Evidence for a New Sub-class of Methionine Sulfoxide Reductases B with an Alternative Thioredoxin Recognition Signature*

Methionine sulfoxide reductases catalyze the reduction of protein-bound methionine sulfoxide back to methionine via a thioredoxin-recycling process. Two classes of methionine sulfoxide reductases, called MsrA and MsrB, exist that display opposite stereoselectivities toward the sulfoxide function. Although they are structurally unrelated, they share a similar chemical mechanism that includes three steps with 1) formation of a sulfenic acid intermediate with a concomitant release of 1 mol of methionine per mole of enzyme; 2) formation of an intradisulfide Msr bond; and 3) reduction of the oxidized Msr by thioredoxin. In the MsrBs that have been biochemically, enzymatically, and structurally characterized so far, the cysteine involved in the regeneration of the catalytic Cys-117 is Cys-63. Cys-117 is located on a β strand, whereas the recycling Cys-63 is on a loop near Cys-117. The distance between the two cysteines is compatible with formation of the Cys-117/Cys-63 intradisulfide bond. Analyses of MsrB sequences show that at least 37% of the MsrBs do not possess the recycling Cys-63. In the present study, it is shown that Cys-31 in the Xanthomonas campestris MsrB, which is located on another loop, can efficiently substitute for Cys-63. Such a result implies flexibility of the MsrB structures, at least of the loops on which Cys-31 or Cys-63 are located. The fact that about 25% of the putative MsrBs have no recycling cysteine supports other recycling processes in which thioredoxin is not operative.

Post-translation oxidation of methionine into methionine sulfoxide (MetSO) 1 in proteins is known to provoke loss of protein function and, in particular, to be involved in the aging process (for a review see Ref. 1). There exist methionine sulfoxide reductases (Msr) that catalyze reduction of free and peptide-bound methionine sulfoxide back to methionine and therefore restore the function of the modified MetSO proteins (2)(3)(4)(5)(6). Two structurally unrelated classes of monomeric Msrs have been described so far (7)(8)(9)(10). MrsAs reduce the S-epimer, whereas MsrBs reduce the R-epimer at the sulfur atom of MetSO (11)(12)(13)(14). Both classes display a similar new catalytic mechanism with at least three steps (13,15,16): 1) a reductase process that includes formation of a sulfenic acid intermediate and a concomitant release of 1 mol of methionine per mole of enzyme, the rate of which is not rate-limiting; 2) a second step that leads to formation of an intradisulfide intermediate with a rate that is at least as fast as that of the sulfenic acid intermediate formation; and 3) a third step that consists of reducing the oxidized Msr under a disulfide state by thioredoxin (Trx) (Scheme 1). This step was shown to be overall rate-limiting for both classes of Msrs (16,17).
Inspection of the alignment of the sequences of MsrBs deduced from the DNA sequences show that, as expected, the catalytic Cys-117 on which the sulfenic acid intermediate is formed, is invariant, including SelR MsrB in which Cys-117 is a Sec residue. In contrast, the recycling Cys-63 from Neisseria meningitidis which was shown to be involved in the regeneration of the Cys-117 through formation of an intradisulfide bond followed by reduction by Trx, is present in only 63% of the MsrB sequences (see Fig. 1). The absence of the recycling Cys-63 in the remaining MsrBs suggests an alternative mechanism of regeneration of the reductase activity for these MsrBs, in vivo. In the crystal structure of the MsrBs from Neisseria gonorrhoeae (9) and N. meningitidis, 2 Cys-117 is situated on a ␤ strand, whereas Cys-63 is located within a loop. The distance between the sulfur atoms of both cysteines is compatible with the catalytic mechanism and in particular with formation of the Cys-117/Cys-63 intradisulfide bond. Two hypotheses can be postulated to explain the activity of MsrBs devoid of the recycling Cys-63. In the first one, the sulfenic acid on Cys-117 is reduced by either Trx or another reductant. An alternative is the participation of another cysteine that would play a role similar to that of Cys-63. In this case, this implies a conformational flexibility of the MsrB structures and a positioning of the recycling cysteine compatible not only with an efficient formation of an intradisulfide bond with Cys-117 but also with an efficient reduction of the disulfide bond by Trx.
