Characterization of the Amino Acids from Neisseria meningitidis Methionine Sulfoxide Reductase B Involved in the Chemical Catalysis and Substrate Specificity of the Reductase Step*

Methionine sulfoxide reductases (Msrs) are antioxidant repair enzymes that catalyze the thioredoxin-dependent reduction of methionine sulfoxide back to methionine. The Msr family is composed of two structurally unrelated classes of enzymes named MsrA and MsrB, which display opposite stereoselectivities toward the S and R isomers of the sulfoxide function, respectively. Both classes of Msr share a similar three-step chemical mechanism involving first a reductase step that leads to the formation of a sulfenic acid intermediate. In this study, the invariant amino acids of Neisseria meningitidis MsrB involved in the reductase step catalysis and in substrate binding have been characterized by the structure-function relationship approach. Altogether the results show the following: 1) formation of the MsrB-substrate complex leads to an activation of the catalytic Cys-117 characterized by a decreased pKapp of ∼2.7 pH units; 2) the catalytic active MsrB form is the Cys-117-/His-103+ species with a pKapp of 6.6 and 8.3, respectively; 3) His-103 and to a lesser extent His-100, Asn-119, and Thr-26 (via a water molecule) participate in the stabilization of the polarized form of the sulfoxide function and of the transition state; and 4) Trp-65 is essential for the catalytic efficiency of the reductase step by optimizing the position of the substrate in the active site. A scenario for the reductase step is proposed and discussed in comparison with that of MsrA.

Methionine sulfoxide reductases (Msrs) are antioxidant repair enzymes that catalyze the thioredoxin-dependent reduction of methionine sulfoxide back to methionine. The Msr family is composed of two structurally unrelated classes of enzymes named MsrA and MsrB, which display opposite stereoselectivities toward the S and R isomers of the sulfoxide function, respectively. Both classes of Msr share a similar three-step chemical mechanism involving first a reductase step that leads to the formation of a sulfenic acid intermediate. In this study, the invariant amino acids of Neisseria meningitidis MsrB involved in the reductase step catalysis and in substrate binding have been characterized by the structure-function relationship approach. Altogether the results show the following: 1) formation of the MsrBsubstrate complex leads to an activation of the catalytic Cys-117 characterized by a decreased pK app of ϳ2.7 pH units; 2) the catalytic active MsrB form is the Cys-117 ؊ /His-103 ؉ species with a pK app of 6.6 and 8.3, respectively; 3) His-103 and to a lesser extent His-100, Asn-119, and Thr-26 (via a water molecule) participate in the stabilization of the polarized form of the sulfoxide function and of the transition state; and 4) Trp-65 is essential for the catalytic efficiency of the reductase step by optimizing the position of the substrate in the active site. A scenario for the reductase step is proposed and discussed in comparison with that of MsrA.
Methionine residues in proteins are susceptible to oxidation to methionine sulfoxide (MetSO) 4 by reactive oxygen species. This post-translational oxidative damage has been implicated in a variety of age-related diseases (1)(2)(3)(4). In this context, the repair of oxidized protein-bound MetSO back to methionine by methionine sulfoxide reductases (Msr) can be considered as a means of antioxidant defense (5)(6)(7)(8)(9)(10). It has also been described to be a component of the mechanisms used by pathogenic bacteria of the genus Neisseria to infect host cells, and thus to be a key factor in the virulence of these bacteria (11).
There exists two structurally unrelated classes of Msrs called MsrA and MsrB (12)(13)(14)(15), which reduce free and protein-bound Met-(S)-SO and Met-(R)-SO, respectively, and display a similar three-step mechanism (16 -19). The first step, i.e. the reductase step that is not rate-limiting, leads in both cases to the formation of a sulfenic acid intermediate and release of 1 mol of Met per mol of Msr. Recently, the role of invariant amino acids of the MsrA domain of the PilB protein of Neisseria meningitidis involved in the chemical catalysis of the reductase step was investigated by comparing the kinetic properties of the wild type to those of several mutated MsrAs (20,21). A scenario of the reductase step was thus proposed in which the substrate binds to the active site with its sulfoxide function largely polarized via interactions with invariant amino acids, the role of which is to stabilize the transition state of the sulfurane type. In this scenario, the stabilization of the thiolate form of the catalytic Cys is proposed to be substrate-assisted by the positive charge borne by the sulfur of a sulfoxide. This scenario implies a transfer of the hydrogen of the catalytic Cys to the oxygen of the sulfoxide function via an acid-base catalyst, which is likely to be a Glu residue. In parallel with this study, recent work was also done to characterize the amino acids involved in MetSO recognition (22). In particular, the hydrophobic pocket composed of the side chains of a Trp and a Phe residue was shown to be essential in positioning the substrate for efficient catalysis of the reductase step via van der Waals interactions with the ⑀-methyl group of the substrate.
