Methionine Sulfoxide Reductase a Is a Stereospecific Methionine Oxidase

Methionine sulfoxide reductase A is a stereospecific methionine oxidase

Jung Chae Lim, Zheng You, Geumsoo Kim, and Rodney L. Levine1

Laboratory of Biochemistry, National Heart, Lung, and Blood Institute, Bethesda, MD 20892-8012

Edited by Irwin Fridovich, Duke University Medical Center, Durham, NC, and approved May 10, 2011 (received for review February 10, 2011)

Methionine sulfoxide reductase A (MsrA) catalyzes the reduction Results of methionine sulfoxide to methionine and is specific for the S epi- Stoichiometry. Branlant and coworkers have studied in careful mer of methionine sulfoxide. The enzyme participates in defense detail the mechanism of the MsrA reaction in bacteria (17, 18). against oxidative stresses by reducing methionine sulfoxide resi- In the absence of reducing agents, each molecule of MsrA dues in proteins back to methionine. Because oxidation of methio- reduces two molecules of MetO. Reduction of the first MetO nine residues is reversible, this covalent modification could also generates a sulfenic acid at the active site cysteine, and it is function as a mechanism for cellular regulation, provided there reduced back to the thiol by a fast reaction, which generates a exists a stereospecific methionine oxidase. We show that MsrA disulfide bond in the resolving domain of the protein. The second itself is a stereospecific methionine oxidase, producing S-methio- MetO is then reduced and again generates a sulfenic acid at the nine sulfoxide as its product. MsrA catalyzes its own autooxidation active site. Because the resolving domain cysteines have already as well as oxidation of free methionine and methionine residues formed a disulfide, no further reaction forms. The mass of the þ16 in peptides and proteins. When functioning as a reductase, MsrA protein has increased by 14 Da, Da from the sulfenic oxygen −2 fully reverses the oxidations which it catalyzes. and Da from the disulfide. When the mouse MsrA was incubated with MetO for 30 min, the stoichiometry of the Met formation appeared at first examination to be identical to that ethionine is an essential amino acid for most mammals for the bacterial enzymes (Fig. 1A). However, when the incuba- including humans. Approximately 99% of cellular methio- M tion time was lengthened, it became clear that three MetO are nine is found in protein, with the remainder in small molecules reduced by each MsrA (Fig. 1B) and that the mass of protein such as free methionine and S-adenosyl-methionine (1). Methio- increased by 12 Da, not 14 Da (from 23,648.6 Da to 23,660.6 Da, nine, as cysteine (Cys), the other sulfur-containing amino acid, is with an error of ∼0.5 Da for each mass). We therefore measured relatively easily oxidized to yield methionine sulfoxide. Oxidation the mass change of MsrA during the reaction and identified four by hydrogen peroxide, hypochlorous acid, and other reactive major species (Fig. 1C). species creates a chiral center at the sulfur so that the methionine sulfoxide (MetO) produced is a mixture of the S and R epimers Characterization of the MsrA Species. The species with different (2). Reduction of the sulfoxide back to methionine is catalyzed by masses could be purified by taking samples at different times of methionine sulfoxide reductases (Msr), enzymes found in almost incubation with MetO and purifiying them by reverse phase chro- all organisms from microbes to humans. Two distinct classes of matography (Fig. S1). These were treated with iodoacetamide to reductases are required, one to reduce the S epimer [methionine alkylate cysteine thiols, digested with trypsin, and the peptides sulfoxide reductase A (MsrA)] and one to reduce the R epimer mapped by reverse phase HPLC and tandem mass spectrometry. [methionine sulfoxide reductase B (MsrB)]. MsrA was described Peptides that were modified during the reaction with MetO were many years before MsrB and as a consequence most studies identified and their posttranslational modifications established focused on MsrA. Although the enzyme is almost ubiquitous (Table 1). Similar to the bacterial MsrA, the reduction of the first and a variety of studies establish its importance, its actual roles MetO of the mouse MsrA leads to disulfide formation in the −2 in cellular function remain unclear. Knocking out MsrA caused resolving domain, with a net mass change of Da. The second MetO further decreases the mass to −4 Da from the native, a increased susceptibility to oxidative stress in mice (3), yeast (4), þ14 and bacteria (5–7). Conversely, overexpressing MsrA conferred distinct difference from the Da observed in the bacterial form. Although the additional decrease of 2 Da could have increased resistance to Drosophila (8), Saccharomyces (9), Arabi- resulted from the formation of a second disulfide bond, it did not. dopsis (10), PC-12 cells (11), and human T cells (9). Overexpres- We then considered that a sulfenic acid did form on the active site sion in Drosophila doubled the lifespan of the flies (8). Cys72 but then reacted with the nitrogen of an adjacent residue’s There is reasonable evidence that cyclic oxidation and reduc- peptide bond to form a sulfenamide, a well-documented reaction of methionine residues in proteins act to scavenge reactive tion of protein sulfenic acids (19, 20). Sulfenamide formation oxygen and nitrogen species (12). Methionine residues can thus −4 “ ” would generate a modified protein of Da from the native, function as macromolecular bodyguards, preventing the oxida- and because the electrospray ionization technique is dehydrating, tion of critical residues in structural proteins and enzymes (13). sulfenamide formation would be favored in the mass spectro- Because methionine oxidation is a reversible covalent modifica- meter. Dimedone and 7-chloro-4-nitrobenzo-2-oxa-1,3-diazole tion, investigators have frequently suggested that it may have (NBD) both react with sulfenic acid to give a stable derivative. a role in cellular signaling and regulation but the experimental evidence to support such a role is limited (14–16). Although MsrA is a stereospecific MetO reductase, no stereospecific Author contributions: J.C.L., Z.Y., G.K., and R.L.L. designed research; J.C.L., Z.Y., G.K., and R.L.L. performed research; J.C.L., Z.Y., G.K., and R.L.L. analyzed data; and J.C.L. and R.L.L. methionine oxidase has yet been described. If such an oxidase wrote the paper. exists, it would complete the enzymatic requirements for a rever- The authors declare no conflict of interest. sible covalent modification. We now show that MsrA is a dual- This article is a PNAS Direct Submission. function enzyme with the ability to stereospecifically catalyze 1To whom correspondence should be addressed. E-mail: [email protected]. both the oxidation of methionine in proteins to S-MetO as well This article contains supporting information online at www.pnas.org/lookup/suppl/ as reduction of S-MetO back to methionine. doi:10.1073/pnas.1101275108/-/DCSupplemental.