In the present study, we show that the Xanthomonas campestris MsrB, which does not possess Cys-63, exhibits an efficient Trx-recycling process. Cys-31, which is situated in a loop different from that on which is located Cys-63, is shown to play a role in the Trx-recycling process similar to that of Cys-63 of the N. meningitidis enzyme. Conversely, the C63S MsrB from N. meningitidis, in which a cysteine is introduced at position 31, is shown to be almost as active in the Trx-recycling process as the wild type. These results and those obtained with * This research was supported by the CNRS (Programme Protéomique et Génie des Protéines 2001), the University Henry Poincaré Nancy I, the ARC (Association pour la Recherche sur le Cancer Grant 5436), and the Institut Fédératif de Recherches 111 Bioingénierie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed.  (13). The other recombinant pSKMrsB plasmids were already described (13). Site-directed mutagenesis were performed using the QuikChange site-directed mutagenesis kit (Stratagene). The Escherichia coli strain used for X. campestris and N. meningitidis MsrB productions was BE002 (MG1655 msrA::spec⍀, msrB::␣3kana), transformed with the plasmidic construction pSKMsrB containing the coding sequence of the corresponding msrB gene under the lac promoter. The BE002 strain was kindly provided by Dr. F. Barras. Its use prevented expression of endogenous wild-type MsrB from E. coli. This is supported by the fact that overexpressed C117S MsrB did not display any detectable MsrB activity.
For MsrBs, excepted N. meningitidis S31C/C63S MsrB, the protocol of purification was carried out as previously described for PILB-MsrB (13). In the case of the N. meningitidis S31C/C63S MsrB, the enzyme was found after sonication and centrifugation in the pellet. The pellet was solubilized in a minimal volume of buffer A (50 mM Tris-HCl, 2 mM EDTA, pH 8.0) containing 20 mM dithiothreitol and 6 M urea. The enzyme was then renatured against a dialysis gradient buffer A containing decreasing concentrations of urea from 3 to 0 M. The solution was then applied onto an ACA 54 resin (buffer A). Purified fractions were pooled and applied onto a Q-Sepharose column equilibrated with buffer A, followed by a linear gradient of KCl (0 -0.4 M) using a fast protein liquid chromatography system (Amersham Biosciences). MsrB was eluted at 100 mM KCl. Purified fractions were pooled and finally applied onto a phenyl-Sepharose column (Amersham Biosciences) equilibrated with buffer A, containing 1 M (NH 4 ) 2 SO 4 . N. meningitidis S31C/ C63S MsrB was eluted with a linear gradient from 1 to 0 M (NH 4 ) 2 SO 4 in buffer A. At this stage, wild-type and mutant MsrBs were pure, as checked by electrophoresis on 12.5% SDS-PAGE gel followed by Coomassie Brilliant Blue R-250 staining. The predicted mass was confirmed by mass spectrometry analyses.
Storage of the enzymes was done as previously described (13). The molecular concentrations were determined spectrophotometrically, using an extinction coefficient at 280 nm deduced from the method of Scopes (18) of 9,710 M Ϫ1 ⅐cm Ϫ1 for X. campestris wild-type and C31S/ T63C MsrBs and 17,330 M Ϫ1 ⅐cm Ϫ1 for N. meningitidis wild-type, C63S, and S31C/C63S MsrBs. In the paper, N. meningitidis and X. campestris MsrB amino acid numbering are based on the E. coli MsrB sequence in which the N-terminal Met is omitted. Trx1 and Trx reductases from E. coli were prepared following experimental procedures already published (19,20).
Determination of Metal Content-The zinc and iron contents of wildtype X. campestris MsrB were analyzed using inductively coupled plasma-atomic emission spectrometry. In parallel, analyses of control buffer samples showed that zinc and iron concentrations were less than 72 g⅐liter Ϫ1 .
