Selenium vitaminology: The connection between selenium, vitamin C, vitamin E, and ergothioneine

Abstract

Selenium is connected to three small molecule antioxidant compounds, ascorbate, α-tocopherol, and ergothioneine. Ascorbate and α-tocopherol are true vitamins, while ergothioneine is a “vitamin-like” compound. Here we review how selenium is connected to all three. Selenium and vitamin E work together as a team to prevent lipid peroxidation. Vitamin E quenches lipid hydroperoxyl radicals and the resulting lipid hydroperoxide is then converted to the lipid alcohol by selenocysteine-containing glutathione peroxidase. Ascorbate reduces the resulting α-tocopheroxyl radical in this reaction back to α-tocopherol with concomitant production of the ascorbyl radical. The ascorbyl radical can be reduced back to ascorbate by selenocysteine-containing thioredoxin reductase. Ergothioneine and ascorbate are both water soluble, small molecule reductants that can reduce free radicals and redox-active metals. Thioredoxin reductase can reduce oxidized forms of ergothioneine, while the biological significance is not yet realized, this discovery underscores the centrality of selenium to all three of these antioxidants.

Graphical Abstract

Introduction:

Selenium is an essential micronutrient for mammals because it is found in important antioxidant enzymes as the 21^st amino acid in the genetic code, selenocysteine [1]. Two of these selenium-containing antioxidant enzymes, glutathione peroxidase-4 (GPX4) and thioredoxin reductase (TXNRD) are connected to α-tocopherol (α-TO), ascorbate (Asc), and ergothioneine (EGT). Both Asc and α-TO are considered to be true vitamins (for humans), while EGT is considered to be a “potential vitamin” or a “candidate vitamin” because it is an organic compound obtained in the diet and has been shown to provide some health benefits [2-4]. A true vitamin must satisfy two criteria: It must be an organic compound that must be obtained in the diet because it cannot be biosynthesized by humans, and its presence in the diet is essential in preventing disease.

Humans are missing the enzyme that catalyzes the last step in Asc biosynthesis and thus must acquire it in the diet to prevent the disease known as scurvy [5]. Scurvy is a disease of impaired collagen biosynthesis because Asc is a cosubstrate for prolyl hydroxylase, which catalyzes the conversion of peptidyl-proline to peptidyl-hydroxyproline. The Asc is needed to keep the enzyme bound iron reduced, preventing self-hydroxylation and inactivation [6]. Without dietary Asc, prolyl hydroxylase ceases to function and collagen cannot be replaced, eventually resulting in hemorrhage. Since Asc is water soluble, it is not stored in the body, and must be continually present in the diet to prevent scurvy.

Similarly, α-TO is acquired in the diet, and since it is lipid soluble, it partitions into cellular membranes where it functions to scavenge lipid radicals and lipid hydroperoxyl radicals, thus helping to prevent lipid peroxidation [7]. Unlike Asc, which is not stored, α-TO can be stored in adipose tissue. It is therefore relatively difficult for a human to become deficient in α-TO, which is why it was not recognized to be essential for humans until 1983 [8]. Deficiency of α-TO only becomes apparent in humans who have abetalipoproteinemia and other fat absorption disorders. Abetaliproteinemia is a genetic disorder in which individuals cannot make chylomicrons, the lipoprotein that binds fat-soluble dietary vitamins such as α-TO, vitamin K, and vitamin A. Chylomicrons transport α-TO to the liver where it is transferred to alpha-tocopherol transfer protein, which distributes it to tissues. Individuals who have abetalipoproteinemia must be supplemented with large doses of α-TO in order to prevent neurological disease, which first manifests as ataxic neuropathy and retinal pigmentation. Left untreated, this progresses to neuronal degeneration in the brainstem, spinal cord, and peripheral nerves [9].

Since the essentiality of α-TO only becomes apparent under a stress condition, this may be a good model for EGT, the deficiency of which is not currently associated with a particular disease. However, evidence of its importance to humans comes from the fact that humans have a specific ergothioneine transporter (ETT) encoded by the slc22a4 gene that enriches tissues with EGT [10^••]. As pointed out by Gründemann, ETT is highly specific for EGT since it binds the structurally similar compound, carnitine, 100-fold less tightly [10^••]. While the ETT is present in many human tissues, it is especially concentrated in the bone marrow, granulocytes, kidney, small intestine, fallopian tubes, eye, and various parts of the brain [11^•]. Donald Melville, a pioneer in the study of EGT who long searched for the biological function of EGT said, “Perhaps its function becomes apparent only under conditions of stress.” [12]. To Melville’s point, low plasma levels of EGT are associated with increased frailty and increased cognitive impairment in the elderly [13^••, 14^•]. Conversely, high plasma levels of EGT are associated with lower levels of cardiovascular disease and lower overall mortality [15^••].

The connection between selenium and these three antioxidants is discussed below.

