Methionine in proteins: The Cinderella of the proteinogenic amino acids

Abstract

Methionine in proteins, apart from its role in the initiation of translation, is assumed to play a simple structural role in the hydrophobic core, in a similar way to other hydrophobic amino acids such as leucine, isoleucine, and valine. However, research from a number of laboratories supports the concept that methionine serves as an important cellular antioxidant, stabilizes the structure of proteins, participates in the sequence‐independent recognition of protein surfaces, and can act as a regulatory switch through reversible oxidation and reduction. Despite all these evidences, the role of methionine in protein structure and function is largely overlooked by most biochemists. Thus, the main aim of the current article is not so much to carry out an exhaustive review of the many and diverse processes in which methionine residues are involved, but to review some illustrative examples that may help the nonspecialized reader to form a richer and more precise insight regarding the role‐played by methionine residues in such processes.

1. WHAT PROPERTIES MAKE METHIONINE UNIQUE?

Methionine, a sulfur‐containing amino acid, is often thought to be a generic hydrophobic residue. In fact, this is the way most Biochemistry textbooks describe this molecule. Not surprisingly, many biochemists regard this amino acid as one functionally replaceable with another hydrophobic residue. Although this might be the case for some methionine positions in some proteins, we should realize that methionine is a unique proteionogenic amino acid, remarkable in many regards. For instance, unlike other hydrophobic residues such as valine, leucine or isoleucine, the side chain of methionine is unbranched providing ample flexibility.

This extra‐flexibility is not an idle property. When several methionines are arranged on one side of an amphiphilic α‐helix, these flexible residues provide a malleable nonpolar surface that can adapt itself to peptide binding partners of varied sequence. This seems to be the case of calmodulin and the signal recognition particle 54 kDa subunit. These proteins share two relevant features: (a) they contain structural domains that are unusually rich in methionine residues, and (b) they are able to interact, in a specific way, with many protein partners that are diverse in sequence among them.

Being important as it is, the lack of branching is not the only property that contributes to the uniqueness of this amino acid. One obvious consequence is derived from the fact that the polarizability of sulfur in its lower oxidation states is substantially larger than the polarizabilities of typical hydrocarbon moieties. Consequently, during protein–protein interactions one can expect that London dispersion forces make greater contributions when sulfur is in place of carbon.1 On the other hand, another consequence, perhaps less obvious but biologically relevant, of the presence of a sulfur atom, has to do with the energetic stability of the side chain conformations.

Indeed, the structural plasticity of methionine also arises from the unusual properties of the torsion angle of the side chain involving a CS—CC unit. This unit shows relatively little energetic preference between conformations. This is best illustrated, as shown in Figure 1, by the comparison between butane and 2‐thiobutane.

Figure 1.

Energetic stability of butane and 2‐thiobutane rotamers. Enthalpic profile of gauche and anti conformations, shown in a Newman projection down the central C—C bond from butane (left), or down the central S—C bond from 2‐thiobutane (right). For butane, the anti conformation is energetically favored by 0.9 kcal/Mol. In contrast, the 2‐thiobutane rotamers show little energetic preference between gauche and anti conformations, with gauche only slightly favored (0.05 kcal/Mol)

A statistical survey of resolved structures revealed that the modest energetic discrimination among rotamers exhibited by 2‐thiobutane, can be extrapolated to what happens to methionine in proteins. Thus, Janin and Wodak showed that the distribution of the methionine ${\mathcal{X}}_3$ torsion angle (which controls the position of the ε‐methyl group) is virtually flat over the entire range of possible values.2 In this way, methionyl residues enjoy a wide freedom to mold themselves to accommodate nonpolar binding partners of differing structure. Hence, the great flexibility of the methionyl side chain possibly depends more on the presence of a S—C bond than on the lack of branching. That may help to explain, at least in part, why methionine was selected as a proteinogenic amino acid while norleucine, an unbranched structural analogous of methionine lacking sulfur, was not.3 Interestingly, norleucine can replace methionine and support normal bacterial protein synthesis under certain conditions. However, norleucine‐substituted cells exhibit a fitness lower than their wild type counterparts under conditions of oxidative stress.4 Under such circumstances, the sulfur present in methionine seems to make a difference.

2. METHIONINE RESIDUES ARE SUBJECTED TO REVERSIBLE OXIDATION AND REDUCTION

Since carbon and sulfur have similar electronegativity, methionine, and norleucine both have apolar side chains, as we have already discussed above. However, the sulfur atom from methionine can be readily oxidized to form the corresponding sulfoxide, converting an apolar side chain into a highly polar one (Figure 2a).