Quantification of the Free Cysteine Content with 5,5Ј-Dithiobis(2nitro)benzoate-Cysteine content of MsrBs and Trx was determined using DTNB under nondenaturing conditions in buffer A and denaturing conditions (final concentration of 1% SDS in buffer A) after incubation, or not, with 150 mM DL-Met-R,S-SO without the addition of any exogenous reducing system as previously described (13).
Characterization of the Sulfenic Acid Intermediate on the X. campestris C31S MsrB-The sulfenic acid intermediate was characterized spectrophotometrically by using thionitrobenzoate (TNB Ϫ ) under nondenaturing conditions. This approach took advantage of the unique reactivity of sulfenic acids toward nucleophiles such as TNB Ϫ . TNB Ϫ was prepared by reducing the corresponding disulfide using the procedure of Silver (21). The progress curve of TNB Ϫ disappearance for X. campestris C31S MsrB was recorded at 412 nm in buffer A. Enzyme concentrations were 7.35 and 14.7 M, and the TNB Ϫ concentration was 60 M. The amount of TNB Ϫ consumed was calculated using an extinction coefficient at 412 nm of 13,600 M Ϫ1 ⅐cm Ϫ1 .
Preparation of MsrBs and Trx in Oxidized Disulfide State-Oxidation of MsrBs was achieved by mixing 100 M reduced MsrB (MsrB red ) with 300 mM MetSO in buffer A. The MetSO used was DL-Met-R,S-SO of which only the R isomer is a substrate for MsrB. Oxidation of Trx was achieved by mixing 500 M reduced Trx (Trx red ) with 1 mM DTNB in buffer A. After 10 min of incubation at room temperature, the oxidized proteins were passed through an Econo-Pac 10 DG desalting column (Bio-Rad) equilibrated with buffer A. The state of oxidation was checked by titration with DTNB.
Preparation of Ac-L-Met-R,S-SO-NHMe-Ac-L-Met-R,S-SO-NHMe was prepared from Ac-L-Met-OMe (Bachem) as previously described (17). The proton NMR spectrum is in accordance with the Ac-L-Met-R,S-SO-NHMe structure: 1

CO).
Enzyme Kinetics-Kinetic parameters were determined under steady-state conditions with the Trx recycling system (1.28 M E. coli Trx reductase and 0.3 mM NADPH), with saturating concentration of Ac-L-Met-R,S-SO-NHMe and by varying the concentrations of Trx red , or with saturating concentration of Trx red and by varying the concentrations of Ac-L-Met-R,S-SO-NHMe. Initial rate measurements were carried out at 25°C in buffer A 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 fit to the Michaelis-Menten relationship using least squares analysis to determine k cat value and K m for Trx red and Ac-L-Met-R,S-SO-NHMe. It should be mentioned that K m for Ac-L-Met-R,S-SO-NHMe has to be divided by 2, because the S isomer is not a substrate.
Fluorescence Properties of Wild-type and C31S MsrBs from X. campestris and Trx-The fluorescence characteristics of 1) the wildtype MsrB in its reduced form and Cys-117/Cys-31 disulfide state, 2) the C31S MsrB in its reduced form and Cys-117 sulfenic acid state, and 3) Trx in its reduced and disulfide state were recorded on a spectrofluorometer (flx SAFAS) thermostatted at 25°C (see legends of Figs. 2 and 3 for more details).

Kinetics of the Formation of the Cys-117/Cys-31 and Cys-117/Cys-63 Disulfide Bonds of Wild-type and C31S/T63C MsrBs from X. campestris, Respectively, in the Absence of Trx by Single Turnover Stopped-flow
Experiment-Kinetics of the emission fluorescence intensity increase associated with the formation of the Cys-117/Cys-31 and Cys-117/ Cys-63 disulfide bonds of the MsrBs from X. campestris were measured at 25°C on an SX18MV-R stopped-flow apparatus (Applied PhotoPhysics) fitted for fluorescence measurements. The excitation wavelength was set at 293 nm, and the emitted light was collected above 320 nm, using a cutoff filter. One syringe contained MsrB in buffer A (10 M final concentration after mixing), and the other one contained Ac-L-Met-R,S-SO-NHMe (100 mM final concentration). An average of six runs was recorded. Rate constants, k obs , were obtained by fitting fluorescence traces with the monoexponential Equation 1 in which c represents the end point, a, the amplitude of the fluorescence increase (Ͻ0) and k obs the rate constant.