10472–10477 ∣ PNAS ∣ June 28, 2011 ∣ vol. 108 ∣ no. 26 www.pnas.org/cgi/doi/10.1073/pnas.1101275108 Downloaded by guest on October 2, 2021 A NBD reacted with the unoxidized Cys of each form. Dimedone 2 did not react with the native nor −2 Da forms. Both dimedone and NBD reacted with the −4 Da form to give the derivative expected from a sulfenic acid (Table 1). Thus, except for the ol of MsrA 1 secondary formation of a sulfenamide by the mouse MsrA, its catalytic cycle thus far was the same as that observed for the bacterial enzyme. 0 Reduction of the third MetO generated an MsrA species which was þ12 Da heavier than the native form. Because this species

Mol of Met/mo was þ16 Da from the −4 Da species, it probably resulted from −4 0 1020304050 the addition of an oxygen to the Da form. The most likely explanation would be the addition of another oxygen to the active MetO (µM) site Cys72, generating either a sulfinic acid or a more complex derivative such as an S-oxide (21). However, the þ12 Da form B 3 still reacted with dimedone and NBD to give sulfenic acid deri- vatives. In addition, the þ12 Da form as well as the −2 Da and f MsrA −4 2 Da forms was converted back to the native mass upon addition of a reducing agent such as dithiothreitol or tris(2-carbox- yethyl)phosphine (TCEP). 1 These characteristics indicated that the oxygen was on a residue other than the active site Cys72. Peptide mapping and 0

l of Met/mol sequencing by mass spectrometry demonstrated that the modification giving rise to the þ12 Da form was conversion of Met229 Mol to methionine sulfoxide, resulting in a 16 Da increase in the mass 0 30 60 90 120 150 180 of the tryptic peptide (Fig. 2 and Table S1). Thus, the mouse Incubation Time (min) MsrA can autooxidize a methionine residue. After reduction of C 100 the second MetO, the sulfenic acid on Cys72 oxidizes Met229,

e and is itself reduced back to a thiol which allows it to react with a third MetO. Oxidation of Met229 is relatively slow (Fig. 1B). 75 16 16 When wild-type MsrA was incubated with Met OinH2 O, the protein mass increased by 12 Da. When incubated with 50 16 18 Met OinH2 O, the mass increased by 14 Da (Table 2). Thus 25 the oxygen in Met229O is not transferred from the substrate MetO ve abundance but is derived from water. When the þ12 Da form was incubated with NBD, the mass of the derivative was 4 Da higher when 0 18 16 carried out in H2 O rather than H2 O, demonstrating that both % relati the oxygen of Met229O and that of the sulfenic acid on Cys72 are 0306090120150180 from water. To rule out possible complexities arising from disul- Incubation Time (min) fide formation between Cys218 and Cys227 and sulfoxide formation on Met229 in wild-type MsrA, we constructed a mutant A Fig. 1. Met formation and MsrA mass changes. ( ) Met production as a MsrA in which formation of sulfenic acid should be the only cova- function of MetO concentration. Wild-type MsrA (1 μM) was incubated in 50 mM sodium phosphate, 1 mM DTPA, pH 7.5 without reductant at 37 °C lent modification induced by exposure to substrate. Cys107, for 30 min. (B) Met production as a function of time. The incubation was Cys218, and Cys227 were mutated to Ser and the six C-terminal as in A with MetO held at 50 μM. (C) Mass analysis from the incubations residues were deleted (C107S/C218S/C227S/Δ228–233). When shown in B. Minus 2 Da species, ◊; −4 Da species, Δ; þ12 Da species, ∇. incubated with MetO, its mass decreased by 2 Da, consistent with

