Characterization of the Amino Acids from Neisseria meningitidis MsrA Involved in the Chemical Catalysis of the Methionine Sulfoxide Reduction Step*

Methionine sulfoxide reductases (Msrs) are ubiquitous enzymes that reduce protein-bound methionine sulfoxide back to Met in the presence of thioredoxin. In vivo, the role of the Msrs is described as essential in protecting cells against oxidative damages and as playing a role in infection of cells by pathogenic bacteria. There exist two structurally unrelated classes of Msrs, called MsrA and MsrB, specific for the S and the R epimer of the sulfoxide function of methionine sulfoxide, respectively. Both Msrs present a similar catalytic mechanism, which implies, as a first step, a reductase step that leads to the formation of a sulfenic acid on the catalytic cysteine and a concomitant release of a mole of Met. The reductase step has been previously shown to be efficient and not rate-limiting. In the present study, the amino acids involved in the catalysis of the reductase step of the Neisseria meningitidis MsrA have been characterized. The invariant Glu-94 and to a lesser extent Tyr-82 and Tyr-134 are shown to play a major role in the stabilization of the sulfurane transition state and indirectly in the decrease of the pKapp of the catalytic Cys-51. A scenario of the reductase step is proposed in which the substrate binds to the active site with its sulfoxide function largely polarized via interactions with Glu-94, Tyr-82, and Tyr-134 and participates via the positive or partially positive charge borne by the sulfur of the sulfoxide in the stabilization of the catalytic Cys.

Methionine sulfoxide reductases (Msr) 3 are enzymes that catalyze the reduction of free and protein-bound methionine sulfoxide (MetSO) back to Met. Two structurally unrelated classes of Msrs have been described so far. MsrAs are stereo specific toward the S isomer on the sulfur of the sulfoxide function, whereas MsrBs are specific toward the R isomer. Both classes share a similar three-step catalytic mechanism (Scheme 1). First, the reductase step leads to formation of a sulfenic acid intermediate on the catalytic cysteine concomitantly with the release of one mole of Met/mole of Msr. Then, an intra-disulfide bond is formed via the attack of a second Cys (called the recycling Cys) on the sulfenic acid intermediate accompanied by release of a water molecule. Finally, the disulfide bond is reduced by thioredoxin (Trx) in the last step. Recently, the kinetics of the three steps have been investigated for MsrA and MsrB domains of the PilB protein of Neisseria meningitidis (1,2). For both classes of Msrs, the rate-limiting step is associated with the Trx recycling process, whereas the rate of formation of the intra-disulfide bond is governed by that of formation of the sulfenic acid intermediate, the rate of which is fast.
The three-dimensional structures of the MsrA from Escherichia coli, Bos taurus, and Mycobacterium tuberculosis have been recently solved by x-ray crystallography (3)(4)(5). The active site can be represented as an opened basin readily accessible to the MetSO substrate in which the catalytic Cys-51 is located at the entrance of the ␣ 1 helix. In all the structures, the active site is occupied by a molecule that is covalently or non-covalently bound to the catalytic cysteine. In the case of E. coli MsrA, a dimethyl arsenate molecule is covalently bound, whereas it is a dithiothreitol molecule in B. taurus enzyme. In the case of M. tuberculosis MsrA, a methionine residue from a neighboring monomer occupies the active site. In all three structures, a water molecule is present, the position of which can mimic the oxygen atom of the sulfoxide function of MetSO. This water molecule is tightly H-bonded to three invariant amino acid residues, i.e. Tyr-82, Glu-94, and Tyr-134. All the three structures also support the involvement of invariant Phe-52 and Trp-53 in the substrate recognition via the formation of a hydrophobic pocket in which the ⑀ methyl group of MetSO can bind.
Study of the reduction mechanism of dimethyl sulfoxide (Me 2 SO) by methanethiol in Me 2 SO solution has recently been investigated by quantum chemistry calculations (6). It was shown that 1) a sulfurane species is formed prior to formation of either a sulfenic acid intermediate or a disulfide species and 2) the rate-limiting step is governed by proton transfer between the thiol and the sulfoxide functions prior to sulfurane formation. Although these conclusions are derived from studies based on a model in solution, they provide a framework for the study of the chemical reductase step occurring within the MsrA active site.