Figure 2.

Reversible enzyme‐catalyzed interconversion of Met and MetO. (a) Structural comparison of methionine and methionine sulfoxide emphasizing the apolar and polar character of these molecules, respectively. (b) Non‐enzymatic oxidation of methionine by ROS yields a racemate of the two diastereomers Met—S—O and Met—R—O. However, the stereospecific enzymes MsrA and Mical only produce the S— and R—epimers, respectively. These epimers can be reduced back to Met by the enzymes MsrA and MsrB that exhibit stereospecificity for their substrates: Met—S—O and Met—R—O, respectively. In the molecular models that we show here, the following color code has been used: Green for carbon, blue for nitrogen, red for oxygen, yellow for sulfur, and light gray for hydrogen

Methionine sulfoxide can then be reduced back to methionine. In this way, the interconversion of Met and MetO, in vivo, constitutes a process involving two different reactions, each one with its own forward and backward unidirectional rates (Figure 2b). These redox reactions can be enzyme‐catalyzed. Thus, Met forms MetO by adding oxygen to its sulfur atom in a reaction that can be catalyzed at least by two different enzymes: methionine sulfoxide reductase A (MsrA, a bifunctional enzyme) that, when operating in the oxidizing direction, yields the S epimer of methionine sulfoxide (Met—S—O)5; and Mical, an enzyme that also exhibits stereospecificity, but, in this case, catalyzing the formation of the R epimer (Met—R—O).6 The reduction back to methionine of these MetO epimers is mediated by MsrA and MsrB, respectively. This reversible posttranslational covalent modification of methionyl residues can have far‐reaching biological implications, as we will outline next.

3. S‐AROMATIC MOTIFS PLAY A UNIQUE ROLE IN STABILIZING PROTEIN STRUCTURE

Next, in the large inventory of the benefits provided by the presence of a sulfur atom in the side chain of methionine, we can find those effects derived from the so‐called S‐aromatic motifs. Indeed, an interaction of methionine and nearby aromatic residues (Phe, Tyr, and Trp) was described as early as in the mid‐80s.7 Even earlier, a frequency of sulfur and aromatic ring in close proximity within proteins higher than expected had been noticed.8 Despite the potential importance of these findings, they went largely overlooked, perhaps because the physicochemical nature of this bond is only poorly understood. Although the strength of these interactions may depend on the conditions of its environment,9, 10 it is accepted that the S‐aromatic interaction occurs at a greater distance (5–7 å) than a salt bridge (< 4 å), while the energies associated with either interaction are comparable. More recently, extensive surveys of the Protein Data Bank have revealed the importance of the methionine‐aromatic motif for stabilizing protein structures and for protein–protein interactions.11, 12

An example illustrating the importance of this motif in the context of cellular signaling is provided by the binding of lymphotoxin‐alpha (LTα) to its receptor, the tumor necrosis factor receptor 1 (TNFR1). In this line, Valley and coworkers showed that both, the Met120 from LTα and Trp107 from TNFR1 are critical for the binding and function of the complex LTα‐TNFR1.12 Interestingly, site‐specific oxidation of Met120 rendered LTα inactive. The authors of this study concluded then that oxidation of LTα inhibited its activity by disrupting the critical Met120‐Trp107 S‐aromatic interaction, thereby preventing binding to TNFR1.13 This observation, leads to an obvious question: how does oxidation of the sulfur moiety of an S‐aromatic motif affect the stability of the bond?

Two groups have independently addressed this question using quantum mechanical calculations. In one of these studies, Lewis and colleagues conclude that oxidation increases the strength of the S‐aromatic interaction.13 To reconcile this conclusion with the observation that oxidation of LTα prevents the formation of the complex LTα‐TNFR1, they speculate that an intra‐peptide interaction between MetO120 and Tyr96 could lock the ligand in a configuration that prevents the inter‐peptide MetO120–Trp107 interaction, thereby yielding an inactive ligand. Although such interpretation cannot be ruled out, a more parsimonious explanation can be given. Thus, in the second of the aforementioned studies, Orabi and English present data supporting a weaker aqueous MetO‐aromatic interaction when compared to the Met‐aromatic one.10 These authors point out that while gas‐phase Me₂SO—toluene is more stable than Me₂S—toluene, the change in stability is the other way around when water is present. Hence, the magnitude and sign of the energy change may be dependent on the sulfur environment.

In any event, we do know that sulfur atoms taking part in S‐aromatic motifs are underrepresented among the sulfur atoms that have been detected as sulfoxides in a proteome wide study.14, 15 Interestingly, the aromatic residues tyrosine and tryptophan were significantly underrepresented in the proximity of oxidized methionines when compared to those methionines oxidation‐resistant (Figure 3).