In contrast, no information is available on the nature of the amino acids and how they operate in catalysis of the reductase step and on the structural factors involved in recognition of the substrate in MsrB. Only the three-dimensional structure of the MsrB domain of the PilB protein of Neisseria gonorrhoeae, which differs from that of the N. meningitidis MsrB domain by only two amino acids, has been determined by x-ray so far (23). One of the MsrB domains in the asymmetric unit has a cacodylate molecule bound. This cacodylate mole-cule is a good mimic of a MetSO-bound active site, which is surface-exposed. As shown in Fig. 1, methyls (designated A and B) represent the ⑀-methyl and the ␥-methylene of MetSO, respectively, whereas the A atom and one of the oxygen atoms mimic the sulfur and the oxygen of the sulfoxide function, respectively. The methyl groups of cacodylate are located in a cavity described to be formed by the side chains of Cys-117 5 and Trp-65, whereas the oxygen atom that mimics the oxygen of the sulfoxide interacts through a hydrogen-bonding network with His-103 and with Thr-26, His-100, and Asn-119 but in this case via a water molecule. Finally, inspection of the x-ray structure shows a Cys-117, Arg-116, Asp-107 triad that was postulated to activate the catalytic Cys-117 (23).
In this study, in view of the structural information, all these residues of the MsrB domain of the N. meningitidis PilB were substituted. The kinetic parameters were determined for all mutated MsrBs and compared with those of the wild type. The pH dependence of the rate constant of the reductase step of H103N and R116L MsrBs was also determined and compared with that of the wild type. Two major results of this study are as follows: 1) the catalytic active form of MsrB in the reductase step is the Cys-117 Ϫ /His-103 ϩ with pK app values of 6.7 and 8.3, respectively; 2) activation of Cys-117 does not involve Arg-116 and Asp-107 as postulated previously but is rather substrateassisted via the polarized form of the sulfoxide function of the substrate. Taking into account all the results, a scenario for the catalysis is proposed in comparison with that of MsrA and is discussed in terms of evolution not only for catalysis but also for substrate recognition.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis, Production, and Purification of Enzymes-The Escherichia coli BE002 strain (MG1655 msrA::spec⍀, msrB::␣3kan) was used for all N. meningitidis MsrB productions, transformed with the plasmidic construction pSKPILBMsrB containing only the coding sequence of msrB under the lac promoter (17). 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 all preparations were devoid of contaminating activity by the Msrs from E. coli. Site-directed mutageneses were performed using the QuikChange site-directed mutagenesis kit (Stratagene).
Purification of the wild-type MsrB was performed as described previously (17). In the case of mutated MsrBs, the enzymes were produced as a nonsoluble form. Pellets obtained after sonication and centrifugation were resuspended in a minimal volume of buffer A (50 mM Tris-HCl, 2 mM EDTA, pH 8) containing 20 mM dithiothreitol and 6 M urea, and then dialyzed twice against buffer A containing 3 and 0 M urea, successively. The solution was then applied successively onto an ACA 54 resin, a Q-Sepharose column, and a phenyl-Sepharose column (Amersham Biosciences) as described previously for the wild type (17).
Purity of wild-type and mutated MsrBs was checked by electrophoresis on a 15% SDS-PAGE followed by Coomassie Brilliant Blue R-250 staining and by electrospray mass spectrometry analyses. Storage of the enzymes was done as described previously (17). The molecular concentration was determined spectrophotometrically, using extinction coefficient of 17,330 M Ϫ1 ⅐cm Ϫ1 at 280 nm for wild-type and mutated MsrBs except for W65A and W65F MsrBs for which an extinction coefficient of 11,580 M Ϫ1 cm Ϫ1 was used. Trx1 and Trx reductase from E. coli were prepared following experimental procedures described previously (24,25).