Determination of the Rate of Ac-L-Met-NHMe Formation by Single Turnover Quenched-flow Experiments-Quenched-flow measurements SCHEME 1. Proposed catalytic mechanism of Msrs. The three-step catalytic mechanism as described by Olry et al. (13) is shown. For MsrB from N. meningitidis, X represents Cys-117 and Y represents Cys-63 or Cys-31.
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 previously described (16). Briefly, equal volumes (60 l) of a solution containing 600 M MsrB in 50 mM MES, pH 5.5 (for the N. meningitidis enzyme), or in buffer A (for the X. campestris enzymes) and a solution containing 700 mM Ac-L-Met-R,S-SO-NHMe in 50 mM MES, pH 5.5 (for the N. meningitidis enzyme), or 200 mM Ac-L-Met-R,S-SO-NHMe in buffer A (for the X. campestris enzymes) were mixed in the aging loop. The mixtures were then allowed to react for 40 -5000 ms before being mixed with an equal volume of a quenching 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. Trifluoroacetic acid (100%, 50 l) was then added to 200 l of the quenched samples to precipitate the protein. Samples were centrifuged at 12,000 ϫ g for 30 min at room temperature. Ac-L-Met-NHMe quantification in the resulting supernatant was carried out by reverse phase chromatography as described previously (17). In parallel, the quenched samples, which were not treated with 100% trifluoroacetic acid, were used to determine the protein concentration from the absorbance at 280 nm.
For Ac-L-Met-NHMe titration, data were plotted as moles of Ac-L-Met-NHMe formed per mole of MsrB as a function of time. The rate of Ac-L-Met-NHMe formation was determined by fitting the curve to the monoexponential Equation 2 in which a represents the maximum fraction of Ac-L-Met-NHMe formed per mole of MsrB and k obs , the rate constant.

Kinetics of Formation of Oxidized Trx (Trx ox ) upon the Reduction of the X. campestris MsrB Cys-117/Cys-31 and Cys-117/Cys-63 Disulfide Bonds and of the N. meningitidis MsrB Cys-117/Cys-31 Disulfide Bond
under Single Turnover Conditions-Two approaches were used depending on the rate of the kinetics. When the rate was rapid as observed with the X. campestris wild-type and N. meningitidis S31C/C63S MsrB, kinetic measurements were determined on a stopped-flow apparatus by following selectively the quenching of the fluorescence emission of Trx associated with the formation of its Cys-32/Cys-35 disulfide bond upon the reduction of oxidized MsrB (MsrB ox ), as previously described for MsrA (16) and more recently for the MsrB from N. meningitidis (17). Briefly, the excitation wavelength was set at 310 nm, and the emitted light was collected above 320 nm, using a cutoff filter. One syringe contained the MsrB ox in buffer A, and the other one contained Trx red at various concentrations in buffer A. An average of six runs was recorded for each concentration of Trx red .
When the rate was low as observed with the X. campestris C31S/ T63C MsrB, the Trx fluorescence quenching associated with formation of its Cys-32/Cys-35 disulfide bond upon reduction of the MsrB Cys-117/ Cys-63 disulfide bond was carried out at 25°C on a spectrofluorometer (flx SAFAS). The excitation wavelength was set at 310 nm, and the emitted light was collected at 340 nm with an adapted voltage.
In both approaches, rate constants, k obs , were obtained by fitting fluorescence traces with the monoexponential Equation 1 with a Ͼ 0. k obs values were fit to Equation 3 using least squares analysis to determine k r , k f , and K for Trx red . S represents the Trx red concentration, K represents the K s values for Trx red , and k r and k f represent the rate constants of the disulfide exchange in the reverse and forward directions, respectively.
was used, assuming that binding of Trx red to MsrB ox is rapid equilibrium.