Table 1. Characterization of MsrA species MsrA species, calculated mass without reagent Intact, 23,648.6 Da −2 Da, 23,646.6 Da −4 Da, 23,644.6 Da þ12 Da, 23,660.6 Da NBD Dimedone NBD Dimedone NBD Dimedone NBD Dimedone Observed mass 24,301.1 23,648.5 23,973.1 23,646.4 23,988.7 23,784.7 24,004.4 23,800.8 Observed mass change +652.5 −0.1 +324.5 −2.2 +340.1 +136.1 +355.8 +152.2 Calculated mass +652.0 0.0 +324.2 −2.0 +340.2 +136.2 +356.2 +152.2 change Assignment* 4 Cys-S-NBD — 1 Cys-S-S-Cys, 1 Cys-S-S-Cys 1 Cys-S-S-Cys, 1 Cys-S-S-Cys, 1 Cys-S-S-Cys, 1 Cys-S-S-Cys, 2 Cys-S-NBD 1 Cys-S-NBD, 1 Cys-S- 1 Cys-S-NBD, 1 Cys-S- 1 Cys-S-O-NBD Dimedone 1 Cys-S-O-NBD, Dimedone,

1 MetO 1 MetO BIOCHEMISTRY

Cys72-SH Cys72-SH Cys72-S-O-H Cys72-S-O-H

Cys218-SH Cys218-S Cys218-S Cys218-S ⏐ ⏐ ⏐ Cys -SH Cys -S 227 Cys227-S Cys227-S 227

Met229-S Met -S Met229-S=O Met229-S 229

*For clarity, the status of Cys107 is not shown in the diagram in the next row.

Lim et al. PNAS ∣ June 28, 2011 ∣ vol. 108 ∣ no. 26 ∣ 10473 Downloaded by guest on October 2, 2021 300 Intact 1

150 2 210 nm

0 300 +12 Da form

3 150 4 210 nm

0 32 34 36 38 (min)

Peptide No Observed Mass (Da) From-To Sequence Assignments 1 2,181.03 213-233 NPDGYCGLGGTGVSCPMAIKK 2 Cys alkylated 2 2,052.91 213-232 NPDGYCGLGGTGVSCPMAIK 2 Cys alkylated - - 3 2,080.94 213-233 NPDGYCGLGGTGVSCPMAIKK= 1 Intra-disulfide bond S S O 1 Oxygen - - 4 1,952.89 213-232 NPDGYCGLGGTGVSCPMAIK= 1 Intra-disulfide bond S S O 1 Oxygen

Fig. 2. Covalent modifications in the C-terminal residues of MsrA. Unmodified and the þ12 Da forms were mapped. Oxidation of Met229 to the sulfoxide was confirmed by sequencing, which also confirmed that the two cysteine residues in peptides 1 and 2 were alkylated while they formed a disulfide in peptides 3 and 4.

sulfenic acid formation followed by conversion to the sulfena- MetO provides a convenient assay for the S and R epimers (23) mide (Table 2). Addition of NBD confirmed the presence of because the more hydrophilic S epimer elutes earlier than the sulfenic acid. When the latter reaction was carried out in R epimer (Fig. 3A, Upper). When MetO produced by peroxide 18 H2 O, the NBD derivative was 2 Da heavier than when per- oxidation of Met was incubated with MsrA and DTT, only the 16 formed in H2 O, confirming that the oxygen of the sulfenic acid S epimer was reduced (Fig. 3A, Lower). When a sulfenic acid is donated by water. form of MsrA was incubated with Met in the absence of DTT, only S-MetO was produced (Fig. 3B, Upper). Addition of DTT Methionine Oxidation is Stereospecific for the S Epimer. Oxidation of supported reduction back to Met (Fig. 3B, Lower). Met by chemical oxidants such as hydrogen peroxide generates In addition to free Met and Met229 of MsrA, we tested two a mixture of the S and R epimers (22). MsrA is able to reduce other peptides and the natively unfolded protein α-synuclein only the S form and MsrB can reduce only the R form (23, 24). as substrates (Table 3). The tripeptide Gly-Met-Gly and the Reverse phase chromatography of o-phthalaldehyde derivatized hexapeptide Pro-Met-Ala-Ile-Lys-Lys were both substrates. The

Table 2. The oxygen of Cys72sulfenic acid and of Met229O is derived A S-MetO R-MetO from water L-MetO from H2O2 MsrA species, C107S/C218S/C227S/ 20 Δ –

calculated mass Wild-type, 23,648.6 Da 228 233, 22,931.5 Da =445 10

NBD NBD 0 − + − + + MsrA + DTT 16 In H2 O 23,660.6 24,004.6 22,929.9 23,110.3 20 Observed mass +12.0 +356.0 −1.6 +178.8 =230, Em=