pH Dependence of MsrB Thiol Reaction Rates with 2,2Ј-Dipyridyl Disulfide (2-PDS)-Kinetic measurements of MsrB reaction with 2-PDS were carried out on an SX18MV-R stopped-flow apparatus (Applied PhotoPhysics). Kinetic reactions were performed at 25°C, under pseudo-first-order conditions at a constant ionic strength of 0.15 M over the pH range 7.25-9.5 with polybuffer B (120 mM Tris, 30 mM imidazole, 30 mM acetic acid). One syringe was filled with wild-type or Cys to Ser mutated MsrBs, and the other was filled with 2-PDS. MsrB and 2-PDS concentrations after mixing were 10 and 310 M, respectively. Release of pyridine-2-thione was followed at 343 nm using an extinction coefficient of 8080 M Ϫ1 ⅐cm Ϫ1 .
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 monoexponential Equation 1, where a is the burst amplitude and c is the end point.
The second-order rate constant k 2 was calculated by dividing the k obs value by the concentration of 2-PDS. k 2 versus pH profiles fitted to Equation 2 for one-pK a profiles, in which k 2(max) represents the second rate constant for the thiolate form.
Kinetics in the Presence of the Trx Recycling System-Kinetics in the presence of the Trx recycling system were carried out with 400 mM Ac-L-Met-(RS)-SO-NHMe (AcMetSONHMe) as a substrate and 100 M E. coli Trx1 as a reductant in the presence of the Trx recycling system (1.28 M E. coli Trx reductase and 0.3 mM NADPH), as described previously (17). AcMet-SONHMe was prepared from Ac-L-Met-OMe (Bachem) as described previously (26). Purity of AcMetSONHMe was assessed by proton NMR analysis.
Kinetics of the First Step in the Absence of Reductant by Single Turnover Fluorescence Experiments at pH 8-Kinetics of the fluorescence decrease of Trp-65 associated with the formation of the disulfide bond were measured for T26A, H100A, H100N, D107A, R116L, and N119A MsrBs at 25°C in buffer A on an SX18MV-R stopped-flow apparatus (Applied PhotoPhysics) adjusted for fluorescence measurements, as described previously (26). The excitation wavelength was set at 291 nm, and the emitted light was collected above 320 nm, using a cutoff filter. One syringe contained mutated MsrB in buffer A (10 M final concentration after mixing), and the other one contained AcMetSONHMe at various concentrations in buffer A (10 -600 mM, final concentration). An average of six runs was recorded at each substrate concentration. Rate constants, k obs , were obtained by fitting fluorescence traces with the monoexponential equation (Equation 3), in which c represents the end point, and a represents the amplitude of the fluorescence decrease.
When saturation is obtained as it is in wild-type, T26A, H100A, H100N, D107A, and R116L MsrBs, data were fit to Equation 4 using least squares analysis to determine k obs max and K S for AcMetSONHMe, where S represents the AcMet-SONHMe concentration, and K S represents the apparent affinity constant for the substrate.
For H103N and H103A MsrBs, kinetics of the slow fluorescence decrease were measured at 25°C in buffer A on an flx spectrofluorometer (SAFAS). The excitation wavelength was set at 291 nm, and the emitted light followed at 340 nm. The assay mixture contained 10 M of enzyme in buffer A with various concentrations of AcMetSONHMe (10 -300 mM). Data were then treated using the same procedure as described above to obtain k obs max and K S values.
To determine the pseudo-second-order constant (k 2 ) values, subsaturating concentrations of AcMetSONHMe were used, from 0.1 to 2 mM for wild-type, T26A, H100A, H100N, D107A, R116L, and N119A MsrBs and from 0.1 to 1 mM for H103A and H103N MsrBs. Kinetics of the rapid fluorescence decrease were measured at 25°C in buffer A with a stopped-flow apparatus for wild-type, D107A, and R116L MsrBs. The kinetics of the slow fluorescence decrease were measured on an flx spectrofluorometer (SAFAS) at 25°C in buffer A for T26A, H100A, H100N, H103A, H103N, and N119A MsrBs. The slope (k 2 ) of the rate constants (k obs ) plotted against substrate concentration was obtained by linear fitting.
pH Dependence of the Reductase Step Rate Constant for Wildtype and H103N MsrBs-Kinetics of the first step were recorded by single turnover fluorescence experiments in the absence of reductant as described above. Determination of k obs max and K S values for AcMetSONHMe as a function of pH was carried out using the same procedure as described above, but replacing buffer A with polybuffer B. k obs max values were plotted against pH and fit for the wild type to Equation 5, derived from a two-pK a model for a bell-shaped curve, and for H103N MsrB to Equation 6, derived from a single pK a model for a monosigmoidal increasing curve.