RESULTS
Biochemical Properties of Wild-type, C31S/T63C, and C31S MsrBs from X. campestris and of S31C/C63S MsrB from N. meningitidis-DTNB thiol titrations on the X. campestris wild-type MsrB red revealed six cysteines under denaturating conditions. These results are in agreement with the X. campestris ATCC 13951 DNA sequence that indicates six cysteines at positions 31, 45, 48, 94, 97, and 117 (Fig. 1). Under native conditions, only two cysteines were reactive (Table I). Metal analyses of the purified enzyme by atomic emission spectrometry showed the presence of zinc in a stoichiometry of 0.9 mol per mole of enzyme. The presence of iron, less than 0.05 mol per mole of enzyme was also detected. As shown from the inspection of the x-ray structure, a putative Zn 2ϩ -binding site is present in the N. gonorrhoeae MsrB (9). Assuming a fold similar to that of the N. gonorrhoeae MsrB, the metal binding site of the X. campestris MsrB is composed of the two CXXC motifs with Cys-45, Cys-48, Cys-94, and Cys-97 and is situated in an opposite direction from the active site. Therefore, the accessible cysteines in the X. campestris MsrB are likely Cys-31 and Cys-117, the latter being the catalytic cysteine on which the sulfenic acid is formed. Also two cysteines were reactive in the C31S/T63C MsrB red indicating that the Cys-63 is accessible to DTNB. In the case of the C31S MsrB red , only one cysteine, Cys-117, is shown to be accessible (Table I). Finally, two cysteines are also reactive in the S31C/C63S MsrB red from N. meningitidis. This showed that the introduced Cys-31 is also accessible, as is the case for Cys-63 in the wild type (13).
Enzymatic Properties of the Wild-type and C31S MsrBs from X. campestris-The activity of the wild type was tested with Ac-L-Met-R,S-SO-NHMe as a substrate to evaluate the kinetics of the reductase step at saturating concentration of the substrate. Surprisingly, a turnover activity was observed in the presence of Trx red and Trx reductase. This raised the question of the Msr intermediate entities reduced by Trx red . A possibility was a reduction of the sulfenic acid intermediate by Trx red . An alternative was the reduction of a disulfide bond. The latter possibility implied that Cys-31 forms a disulfide bond with Cys-117.
To better characterize the catalytic mechanism, the kinetic constants were first determined under steady-state conditions using the Trx reductase recycling system at pH 8.0. K m for Ac-L-Met-R,S-SO-NHMe and Trx red , and k cat values were 6 mM, 5.7 M, and 2.5 s Ϫ1 , respectively (Table II). Some significant differences are observed from what was published for the N. meningitidis MsrB (13). In particular, K m for Trx red is 6-fold lower, and the k cat value is increased by a factor of 5. The rate of Ac-L-Met-NHMe formation was then studied under single turnover conditions in the absence of Trx using the same approach as already described for MsrA and MsrB from N. meningitidis (16,17). The study was done at 100 mM Ac-L-Met-R,S-SO-NHMe at pH 8.0. A k obs value of 8 s Ϫ1 and a stoichiometry of 0.8 mol of Ac-L-Met-NHMe per mole of enzyme were determined (Table III).
In parallel, thiol titration was done by DTNB. A loss of two cysteines was observed upon reduction of DL-Met-R,S-SO (Table I). This result strongly supported the formation of a Cys-117/Cys-31 disulfide bond but did not give any information on the rate of its formation. In that context, it could not be excluded that the sulfenic acid intermediate is reduced by Trx red and that its reduction is kinetically favored against the disulfide bond reduction. To address this question, the activity of C31S MsrB was studied in the presence of Trx red . First, no Trx-recycling reductase activity was detected, and a stoichiometry of 1.0 mol of Ac-L-Met-NHMe formed per mole of enzyme was observed (Table III). Second, formation of the sulfenic acid intermediate was proved by its titration by TNB Ϫ , with a stoichiometry of 0.8 per mole of enzyme. Third, the rate of Ac-L-Met-NHMe formation, which is a means to attain the rate of formation of the sulfenic acid intermediate, was determined using a stopped-flow apparatus adapted in a quenched-flow mode (17) and was shown to be similar to that measured with the wild type (Table III). Therefore, Trx red likely reduces only the disulfide intermediate in the wild type.