In the present study, the role of Glu-94, Tyr-82, and Tyr-134 residues and how the catalytic Cys-51 is stabilized in the reductase step of the MsrA from N. meningitidis have been investi-gated. For that, the kinetic parameters and the pH dependence of the rate constant of the reductase step of mutated MsrAs at positions 82, 94, and 134 were determined and compared with those of the wild type. The pK app of Cys-51 in the free enzyme was also determined. The results show that Cys-51 is activated upon substrate binding to the active site with a shift of its pK a from 9.5 to 5.7. Substitutions at positions 82, 94, and 134 do not modify the apparent affinity for the substrate in the reductase step. In contrast, drastic decrease of the reductase step rate is observed for the E94A and Y82F/Y134F MsrAs, whereas E94Q MsrA displays only a small decrease. Moreover, each mutated MsrA is characterized by a shift of the pK app of its Cys-51 to higher values compared with wild type. Taking into account all the results, a scenario for the catalysis of the sulfoxide reductase step is proposed in which Glu-94, Tyr-82, and Tyr-134 stabilize the sulfurane transition state formed. In this scenario, the substrate binds to the active site with its sulfoxide function largely polarized via interactions with the side chains of Glu-94, Tyr-82, and Tyr-134 and plays a major role in stabilizing Cys-51 via the positive, or partially positive, charge borne by the sulfur of the sulfoxide function.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis, Production, and Purification of Wild-type and Mutated N. meningitidis MsrAs-The E. coli strain used for all N. meningitidis MsrA productions was BE002 (MG1655 msrA::spec⍀, msrB::␣ 3kana), transformed with the plasmidic construction pSKPILBMsrA containing only the coding sequence of msrA from pilB, under the lac promoter (7). The BE002 strain was kindly provided by Dr. F. Barras. Its use prevented expression of endogenous wild-type MsrA and MsrB from E. coli and thus avoided any contamination of the activity of the N. meningitidis MsrA by the Msrs from E. coli. Site-directed mutageneses were performed using the QuikChange site-directed mutagenesis kit (Stratagene).
Purifications were realized as previously described (1). Wildtype and mutated MsrAs were pure, as checked by electrophoresis on 12.5% SDS-PAGE gel followed by Coomassie Brilliant Blue R-250 staining and by electrospray mass spectrometry analyses. Storage of the enzymes was done as previously described. The molecular concentration was deter-mined spectrophotometrically, using extinction coefficient at 280 nm of 26,200 M Ϫ1 ⅐cm Ϫ1 for wild-type and mutated MsrAs. In this report, N. meningitidis MsrA amino acid numbering is based on E. coli MsrA sequence.
pH Dependence of MsrA Thiol Reaction Rates with 2,2Ј-Dipyridyl Disulfide (2PDS)-Because of the high reactivity of Cys-51 and Cys-198 in MsrA, fast kinetic measurements were carried out on an Applied PhotoPhysics SX18MV-R stoppedflow apparatus. MsrA reactions with 2PDS were performed at 25°C under pseudo-first-order conditions in 30 mM acetic acid, 30 mM imidazole, 120 mM Tris/HCl buffer at constant ionic strength of 0.15 M over a pH range of 6 to 10 (polybuffer B). MsrA and 2PDS concentrations after mixing were 6.2 and 310 M, respectively. The pseudo-first-order rate constant k obs was determined at each pH by fitting the absorbance (A) at 343 nm versus time (t) to mono-exponential Equation 1, where a is the burst amplitude and c is the end point. The second-order rate constants k 2 were calculated by dividing k obs by 2PDS concentration and then fitted to Equation 2, in which k 2max represents the second rate constant for the thiolate form.

Measurement of the Thiol Ionization by Ultraviolet
Absorbance-Absorbance spectra were measured for all enzymes in 1.0-cm path length quartz cuvettes in a SAFAS UVvisible absorbance spectrophotometer. The protein samples were diluted to 23 M in polybuffer B. Spectra were recorded at 25°C in 0.5-nm steps from 300 to 200 nm over a pH range of 7 to 10. The buffer solution was scanned relative to air, followed by a protein solution in the same cuvette versus air. The two spectra were then subtracted and the difference converted to molar absorption coefficients at 240 nm (⑀ 240 nm ). Data were fitted to a model derived from the Henderson-Hasselbach equation as shown in Equation 3 for one apparent pK a .