Ex 10 Change in mass 18 In H2 O 23,662.5 24,008.3 22,929.9 23,112.3 B 0 Observed mass +13.9 +359.7 −1.6 +180.8 L-MetO from MsrA 40 Change in mass -DTT Difference +1.9 +3.7 0.0 +2.0 18 20 between H2 O 16 and H2 O 0 † Assignment* 1 Cys-S-S-Cys, 1 Cys-S- OH 0, Em=445 40 1 Cys-S-†OH, + DTT 1 Met†O 20 Ex=230

† † 0 Cys -S- O-H Cys72-S- O-H 72 5.6 5.8 6.0 6.2 6.4 6.6 6.8 7.0 (min) Cys -S 218 Fig. 3. MsrA is stereospecific for the S epimer. (A) Reduction. A mixture of ⏐ the S and R epimers of MetO was prepared by incubation of 500 μM L-Met Cys227-S with 1 M hydrogen peroxide for 2 h at 37 °C. The Upper tracing shows the † chromatogram of the untreated product. The Lower tracing shows that only Met229-S= O the S epimer was reduced when 20 μM L-MetO was incubated with MsrA and *For clarity, the status of Cys107 is not shown in the diagram in the next row. 10 mM DTT. (B) Oxidation. 20 μM of the sulfenic acid form of MsrA (C107S/ † Oxygen incorporated from water. Two micrograms of MsrA were C218S/C227S/Δ228–233) was incubated with 20 μM L-Met for 6 h at 37 °C. 16 18 lyophilized, redissolved in 50 μLH2 OorH2 O containing 10 mM L- The product is shown in the Upper tracing, and its complete reduction upon MetO, incubated for 30 min at 37 °C, and then incubated another 60 min addition of 10 mM DTT in the Lower tracing. For both A and B, incubation with or without 0.5 mM dimedone or NBD. with 10 mM DTT alone had no effect.

10474 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1101275108 Lim et al. Downloaded by guest on October 2, 2021 Table 3. Substrates of the MsrA oxidase Substrate Unoxidized mass* Oxidized mass* Change in mass Assignment

Gly-Met-Gly 263.09 279.09 +16.0 Met2O Pro-Met-Ala-Ile-Lys-Lys 686.41 702.41 +16.0 Met2O C72A MsrA 23,616.6 23,631.2 +14.6 Met229OCys218S-SCys227 α-Synuclein 14,460.2 14,476.4, 14,492.8 +16.2, +32.6 Met1O, Met1O, and Met5O *The measured masses are monoisotopic for the peptides and average for the proteins. The Met residues which were oxidized were determined by tandem mass spectrometric analysis of the peptides and of tryptic digests of the proteins.

hexapeptide sequence is that of the C terminus of MsrA. In Met to R-MetO (27), it does not oxidize methionine residues addition to the Met form of the peptide, we also synthesized the whose amino group is in peptide linkage (28). MetO form. They were used as substrates for the wild-type MsrA In the course of an investigation of the catalytic cycle of the to determine kinetic parameters for the oxidation and reduction mouse MsrA we discovered that the enzyme mediated autooxi- reactions. From two separate determinations, the Km for the dation of its Met229, analogous to the autophosphorylation reduction of the S-MetO-containing peptide was 0.12 mM and mediated by many protein kinases. Unlike the kinases, a second V 1 4 μ −1 −1 the max was . mol · min mg protein. In the oxidation enzyme is not required to reverse the covalent modification. K V 0 25μ reaction, the m was 3.5 mM and the max was . mol · MsrA functions as both a stereospecific methionine oxidase and min−1 mg−1 protein. α-Synuclein was oxidized at both Met1 a stereospecific methionine sulfoxide reductase. Dual-function and Met5 but not at Met116 or Met127. Addition of DTTallowed mediators of reversible covalent modification in proteins have MsrA to completely reduce the MetO it produced, confirming been known for many years. For example, bacterial glutamine that oxidation of Met in peptide linkage is also stereospecific. synthetase, a key enzyme in intermediary metabolism, is regu- The bacterial MsrA oxidized the hexapeptide and α-synuclein lated by reversible adenylylation of a tyrosine residue (29). It although the rate of oxidation was much slower than that was initially thought that adenylylation and deadenylylation were catalyzed by mouse MsrA. The bacterial MsrB and free methio- mediated by separate enzymes, but both reactions are catalyzed P nine-R-sulfoxide reductase did not undergo autooxidation nor by the same protein, initially named I (30). No futile cycle occurs did they oxidize the hexapeptide or α-synuclein. Thus, of the because the direction of the reaction is regulated by a second P P enzymes tested, only MsrA can function as a stereospecific protein, II, which interacts with I (31). We propose that MsrA methionine oxidase. There is a Val rather than Met at position mediates both the oxidation and reduction of methionine resi- 231 in human MsrA and no Met in the C-terminal region of dues within certain proteins. The direction of the reaction would Escherichia coli MsrA, and neither enzyme had significant auto- be controlled either by a reversible covalent modification of oxidation (Table 4). Introduction of a Met in the C-terminal MsrA or, perhaps more likely, by interaction with another pro- region rendered both human and E. coli MsrA autooxidizable. tein. The crystal and solution structures of the microbial and mammalian MsrAs reveal a common domain structure (32–34) Discussion and suggest a simple mechanism for regulation by protein–pro- Methionine and cysteine are the two sulfur-containing amino tein interaction (Fig. 4). In mouse MsrA, the active site is distant acids present in proteins. Well-known roles of cysteine include from the carboxyl domain which contains Cys218 and Cys227, catalysis, structure, redox sensing and regulation, and antioxidant which are required to reduce the sulfenic acid formed when MsrA defense. We have noted that methionine residues share at least acts as a reductase. They form a disulfide which in vivo is reduced two functions with their sulfur-containing sibling: antioxidant by thioredoxin to complete the reductase cycle. In vitro, investi- defense and regulation of cellular function through reversible gators typically replace the thioredoxin thioredoxin reductase oxidation and reduction (25). Although cysteine is well-recog- system by a low molecular weight reductant such as DTT or nized for the ease of its oxidation, it is often not appreciated that TCEP. Both DTT and TCEP can directly reduce the active site methionine is quite readily oxidized to MetO (22, 26). Conversion sulfenic acid, obviating the need for the resolving cysteines in the of methionine to its sulfoxide is mediated by a host of reactive species over a broad pH range. For example, hydrogen peroxide Reductase Oxidase readily reacts with certain methionine residues in proteins even at pH 5 or lower. Cysteine residues are typically quite resistant to C-terminal domain C-terminal domain not oxidation under these conditions because the thiol of cysteine accessible to thioredoxin accessible to thioredoxin must be ionized for oxidation to occur, and the pKa of cysteine is over eight unless altered by other residues within the protein. Although nonenzymatic oxidation of methionine could drive one