Figure 3.

Tyrosine is excluded from the environment of MetO. The frequencies at which tyrosine (Y) is present at different positions centered either around MetO or non‐oxidized Met were independently computed. The differences between these two types of frequency sets (MetO vs. Met) were used to determine the standard Z score. Thus, Z values that are much less than zero indicates that tyrosine is underrepresented in the environment of MetO with respect to an environment of Met. The discontinuous horizontal lines correspond to the critical Z values at a significance level of α = 0.01. For comparative purposes, the Z‐plot for the amino acid asparagine (N), which is not differentially distributed among MetO and Met, is also included

Although we do not know the underlying cause of this amazing observation, one hypothesis that might be worth investigating is that methionine, when found in spatial proximity of a tyrosine residue, may participate in a one‐electron redox reaction as a relay amino acid. Given that the rate for a single‐step electron‐transfer reaction between an electron donor and an electron acceptor decreases exponentially with the distance, long range electron transfer is fast only if a multistep hopping device is deployed in place, splitting the global process into shorter and, therefore, faster steps.16 In this context, methionines may operate as stepping‐stones for these multistep reactions by acting as intermediated charge carries. Figure 4 shows how methionine can be oxidized to MetO in a two‐electron oxidation reaction. Nevertheless, this oxidation can also take place in two sequential steps. Whenever a tyrosine residue is close enough to the oxidized sulfur, the molecule can undergo an intramolecular electron transfer (intramolecular redox reaction) yielding back methionine and leaving behind a tyrosyl radical. On the contrary, if there were not a tyrosine residue close to the methionine, the most likely end product of the redox process would be MetO. Indeed, a mechanism like that would explain the stunning observation summarized in Figure 3. However, whether that is or not the case remains to be elucidated.

Figure 4.

Methionine as a redox relay during long‐range electron transfers. Met can be oxidized to MetO in a two‐electron redox reaction (top horizontal chemical equation). Alternatively, this oxidation can take place by means of a reaction mechanism involving two elementary one‐electron oxidation reactions (triangular chemical equation). In the latter case, and whenever a tyrosine residue is close enough to the oxidized sulfur, an intramolecular collateral redox reaction that can take place is the reduction back to methionine at the expense of yielding a tyrosyl radical (vertical chemical equation)

4. METHIONINE RESIDUES IN PROTEINS CAN ACT AS ENDOGENOUS ANTIOXIDANTS

The oxidation of certain methionine residues of a protein may have no effect on its activity. A historical and paradigmatic example of that is provided by α₂‐macroglobulin (A₂M), a high molecular weight (~725 kDa) plasma proteinase inhibitor that targets a wide variety of proteinases. Each subunit of A₂M consumes eight equivalents of chloramine (an oxidant) without any loss of anti‐proteinase function. During the second phase of oxidative modification, the A₂M is inactivated with loss of activity proceeding in a manner directly proportional to the consumption of chlorinated oxidants. At this point, each subunit consumed 16 mol of chloramine, but only 14 Met residues were oxidized.17 Further studies demonstrated that a single tryptophan residue in each subunit was being oxidized by the remaining chloramine and that the decrease in total tryptophan residues (from 11 to 10) was directly proportional to loss of anti‐proteinase activity in tandem with the dissociation of the tetrameric A₂M into dimers.17

The ability of A₂M to tolerate oxidation of eight methionyl residues without loss of activity led to the hypothesis in which these residues functioned as antioxidants that protected the critical tryptophan residue from oxidation. The subsequent observation that the sulfoxidation of certain methionine positions in other proteins had not apparent effect on the activity of these proteins, led to Levine and coworkers to the more general hypothesis that methionine residues on the protein surfaces may act as reactive species of oxygen (ROS) sinks, and therefore function as antioxidants.18

5. EVOLUTIONARY STUDIES PROVIDE ADDITIONAL SUPPORT TO THE VIEW OF METHIONINE RESIDUES AS ANTIOXIDANTS

Although ROS are generated in diverse cellular compartments, the vast majority of ROS production can be traced back to the mitochondrion. Not surprisingly, this organelle displays a variety of ROS scavenging systems. In this line, the reassignment of the AUA codon from isoleucine to methionine, observed in the genetic code of mammalian mitochondria, has been interpreted as an adaptive process leading to antioxidant methionine accumulation in respiratory chain complexes.19 Indeed, while the average methionine content of the proteins encoded by the nuclear genome (nDNA) is around 2%, this percentage rises to 6% when the proteins analyzed are those encoded by the mitochondrial genome (mtDNA).

