Understanding Hydrogen: Lessons to Be Learned from Physical Interactions between the Inert Gases and the Globin Superfamily

Abstract

Hydrogen gas (molecular hydrogen, H₂) has significant effects in a range of organisms, from plants to humans. Many inert gases have been reported to have similar effects, and such responses may be most pronounced when cells are stressed. Xenon (Xe), for example, is a well-known anesthetic. The direct targets of these gases, in most cases, remain elusive. Myoglobin and hemoglobin are known for their roles in the transport of gases through coordinate interactions with metals (O₂, NO, CO) and covalent modifications of thiols (NO, H₂S) and amines (CO₂). These are well exemplified in biotrophic reactions of NO with heme iron (to form iron nitrosyl heme) and cysteine (to form bioactive S-nitrosothiols) essential for tissue oxygenation. Here, we consider an alternative “third mode” of gas transport in what have been dubbed “Xenon pockets”, whereby inert gases may have functional effects. Many proteins have similar cavities, and possible effects include alterations in allosteric properties of proteins (potentially altering protein hydration). Here, it is suggested that similar to other inert gases, H₂ also has biological effects by utilizing these protein structures. This ought to be investigated further, in a range of species, to determine if this is the mode of action of H₂.

1. Introduction

There has been an escalating interest in the effect of hydrogen gas (molecular hydrogen: H₂) on biological systems, and it has been suggested that H₂ can be used as a therapeutic in biomedicine [1] and in agriculture [2]. For example, it has been suggested that H₂ treatment may be useful for mitigating the effects of neurodegenerative disease [3] and as a treatment for COVID-19 [4,5]. In plant sciences, field trials show that H₂ can improve the quality of rice [6].

H₂ can be supplied to biological systems in a variety of ways. For in-hospital treatment, it can be given as a gas, often in combination with oxygen in what is referred to as oxy-hydrogen (HHO: H₂/O₂) [7]. Alternatively, H₂ gas can be bubbled into water to create a solution which is enriched in H₂, referred to as hydrogen-rich water (HRW), as used by Lin et al. [8] when studying fatty liver caused by alcohol in mice. A variation of this is hydrogen-rich saline (HRS), for example, in [9,10]. Other variations include the use of hydrogen nanobubble water (HNW) [11], which is reported to have more hydrogen dissolved and to retain the H₂ in solution for longer. One of the issues with many H₂ treatments is that the H₂ gas quickly moves into the atmosphere and therefore has limited biological activity. It tends to provide a bolus effect. HNW may aid in transforming it into a more physiological process, a slower and longer H₂ release into biological materials. It appears that in humans H₂ can be administered to the lungs [12], as a drink [13], or can even be used as a topical treatment [14]. For plants, HRW can be supplied to the soil [15], the feed water [16] or foliage [17], and the atmosphere can be augmented with H₂ gas [18]. Alternatively, H₂ can be supplied to biological tissues via a donor molecule, which will release H₂ in the needed location. One such donor is magnesium hydride (MgH₂), as used by Li et al. [19]. Such donors, in similarity with HNW, allow for prolonged diffusion, and therefore greater/sustained physiological exposure of the organism or tissue to H₂.

With so many ways to treat biological materials with H₂, a range of physiological effects have been reported, but the underlying mechanisms remain somewhat controversial [20]. It has been widely reported that H₂ increases the antioxidant capacity of cells [21,22,23]. Enzymes such as superoxide dismutase (SOD), ascorbate peroxidase and catalase (Cat) have increased gene expression and increased activity [23], for example, whilst changes in glutathione (GSH) metabolism have also been noted [24]. However, the direct actions of H₂ which lead to such changes in the endogenous antioxidants of the cell, and therefore the intracellular redox status, have not been defined.

Others have noted that there are changes in the activity of heme oxygenase during H₂ treatment. This was reported in cucumber adventitious root development [25] and in treatment of inflammatory bowel disease [26]. In the latter paper, HRW induced gene expression of HO-1, as well as the lowering of oxidative stress, endoplasmic reticular stress, and inhibiting the immune response, all mitigating the effects of the disease [26]. However, as before, the direct action of H₂ to bring these changes about was not unraveled.