Selenium connection to vitamin E:

The connection between vitamin E and selenium was established by Karl Schwarz while studying dietary induced liver necrosis in rats. He found that three substances could prevent liver necrosis: high concentrations of cystine (termed “factor 1”), vitamin E, and a third substance (initially unknown) that he labeled as “factor 3” since it was the third dietary factor that could prevent liver necrosis [16]. Later, this third factor was discovered to be an organic form of selenium [17]. To the best of my knowledge, the specific form of selenium that Schwarz discovered that prevented liver necrosis has not been identified, but could potentially be selenomethionine [18]. Vitamin E as α-TO could prevent liver necrosis in rats if factor 3 (selenium) was missing in the diet. This is referred to as a “mutually sparing” effect [19].

Selenium and α-TO work together as partners to prevent lipid peroxidation in biological cell membranes [20^••]. Lipid hydroperoxyl radicals are quenched to lipid hydroperoxides by α-TO. The resulting lipid hydroperoxide is then reduced to a lipid alcohol and water by the action of phospholipid glutathione peroxidase (GPX4), using glutathione (GSH) as a coenzyme. GPX4 is essential in protecting against ferroptosis [21]. Although lipid peroxidation is most effectively inhibited when both α-TO and GPX4 (containing selenium) are present in the membrane, the presence of one of these is sufficient to prevent liver necrosis by protecting the membrane from lipid peroxidation.

The connection between α-TO and selenium is further strengthened by the fact that selenium deficiency appears to enhance loss of α-TO [22]. As noted by Burk, this is likely due increased α-TO oxidation when bodily selenium content is low [23].

Selenium connection to vitamin C:

It was discovered that TXNRD was a selenocysteine-containing enzyme in 1996 by Stadtman [24]. Soon after Burk, May, and coworkers discovered that TXNRD could reduce both dehydroascorbic acid (DHAA) and ascorbyl free radical (AFR) to Asc [25, 26]. This shows a direct connection between selenium and vitamin C. The AFR is recycled back to Asc by NADH- and NADPH-dependent microsomal and mitochondrial reductases as well as by TXNRD, which appears to contribute ~20% of the total AFR-recycling activity in rat liver based upon the decrease in activity under conditions of selenium deficiency [26]. Recycling of DHAA to Asc occurs both by a non-enzymatic pathway via GSH and multiple enzymatic pathways including glutaredoxin (Grx), protein disulfide isomerase (PDI), and 3α-hydroxysteroid dehydrogenase [27]. The data indicates that TXNRD contributes ~20% of the total DHAA-recycling activity in rat liver [25].

The fact that Grx2 and PDI can recycle DHAA to Asc highlights another way in which selenium is connected to vitamin E [28]. In vitro experiments have shown that the α-tocopheroxyl radical (α-TO•) can be reduced back to α-TO by Asc, although whether this occurs in vivo is a subject of debate [29]. If biological reduction of α-TO• by Asc becomes firmly established, then selenium, vitamin C, and vitamin E are all connected since TXNRD recycles DHAA to Asc as shown in Figure 1. As pointed out by Stadtman, this is another way in which selenium and vitamin E can be “mutually sparing” [19].

Figure 1: Connection between selenium and vitamin C and E.

Recycling of DHAA to Asc can occur by both Grx2 and PDI, which are substrates of selenocysteine-containing TXNRD. Asc may recycle α-TO• to α-TO in vivo.

Selenium connection to ergothioneine:

While the biological function of EGT in animals is yet to be fully understood, an antioxidant role similar to Asc and α-TO seems likely [30^••]. EGT is biosynthesized in fungi and some bacteria by combining cysteine and histidine to yield the betaine of 2-thiohistidine (Figure 2) [31]. EGT shares some physico-chemical properties with Asc. Both are highly water soluble, quench free radicals, reduce metals, and react with singlet oxygen (¹O₂) [30^••]. When acting as antioxidants, they both readily undergo one-electron reductions, though EGT can also participate in two-electron reductions. A notable difference is that EGT binds divalent metal cations such as Fe²⁺, Cu²⁺, and Zn²⁺ very tightly [30^••]. The thione of EGT is in equilibrium with the thiol form (Figure 2), with the thione form predominating. Another difference with Asc is that the thiol form (ESH) can be oxidized to a disulfide, a form not available to Asc [31].

Figure 2: Structures of Asc and EGT and comparison of their oxidized forms.

When Asc and EGT undergo one-electron redox reactions, a resonance stabilized radical species is formed. Note the structural similarity of the radical species. The ascorbyl free radical (AFR) forms from an enol, while the ergothioneine radical (ES•) forms from the structurally isologous thiolimine. Reaction of Asc with ¹O₂ results in the formation of DHAA, while the same reaction with EGT yields the structurally similar 5-oxo-EGT species (major) and the EGT-sulfine (minor). The ESH form of EGT can oxidize to the disulfide ESSE, which is an unstable disulfide that is similar to the reactive disulfide DTNB. Both ESSE and DTNB are substrates for TXND.

Since its discovery in 1909, the biological function of EGT in animals has been long sought [32]. Outstanding work by Gründemann and coworkers has suggested that the specific redox niche that EGT occupies is that of a ¹O₂ scavenger [3, 33]. Singlet oxygen is an excited form of oxygen in which one of the unpaired electrons in a π* orbital flips its spin and becomes paired with the remaining unpaired electron in the degenerate π* orbital, leaving an empty π* orbital. This excited form of oxygen is a strong oxidizing agent. The major reaction pathway between EGT and ¹O₂ produces 5-oxo-EGT, while a minor reaction pathway yields EGT-sulfine (structures are shown in Figure 2) [34^••].