In this context, we reasoned that if, as proposed by Bender and colleagues, the adaptive function of methionyl residues as ROS scavengers has been an evolutionary driven force, then those mammalian lineages that exhibit higher rates of ROS generation might have been subjected to higher selective pressures to increase the methionine content of their mitochondrial proteins. In other words, if mitochondrial methionine residues serve as a ROS sink, then the proteins from animals subjected to high oxidative stress should accumulate methionine more effectively than their orthologous proteins from species exposed to lower oxidative stress. To address this hypothesis, we carried out a meta‐examination of mitochondrial genomes from mammalian species using longevity as an inverse proxy of the ROS production rate. Our analyses unveiled a hitherto unnoticed observation: mitochondrially encoded polypeptides from short‐lived species were enriched in methionine when compared with their long‐lived counterparts.20 Furthermore, mutational bias and nucleotide bias could be ruled out as driving forces, suggesting a role for natural selection in determining the methionine content of proteins.21

6. METHIONINE IS A KEY PLAYER IN ADAPTIVE MISTRANSLATION DURING OXIDATIVE STRESS

Recent works from several laboratories have consolidated the idea that mistranslation can be actively regulated to elaborate an adaptive response to stress conditions. In this context, methionine is emerging as a key player in the so‐called adaptive mistranslation, a phenomenon present in a wide variety of eukaryotic organisms.22, 23 For instance, around 1% of methionine residues used in protein synthesis have been previously aminoacylated to non‐methionyl tRNAs and then incorporated to proteins. However, this figure goes to 10% when cells are exposed to oxidative stress.23

Therefore, it seems that under certain conditions Met‐specific misacylation functions adaptively to increase methionine incorporation into proteins, providing in this way protection against oxidative stress. The mechanism by which the cells couple the stimulus (oxidative stress) to the response (increase methionine incorporation into proteins) has been more recently unraveled by Lee and colleagues who showed that in cells challenged with an oxidative stimulus, methionyl‐tRNA synthetase is phosphorylated by ERK1/2. This phosphorylation renders the synthetase promiscuous, so that it acylates non‐methionine tRNAs with methionine, which is eventually incorporated into the proteins at positions that, under no stress conditions, should be occupied by other amino acids according to the genetic code.24

As methionine residues can act as endogenous antioxidants, raising the methionine content in proteins via mistranslation has been interpreted as a strategy to consume and eliminate ROS. Nevertheless, we would like to note that interpretations alternative and complementary to the sink role assigned to these extra‐methionines, should not be ruled out. For instance, it may well be that methionine mistranslation bypasses the restraint of the genetic code to generate mutated proteins with new signaling properties.25 It is interesting to note that tRNA^Lys(CTT), which decodes AAG codons, seems to be mismethionylated more often than tRNA^Lys(TTT), which decodes AAA codons.23, 26 In other words, although under most situation both codons (AAG and AAA) are synonymous, they could stop being synonymous under conditions of oxidative stress. In this scenario, it is feasible that natural selection could have favored the appearance and maintenance of AAG codons at strategic positions. In this context, the change of lysine by methionine triggered by ERK1/2 under stress conditions, may have a signaling effect and contribute to the response beyond the ROS quenching potential of methionine residues.

7. METOSITE: AN INTEGRATED RESOURCE FOR THE STUDY OF METHIONINE OXIDATION AND ITS BIOLOGICAL IMPLICATIONS

Hitherto, we have reviewed the evidences supporting methionine oxidation as a reversible covalent modification. We have also shown that for certain positions within certain proteins, the oxidation of methionine has little, if any, effect on the protein activity. On the other hand, the mere addition of an oxygen atom to the sulfur atom of a single specific methionine residue can cause, as we will illustrate in the next section, drastic changes in the physicochemical properties of the complete protein, which, in turn, can affect the stability, subcellular localization and/or the activity of the modified protein. In other words, like phosphorylation, methionine sulfoxidation can be regarded as a posttranslational modification able to play a role in cell signaling and regulating cellular processes. Thus, we should distinguish between those methionines whose oxidation does not impact the properties of the protein, and consequently they may play a role as endogenous antioxidants, from those other methionines that may be signaling competent because when oxidatively modified they affect at least some property of the protein. This last group of methionyl residues are potentially signaling‐competent and collectively they may play a relevant role in adaptation to oxidative conditions.