half of a reversible covalent modification cycle, the product would C72 C218 C72 C218 be a mixture of the S and R epimers of MetO. Reversal would C227 C227 require two enzymes, either MsrA and MsrB or one reductase and an epimerase (which has not yet been described). Although such a mechanism seems possible, the requirement for two Thioredoxin separate enzymes to reverse the oxidation is not appealing. A Postulated

Regulatory BIOCHEMISTRY stereospecific methionine oxidase would obviate that problem. Protein Although flavin monooxygenase 3 stereospecifically oxidizes free Thioredoxin

Table 4. Ability of recombinant MsrA to autooxidize Met residues Fig. 4. Potential mechanism for regulation of the reductase and oxidase Mouse Human E. coli reactions. When thioredoxin has access to the carboxy-domain resolving cysteines, MsrA functions as a reductase. When access of thioredoxin is Wild-type M229L Wild-type V231M Wild-type P209M blocked as a consequence of a conformational change induced by a reversible covalent modification or, as shown, by the binding of a another protein, Autooxidation Yes No No Yes No Yes MsrA functions as an oxidase.

Lim et al. PNAS ∣ June 28, 2011 ∣ vol. 108 ∣ no. 26 ∣ 10475 Downloaded by guest on October 2, 2021 carboxy domain. However, when MsrA was incubated with MetO but it is questionable whether this reaction occurs in vivo (38). in the absence of these reducing agents, it acted as a methionine Whether these or other small reactive species act on MsrA in vivo oxidase. In vivo, a protein which binds to the carboxy domain remains to be investigated. could prevent the resolving cysteines from accessing the active We suggest that the catalytic mechanism for oxidation of site, thus allowing the oxidase reaction to proceed. The same methionine is simply a reversal of that for reduction of methio- functional result would be obtained if an interacting protein nine sulfoxide. Based on experimentally derived data, a detailed blocked the access of thioredoxin to the disulfide bond in the mechanism for the catalytic cycle of Msr acting as a reductase has carboxy domain. MsrA could therefore mediate a regulated, re- been proposed by Branlant and coworkers (18). Both MsrA and versible oxidation of methionine residues in proteins and provide MsrB are thought to form sulfurane transition state with bipyr- a mechanism for control of cellular functions, including signaling. amidal geometry. A rearrangement of this transition state was In our assays, the hexapeptide substrate’s kinetic parameters hypothesized to occur, leading to formation of the sulfenic acid were an order of magnitude more favorable in the reduction on the catalytic cysteine and release of the methionine product. direction than in the oxidation direction, but it is probably not In our current studies, when MsrA acted as a methionine oxidase, appropriate to compare them. The reason is that the reduction the sulfoxide oxygen was derived from water and not from the reaction requires binding of only one substrate at the active site, substrate methionine sulfoxide. Because the sulfenic acid of the oxidized peptide. However, the oxidase reaction requires MsrA can form a sulfenamide, reversal of sulfenamide formation binding of MetO to the active site, oxidation of Cys72 to the by hydrolysis would introduce an oxygen from water in the sul- sulfenic acid, release of Met, and finally binding of the reduced fenic acid. However, direct participation of water in the catalytic peptide to the active site. Identification of in vivo substrates of reaction is consistent with the relationship of the substrate and the MsrA oxidase and assessment of the role of their oxidative the catalytic cysteine seen in crystal structures of MsrA (32–34). modification in cell signaling will be important in determining The sulfoxide oxygen points away from the cysteine thiolate and the physiological significance of the proposed mechanism. is too far away for direct transfer to the thiolate. Instead, the For experimental convenience we used only MetO to drive thiolate attacks the sulfur of the methionine sulfoxide as part of the oxidase reaction in these studies. Reactive species such as a catalytic mechanism which requires participation of water as a hydrogen peroxide, hypochlorous acid, chloramines, and perox- reactant. One possible scheme is shown in Fig. 5. ynitrite can also oxidize cysteine to the sulfenic acid (26). Hydro- Materials and Methods gen peroxide is well-recognized to participate in cell signaling Nonmyristoylated wild-type and mutant MsrA were expressed and purified through oxidation of active site thiols of enzymes (35). Hypo- as described (39). For each preparation the mass was checked by HPLC-mass chlorous acid and chloramines are relatively specific oxidizing spectrometry and was always within 0.5 Da of the calculated mass. The agents for the sulfur-containing amino acids (36). In vitro, perox- methodology was as described except that the mass spectrometer was an ynitrite readily oxidizes methionine residues to the sulfoxide (37), Agilent model 6200, which was also utilized for peptide mapping and