Steady-state MsrA Kinetics in the Presence of the Trx Recycling System-Steady-state kinetic parameters were determined with the Trx reductase recycling system (E. coli Trx (100 M), E. coli Trx reductase (4.8 M), NADPH (1.2 mM)) and by varying the concentrations of AcMetSONHMe. AcMetSON-HMe was prepared and purified as previously described (2). Initial rate measurements were carried out at 25°C in buffer A or polybuffer B on a Kontron Uvikon 933 spectrophotometer by following the decrease of the absorbance at 340 nm due to the oxidation of NADPH. Initial rate data were fitted to the Michaelis-Menten relationship using least squares analysis to determine k cat and K m for AcMetSONHMe. E. coli Trx1 and SCHEME 1. Schematic representation of the catalytic mechanism of MsrA and MsrB from N. meningitidis. The mechanism consists of three steps. In step 1, called the reductase step, a sulfenic acid intermediate is formed on the catalytic Cys-X with a concomitant release of one mol of Met/mol of enzyme. In step II, a disulfide bond is formed between the Cys-X and the recycling Cys-Y with a release of a water molecule. In step III, return of the active site to a fully reduced state proceeds via reduction of the Msr disulfide bond by reduced Trx. RSOCH 3 and RSCH 3 represent MetSO and Met, respectively. For N. meningitidis MsrA, Cys-X ϭ Cys-51 and Cys-Y ϭ Cys-198.
Trx reductase were prepared following experimental procedures already published (7).
Preparation of MsrA under Oxidized Disulfide State-MsrA oxidation was achieved by mixing 100 M MsrA with 100 mM MetSO in buffer A. The MetSO used was DL-Met-R,S-SO of which only the S isomer is a substrate for MsrA. After 10 min of incubation at room temperature, oxidized proteins were passed through an Econo-Pac 10 DG desalting column (Bio-Rad) equilibrated with buffer A. Oxidation of MsrA in the disulfide state was checked by titration with 5,5Ј-dithiobis(2-nitro)benzoate.
Fluorescence Properties of Wild-type and Mutated MsrAs-The fluorescence excitation and emission spectra of wild-type and mutated MsrAs in their reduced and Cys-51/Cys-198 disulfide state were recorded on a flx spectrofluorometer (SAFAS) thermostated at 25°C in buffer A with 10 M of each protein as previously described (1).

Determination of the Rate of Met Formation and of Thiol Loss by Single Turnover Quenched Flow Experiments-Quenched
flow measurements were carried out at 25°C on a SX18MV-R stopped-flow apparatus (Applied PhotoPhysics) fitted for double mixing and adapted to recover the quenched samples as previously described (1). The apparatus worked in a pulsed mode. Under the conditions used, a minimum aging time of ϳ25-40 ms was determined. Equal volumes (57.5 l) of a solution containing 550 M Glu-94-mutated MsrA in buffer A and a solution containing AcMetSONHMe in buffer A were mixed in the aging loop. The mixture was then allowed to react for the desired time before being mixed with 115 l of a quenched aqueous solution containing 2% of trifluoroacetic acid. Quenched samples were then collected in a 200-l loop. For each aging time, four shots were done and the four corresponding quenched samples were pooled in a volume of 700 l and then analyzed.
After protein precipitation and centrifugation, Ac-L-Met-NHMe (AcMetNHMe) quantification in the resulting supernatant was carried out by reverse phase chromatography as previously described (2): 100 l were injected onto a 4.6 ϫ 250-mm Atlantis dC18 reverse phase column (Waters) on an AKTA explorer system (Amersham Biosciences) equilibrated with H 2 O/0.1% trifluoroacetic acid. AcMetNHMe was eluted after AcMetSONHMe with a linear gradient of acetonitrile.
The other part of the quenched samples that was not treated with 100% of trifluoroacetic acid was used to 1) determine the protein concentration from the absorbance at 280 nm and 2) quantify the free cysteine content, using 2PDS as a thiol probe, in the presence of urea to avoid precipitation of the protein in the cuvette. Progress curves of pyridine-2-thione production were recorded at 343 nm in 1.1 M urea, buffer A. Enzyme concentration was 6.19 M, and 2PDS concentration was 665 M. The amount of pyridine-2-thione formed was calculated using an extinction coefficient at 343 nm of 8,080 M Ϫ1 ⅐cm Ϫ1 .