Given the biological relevance of methionine modification, it is of great interest to identify classify and document sulfoxidized proteins/sites in different organisms and under different conditions. To this end, we have recently published a database (MetOSite, https://metosite.uma.es) that provides easy access to information on experimentally sulfoxidized methionine sites. MetOSite currently contains 7242 methionine sulfoxide sites found in 3562 different proteins belonging to 23 species, with Homo sapiens, Arabidopsis thaliana, and Bacillus cereus as the main contributors.27 Annotations are provided at both protein and residue levels, along with links to the relevant references.

Each MetO site present in the database has been assigned to one of three possible groups. Group 1 is formed by all the sites detected as MetO in high‐throughput studies and therefore the effect of such modification on the biological properties of the oxidized protein remains to be characterized. Group 2 is composed by all those methionine sites whose oxidation has no apparent effect on the protein properties, for which reason a ROS sink role has been assigned to them. Finally, the sulfoxidation of any site from Group 3 leads to changes in the properties of the target protein.

8. METHIONINE SULFOXIDATION MEDIATES THE REGULATION OF DIVERSE CELLULAR PROCESSES

Many examples of regulation by methionine oxidation have been described in the literature. However, the aim of the current article is not so much to be comprehensive as to be illustrative of the regulatory importance, somewhat overlooked, of this modification. Hence, in the remainder we will present only a few examples that together outline the evolutionary potential of methionine sulfoxidation as a regulatory strategy. A more extensive review can be found at Reference 28.

8.1. HypT, a transcription factor that becomes active when oxidized

Gene transcription is among the many cellular processes susceptible to be regulated by sulfoxidation of specific methionine residues. Thus, in prokaryotes, the transcription factor HypT (hypochlorite‐responsive transcription factor), which is involved in the defense against oxidative stress triggered by hypochlorous acid (HClO), presents three methionines (Met123, Met206, and Met230) highly reactive against this oxidant. These methionyl residues when oxidized yield the active form of HypT. Once the oxidative stress conditions cease, the oxidized protein is reduced by means of the MsrA/B activity of the bacteria, inactivating HypT.29 Furthermore, this active state can be mimicked by methionine‐to‐glutamine substitution, leading to the constitutively active mutant HypT^{M123,206,230Q}. Here, it is important to realize that glutamine and methionine sulfoxide exhibit a remarkable structural similarity and almost identical polarity, reasons for which glutamine is often used in mutagenesis experiments to mimic the effect of constitutive MetO at the chosen position.30

8.2. Stereospecific oxidation of Met‐76 in calmodulin impacts growth and behavior

MsrA is a bifunctional enzyme capable of stereospecifically oxidizing Met and reducing MetO residues.5 In an early work, Levine and colleagues showed that calmodulin was subject to oxidation mediated by MsrA.31 Furthermore, in this study they described an intriguing specificity, when calmodulin was in its calcium bound form, Met76 was the only residue oxidized to MetO of the nine methionyl residues present in the protein. At that point, these authors concluded that the reversible covalent modification of Met76 (Met77 if we count the initiating methionine, which is absent in the mature form of the protein) may regulate the interaction of calmodulin with one or more of its many targets. However, at that time there was no basis for deducing which of the hundreds of known calmodulin interacting proteins would be regulated by such a redox mechanism.

To approach this question, more recently Levine's group has created a mutant mouse in which wild type calmodulin‐1 was replaced by a mutant form M76Q of calmodulin, which mimics a constitutive MetO at residue 76. As the mutant mice exhibited growth and behavior phenotypes reminiscent of that observed in mice lacking calcium/calmodulin kinase IIα (CaMKIIα), the authors assessed in vitro the ability of MetO76 calmodulin to activate CaMKIIα. They found that MetO76 calmodulin was indeed less effective in activating CaMKIIα when compared to Met76 calmodulin.30

8.3. MsrB1 and Mical regulate actin assembly via reversible methionine oxidation

The cytoskeleton participates in a plethora of cellular phenomena, which makes its polymerization/depolymerization dynamic to be a complex and highly regulated process. Actin's polymerization properties are dramatically altered by sulfoxidation of a well‐conserved methionine at position 44 in the protein. This methionine residue was identified by the Terman laboratory as a substrate of Mical, an NADPH oxidoreductase that specifically oxidized this residue. Oxidation by Mical of actin Met44 leads to filament severing and decreased polymerization.32 Independently, that group and Gladishev's team, proved that the oxidation was stereospecific and generated only the R‐MetO. Even more, the modification was fully reversible by MsrB1, known to exhibit specificity for R‐MetO residues.33, 34 Together, these results draw a beautiful picture in which oxidation of Met44 in actin by Mical induces depolymerization of F‐actin and reduction of MetO44 by MsrB1 restores the ability of G‐actin to polymerize.