Fig. 5. Proposed mechanisms for the oxidase and reductase reactions. The reductase mechanism shown on the right is an extension of that proposed by Branlant and coworkers (18). In particular, we propose that the oxygen atom on the sulfenic acid is derived from a water molecule rather than methionine sulfoxide. At the initiation of the reductase reaction (II), the oxygen of the sulfoxide substrate is equidistant (2.6 Å) from the carboxylate oxygen of Glu115 and the hydroxyl of Tyr103 (II) (34). Glu115 acts as a proton donor to the sulfoxide, and the protonated hydroxyl sulfonium intermediate (III) is stabilized by hydrogen bonding with Tyr103. Nucleophilic attack of the hydroxyl sulfonium intermediate by the thiolate of Cys72 forms the sulfurane intermediate (IV) (42–44). It has been suggested that the sulfurane hydroxyl group could be protonated by a second methanthiol molecule and form the sulfonium intermediate upon loss

of water. The pKa of methanthiol and the hydroxyl of tyrosine are both about 10.2, leading us to suggest that Tyr103 is the second proton donor, facilitating formation of the sulfonium intermediate upon leaving of a water molecule which would be exchangeable with solvent water (V). Another water molecule, seen in several crystal structures of MsrA is sited 2.8 Å from Cys72 on the opposite side from the bound substrate. This location places it in perfect position to react with the sulfonium intermediate to form the sulfenic acid on Cys72 and release Met as a product (VI). After the release of Met, the Cys72 sulfenic acid form (VII) can be reduced back to the thiolate (I) by reducing agents, including the C-terminal resolving cysteines or DTT. The sulfenic acid can also undergo reversible dehydration to form a sulfenamide by reaction with either a side chain amino group or an amide group in the backbone, a reaction favored in the electrospray mass spectrometer. The left side of the figure shows the proposed mechanism for the MsrA oxidase. Met binds to the sulfenic acid form of MsrA in the same orientation as MetO does in the reductase cycle and with water bound to Glu115 and Tyr103 (VIII). A lone pair of electrons on the Met sulfur attacks the sulfenic sulfur in a SN2 (bimolecular nucleophilic substitution) manner to form the positively charged sulfonium intermediate (IX) and release a molecule of water. Another molecule of water, presumably again bound in the pocket with Glu115 and Tyr103, attacks the sulfonium sulfur to form the sulfurane intermediate (X). Glu115 then acts as a general base to abstract a proton from the hydroxyl group of the sulfurane intermediate. The sulfur–sulfur bond breaks to form MetO and Cys72 is regenerated (XI). With release of MetO, the reduced form of MsrA is formed (I).