It has been suggested that H₂ acts directly as an antioxidant, particularly scavenging the hydroxyl radical (•OH), a reactive oxygen species (ROS), and peroxynitrite (ONOO⁻), a reactive nitrogen species (RNS). However, other ROS and RNS which can act in cell signaling roles, such as the superoxide anion (O₂^•−), hydrogen peroxide (H₂O₂) and nitric oxide (•NO), are all arguably more important in controlling cell function than •OH or ONOO⁻ [27,28]. However, even the action of H₂ against •OH or ONOO⁻ has been disputed. With a more in-depth look at the kinetics involved Penders et al. [29] arguing that these reactions of H₂ are not significant under physiological conditions.

If it is not known how H₂ controls gene expression, the antioxidant capacity of the cell, or the activity of heme oxygenase, and the direct interaction with ROS and RNS is questioned, alternative mechanisms of H₂ action are required.

It has been suggested that the redox midpoint potential for the H₂/H⁺ couple is low enough to drive the change of the redox status of some biomolecules [30], perhaps those containing heme and those involved in mitochondrial function. There appears to be a precedent of this in bacterial systems, with cytochrome c₃ being reduced by H₂ [31]. There is no experimental evidence of similar reactions taking place in animals or plants, but it seems premature to rule this out.

An alternative mechanism was mooted because of the spin states of H₂ [32]. Again, there is no experimental evidence given for this mode of action of H₂, but the idea of atomic states of inert gases will be revisited below, when discussing Xe, so perhaps we should not rush to rule this out. Interestingly, the Fe of hemoglobin (Hb) has been known for a long time to have paramagnetic properties, as more recently discussed [33]. According to the original paper, the oxygen molecule undergoes “a profound change in electronic structure” when it interacts with the hemoglobin [34]. Recently it has been suggested that it is possible for H₂ to directly interact with the Fe in the heme of hemoglobin [35]. In this work, the authors theoretically explore the manner in which the heme may alter the electronic configuration of the H₂ molecule, depending on whether the interaction is symmetrical or asymmetrical. They further suggest that there is the possibility for the production of hydrogen radicals (H), which then would have the capacity to react with hydroxyl radicals or peroxynitrite, and hence this may be the mechanism by which these reactive compounds are scavenged from the cell. However, as intriguing as this is, it does pose many questions. It is not reported here if H₂ binding to the hexa-position of the Fe²⁺ in deoxyhemoglobin alters the overall capacity for O₂ transport in vivo. When oxygen binds to deoxyhemoglobin, the Fe transitions to the Fe³⁺ state with the oxygen being a bound superoxide anion. Kim et al. [35] also suggest loss of negative charge on the Fe on H₂ binding, akin to the mechanism observed with oxygen. It would be interesting to observe a full set of visible wavelength spectra for hemoglobin when various forms of the protein are treated with hydrogen gas. If H₂ can bind to heme prosthetic groups as suggested by Kim et al. [35], it would be interesting to explore the way in which this could be extrapolated to a whole range of other heme-containing proteins. Some of such proteins also bind oxygen, such as nitric oxide synthase (NOS) and NADPH oxidase. Others do not react with oxygen, such as cytochrome c. In some proteins, the heme is covalently bound to the polypeptide (e.g., cytochrome c) whilst in others it is not, such as cytochrome b. Such questions are pertinent not just to animals, but to plants, too, where a true hemoglobin does not exist, although homologues do [36]. Will they be able to undergo a similar mechanism as mooted by Kim et al. [35]? In posing this question, the present work is interesting and may inform a range of experiments in the near future.