Data were plotted as mole of AcMetNHMe formed/ mole of MsrA and as free remaining thiols/mole of MsrA, both as a function of time. The rate of Met formation was determined by fitting the curve to the monoexponential Equation 4 in which a represents the fraction of Met formed/mole of MsrA and k Met represents the rate constant.
The rate of loss in free thiols was determined by fitting the curve to the monoexponential Equation 5 in which y0 represents the number of free remaining thiols, a the number of oxidized thiols, and k SS the rate constant.
Kinetics of the Formation of the Cys-51/Cys-198 MsrA Disulfide Bond in the Absence of Reductant by Single Turnover Stopped-flow Experiment at pH 8-Kinetics of the Trp-53 fluorescence variation associated with the formation of the Cys-51/Cys-198 disulfide bond were measured for E94Q, Y82F, and Y134F MsrAs at 25°C on a SX18MV-R stopped-flow apparatus (Applied PhotoPhysics) fitted for fluorescence measurements as described previously (1). The excitation wavelength was set at 284 nm, and the emitted light was collected using a 320-nm cutoff filter. One syringe contained MsrA in buffer A (10 M final concentration after mixing), and the other one contained AcMetSONHMe at various concentrations in buffer A. An average of at least six runs was recorded for each AcMetSONHMe concentration. Rate constants, k obs , were obtained by fitting fluorescence traces with the monoexponential Equation 6 in which c represents the end point, a the amplitude of the fluorescence increase (Ͻ0), and k obs the rate constant.
Data were fitted to Equation 7 using least square analysis to determine k max and K S for AcMetSONHMe. S represents the AcMetSONHMe concentration and K S the apparent affinity constant.
Kinetics of the Trp-53 fluorescence variation associated with the formation of the Cys-51/Cys-198 disulfide bond were measured for Y82F/Y134F and Y82F/Y134F/E94Q MsrAs at 25°C on a flx spectrofluorometer (SAFAS). The excitation wavelength was set at 284 nm, and the fluorescence emission at 340 nm was recorded versus time after enzyme addition. Data were then treated as described above to obtain k obs , k max , and K S values.
pH Dependence of the Reductase Step Rate Constant-Determination of k max and K S as a function of pH was carried out for wild-type MsrA by single turnover pre-steady-state fluorescence stopped-flow spectroscopy, using the same procedure as described in the previous section but replacing buffer A with polybuffer B. k obs values for E94Q, Y82F, Y134F, Y82F/Y134F, and Y82F/Y134F/E94Q MsrAs were determined at saturating concentration of AcMetSONHMe as a function of pH. Kinetics of Trp-53 fluorescence variation were recorded either with the stopped-flow apparatus or the spectrofluorometer depending on the mutated MsrA, as described in the previous section. The pH dependence of the reductase step rate constant for E94A and E94D MsrAs was determined under steady-state conditions using the Trx recycling system. k max (or k obs ) values were plotted against pH and fitted to Equation 8, deriving from a one-pK a model, where k max opt represents the maximum pHindependent rate constant.

Determination of pK app of the Cys Residues
The pK app of both Cys-51 and Cys-198 were determined in the reduced free enzyme by two methods. The first one involved determining the second-order rate constant of the reaction with the Cys-specific reactivity probe 2-PDS as a function of pH. The second one took advantage of the variation of the thiolate UV absorbance as a function of pH.
Kinetics of Reaction of Reduced Wild-type, C51S, and C198S MsrAs with 2PDS-Reaction of 2PDS with wild-type MsrA obeyed pseudo-first-order kinetics, with formation of 2 mol of pyridine-2-thione/mol of MsrA as determined from the absorbance change at 343 nm. This result was expected as two Cys are present in N. meningitidis MsrA at positions 51 and 198. For all pH used, stopped-flow traces fitted to monoexponential Equation 1, with amplitude corresponding to the release of 2 mol of pyridine-2-thione. pH-k 2 profile fitted to monosigmoidal Equation 2 with a pK app value of 9.7 and k 2max value of (2.4 Ϯ 0.3)⅐10 5 M Ϫ1 s Ϫ1 (Fig. 1A). The product of 2PDS reaction with wild-type MsrA is the disulfide-oxidized enzyme and not the thiopyridine adducts. Indeed, no release of pyridine-2-thione was observed when 10 mM dithiothreitol was added to the purified product (data not shown).