10476 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1101275108 Lim et al. Downloaded by guest on October 2, 2021 sequencing by tandem mass spectrometry (40). To determine the stoichio- six residues at the C terminus and the three Cys other than the active site metry of Met formation, 1 μM MsrA was incubated with different concen- Cys72 were mutated to Ser. Kinetic parameters were determined with trations of L-methionine sulfoxide (MetO) without reducing agents in 50 mM wild-type MsrA using the hexapeptide as substrate with Met or MetO (Amer- sodium phosphate pH 7.5 with 1 mM diethylenetriaminepentaacetic acid ican Peptide), varying the concentrations from 0.25 to 10 mM. The initial rate (DTPA). The reaction was stopped by adding 10% trifluoroacetic acid to a at different substrate concentrations was determined after 5 (reductase) or final concentration of 0.5%. Met was determined by amino acid analysis 15 min (oxidase) incubation at 37° from the area of the hexapeptide on the (41). S-MetO and R-MetO were separated on an APEX column after deriva- – tization (1). For detection of sulfenic acid, 1 μM MsrA was incubated with HPLC reverse phase chromatogram at 210 nm and fit to Michaelis Menton 0.5 mM 5, 5-dimethyl-1, 3-cyclohexanedione (dimedone) or NBD-chloride kinetics. (NBD) for 1 h at 37 °C. To determine susceptibility to oxidation, peptides and proteins were incubated with the C107S/C218S/C227S/Δ228–233 MsrA ACKNOWLEDGMENTS. This research was supported by the Intramural Research in the presence of 10 mM MetO for 6 h at 37 °C. This mutant lacks the last Division of the National Heart, Lung, and Blood Institute.