In Equation 5, pK a1 and pK a2 represent the pK app values governing the acid and the basic limb, respectively, and kЈ and kЉ represent the pH-independent rate constants of the monodeprotonated and double deprotonated forms of enzyme, respectively.
In Equation 6, pK a represents the pK app , and kЈ represents the maximum pH-independent rate constant of the deprotonated form of enzyme.
Determination of the Rate of Ac-L-Met-NHMe (AcMetNHMe) Formation by Single Turnover Quenched-flow Experiments at pH 8 for W65A and W65F MsrBs-In the case of W65F MsrB, quenched-flow measurements were carried out at 25°C on a SX18MV-R stopped-flow apparatus (Applied PhotoPhysics) fitted for the double-mixing mode and adapted to recover the quenched samples as described previously (20,26). Equal volumes (60 l) of a solution containing 200 M W65F MsrB were mixed with various concentrations of AcMetSONHMe in buffer A (from 50 to 400 mM, final concentration) and introduced in the aging loop. The mixtures were allowed to react for 40 -5000 ms and quenched with 2% trifluoroacetic acid, and samples were collected in a 200-l loop. At each aging time, four data runs were collected, and the corresponding quenched samples were pooled in a volume of 700 l. Trifluoroacetic acid (100%, 50 l) was subsequently added to 200 l of the quenched samples to obtain precipitated protein, and in parallel, quenched samples not treated with 100% trifluoroacetic acid were used to determine the protein concentration from absorbance readings at 280 nm. In the case of W65A MsrB, quenching was carried out by manual addition of trifluoroacetic acid (100%, 50 l) in 200 l of mixture containing 200 M of enzyme and 200 -800 mM of AcMetSONHMe in buffer A at different time intervals (30 -3600 s).

JOURNAL OF BIOLOGICAL CHEMISTRY 32399
Samples treated with acetic acid were centrifuged at 12,000 ϫ g for 30 min at room temperature. AcMetNHMe quantification in the resulting supernatant was carried out by C18 reverse phase chromatography as described previously (20,26). Data were plotted as moles of AcMetNHMe formed per mol of MsrBs as a function of time (t). The rate of AcMetNHMe formation was determined by fitting the curve to the monoexponential Equation 7 in which a represents the fraction of AcMetNHMe formed per mol of MsrB, and k obs is the rate constant.
For W65F MsrB, k obs versus AcMeSONHMe concentration profile was fit to Equation 4 using least squares analysis to determine k obs max and K S for AcMetSONHMe. For W65A MsrB, the pseudo-second-order rate constant (k 2 ) was obtained from the slope of rate constant (k obs ) plotted against substrate concentration by linear fitting.

Rationale for the Kinetic Characterization of the Mutated
MsrBs with AcMetSONHMe as Substrate-The rate-limiting step of wild-type MsrB is associated with the Trx-recycling process (26). Therefore, to interpret the consequences on the kinetics of the reductase step of the mutated MsrBs, it was necessary to attain the rate of this step. This was done by following the decrease of the Trp-65 fluorescence emission signal intensity upon going from the reduced form to the disulfide oxidized form as described recently for the wild type (26). In this context, we assumed that the reductase step for all mutated MsrBs remain rate-determining in the process leading to the formation of the MsrB disulfide bond as demonstrated for the wild type (26). This was verified for H100A, H100N, H103A, and H103N MsrBs by measuring directly the rate of formation of AcMetNHMe. In all cases, the rate of formation of AcMetNHMe was similar to that determined by following the decrease of the Trp-65 fluorescence intensity in the presence of 150 mM of Met-(RS)-SO (data not shown). As shown below, two types of mutated MsrBs were observed in terms of rate-limiting step as follows: those for which the rate-limiting step remains associated with the Trx-recycling process and those whose rate-limiting step precedes the Trx-recycling process. To demonstrate this, the rate of decrease in NADPH followed at 340 nm was determined under saturating concentration of Trx-TrxR-NADPH in a coupled enzymatic system and compared with that of the decrease of the Trp-65 fluorescence intensity under the same conditions of pH (i.e. 8) and concentration of AcMetSONHMe.