C51S and C198S MsrAs behaved similarly to wild-type MsrA, except that only 1 mol of pyridine-2-thione/mol of MsrA was formed. pK app value of 9.3 Ϯ 0.1 and a k 2max value of (3.1 Ϯ 0.7)⅐10 4 M Ϫ1 s Ϫ1 for Cys-51 and a pK app of 9.8 Ϯ 0.1 and a k 2max value of (2.6 Ϯ 0.6)⅐10 4 M Ϫ1 s Ϫ1 for Cys-198 were determined (Fig.  1B). Altogether, the data support a pK app value of both Cys-51 and Cys-198 in the reduced free wild-type enzyme close to 9.5.
Direct Thiolate UV Absorbance of Reduced Wild-type, C51S, and C198S MsrAs-The thiolate absorbance of wild-type, C51S, and C198S MsrAs was monitored between pH 6 and 10. Analysis of the spectra and of the ⑀240 nm as a function of pH yielded monosigmoidal plots for all three MsrAs. Data fitted to pK values of 9.7, 9.8, and 9.7, associated with ⌬⑀240 nm of 3.1⅐10 4 , 2.3⅐10 4 , and 2.6⅐10 4 M Ϫ1 ⅐cm Ϫ1 for wild-type, C51S, and C198S MsrAs, respectively (Fig. 2). These pK app values of Cys-51 and Cys-198 in the reduced free enzyme are in good agreement with those obtained with 2PDS.

Kinetic Characterization with Identification of the Rate-limiting Step of the Mutated MsrAs at pH 8
Steady-state catalytic constants of mutated MsrAs at positions 82, 94, and/or 134 were determined at pH 8, which is the optimum pH for the wild type (1). AcMetSONHMe was used instead of MetSO because MsrA displays a better affinity for AcMetSONHMe (9). As shown in Table 1, Y82F, Y134F, and E94Q MsrAs exhibited slight modifications of k cat compared with wild-type MsrA, with k cat values from 0.9 to 2.2 s Ϫ1 and a K m increase from 0.8 to 25 mM. In contrast, E94A, E94D, Y82F/ Y134F, and Y82F/Y134F/E94Q MsrAs showed strongly decreased k cat values, from 1⅐10 Ϫ3 to 2.5⅐10 Ϫ1 s Ϫ1 and increased K m from 24 to 161 mM. Values of k obs were determined using nonlinear regression analysis, and second-order rate constants k 2 were fitted to Equation 2 (solid line) (see also "Experimental Procedures"). The plateau of the sigmoidal plot was not attained at the pH tested. Therefore, the pK app values obtained by fitting could only be taken as estimates. In the wild type, the rate of the reductase step is largely higher than the k cat value (1). Therefore, to interpret the eventual kinetic consequences of the substitutions at positions 82, 94, and 134 at the level of the reductase step, it was first necessary to attain this rate. This was determined for E94Q, Y82F, Y134F, Y82F/Y134F, and Y82F/Y134F/E94Q MsrAs by following the variation of the Trp-53 fluorescence intensity under single turnover conditions, i.e. in the absence of reductant (1). In that context, it was assumed that the reductase step of the mutated MsrAs is still rate-determining in the process leading to formation of the Msr disulfide bond, 4 as previously shown for the wild type (1). In the case of E94Q, Y82F, and Y134F MsrAs, formation of the disulfide bond led to an increase in the Trp-53 fluorescence emission similar to that described for the wild type, whereas a quenching of the fluorescence was observed for Y82F/ Y134F and Y82F/Y134F/E94Q MsrAs.Structural factors responsible for this different behavior remain unknown. For all mutated MsrAs, the variation of the fluorescence signal in function of time is of monoexponential type whatever the AcMetSONHMe concentration. The kinetic parameters k max and K S values are summarized in Table 1. No strong K S effect was observed for any of the five mutated MsrAs (K S values ranging from 26 to 151 mM). The k max value for E94Q MsrA is still high compared with that of the wild type and is 14-fold higher than k cat value, which is indicative of a rate-limiting step still associated with the Trx-recycling step. This is also the case for Y82F and Y134F MsrAs for which the k max values are 23-and 160-fold higher than k cat values, respectively. In contrast, drastic effects on k max , i.e. at least 2⅐10 4 -fold decrease, were observed for Y82F/Y134F and Y82F/Y134F/E94Q MsrAs. Moreover, k max and k cat values are similar, showing that the reductase step is the rate-limiting step.