It is an under recognized fact that ¹O₂ is produced in the human body by multiple different pathways [35^••]. One of the most important of these pathways is the production of ¹O₂ in immune cells such as macrophages and neutrophils, where it is generated by an enzymatic pathway that utilizes myeloperoxidase and superoxide dismutase [36, 37]. The ¹O₂ generated in this pathway is used to kill invading bacteria. Another important ¹O₂-generating pathway is the reaction between protoporphyrin IX, the iron-free precursor of heme, in erythropoietic cells [38]. This pathway is important in erythropoietic protoporphyria, a genetic disorder due to a defect in ferrichetalase. Affected individuals accumulate protoporphyrin IX in erythrocytes, skin, and liver and are sensitive to sunlight. The production of ¹O₂ in the afflicted results in increased oxidative stress which contributes to the pathology of the disease [39, 40^•]. Other notable ways in which ¹O₂ is generated in the body include the decomposition of lipid hydroperoxides and the reaction of GSH with superoxide [41, 42].

While it is true that Asc is an excellent quencher of ¹O₂, (the rate constant for the reaction is 3 x 10⁸ M⁻¹s⁻¹), [43] EGT reacts ~4-fold faster based upon the data of Gründemann and coworkers. [33] Moreover, the reaction of Asc with ¹O₂ produces H₂O₂, which then must be detoxified by the consumption of additional reducing equivalents. The reaction of EGT with ¹O₂ produces 5-oxo-EGT by the major pathway, which can be rapidly reduced by four equivalents of GSH [33]. The resulting GSSG can then be recycled back to GSH by glutathione reductase with the consumption of NADPH.

My group has had a long standing interest in the catalytic mechanism of TXNRD. Recently, we noticed the structural similarity between the oxidized forms of EGT and Asc [44^••]. Since AFR and DHAA are substrates for TXNRD, we wondered if ES• and 5-oxo-EGT could be substrates for TXNRD. We also wondered whether the disulfide form of EGT (ESSE), could also be a substrate since it is a highly reactive, electrophilic disulfide, similar to 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB), which is often used as a model substrate to test the reactivity of TXRND.

We were able to produce ESSE by reaction of EGT with H₂O₂ in situ and then add TXNRD2 from mouse and 10 μM E. coli thioredoxin (Trx). The catalytic efficiency (k_cat/K_M) of this system was 1.34 x 10⁷ min⁻¹ M⁻¹, which was 10-fold lower compared to using glutathione reductase and GSH [44^••]. Currently, we have not investigated the kinetics of this reaction with human TXNRD1 and human Trx, which may give a higher value. At the moment we can only speculate that the human thioredoxin system may make an equal contribution to the recycling of ESSE to EGT. We were also able to show that 5-oxo-EGT could be reduced by mouse TXNRD2, but the kinetics are complicated by the fact that we generated both 5-oxo-EGT and ESSE with our ¹O₂ generating system [44^••]. We were not able to show that ES• was a substrate for TXNRD2, probably because its lifetime is too short as it can rapidly form ESSE. I end this section by noting that the role of the thioredoxin system in the redox biology of EGT is currently unexplored.

Could EGT, Asc, and selenium all be connected?:

In Figure 1, we show a potential way in which selenium (as part of TXNRD), Asc, and vitamin E are connected. I conclude this review with the musing that EGT, Asc, and selenium could be connected together by the relationship depicted in Figure 3. EGT is reported to have a redox potential of −0.06 V. I recently questioned whether this value was correct because it was reported in 1944 along with the redox potential of GSH. The value reported for GSH using the same method to determine the redox potential of EGT was reported as +0.03 V [45]. The widely accepted value for the redox potential of GSH is in the range of −0.24 V to −0.26 V [46, 47]. Thus, there is reason to be skeptical of the value reported for EGT. The redox potential of Asc was recently measured as +0.35 V, though it has been reported to be in the range of +0.06 V to + 0.40 V [48, 49]. If the reported redox potential of EGT is correct, then the more negative value for EGT indicates it should be able to reduce DHAA to Asc, thus have I indicated the direction of electron flow in Figure 3. However, we were able to reduce the disulfide bond in peptides containing 2-thiohistidine very readily with Asc, [50^•] which indicates that the positions of EGT/ESSE and Asc/DHAA should be switched in Figure 3. Measurement of the redox potential is an ongoing project in my lab at the moment [51^••].

Figure 3: Potential connection between selenium, EGT, and vitamin C.

As shown here, EGT reduces DHAA to Asc with concomitant production of ESSE, which can be recycled by TXRND. This relationship depends on whether EGT is more reducing than Asc. The question mark indicates that the positions of EGT/ESSE and Asc/DHAA could be reversed. There are some tissues where both Asc and EGT are present such as granulocytes, where both are found in the mM range.