For all mutated MsrBs, k 2 values of the reductase step that represent the k obs max /K S values were determined under subsaturating concentrations of AcMetSONHMe. When saturating kinetics of AcMetSONHMe was observed, K S and k obs max values were determined, and k 2 values were also calculated directly from these values.
In the case of mutated MsrBs at position 65, in the absence of the fluorescent Trp-65 probe, the rate of the reductase step was determined by following the rate of formation of AcMetNHMe under single turnover conditions, i.e. in the absence of reductant. This was done at several concentrations of AcMetSONHMe for W65A and W65F MsrBs (see Fig. 4C). In all cases, formation of ϳ0.8 -0.9 mol of AcMetNHMe per mol of mutated MsrB was observed for all the substrate concentrations used.
The kinetic experiments were performed with AcMet-SONHMe because of the better affinity of MsrB for this substrate compared with MetSO (26). Thus, the effect of the various substitutions made could be more easily quantified in terms of K S and k obs max values and not only in terms of k 2 values. This proved particularly important in obtaining the pH dependence of k obs max of the reductase step for wild-type, R116L, and H103N MsrBs. Indeed, to interpret the pH profile in terms of contribution of ionizable groups within the Michaelis complex, the concentration of AcMetSONHMe has to be saturating over the pH range investigated.
Determination of the pK app Value of the Catalytic Cys-117 in the Reduced Form of MsrB-The method consisted of determining the second-order rate constant of the reaction with the Cys-specific reactivity probe 2-PDS as a function of pH, by following the formation of pyridine-2-thione at the max of 343 , with amplitude corresponding to the release of 2 mol of pyridine-2thione. The pH-k 2 profile fitted to monosigmoidal Equation 2 resulting in a pK app value higher than 9.3 and a k 2(max) value higher than 1.5 ϫ 10 4 M Ϫ1 ⅐s Ϫ1 (Fig. 2). The fact that no release of pyridine-2-thione was observed when 10 mM dithiothreitol was added showed formation of the disulfide-oxidized MsrB and not the thiopyridine adduct on each Cys (data not shown). Therefore, the pK app value may be assigned to either the catalytic Cys-117, the recycling Cys-63, or to both Cys-63/ Cys-117. This is dependent on the Cys that is first titrated by 2-PDS and the rate-limiting step in formation of the disulfide bond. Thus, to attain the pK app of Cys-117, reactions of 2-PDS with C117S and C63S MsrBs were also carried out. Both mutated MsrBs behaved similarly to the wild-type MsrB, except that only 1 mol of pyridine-2thione/mol of enzyme was released (curves not shown). Both pH-k 2 profiles were of monosigmoidal type with no saturation at high pH, resulting in the estimation of a pK app and k 2(max) values higher than 9.3 and 2.1 ϫ 10 4 M Ϫ1 ⅐s Ϫ1 and 9.0 and 0.8 ϫ 10 3 M Ϫ1 ⅐s Ϫ1 for C63S and C117S MsrBs, respectively (Fig. 2). Altogether, the data favor a pK app value of Cys-117 close to 9.3.