The fact that the variation of the Trp-53 fluorescence message is, for unknown reasons, not of monoexponential type for E94A and E94D MsrAs (curve not shown) led us to use another method to attain the kinetics of the reductase step for these two mutated MsrAs. This method, which was already used for the wild type, consists of following the rate of formation of AcMet-NHMe and of the Cys-51/Cys-198 disulfide bond under single turnover conditions, i.e. in the absence of Trx. This was done at a saturating concentration of 300 mM AcMetSONHMe. E94D analysis required the use of a rapid mixing apparatus, whereas E94A study was possible by manual mixing. Formation of 0.9 mol of AcMetNHMe/mol of enzyme was observed for both E94D and E94A MsrAs. Rate constant for AcMetNHMe formation (k Met ) was 0.27 s Ϫ1 and 2.9⅐10 Ϫ2 s Ϫ1 for E94D and E94A MsrAs, respectively. The free thiol content profile fitted to a monoexponential model, with concomitant loss of Cys-51 and Cys-198 thiols, with rate constant (k SS ) of 0.31 s Ϫ1 and 4.10 Ϫ2 s Ϫ1 for E94D and E94A MsrAs, respectively. The fact that k Met , k SS , and k cat values are similar indicated that the reductase step is rate-limiting for both mutated MsrAs. Therefore, the K m value determined under steady-state conditions represents the K S value of the reductase step. As observed for the other mutated MsrAs, the K S values are similar to that of the wild type (see Table 1).

pH Dependence of the Kinetic Parameters of the MetSO Reductase Step
The kinetic parameters k max and K S of the reductase step for the wild type were determined at different pH values by fluorescence stopped-flow spectroscopy under single turnover conditions, i.e. in the absence of Trx. The pH-k max plot, presented in Fig. 3, exhibits a monosigmoidal profile governed by the contribution of an ionizable group of pK app 5.7 Ϯ 0.1 that must be deprotonated for efficient MetSO reduction. This ionized species is characterized by a k max opt value of 730 s Ϫ1 .
For E94Q, Y82F, Y134F, Y82F/Y134F, and Y82F/Y134F/ E94Q MsrAs, the rate of the reductase step was also measured by fluorescence spectroscopy under single turnover conditions, either on a stopped-flow apparatus or a conventional spectrofluorometer, depending on the value of the rate. In the case of E94A and E94D MsrAs, the rate of the reductase step was determined under steady-state conditions. For each substituted MsrA, the observed rate constant k obs of the MetSO reductase step was measured, at each pH, with only one concentration of AcMetSONHMe (300 mM). This concentration was shown to be saturating over the pH range investigated except for E94Q MsrA (data not shown). Therefore, k obs value can be considered 4 For each Glu-94-mutated MsrA, the rate constant of the reductase step was shown to be rate-determining in the two-step process leading to the formation of the disulfide bond by following the rate of AcMetNHMe and disulfide bond formation directly by acid-quenching of the reaction, followed by quantification of product and MsrA remaining free thiols (see "Results" for E94A and E94D MsrAs, data not shown for E94Q). For each Tyr-substituted MsrA, the rate of the reductase step was shown to be too fast at pH 8 to allow direct determination of the rate of AcMetNHMe and disulfide bond formation (data not shown).
a Steady-state parameters were deduced from nonlinear regression of initial rates to the Michaelis-Menten relationship (see "Experimental Procedures"). b Kinetic parameters of the reductase step were obtained from nonlinear regression of k obs to Equation 7 (see "Experimental Procedures"), except for E94A and E94D MsrAs. For these two latter substituted MsrAs, k max and K S values correspond to k cat and K m values determined under steady-state conditions (see "Results"). c Kinetic parameters, k max opt at optimum pH and pK app values, were deduced from nonlinear regression of k obs to Equation 8 (see also Fig. 4). as a k max value. In the case of E94Q MsrA, 300 mM AcMetSON-HMe was not saturating at pH Ͼ8, and consequently k obs values were determined only up to pH 8. The k obs profile remains monosigmoidal for all substituted MsrAs, with increasing k obs value with increasing pH. Data fitted to a single pK a model (see Fig. 4). E94D MsrA has a pK app of 6.7 with a k max opt value of 0.19 s Ϫ1 . E94A, E94Q, Y82F, Y134F, and Y82F/Y134F MsrAs displayed a more pronounced pK app shift, with values ranging from 7.5 to 8.0 and with k max opt values of 2⅐10 Ϫ2 , 28, 46, 380, and 7⅐10 Ϫ2 s Ϫ1 , respectively (see Fig. 4 and Table 1). It is noteworthy that substitutions of Tyr-82 or Tyr-134 induced a similar pK app shift to 7.6 and that double substitution of these two Tyrs led to a higher shift to 8.0. The triple substituted MsrA Y82F/Y134F/E94Q displayed the most highly shifted pK app , with a value of 9.5 and the lowest k max opt value of 1.1⅐10 Ϫ2 s Ϫ1 (see Fig. 4 and Table 1).