It is clear, therefore, that H₂ has biological effects, but the way it is acting is still unclear. However, several other inert gases also have biological effects, including argon (Ar), Xe, helium (He), krypton (Kr) and neon (Ne). It is unlikely that such gases have chemical effects on biological materials; being noble gases, they are chemically inert, but they may have physical effects. Here, it is proposed that the effects of H₂ may be similar, and the biological responses to H₂ may be downstream of how H₂ interacts with proteins, and this may have direct comparisons with the action of the noble gases. Several significant reactive signaling molecules, including the gas NO, display effects through the covalent modification of proteins and other bio-molecules. Such protein modifications include oxidation (by ROS) [37], S-nitrosylation (RNS S-nitrosylation is alternatively referred to as S-nitrosation) [38], nitration (by RNS) [39] or persulfidation (by hydrogen sulfide (H₂S)) [40]. However, it is very unlikely that noble gases partake in such reactions, and it is also not likely that H₂ could be involved in the catalysis of such biomolecule alterations. Therefore, if direct chemical reactions are unlikely, a more physical interaction may be responsible for the action of H₂ and other inert gases. Below the evidence is reviewed.

2. Biological Effects of Argon

Ar is an inert gas. It is a single atom with an atomic number of 18 and atomic mass of 39.948. It is the third most abundant gas in the atmosphere, at just under 1% (9340 ppm) of the composition of air. Even though it is described as “extremely inert” [41], being a noble gas, it has still been shown to possess bioactivity. Ar has been shown to have anesthetic properties at high pressure—interestingly, Kr shows the same properties [42]. Ar effects have been reported in an ischemic stroke model and to produce neuro-protective effects [43,44]. Ar also gave protection to the myocardium against infarction [45], but failed to protect kidney tissues if they are deprived of oxygen and glucose [46]. Ar was explored as a gas useful for diving in the 1930s, but was discarded because of its anesthetic effects [47].

Binding of Ar to deoxyhemoglobin was suggested in a study of nitrogen (N₂) binding [48]. It was thought that Ar should be able to bind the same hydrophobic pockets as N₂ in the proteins, and that this may have some influence on O₂ binding, but because of likely differences in 2,3-diphosphoglycerate (2,3-DPG) concentrations in the experiments of others it was hard for the authors to be confident on the exact effects.

Very recently, effects of argon in plants have been reported [41]. In this study, argon was delivered to the plants as argon-rich water (ARW) which had been created by bubbling 99.9% pure argon into distilled water. Once diluted, this gave a range of argon concentrations up to 0.750 mmol L⁻¹. This argon-containing solution increased germination rates and seedling growth when the plants (alfalfa; Medicago sativa L. “Victoria”) were under salinity stress. NaCl-induced lowering of α/β-amylase activities were abolished by argon, and gene expression studies showed that relevant genes were affected, for example, ARW increased the level of transcripts for NHX1, a Na⁺/H⁺ antiporter. Interestingly, the authors also noted an increase in the antioxidant capacity of the plants once treated with ARW, an effect often reported with HRW.

Clearly, therefore, Ar produces biological effects in a range of organisms. Being inert, it is unlikely that argon is involved in the direct chemical alteration of biological molecules, but more likely there is a physical interaction which is mediating the effects observed. Therefore, it is proposed here that H₂ may also have a similar physical interaction, and in order to understand what might be happening, turning to what is known about Xe may be a way forward.

3. Biological Effects of Xenon

Xe is also a noble gas, with an atomic number of 54 and an atomic mass of 131.293. Even though it is an inert gas, it has been shown to possess a range of bioactivities.

Xe has long been known to be an anesthetic agent [49]. Such ideas were reported by J.H. Lawrence and colleagues in the 1940s [50]. Xenon is known to possess effects as an antagonist of the N-methyl-D-aspartate (NMDA)-type glutamate receptor, and as such has been studied for its neuro-protective effects [51]. Xe has also been found to elevate hypoxia-inducible factor-1α (HIF-1α) and hence vascular endothelial growth factor (VEGF), leading to an increase in expression of inducible nitric oxide synthase (iNOS) [52]. Other biological effects include the inhibition of the migration of breast adenocarcinoma cells and decreased release of “regulated on activation normal T cell expressed and secreted” (RANTES), a pro-angiogenesis factor [53]. In endothelial cells, Xe decreased Ca²⁺ signaling during the cell cycle [54]. Xenon has also been shown to have protective effects during hypoxia [55] and during hypothermic conditions [56], although, when such work was repeated more recently, the authors stated that “Xenon gas did not affect cell function” [57]; however, the conditions of cell growth used here were not as harsh as used by others. Finally, Xe has been shown to produce anti-apoptotic effects, maintain mitochondrial integrity and inhibit the activity of caspase-3 [58].