Kinetic Investigations on the Role of Cys-117 and His-103 Directly Involved in Catalysis of the Reductase
Step-The kinetic parameters k obs 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. As shown in Fig. 3, the pHk obs max plot is characterized by a bell-shaped profile that indicates that two groups are required to be in the proper state of ionization for maximum catalytic activity at pH 7.5. One of these groups characterized by a pK app of 6.6 must be ionized, whereas the other with a pK app of 8.3 has to be protonated. Substitution of His-103 by Asn and Ala resulted in a drastic effect on the rate of the reductase step. At pH 8, the k obs max value, which is 10 4 -fold lower compared with that of the wild type (Table 1 and Fig. 4A), is similar to the rate determined under steady-state conditions (13 ϫ 10 Ϫ4 and 40 ϫ 10 Ϫ4 s Ϫ1 at 400 mM of substrate for H103A and H103N, respectively) indicating that the reductase step is now rate-limiting. The K S value for AcMetSONHMe is in a similar range or slightly lower than that of the wild type (Table 1 and Fig. 4A). The pH-k obs max plot of H103N MsrB is of monosigmoidal type and showed a drastic shift in the pH profile toward the basic pH (Fig. 3). The pK app value of the ascending limb is at least 9.0. Such results support the interpretation that the pK app values of 6.6 and 8.3 observed for the wild type correspond to those of Cys-117 and His-103,   Fig. 4). Data for wild-type, T26A, H100A, H103A, H103N, D107A, and R116L MsrBs were fit to Equation 4 by least squares regression to provide k obs(max) and K S values (see Fig. 4). For the wild-type and mutated MsrBs, pseudo-second-order rate constant k 2 values were determined from the slope of the rate constant (k obs ) plotted against a subsaturating substrate concentration by linear fitting (see Fig. 5). It is important to note that the K S values have to be divided by 2 and the k 2 values multiplied by 2, taking into account the fact that the S isomer of the sulfoxide function is neither a substrate nor an inhibitor of MsrB (17 respectively. This also indicates the following: 1) formation of the MsrB-MetSO Michaelis complex leads to an activation of Cys-117 characterized by a decreased pK app of ϳ2.7 pH units, whereas that of His-103 is increased by at least ϳ1.6 pH units if one considers a pK app of 6.7 for His-103 in the MsrB reduced form; and 2) His-103 plays an acid/base catalyst role. This is illustrated by the drastic k obs max decrease observed not only when His-103 is substituted by Ala but also by Asn, which is an isostere of His. As mentioned in the Introduction, Lowther et al. (23) proposed, from the inspection of the active site in which a cacodylate molecule is bound, an activation of Cys-117 through the catalytic triad Cys-117, Arg-116, and Asp-107. In fact, when Arg-116 was substituted by Leu, the pH-k obs max profile was similar to that of the wild type (curve not shown). Only a decrease in the k obs max value was observed over the pH range investigated. At pH 8, the k obs max value was decreased by a factor of 7-fold (Table 1). For D107A MsrB, the k obs max and K S values determined solely at pH 8 are similar to those determined for the wild type (Table  1). Altogether, these results exclude any role of Arg-116 and Asp-107 in the activation of Cys-117. Kinetic Investigations on the Role of Thr-26, His-100, and Asn-119 in Catalysis of the Reductase Step-Substituting Ala or Asn for His-100 led to a reductase step that becomes rate-limiting. Indeed, at pH 8, the k obs values of the reductase step of H100A and H100N MsrBs determined at 400 mM of substrate, i.e. 0.49 and 0.21 s Ϫ1 (Fig. 4B), are similar to those determined under steady-state conditions at 400 mM of substrate i.e. 0.24 s Ϫ1 and 0.10 s Ϫ1 , respectively. The k 2 values were ϳ260and 530-fold decreased for H100A and H100N MsrBs, respectively, compared with that of the wild type (Table 1 and Fig. 5B). These observed decreases in k 2 reflect essentially a k obs effect.
A pH-k obs max plot of the H100N MsrB was then done. But at a pH higher than 8, the K S value for the substrate increased above 400 mM (data not shown). This prevented us from attaining a k obs max value and thus determining a pH-k obs max plot. As a consequence, it was not possible to quantify, if any, the effect of substituting His-100 by Asn on the pK app of Cys-117 and His-103.
Substituting Thr-26 by Ala led to a 190-fold decrease in the k 2 value of the reductase step compared with that of the wild type, at pH 8 (Table 1 and Fig. 5B). This decrease is essentially because of a k obs effect (Table 1 and Fig. 4B). At a pH higher than 8, K S values increased significantly and therefore, as for H100N MsrB, prevented the determination of a pH-k obs max profile (data not shown). The fact that at pH 8 at 400 mM of substrate, the k obs value of the reductase was 0.47 s Ϫ1 whereas under steady-state conditions the rate was 0.2 s Ϫ1 did not permit us to conclusively determine the nature of the rate-limiting step.
