enzymes modifying proteins/DNA via the hydroxylation and demethylation reactions have crucial cellular functions, e.g., the control of epigenetic gene expression, ribosomal protein translation, and collagen metabolism as well as the capacity to induce adaptation against hypoxia. These enzymes are evolutionarily conserved, which implies that they have fundamental functions in cells. Interestingly, it is known well that the functions driven by TETs, KDMs, PHDs, and OXFOD1 are significantly affected during the aging process. It seems that energy metabolism has a critical role in their regulation, since 2-oxoglutarate is an obligatory substrate for these enzymes, whereas succinate and fumarate are potent inhibitors. This underlines the important role of Krebs cycle intermediates in many vital cellular functions. These enzymes also require O₂ for the decarboxylation reaction and thus they are sensors for O₂ level in cells, especially PHDs. On the other hand, hypoxia affects the Krebs cycle enhancing succinate production, which induces pseudohypoxia in cells. The third important component in the catalytic activity of 2-OGDO enzymes is iron and its redox regulation by ascorbate. Iron homeostasis is disturbed in many age-related diseases, which could impair the function of 2-OGDO enzymes and aggravate the pathology. It seems that the 2-OGDO enzymes discussed above are at the nexus of age-related regulation, i.e., (i) they are controlled by energy metabolism, O₂ availability and Fe²⁺ redox regulation, and on the other hand, (ii) disturbances in their function can trigger many hallmarks of the aging process (Fig. 2).