Xenon effects have been observed in plants, too. Chlorophyll content, as well as the properties of membranes and vesicle trafficking in root cells, was affected by the treatment of xenon gas (80% Xe, 20% O₂) [59]—chlorophyll was reduced, but using root epidermal cells, it was found that Xe treatment increased the size of Brefeldin A-induced compartments (Brefeldin A inhibits vesicle recycling).

Although not an exhaustive list of the biological effects of xenon, clearly it does have an influence on cell function and is likely to impinge on cell signaling mechanisms in a wide range of organisms, from plants to humans.

4. Biological Effects of Other Noble Gases

Other inert gases also have biological effects. Helium (He) and Neon (Ne), for example, provide cardioprotection [45].

He is a non-anesthetic gas [60], but it is used for the treatment of airway obstruction and for other ventilation problems [61]. He is known to reduce ischemia–reperfusion damage, as well as affect lung tissues, blood vessels and the immune system [62]. Cell signaling components implicated in having such effects include ion channels, kinases, ROS and NO. Others too have listed signaling molecules downstream of He effects. In cardiac tissue, this included extracellular signal-regulated kinase 1/2 (ERK-1/2), p38 mitogen activated protein kinase (p38 MAPK), protein kinase C-epsilon (PKC-ε), and heat shock protein 27 (HSP27) [63].

Not all studies are supportive of noble gases having positive effects. In a study using neuronal cultures from mice, neither Krypton (Kr) nor Ne gases had any protective effects, and He was found to be detrimental to cells [64]. Argon (Ar) was found to be the best noble gas tested when it came to cellular protection following O₂ and glucose deprivation. As mentioned above, Ar failed to protect renal cells from deprivation of oxygen/glucose, and Kr and Ne had similar null effects [46]. Even so, it can be observed that a range of inert gases do have biological effects, and for all of them direct chemical reactions with biomolecules are unlikely. Therefore, do they have similar modes of action and even similar targets in cells?

5. Bioactivity Action of Noble Gases

As discussed, there are a range of noble gases which have biological effects and there are a wide range of proteins involved in mediating the downstream effects. For example, the action of He in cardioprotection was suggested to be mediated by kinases, in particular the inhibition of PI3K, Erk1/2, and p70s6K [45], and to have effects through inhibition of the mitochondrial permeability transition pore (mPTP). Interestingly, the authors stated that they had not investigated the “biochemical actions of helium” on the proteins which they had identified as important for mediating the effects observed. Rizvi et al. [46] in their work on human renal cells (HK2) found that Xe caused an increase in phospho-Akt (p-Akt), hypoxia-inducible factor-1α (HIF-1α) and Bcl-2 levels, the latter instrumental in one of the apoptosis initiation pathways. In rat heart tissue, Xe effects were reported to be mediated by protein kinase C (PKC)-ɛ and p38 mitogen-activated protein kinase (MAPK), but downstream of these signaling components it was found that MAPK-activated protein kinase-2 (MAPKAPK-2) and HSP27 were involved, leading to signaling to the actin cytoskeleton [65]. Winkler et al. [66] highlight the vast array of proteins which have been identified in silico as potentially able to bind to five noble gases, that is, He, Ne Ar, Kr, Xe. The authors downloaded 127,854 protein structures from the Protein Data Bank, and then used a computational approach to estimate the way in which the noble gases might interact with the polypeptide structures. Their analysis included the solvent-accessible surface area (SASA) of the gas atoms and the manner in which this matched the hydrophobicity of the protein contact points. In a further paper, the emphasis was not only focused on binding strength but also the location where there would be an alteration of protein function in which such functional difference would also have clinical relevance [67]. Some notable examples of proteins pulled out of the analysis include kinases (both serine/threonine and tyrosine, including MAPKs), phosphatases, carbonic anhydrase, phosphodiesterases, caspases, and nitric oxide synthase. Therefore, a wide range of cell signaling components are potentially altered by the presence of noble gases. It would be interesting to take this approach with H₂ as well.