Substituting Asn-119 by Ala led to a k 2 value of the reductase step, which decreases by a factor of ϳ600, at pH 8 (Table  1 and Fig. 5B). No significant saturating effect was observed up to 100 mM of substrate (Fig. 4B). At 400 mM of substrate, the k obs value was 0.45 s Ϫ1 that has to be compared with the k obs max value of 49 s Ϫ1 for the wild type at pH 8. As a consequence, the k 2 decrease reflects essentially a k obs effect accompanied by a small K S contribution. As for the T26A MsrB, the fact that the k obs of 0.2 s Ϫ1 determined under steady-state conditions at pH 8 is in the range of that deter-mined for the reductase step at 400 mM of substrate does not allow the nature of the rate-limiting step to be interpreted conclusively. Again, it was not possible to obtain a pH-k obs max profile because of a significant increase in the K S value above pH 8. This significant K S increase observed for H100N, T26A, and N119A MsrBs at a pH higher than 8 compared with that of the wild type remains to be explained.
Kinetic Investigations on the Role of the Trp-65 in the Recognition of the Substrate-Substituting Ala for Trp-65 led to a drastic effect on the reductase step as shown by the 10 5 -fold decrease in k 2 value determined at pH 8 (Fig. 4C). The ratelimiting step is now associated with the reductase step as supported by the fact that the k obs value determined under steadystate conditions at 400 mM of substrate, i.e. 1.8 ϫ 10 Ϫ3 s Ϫ1 , is similar to the 2.5 ϫ 10 Ϫ3 s Ϫ1 value found for the reductase step at 400 mM of substrate. The k 2 decrease both reflects a strong effect on k obs max and K S values. In contrast, substitution of Trp-65 by Phe led to smaller kinetic effects. The Trx-recycling process remains rate-limiting as supported by the value of k obs of 0.5 s Ϫ1 at 400 mM of substrate that is significantly lower than the 3.5 s Ϫ1 k obs value determined for the reductase step at the same substrate concentration. The saturating kinetic effect with respect to the substrate was observed for the reductase step with k obs max and K S values of 6 s Ϫ1 and 290 mM values (Table 1 and Fig. 4C). Thus, substituting Phe for Trp only leads to a 8-fold decrease in k obs max and 3-fold increase in K S value at pH 8.  Table 1, except for N119A MsrB. Indeed for this latter mutated MsrB, the k obs max and K S values obtained (1.8 Ϯ 0.3 s Ϫ1 and 1200 Ϯ 250 mM, respectively) could only be taken as estimated due to the fact that substrate saturating concentration was not reached at 400 mM of substrate.  Table 1.

DISCUSSION
As indicated in the Introduction, MsrB from N. gonorrhoeae in complex with a cacodylate molecule is a good model of a Met-(R)-SO-bound Michaelis complex. As shown in Fig. 1, modeling of the active site with Met-(R)-SO bound shows that methyl A of cacodylate, which is located at the bottom of the hydrophobic pocket formed by the side chains of Cys-117 and Trp-65, mimics the ⑀-methyl, whereas the methyl B that is more toward the surface of the pocket corresponds to the ␥-methylene group of the side chain of MetSO. The ⑀-methyl and the ␥-methylene groups are at a distance of ϳ4 Å from the indole ring of Trp-65, and the orientation relative to the indole ring is compatible with van der Waals interactions. Modeling of the active site in which Trp-65 is substituted in silico by Phe suggests that the ⑀-methyl of MetSO remains in interaction with the phenyl ring, whereas the ␥-methylene group is too distant from the phenyl ring to provide substantial van der Waals interactions (Fig. 1). In contrast, changing Trp-65 into Ala increases the volume of the active site pocket. The absence of a hydrophobic pocket as a result of this substitution excludes any possibility of stabilization of the side chain of Met-(R)-SO, i.e. via its ⑀-methyl and its ␥-methylene groups. Therefore, no MsrB-MetSO complex productive for catalysis is predicted to be formed. That is illustrated by the drastic Ϫ10 5 -fold decrease in k 2 observed for W65A MsrB, which includes drastic effects both on K S and k obs max . In contrast, more productive binary complexes can be formed with W65F MsrB because of the presence of a phenyl ring that could interact with at least the ⑀-methyl group. This is attested by the k 2 value that showed only a 22-fold decrease that includes both K S and k obs max contributions. Altogether, the results show that Trp-65 is essential for the catalytic efficiency of the reductase step by optimizing the position of MetSO within the active site relative to that of the catalytic amino acids.