DISCUSSION
The methionine sulfoxide reductase step of the MsrA mechanism was previously shown to be very fast (1) and postulated to imply the formation of a sulfurane transition state of bipyramidal geometry on the basis of the theoretical chemistry study (6). According to these data, the reductase step should imply 1) the formation of an enzyme-substrate complex, 2) the deprotonation of Cys-51, 3) the involvement of an acid catalyst to protonate the sulfoxide substrate and favor the sulfurane-type transition state formation, and 4) the rearrangement of the sulfurane-type transition state to obtain Met and sulfenic acid. The invariant residues Glu-94, Tyr-82, and Tyr-134 are correctly positioned in the three available x-ray MsrA structures to interact via H-bond with a water molecule that is located at the place of the oxygen of the sulfoxide function (Fig. 5). Based on these structural features, a reasonable hypothesis supports Glu-94 as the presum-ably required acid catalyst. This proposition is reinforced by the observation that no other acidic residue is present in the close proximity of the sulfoxide function. The two phenolic side chains of Tyr-82 and Tyr-134 could be involved in substrate binding (i.e. an affinity contribution) and/or in substrate positioning and transition state stabilization (i.e. a chemical catalysis contribution).
Cys-51 Activation within the Active Site-In the free wildtype enzyme, 2-PDS titration and direct thiolate UV absorbance titration revealed a single pK app ϳ9.5 for both Cys-51 and -198. Moreover, the 2PDS chemical reactivity of MsrA Cys is   Fig. 3. For all substituted MsrAs, the rate of the reductase step was determined with 300 mM AcMetSONHMe at various pH values in polybuffer B. The independence of the rate constant on AcMetSONHMe concentration was verified at two extreme pH values (around 5 and 9). For E94Q, Y82F, and Y134F MsrAs, the increase of Trp-53 fluorescence intensity was followed after excitation at 284 nm on a stopped-flow apparatus. For Y82F/Y134F and E94Q/Y82F/Y134F MsrAs, variation of fluorescence emission was recorded on a spectrofluorometer with excitation and emission wavelength set at 284 and 340 nm, respectively. For E94A and E94D MsrAs, the reductase step rate constant was followed using the Trx-recycling system. Experimental k obs data obtained at each pH were analyzed by nonlinear regression against Equation 8 to obtain k max opt and pK app values (see also "Experimental Procedures"). k max opt and pK app values are presented in Table 1. Experimental k obs data were normalized, taking k max opt value as 100% for each enzyme, for the clarity of the representation. Symbols represent normalized data, and lines represent the fit. comparable with that of the thiolate of glutathione (10), arguing for two Cys residues not activated but accessible in the free wild-type MsrA. The fact that 2PDS mediates oxidation of reduced MsrA to its disulfide state with simultaneous liberation of two moles of pyridine-2-thione/enzyme, via an apparent one-step kinetic mechanism, can be explained by a mechanism reminiscent of that of the Msr mechanism. First, the rate-limiting reaction between one Cys and one molecule of 2PDS would lead to a transient mixed disulfide. This disulfide would be rapidly attacked by the second Cys of the enzyme, leading to the disulfide-oxidized MsrA. Such a mechanism would imply that the measured pK app only reflects ionization of the first Cys. It is tempting to assign Cys-51 as this "2-PDS-reactive" Cys. 2-PDS titration of C51S and C198S MsrAs supports this assignment, as Cys-51 (C198S MsrA) exhibits a stronger reactivity than Cys-198 (C51S MsrA) between pH 8 and 9 (see Fig. 1B). The measured pK app values around 9.5 for Cys-51 and Cys-198, respectively, are significantly higher than the value of 8.8 for the model Cys of glutathione (10). As MsrA function is supposed to be widely used during oxidative stress burst, high pK a could serve as a protection against oxidative modification of both Cys by restricting their reactivity toward reactive oxygen or nitrogen species. Stabilization of the thiol state of Cys-51 could be provided by the hydrophobic character of the active site.