It has been suggested that the spin state of H₂ may have an influence on how it interacts with biological molecules, be that small signaling components such as NO or larger entities like proteins [32]. A similar mechanism has been proposed for Xe. Xu et al. [68] looked at nuclear Overhauser effects and interactions with lipids, and the authors suggested that this may account for the molecule anesthetic action. Smith et al. [69] suggest that the nuclear spin of Xe can influence other radical electron pairs, so hinting at a possible mechanism. Investigating spin polarization-induced nuclear Overhauser effect (SPINOE), it was suggested that the induced spin polarization of Xe and He could be transferred to other nuclei [70]. The significance of any of these physical effects to biological systems has yet to be determined, but the literature suggests that it is worth exploring and not instantly dismissing by assuming they are inert and therefore inactive.

6. Xenon Pockets in the Globins

Kim et al. [35] suggests that H₂ directly interacts with hemoglobin through the Fe²⁺ of the heme prosthetic group; this manifests itself as scavenging hydroxyl radicals and peroxynitrite, and hence cellular effects are observed. Both hydroxyl radicals [71] and peroxynitrite [72] are reactive small molecules, so any influence in their accumulation and action can lead to cellular effects. However, this is not the only manner in which hemoglobin will interact with gases.

Hemoglobin is a tetrameric protein responsible for the transport of oxygen in many species including humans, as well as buffering the blood and acting as a reservoir for NO and influencing NO metabolism in animals and plants [73,74,75,76]. There is significant structural homology between humans and other higher order species in key areas of the globin chains responsible for binding oxygen and NO [76], as well as homologues in many other organisms, including plants [36]. Hemoglobin is one of the most abundant proteins in humans, accounting for approximately 1% of total body weight [77].

The Fe²⁺ of the heme prosthetic group (protoporphyrin IX) of the hemoglobin subunits are hexa-coordinate. Four of the coordinates are used to hold the iron atom to the porphyrin rings, whilst a fifth is coordinated to a histidine in the polypeptide chain. However, this allows the protein to bind and release O₂ at the sixth coordinate position—it is here that Kim et al. [35] suggest that H₂ may bind. The binding of O₂ is allosteric [78], affected by many factors such as oxygen, carbon dioxide (CO₂), and proton concentrations (hydrogen, chloride), as well as blood pH, blood flow, ATP and 2,3-DPG concentrations. Upon binding oxygen, hemoglobin undergoes conformational changes to increase oxygen affinity as hemoglobin subunits progressively bind, leading to a sigmoidal oxygen dissociation curve. For a review on hemoglobin, see [79].

The heme group can exist in different electronic states, including oxygenated (R, relaxed form), deoxygenated (T, tense form) and methemoglobin forms (oxidation of heme iron to Fe³⁺: MetHb). Each of these states has a different absorbance spectrum in the visible wavelengths making them relatively easy to study [80]. O₂ is not the only gas which can significantly affect hemoglobin structure and therefore function; CO_2, NO and CO are also similarly impactful. CO, for example, stabilizes Hb in the oxygenated R form, and has 200 times greater affinity for Hb than O₂. CO₂ has perhaps the most significant impact on oxygen binding affinity of Hb through the Bohr effect. CO₂ in humans is transported via binding the terminal amino groups in the alpha chains, with increasing CO₂ concentration shifting the oxygen dissociation to the right, decreasing oxygen binding affinity [81]. The Bohr effect therefore maximizes binding capacity of oxygen in the lungs whilst also optimizing delivery of O₂ to tissues with greatest need. The importance of regulating the oxygen dissociation curve via mechanisms such as CO₂ concentration can be observed by the clinical significance of conditions affecting this, such as the hemoglobinopathies, of which over 1000 variants have been reported [82]. Clinically, drugs such as Voxelotor are employed to treat conditions such as sickle cell anemia, shifting the HbO₂ dissociation curve; however, many hemoglobinopathies are still much more effectively treated by regular RBC replacement via transfusion [77].