As proposed for the catalytic mechanism of the reductase step for MsrA, formation of an Msr-substrate complex implies a concomitant deprotonation of the catalytic Cys via an acid/ base catalyst that in turn protonates the sulfoxide function and favors formation and rearrangement of the transition state of the sulfurane type, which leads to Met and sulfenic acid intermediate formation. As shown under "Results," the catalytic active MsrB form is the Cys-117 Ϫ /His-103 ϩ species with pK app of 6.6 and 8.3, respectively. A scenario for reduction of MetSO by MsrB can thus be proposed. Under reduced form, MsrB has its catalytic Cys-117 protonated with a pK app of ϳ9.3 and its His-103 under deprotonated form. Diffusion of the substrate into the active site leads to activation of Cys-117 and His-103 with transfer of the hydrogen of Cys-117 to the oxygen of the sulfoxide via His-103. Such a transfer is likely to be substrateassisted and permits stabilization of the thiolate form of Cys-117 via the positive or partially positive charge borne by the sulfur of the sulfoxide function, which is in close spatial proximity to the sulfur of Cys-117. Clearly, our results exclude any activation of Cys-117 through a catalytic triad formed by Cys-117, Arg-116, and Asp-107. The protonated form of the sulfox-ide function interacts by a hydrogen bonding network not only with His-103 via a strong hydrogen bond as suggested by the distance of 2.5 Å observed between the N-␦ of His-103 and one of the oxygens of cacodylate but also with His-100, Asn-119, and Thr-26 via a water molecule (Fig. 1). The catalytic mechanism is likely to be concerted and leads to nucleophilic attack of the thiolate of Cys-117 on the polarized form of the sulfoxide function with formation of a sulfurane-type transition state that is stabilized by His-103 and indirectly by His-100, Asn-119, and Thr-26 via a water molecule. This transition state then evolves irreversibly to form the sulfenic acid intermediate and with subsequent Met release. The low affinity of MsrB for Met (data not shown) likely participates in the irreversibility of the reduction process.
However, major points remain to be addressed. In particular, one question concerns the way by which the proton from Cys-117 is transferred to His-103. Indeed, the distance between Cys-117 (for one of the two conformations) and His-103 in the MsrB-MetSO complex is ϳ6.9 Å, considering that the MsrBcacodylate complex is representative of the MsrB-MetSO complex. This distance is too large for direct proton transfer from Cys-117 to His-103 unless a shortening of the distance occurs upon entry and the diffusion of MetSO up to the active site or via eventually the intervention of a water molecule. Another question concerns the nature of the sulfurane transition state and how it evolves toward formation of sulfenic acid intermediate and Met.
From an evolution point of view, it is interesting to compare MsrB to MsrA in terms of catalysis and of substrate recognition. In fact, the reduction mechanism is similar for both classes of Msrs, although the active site geometries and the amino acids involved in the catalysis are different except for the catalytic Cys. In MsrB, His-103 plays an acid/base catalyst role that likely involves Glu-94 in MsrA. Both residues should also stabilize the polarized form of the sulfoxide function and the transition state. A significant difference concerns the pH-k obs max profile, which is of monosigmoidal type for MsrA with assignment of the pK app in the ascending limb to the catalytic Cys-51, whereas apparently Glu-94 is not titrable. In contrast, in MsrB the curve displays a bell-shaped profile, and Cys-117 but also His-103 are titrable. Such a difference remains to be explained. Differences also concern the nature of the other amino acids that stabilize the polarized form of the sulfoxide function and the transition state, the nature of the amino acids that form the hydrophobic pocket, and the 16-fold decrease in k obs max of the reductase step 6 in MsrB when compared with MsrA. Finally, the structural/molecular factors that contribute to better binding of protein-bound MetSO rather than free MetSO remain to be characterized for both classes of Msr. 6 The catalytic efficiency of the reductase step of MsrB is low, i.e. 4.7 ϫ 10 2 M Ϫ1 ⅐s Ϫ1 compared with that of MsrA, which is 1.7 ϫ 10 4 M Ϫ1 ⅐s Ϫ1 at pH 8.0. It is because of a k obs max effect that likely results from differences in the nature of the amino acids involved in catalysis. In both cases, the affinity for the substrate is similar but low even with a substrate that mimics a protein MetSO-bound. This is the cause of the low catalytic efficiency of MsrB and even of MsrA. However, it is important to note that the rate-limiting step of both Msrs is associated with the Trx-recycling process. Therefore, saturation of the reductase step is not necessary to reach steady-state saturating conditions.