8.4. The redox state of a specific methionine sulfur atom acts as a switch for nitrate assimilation

In the fungus Arpegillus nidulans, the regulation at the transcriptional level of the assimilation of nitrate is mediated by the transcription factor NirA. In the absence of nitrate, when NirA is inactive and predominantly cytosolic, Met169 belonging to the nuclear export sequence (NES) domain of NirA is oxidized to MetO169 in a reaction depending on the enzyme FmoB, a flavin‐containing monooxygenase.35 This oxidation causes the protein to be retained in the cytoplasm, thus preventing the transcription of its target genes. On the contrary, the presence of nitrate or nitrite inside the cells of this ascomycete, leads to the reduction of MetO169 back to Met169, allowing in this way the relocation of NirA in the nucleus where it will perform its function.35

8.5. Methionine modification can affect cellular responses by controlling the vulnerability to proteolysis

The nuclear factor kappa‐light‐chain‐enhancer of activated B cells (NF‐κB), is a transcription factor present in most animal cells where it controls the transcription of DNA in response to various types of stress. NF‐κB is a heterodimer formed by RelA and p50. In the absence of stimuli, this heterodimer forms, together with the inhibitory protein IκBα, an inactive ternary complex that is found in the cytoplasm. After the appropriate stimuli, the IκB kinase (IKK) is activated and phosphorylates two serine residues from IκBα (Ser32 and Ser36). These modifications enable the subsequent ubiquitination of residues Lys21 and Lys22, signaling the degradation of IκBα and the release of NF‐κB, which now can migrate to the nucleus where it will perform its function as transcription factor (Figure 5). In this context, taurine‐chloramine, an oxidant produced by activated macrophages, inhibits the response of NF‐κB to cytokines, inhibition that is mediated by the oxidation of Met45 from IκBα. The oxidized protein is now resistant to degradation and the undegraded protein retains the transcription factor NF‐κB into the cytoplasm avoiding, in this way, an excessive inflammatory response in neutrophils.36

Figure 5.

Methionine‐mediated redox regulation of IκBα turnover. In the absence of stimuli, the NF‐κB, formed by the proteins RelA and p50, is retained in the cytoplasm by the inhibitory protein IκB, with which forms a ternary complex. After the appropriate stimuli, the IκB protein is phosphorylated at Ser32 and Ser36, which signals for degradation of IκB and the release of NF‐κB (RelA‐p50), which now can migrate to the nucleus. However, if Met45 is oxidized to MetO45, then the NF‐κB inhibitory protein IκB becomes resistant to degradation and the undegraded protein retains the transcription factor NF‐κB into the cytoplasm

9. CROSS‐TALK BETWEEN METHIONINE SULFOXIDATION AND O‐PHOSPHORYLATION

In addition to those cases in which the sulfoxidation of methionine residues seems to have a direct effect on the activity of the modified protein, there are other cases in which methionine oxidation may impact the function of the target protein acting as a coupling agent between the oxidative signal and the phosphorylation status of that protein. For instance, it has been reported that the oxidation of Met538 within the phosphorylation motif of nitrate reductase inhibits the phosphorylation of the nearby site Ser534, which would avoid the inactivation of nitrate reductase.37

This isolated observation prompted us to test the hypothesis that reversible methionine oxidation might be used in a generalized way as a rheostat to control the phosphorylation of proximal phospho acceptors. To this end, we analyzed the co‐occurrence of these two types of PTMs (sulfoxidation and phosphorylation) within the human proteome.38 In that study we showed that as many as 98% of the proteins containing oxidized methionine after an oxidative insult were also phosphoproteins. Interestingly, the chance of this high proportion being due to a mere sampling effect was calculated to be as low as one in one million. Furthermore, phosphoserine (pSer) and MetO sites cluster together in a statistically significant way when compared to Ser and Met. This proximity between modification sites could not be accounted for by their colocalization within unstructured regions because it was faithfully reproduced in a smaller sample of structured proteins.38 Overall, the results of that study strongly suggest that methionine sulfoxidation in the phosphorylation motifs it is not a random process directed by mass action, but on the contrary, it is a highly specific and selective process.

10. PROTEIN KINASES AND PROTEIN PHOSPHATASES CAN BE DIRECTLY REGULATED BY METHIONINE MODIFICATION

The phosphorylation status of a given protein can be modified through methionine oxidation in several ways. For instance, the oxidation of methionine residues around the phosphoacceptor can alter the affinity of protein kinases and protein phosphatases by the sulfoxidized substrate. In these cases, is the substrate that undergoes the methionine modification. A different alternative is to modify the activities of protein kinases and protein phosphatases directly by methionine sulfoxidation.