1. Luo S, Levine RL (2009) Methionine in proteins defends against oxidative stress. FASEB 24. Moskovitz J, et al. (2002) Purification and characterization of methionine sulfoxide J 23:464–472. reductases from mouse and Staphylococcus aureus and their substrate stereospecifi- 2. Toennies G, Kolb JJ (1939) Methionine studies. J Biol Chem 128:399–405. city. Biochem Biophys Res Commun 290:62–65. 3. Moskovitz J, et al. (2001) Methionine sulfoxide reductase (MsrA) is a regulator of anti- 25. Levine RL, Moskovitz J, Stadtman ER (2000) Oxidation of methionine in proteins: Roles oxidant defense and lifespan in mammals. Proc Natl Acad Sci USA 98:12920–12925. in antioxidant defense and cellular regulation. IUBMB Life 50:301–307. 4. Moskovitz J, Berlett BS, Poston JM, Stadtman ER (1997) The yeast peptide-methionine 26. Vogt W (1995) Oxidation of methionine residues in proteins: Tools, targets, and Proc Natl Acad Sci USA sulfoxide reductase functions as an antioxidant in vivo. reversal. Free Radical Biol Med 18:93–105. – 94:9585 9589. 27. Duescher RJ, Lawton MP, Philpot RM, Elfarra AA (1994) Flavin-containing monooxy- 5. Moskovitz J, et al. (1995) Escherichia coli peptide methionine sulfoxide reductase genase (FMO)-dependent metabolism of methionine and evidence for FMO3 being gene: Regulation of expression and role in protecting against oxidative damage. the major FMO involved in methionine sulfoxidation in rabbit liver and kidney micro- J Bacteriol 177:502–507. somes. J Biol Chem 269:17525–17530. 6. Douglas T, et al. (2004) Methionine sulfoxide reductase A (MsrA) deficiency affects the 28. Elfarra AA, Krause RJ (2005) Potential roles of flavin-containing monooxygenases survival of Mycobacterium smegmatis within macrophages. J Bacteriol 186:3590–3598. in sulfoxidation reactions of l-methionine, N-acetyl-l-methionine and peptides 7. St. John G, et al. (2001) Peptide methionine sulfoxide reductase from Escherichia coli containing l-methionine. Biochim Biophys Acta 1703:183–189. and Mycobacterium tuberculosis protects bacteria against oxidative damage from 29. Shapiro BM, Stadtman ER (1970) The regulation of glutamine synthesis in microorgan- reactive nitrogen intermediates. Proc Natl Acad Sci USA 98:9901–9906. isms. Annu Rev Microbiol 24:501–524. 8. Ruan H, et al. (2002) High-quality life extension by the enzyme peptide methionine 30. Anderson WB, Hennig SB, Ginsburg A, Stadtman ER (1970) Association of ATP: Gluta- sulfoxide reductase. Proc Natl Acad Sci USA 99:2748–2753. 9. Moskovitz J, et al. (1998) Overexpression of peptide-methionine sulfoxide reductase mine synthetase adenylyltransferase activity with the P1 component of the glutamine Proc Natl Acad Sci USA – in Saccharomyces cerevisiae and human T cells provides them with high resistance to synthetase deadenylylation system. 67:1417 1424. oxidative stress. Proc Natl Acad Sci USA 95:14071–14075. 31. Brown MS, Segal A, Stadtman ER (1971) Modulation of glutamine synthetase 10. Romero HM, et al. (2004) Investigations into the role of the plastidial peptide adenylylation and deadenylylation is mediated by metabolic transformation of the Proc Natl Acad Sci USA – methionine sulfoxide reductase in response to oxidative stress in Arabidopsis. Plant PII-regulatory protein. 68:2949 2953. Physiol 136:3784–3794. 32. Lowther WT, Brot N, Weissbach H, Matthews BW (2000) Structure and mechanism of 11. Yermolaieva O, et al. (2004) Methionine sulfoxide reductase A protects neuronal peptide methionine sulfoxide reductase, an “anti-oxidation” enzyme. Biochemistry cells against brief hypoxia/reoxygenation. Proc Natl Acad Sci USA 101:1159–1164. 39:13307–13312. 12. Stadtman ER, Moskovitz J, Berlett BS, Levine RL (2002) Cyclic oxidation and reduction 33. Taylor AB, Benglis DM, Jr, Dhandayuthapani S, Hart PJ (2003) Structure of Mycobac- of protein methionine residues is an important antioxidant mechanism. Mol Cell terium tuberculosis methionine sulfoxide reductase A in complex with protein-bound Biochem 234/235:3–9. methionine. J Bacteriol 185:4119–4126. 13. Levine RL, Mosoni L, Berlett BS, Stadtman ER (1996) Methionine residues as endogen- 34. Ranaivoson FM, et al. (2008) A structural analysis of the catalytic mechanism of ous antioxidants in proteins. Proc Natl Acad Sci USA 93:15036–15040. methionine sulfoxide reductase A from Neisseria meningitidis. J Mol Biol 377:268–280. 14. Ciorba MA, et al. (1997) Modulation of potassium channel function by methionine 35. Rhee SG (2006) H2O2, a necessary evil for cell signaling. Science 312:1882–1883. oxidation and reduction. Proc Natl Acad Sci USA 94:9932–9937. 36. Winterbourn CC (1985) Comparative reactivities of various biological compounds 15. Erickson JR, et al. (2008) A dynamic pathway for calcium-independent activation of with myeloperoxidase-hydrogen peroxide-chloride, and similarity of the oxidant to Cell – CaMKII by methionine oxidation. 133:462 474. hypochlorite. Biochim Biophys Acta 840:204–210. 16. Cao G, et al. (2010) Methionine sulfoxide reductase B1 (MsrB1) recovers TRPM6 37. Moreno JJ, Pryor WA (1992) Inactivation of alpha 1-proteinase inhibitor by peroxyni- J Biol Chem – channel activity during oxidative stress. 285:26081 26087. trite. Chem Res Toxicol 5:425–431. 17. Antoine M, Gand A, Boschi-Muller S, Branlant G (2006) Characterization of the 38. Berlett BS, Levine RL, Stadtman ER (1998) Carbon dioxide stimulates peroxynitrite- amino acids from Neisseria meningitidis MsrA involved in the chemical catalysis of mediated nitration of tyrosine residues and inhibits oxidation of methionine residues the methionine sulfoxide reduction step. J Biol Chem 281:39062–39070. of glutamine synthetase: Both modifications mimic effects of adenylylation. Proc Natl 18. Boschi-Muller S, Gand A, Branlant G (2008) The methionine sulfoxide reductases: Acad Sci USA 95:2784–2789. Catalysis and substrate specificities. Arch Biochem Biophys 474:266–273. 39. Kim G, et al. (2010) Dual sites of protein initiation control the localization and 19. Fu X, Mueller DM, Heinecke JW (2002) Generation of intramolecular and intermole- myristoylation of methionine sulfoxide reductase A. J Biol Chem 285:18085–18094. cular sulfenamides, sulfinamides, and sulfonamides by hypochlorous acid: A potential 40. Stevens LA, Levine RL, Gochuico BR, Moss J (2009) ADP-ribosylation of human defensin pathway for oxidative cross-linking of low-density lipoprotein by myeloperoxidase. Biochemistry 41:1293–1301. HNP-1 results in the replacement of the modified arginine with the noncoded amino Proc Natl Acad Sci USA – 20. Salmeen A, et al. (2003) Redox regulation of protein tyrosine phosphatase 1B involves acid ornithine. 106:19796 19800. a sulphenyl-amide intermediate. Nature 423:769–773. 41. Daly MJ, et al. (2010) Small-molecule antioxidant proteome-shields in Deinococcus PLoS One 21. Imai S, et al. (2002) Plant biochemistry: An onion enzyme that makes the eyes water. radiodurans. 5:e12570. Nature 419:685. 42. Wallace TJ (1964) Reactions of thiols with sulfoxides. I. Scope of the reaction and 22. Lavine TF (1947) The formation, resolution, and optical properties of the diasterio- synthetic applications. J Am Chem Soc 86:2018–2021. meric sulfoxides derived from L-methionine. J Biol Chem 169:477–491. 43. Wallace TJ, Mahon JJ (1964) Reactions of thiols with sulfoxides II. Kinetics and mechan- 23. Sharov VS, Ferrington DA, Squier TC, Schöneich C (1999) Diastereoselective reduction istic implications 1. J Am Chem Soc 86:4099–4103. of protein-bound methionine sulfoxide by methionine sulfoxide reductase. FEBS Lett 44. Balta B, et al. (2006) Theoretical study of the reduction mechanism of sulfoxides by 455:247–250. thiols. J Phys Chem A 110:7628–7636. BIOCHEMISTRY

Lim et al. PNAS ∣ June 28, 2011 ∣ vol. 108 ∣ no. 26 ∣ 10477 Downloaded by guest on October 2, 2021