In addition to impacting Hb individually, many gases work cooperatively and competitively to bind Hb and affect its function. For example, circulating levels of NO bound to Hb are dependent on Hb oxygen saturation, with binding of NO to cysteine residues mediating physiological response of vasodilation in hypoxic conditions [75,76]. MetHb is thought to be able to carry hydrogen sulfide (H₂S) in the vasculature, and H₂S is another gas which has significant signaling roles, including regulating blood flow. Therefore, understanding the way in which gases affect hemoglobin is important.

NO can bind both hemes in hemoglobin and covalently modify cysteine residues. However, whereas heme sequesters NO, cysteine 93 on β-globin (βCys93) is S-nitrosylated, creating a vasodilatory S-nitrosothiol (SNO) [82]. This has the potential to alter blood flow and hence O₂ delivery while serving to transport NO around the body through the vascular system. However, as mentioned, H₂ is unlikely to be involved in this type of chemistry.

It has been known for a long time that Xe can associate with hemoglobin, with an association curve being published about sixty years ago [83]. Others followed with the solubility of both Xe and Kr being reported in the presence of hemoglobin and albumin [84]. Therefore, the interactions of these noble gases with proteins needed to be understood.

Using Metmyoglobin as a model, cavities in the polypeptide chain were resolved to 1.9 Angstroms [85], and correction of this paper [86]. Using X-ray crystallographic techniques on sperm whale myoglobin at 7 atm Xe, it was found that there were four Xe binding sites. The authors concluded that these bindings were in polypeptide cavities, that such structures were probably present in native protein, and that there could be an influence on protein conformation. The same research group went on to investigate what they described as the “energies of xenon binding” to myoglobin. They suggest that there is in fact a fifth potential binding where water would normally coordinate the iron and describe a “connecting network of channel-like pathways through the static protein structure”. They also suggest that there is an easy route from internal binding cavities to the protein surface [87]. As Xe can bind myoglobin, ¹²⁹Xe NMR chemical shifts could be used by others to determine the differences between the MetHb from pig and horse. These authors stated that the Xe binds with a 1:1 ratio with respect to the protein, and that the binding constant was such that the Xe freely exchanges with the free Xe on the outside of the protein structure. It was also reported that the Xe-Fe distance in the two proteins was different: xenon–iron distance 7.4 Å for pig, and 5.3 Å for horse [88].

Therefore, the existence of Xe cavities in protein structures can be observed also in proteins related to myoglobin, i.e., hemoglobin. This was reported over sixty years ago [83,89]. More recently, Savino et al. [90] investigated the Xe binding of deoxygenated wild-type human hemoglobin as well as mutant variants. They also compared the data with that found with sperm whale myoglobin. It was reported that the binding in the a and b subunits of hemoglobin was different, with the b subunit being most similar to myoglobin. Tilton and Kuntz had previously noted differences of Xe between the a and b subunits of hemoglobin [91]. They also stated that “One of the binding sites in metmyoglobin is associated with a cavity on the proximal side of the porphyrin ring, opposite the O₂ binding site” (the authors cite Schoenborn here [89]). It has also been suggested that the ¹²⁹Xe chemical shift in myoglobin is dependent on the oxidation state and spin state of the iron in the heme of myoglobin [91,92]. Leading on from this and perhaps of more pertinence to the argument here are the effects of Xe on oxygen transport of hemoglobin. In the a subunit, O₂ may be leaving through a Xe binding cavity, and the treatment of hemoglobin with Xe decreases the efficiency of O₂ leaving the molecule. The same was not true for the b subunits [93]. However, it has been suggested that the binding of ¹²⁹Xe can be used as a probe for the measure of the oxygenation of blood [92].