As stated above, Cys-51 has to be deprotonated to allow its efficient attack on the sulfoxide function. The observed high rate constant of the reductase step at pH 8.0 implies a large, mandatory shift in Cys-51 pK a , from near 9.5 to at least somewhere below 7. The pH dependence of k max displays a single pK app of 5.7, the rate of sulfoxide reduction increasing with pH. Moreover, the E94Q and Y82F/Y134F/E94Q MsrAs still show the contribution of a single pK of 8 and 9.5, respectively. Such results support the attribution of the pK app of 5.7 to Cys-51 and not to Glu-94. Thus, formation of the MsrA-AcMetSONHMe complex provokes an activation of Cys-51 by decreasing its pK app by 3.8 units.

Glu-94, Tyr-82, and Tyr-134 Contributions in the Reductase
Step-Substitution of Glu-94 by Ala or Asp drastically decreased the k max opt rate of the reductase step of the MsrA mechanism by factors of 3.6⅐10 4 and 3.8⅐10 3 , respectively, with no significant K S effect at pH 8 and caused the shift of the rate-limiting step from the Trx-recycling process to the sulfoxide reduction step. These data support the implication of Glu-94 in the catalysis of sulfoxide reduction, but not in substrate binding, and identify its side chain as a critical catalyst. However, the kinetic parameters of the reductase step obtained with E94Q MsrA revealed a rather minor k max opt decrease of only 26-fold compared with that of the wild type. As Gln-94 cannot be a proton donor, it is tempting to conclude that Glu-94 does not directly play a role as general acid catalyst but likely stabilizes, via H-bonding, the sulfurane transition state leading to sulfenic acid formation (see the last paragraph under "Discussion"). Indeed, substitution of Glu by Gln retains an H-bonding ability. This is reinforced by the fact that no pK app of an acidic catalyst is observed on the pH-k max profile of the wildtype MsrA. In addition to this effect on the rate constant of the sulfoxide reduction, substitution of Glu-94 induced an increase of the pK app governing this step and assigned to Cys-51. Thus, Glu-94 is, directly or not, also involved in the activation of Cys-51 upon substrate binding.
Substitution of Tyr-82 and Tyr-134, or both, by Phe did not affect K S values. Moreover, the k max opt constant of the sulfoxide reduction was slightly decreased by the absence of one of the phenolic hydroxyls, in particular of Tyr-134, but drastically decreased when both were removed. The fact that substitutions of both Tyrs is required to observe a strong decrease of k max opt suggests that in single Tyr-substituted MsrA the remaining Tyr in the presence of Glu-94 compensates the absence of the second phenolic hydroxyl. Similarly to Glu-94, substitution of Tyr-82 and/or Tyr-134 also leads to an increase of the pK app governing this step, assigned to Cys-51. Importantly, in contrast to the catalytic role of Tyr-82 and Tyr-134, which can be fairly well assumed by only one Tyr, Cys-51 activation is more sensitive to the absence of one of the two phenolic side chains.
Taken together, these data strongly suggest that Glu-94, Tyr-82, and Tyr-134, in addition to their role in favoring the polarized form of the sulfoxide function of the substrate bound, are involved in the stabilization of the sulfurane transition state (see the last paragraph under "Discussion") that precedes the sulfenic acid formation. They also participate, directly or indirectly, in the Cys-51 activation upon formation of the MsrAsubstrate complex. As a confirmation, the triple substituted MsrA Y82F/Y134F/E94Q displays the most highly shifted pK app with a value of 9.5, close to the pK app of Cys-51 determined in the free enzyme, and the biggest reduction of k max opt , by a factor of 6.6⅐10 4 compared with the wild type.
Structure/Function Relationships within the MsrA Active Site-The available x-ray structure of MsrA from M. tuberculosis in complex with a Met from a neighboring monomer pro-