In this regard, the activities of calcineurin and CaMKII, a protein phosphatase and a protein kinase, respectively, are both regulated by the redox state of specific methionines present in their sequences (see MetOSite entries Q08209 and Q6PHZ2, respectively). It is interesting to note that both activities are dependent on Ca²⁺ cations and calmodulin, and also that both are modulated, in addition, by oxidative signals, but they do so in the opposite direction. That is, while the oxidation of Met406 of calcineurin interferes with the binding of calmodulin and its subsequent activation,39 the oxidation of the pair Met281, Met282 in the regulatory domain of CaMKII entails an activation of the kinase activity independently of the calcium levels and of binding to calmodulins.40

11. REVERSIBLE OXIDATION OF METHIONINE AND THE REGULATION OF LIQUID–LIQUID PHASE SEPARATION PROCESS

Numerous membraneless organelles, including stress granules, exist as liquid droplets within the cell. These dynamic structures are condensates of macromolecules that are reversibly assembled in a process termed liquid–liquid phase separation. In recent years, phase separation has gain interest as a novel principle of cellular organization and regulation.41 The resulting assemblies concentrate certain molecules while excluding others, and hence, can speed up or slow down biochemical reactions and contribute to compartmentalize the biological processes taking place within the living cell.

A relevant constituent of the stress granules is ataxin‐2, an intrinsically disordered protein with an increasingly larger number of known molecular functions. In a previous work, ataxin‐2 was identified as a putative candidate to be regulated via cross talk between methionine sulfoxidation and serine phosphorylation in response to oxidative stress.38 More recently, the group led by Benjamin Tu has unraveled a prominent role for methionyl residues from Pbp1 (the yeast orthologous of ataxin‐2) in the formation of intracellular drop‐like condensates required for the inhibition of TORC1 signaling during respiratory growth.42, 43

In the presence of fermentable carbon sources, yeast cells activate anabolic pathways and growth. On the other hand, under conditions of limiting carbon source, yeast mitochondria adopt a state of intense respiration. Under these conditions of intense respiration, the balance between growth and autophagy must be decided in favor of the latter. This is achieved, at least in part, by inhibition of TORC1. Inhibition of TORC1 leads to the induction of autophagy and maintain cell viability during nutrient starvation. However, treatment of the cells with either low concentrations of H₂O₂, antimycin A, or cyanide each led to the inhibition of autophagy.43 At this point, we could ask ourselves: what is the role fulfilled by reversible methionine sulfoxidation in this regulatory picture?

Tu and coworkers have presented convincing evidences that methionine residues from the low‐complexity domain of Pbp1 are components of a redox sensor, which constitutes a receptor for ROS produced by mitochondria in response to the nutrient environment (Figure 6). These authors showed that (a) these methionine residues are required, both in vivo and in vitro, for the formation of condensates; (b) oxidation, in both test tube reactions and living cells, of these residues leads to the melting of these liquid‐like droplets; (c) in vitro, the H₂O₂‐mediated melting of these structures is fully reverted by the enzymatic reduction of methionine oxidation in the presence of MsrA and MsrB.

Figure 6.

Liquid–liquid phase separation can be regulated by reversible methionine oxidation. A methionine‐rich low complexity region of ataxin‐2 mediates formation of condensates. One consequence of the formation of these droplet structures is the inhibition of TORC1 and the subsequent stimulation of autophagy. Under oxidative stress conditions, ataxin‐2 methionines are oxidized and the membraneless organelles melt releasing TORC1 from its inhibition and stimulating autophagy

Owing to the fact that the low‐complexity domain of human ataxin‐2 presents 18 evolutionary conserved methionines, one can speculate that the human protein may perform an analogous mechanistic function to that described above for Pbp1 in the yeast. Resorting to the arguments developed by the great biologist François Jacob, evolution always starts with what is already available and reuses successful designs repeatedly in slightly modified variations.44 Once we know that reversible methionine sulfoxidation is a successful solution to assemble tailored microenvironments that are established and maintained without the need of membranous structures,43 it is very likely that new and exciting findings are awaiting us ahead.

12. DO THE SIGNALING‐COMPETENT METHIONINES HAVE A TRACEABLE EVOLUTIONARY ORIGIN?

Oxidation of methionine adds polarity to an otherwise apolar side chain. More concretely, the side chain hydrophobicity index decreases from 0.738 (Met) to 0.238 (MetO) after the sulfoxidation reaction.45 Not surprisingly, polar amino acids such as Gln (hydrophobicity index of 0.251) and Thr (0.450) can sometimes mimic the sulfoxidized state of a protein.30, 46 Herein, we would like to put forward the idea that nature may also employ this trick, but in reverse, evolving methionine sites from glutamine and threonine residues. We shall elaborate a bit more this idea in the remainder of the current review.