7. Hydrophobic Cavities in Other Proteins

It seems clear from the work on the globins that inert atoms such as Xe can interact with proteins and have an influence on their activity. Can this be extended to other proteins, too? If so, can H₂ be acting through interaction with such cavities?

Prangé et al. [94], using X-ray diffraction, showed the interaction of Xe with a range of proteins, including elastase, subtilisin, cutinase, collagenase, lysozyme, urate oxidase and nuclear retinoid-X receptor. These were sourced from a range of organisms from Bacillus licheniformis to humans. Although Xe was used at above atmospheric pressures (8 to 20 bar), they concluded that Xe could bind within discrete pockets and channels embedded in proteins. Using ¹²⁹Xe NMR and a range of proteins including metmyoglobin, methemoglobin, lysozyme, and lipoxygenase the binding of Xe to protein structures was further reported [95]. Lipoxygenase was found to bind gas molecules with the highest affinity, therefore more effectively than the globins.

Using X-ray diffraction, the serine protease, subtilisin, was shown to bind Xe in the protein active site. It was suggested because the active site is a common feature of this class of polypeptide as it is relatively conserved. [96]. However, no major structural changes were observed on Xe binding. In a study also using X-ray diffraction on monooxygenase hydroxylase (MMOH) from Methylococcus capsulatus, it was concluded that hydrophobic cavities that could bind Xe may have a functional role in sequestering and making available the substrates for subsequent catalytic activity [97].

Xenon was used as probe to investigate the dioxygen binding of copper amine oxidase [98]. The rationale given included the knowledge that Xe can bind to hydrophobic cavities in proteins, and also that it is similar in size to O₂ but easier to detect. One binding pocket was common to three oxidases studied, and it was suggested that this O₂ binding region was important for catalysis. With a focus on anesthesia, four proteins (urate oxidase, lysozyme, neuroglobin, myoglobin) were studied by X-ray diffraction to compare the binding of Xe and nitrous oxide. The authors concluded that hydrophobicity alone could not account for the binding of the gases to the proteins, that volume needs to be considered, and that the gases could bind in a fully reversible manner [99].

Recently, two different methods were used to study Xe binding to proteins [100]. These were stability of proteins from rates of oxidation (SPROX) and limited proteolysis (LiP). Using a proteomic approach focused on the yeast genome, SPROX identified 31 novel proteins, while LiP identified 60. Further bioinformatic analysis showed that many of these proteins were involved in mechanisms which involved ATP. These include ATPase pumps and ATP synthase. Therefore, Xe seems to potentially influence ATP metabolism and cellular energy supplies.

There are clearly a range of proteins which can interact with Xe, even though it is inert. Many of the reports are based on using Xe at above atmospheric pressures, but it still shows that such interactions are possible and may account for changes in cellular activity. Oxygen binding and ATP metabolism may be affected, for example. The conclusion of the work with Xe binding to proteins and the investigations of hydrophobic cavities and channels on polypeptide structures is the question of whether other inert molecules can interact with proteins in the same manner. The real question is: Can H₂ interact in this way?

It is known that H₂ can increase O₂ saturation in blood and therefore be beneficial for exercise routines [101]. Perhaps, therefore, the work by Kim et al. [35] where H₂ may sterically hinder O₂ binding makes little sense, but on the other hand, if Xe, and H₂, by extension, decreases the O₂ release from hemoglobin [93], then this may account for higher O₂ saturation in red cells.

8. Conclusions

There is no doubt that H₂ has profoundly affects a range of organisms, from plants to humans, and its application has been suggested for both medical purposes [1] and agriculture [2]. However, even though there are a multitude of effects, the direct action of H₂ remains somewhat elusive. Many of the effects reported such as increases in the cell antioxidant capacity (e.g., increased SOD or CAT) or increased gene expression would need an increase in the activity of signaling components, but this would require some direct interaction of H₂ with those signaling molecules, most likely proteins. As H₂ is not able to directly facilitate post-translational modifications (PTMs) of proteins such as S-nitrosylation, there must be a different interaction involved. Rather than PTMs affecting directly polypeptide topology, very recent work on the hydration of proteins such as hemoglobin suggests that the interaction with water may affect the allosteric nature of the polypeptides, and hence their function [102]. It is possible that inert gases such as Xe (or H₂) may have an influence on such hydration by occupying polypeptide cavities. In another recent paper by the same group, the migration of Xe was studied in Mb [103]. Here, the transition energies were dependent on the occupation state of other cavities in the polypeptide. Certain key amino acids, such as Phe138, were suggested to gate the migrations of Xe, too. The authors suggested that the work shows that Mb is in fact an allosteric protein. Therefore, it appears that inert gases can influence the structure or the dynamic nature of topological changes; they could have an influence on function, which may influence downstream signaling or gene expression if affecting the right proteins. The question here is whether H₂ can partake in such interactions as suggested for Xe. Even if H₂ disrupted the interaction of other gas molecules with polypeptides, it could be significant. It certainly seems worth investigating, perhaps from a theoretical/mathematical perspective, as well as eventually using an experimental approach.

There is potential for H₂ to act as a direct antioxidant, and it is thought that it can scavenge both hydroxyl radicals and peroxynitrite [27]. This has been disputed, mainly because of the kinetics of the reactions [29]. More recently, a mechanism whereby the H₂ molecule interacts with the Fe of heme prosthetic groups was suggested to account for the direct scavenging activity of H₂, mediated by the generation of a hydrogen radical [35]. However, the kinetics of the reaction and its scope, accounting for the wide range of effects observed, needs to be determined—the model protein used was hemoglobin which does not exist in plants per se, for example, although homologues do exist. This possible action of H₂ would be worth investigating further in a range of organisms, including plants, nematodes, and higher animals.

With the above in mind, a different mechanism considering the relative inert nature of H₂ would bring a direct interaction with the amino acid chain of a protein. As discussed, precedents for this can be observed in the action of noble gases on proteins. Biological effects of noble gases are widely known. Xenon is a well-known anesthetic, for example, whilst argon, neon, helium and krypton have all been studied for their effects. Should H₂ be added to this group? Not all the noble gases have the same effects, but their mode of action may be informative to those studying H₂ effects. The work by Winkler and colleagues [66,67] highlighted the way in which in silico approaches can be used to investigate the way inert molecules can interact directly with proteins. Therefore, it would be interesting to see this approach used for H₂, too.

Here, similar to the work of Kim et al. [35], it is suggested that hemoglobin may be used as a model for investigating the ways in which H₂ may have a physical interaction with proteins. As well as oxygen, hemoglobin can interact with a range of gases, including CO, CO₂, H₂S and NO. Again, should H₂ be added to this mix? Different gases interact with hemoglobin in different ways. Kim et al. [35] suggest that H₂ takes the hexa-position of the Fe²⁺ in lieu of oxygen. Here, we are suggesting that the manner in which Xe interacts with proteins such as hemoglobin could be a good model for understanding the direct action of H₂. As well as gases interacting with proteins through the classical manners of binding to heme or causing PTMs such as S-nitrosylation, a third way could be through “Xenon pockets”, and this may account for the reason why H₂ seems to have such wide-ranging effects in a variety of organisms, from plants to humans. Future investigations could take an in silico approach, as used by Winkler et al. [66,67], but needs to be subject to experimentation, especially as hemoglobin is so tractable, having distinct spectra.

It is not suggested here that a direct physical interaction with proteins is the only mechanism of H₂ action in cells. The work by Kim et al. [35] may open the door on our understanding to the scavenging effect of H₂. The redox poise of the H₂ couple may allow a direct mechanism on some heme-containing proteins [30]. In different cellular environments, the redox state or pH are different and may dictate different modes of H₂ action, but it seems timely for this aspect of H₂ biology to be thoroughly investigated.