We already know that, like phosphorylation, methionine sulfoxidation can lead to either inactivation or activation of the target protein. Therefore, when we talk about signaling‐competent methionines we must distinguish between these two types of methionine target. Let us start by imagining a protein with a methionyl residue of the first type (oxidation leads to inactivation). Better said, let us think in its ancestral protein in which, at the position corresponding to the methionine in question, we find any other amino acid. What could that amino acid be? Under the very reasonable assumption that the ancestral protein was constitutively active, the ancestral amino acid should be one with physicochemical properties similar to those of methionine. In this respect, hydrophobic amino acids such as leucine (0.943), isoleucine, and valine (0.825) are obvious candidates. The extant protein, just like the ancestral one, will be active and only will become inactive after the posttranslational modification (methionine sulfoxidation).

Now, we shall address the other case, that is, when the methionine PTM leads to activation. Given that the extant protein is inactive with Met but active with MetO, under the same sensible assumption that the ancestral protein was constitutively active, we must look at amino acids quite different in properties to Met but similar to MetO, and we already know that the obvious candidates are glutamine30 and threonine.46

Interestingly, when we calculated the probability of the different proteinogenic amino acids giving rise to methionine after an evolutionary time equivalent to a PAM unit, we observed that the apolar amino acids Ile, Leu, and Val together with the polar Gln, Lys, and Thr are the most probable predecessors of methionine residues in general (Figure 7). Few readers will be surprised to find Ile, Leu, and Val in this list, however, a priori, Gln, Lys and Thr would not be so predictable as predecessors of methionine residues.

Figure 7.

Likelihood of each amino acid of being replaced by methionine. For each proteinogenetic amino acid (abscissa) the figure shows the likelihood (ordinate) of being replaced by methionine after one PAM unit time (an evolutionary time long enough to allow 1% of the protein residue to be changed). To obtain the likelihood shown on ordinate, the probability of each amino acid being replaced by methionine has been relativized. To this end, all the probabilities were divided by the highest probability, corresponding to that of leucine (0.0008)

The change of Lys to Met has already been discussed above in the context of the adaptive value of mistranslation. On the other hand, our hypothesis that activating modification might have evolved from these polar amino acids (Gln and Thr), provides a rationale for the observation that a number of methionine residues in current proteins come from glutamine and threonine, as indicated by the PAM matrix.47 Hopefully, a resource such as MetOSite may help to address this working hypothesis in a close future.

13. CONCLUDING REMARKS

Methionine residues within proteins are mainly buried in the hydrophobic core.48 Therefore, except for its role in protein initiation, methionine has traditionally been regarded as a generic hydrophobic residue, readily interchangeable with other hydrophobic residues, all of them involved in van der Waals bonding and hydrophobic interactions that contribute to the protein stability, but with little functional/regulatory relevance. This is a woefully poor and outdated view that should be fought. Now, an overwhelming evidence places methionine in a prominent position among the proteinogenic amino acids having functional roles and being involved in regulatory processes.

This new and richer picture of methionine in proteins should come as no surprise, given the fact that methionine presents properties that make this molecule unique among the proteinogenic amino acids. Thus, the unbranched character of its side chain, the modest energetic discrimination among different conformations, the polarizability and redox versatility of its sulfur atom, are all properties that makes methionine an optimal amino acid for transient protein–protein interactions, central to molecular recognition.

On the other hand, the redox properties of the sulfur atom of methionine are difficult to overstate. Met can be oxidized to MetO, which, in turn, can be reduced back to Met. All these redox reactions are susceptible to be fine regulated as they can be enzyme‐catalyzed. At this point, we must distinguish between those methionines whose oxidation does not impact the properties of the protein, and consequently they may play a role as endogenous antioxidants (removing ROS), from those other methionines that may be signaling competent, because when oxidatively modified they affect the activity of the target protein. To this respect, MetOSite is a database that allows the user to obtain information regarding which methionine residues of a given protein of interest have been described as susceptible of oxidation, as well as the functional implications of such modifications.

In summary, methionine in proteins fulfils an important antioxidant role, stabilizes the structure of proteins, participates in the sequence‐independent recognition of protein surfaces, and can act as a regulatory switch through reversible redox reactions. It is time to transform Cinderella into Princess.

Aledo JC. Methionine in proteins: The Cinderella of the proteinogenic amino acids. Protein Science. 2019;28:1785–1796. 10.1002/pro.3698