The next step was to determine the rate of formation of the Cys-117/Cys-31 disulfide bond in the wild type by following the increase of the emission fluorescence intensity upon going from the reduced to the oxidized disulfide form using the same approach as that described for MsrA from N. meningitidis (17). The interpretation of the fluorescence message was not ambiguous, because the fluorescence message due to formation of the sulfenic acid intermediate, which can be attainable through the C31S MsrB, decreased and was moreover insignificant compared with the increase observed upon the disulfide formation ( Fig. 2). At pH 8.0 a k obs value of 11 s Ϫ1 was found that is in the range of the k obs value of 8 s Ϫ1 determined for Ac-L-Met-NHMe formation (at a concentration of 100 mM Ac-L-Met-R,S-SO-NHMe, which is saturating; data not shown) (Table III). Therefore, similar to that found for MsrA and MsrB from N. meningitidis, the sulfenic acid intermediate does not accumulate. As soon as it is formed on Cys-117, it is attacked by Cys-31. Together, the data obtained for the wild-type MsrB from X. campestris supported: 1) a rate-determining step associated with formation of the sulfenic acid intermediate within the two-step process leading to formation of the Cys-117/Cys-31 disulfide bond and 2) an overall rate-limiting step associated with the Trx-recycling process. To better characterize the ratelimiting step, the rate of formation of Trx ox upon reduction of MsrB ox was evaluated by following selectively the quenching of the fluorescence emission intensity of Trx upon going from the reduced to the oxidized form under single turnover conditions as already done for the N. meningitidis MsrB (17) (Table II). This is only possible provided that the change in the fluorescence intensity of MsrB upon going from the oxidized to the reduced forms does not significantly interfere with the Trx message (Fig. 3). As shown in Fig. 3, when the excitation was done at 310 nm, the contribution of MsrB to the quenching of the fluorescence signal is indeed small compared with that of Trx. For each concentration of Trx red , a k obs value was determined. Assuming that binding of Trx red to MsrB ox is at a rapid equilibrium and using Equation 3, a K s value of 30 M for Trx red and k f and k r values of 13.8 s Ϫ1 and ϳ 0 s Ϫ1 can be determined from the curve k obs versus Trx red concentration, respectively (Table II). The fact that the curve passes through the origin indicates that the recycling process going up to    Enzymatic Properties of the C31S/T63C MsrB from X. campestris-In the presence of Trx red , a recycling activity was observed but with a 1000-fold decrease in the k cat compared with the wild type (Table II). This value is, however, significant because no turnover activity was observed in the absence of Trx and the activity is proportional to enzyme concentrations. Due to the high concentration of MsrB needed to measure activity, the K m value for Trx red was not attainable. In the absence of Trx, under single turnover conditions, the rate of Ac-L-Met-NHMe formation was studied using the same approach as already described for MsrA and MsrB from N. meningitidis (16,17). The study was done at 100 mM Ac-L-Met-R,S-SO-NHMe in buffer A at pH 8.0. A k obs value of 1.1 s Ϫ1 and a stoichiometry of 1.0 mol of Ac-L-Met-NHMe per mole of enzyme were determined (Table III). The interpretation of the fluorescence message was not ambiguous because the fluorescence message due to formation of the sulfenic acid intermediate, which can be attainable through the C31S MsrB, decreased and was moreover insignificant compared with the increase observed upon Cys-117/Cys-63 disulfide bond formation (spectra not shown). A k obs value of 1.9 s Ϫ1 was found that is in the range of the k obs value of 1.1 s Ϫ1 determined for Ac-L-Met-NHMe formation (at a concentration of 100 mM Ac-L-Met-R,S-SO-NHMe concentration for the two experiments) (Table III). Therefore, the data demonstrated that the rate-limiting step in the C31S/T63C MsrB is again associated with the Trx-recycling process as in the wild type. The fact that the rate of formation of Ac-L-Met-NHMe is 7-fold decreased compared with the wild type remains to be explained.
The kinetics of the Trx-recycling process was then studied under single turnover conditions. Again, the rate was determined by following the rate associated with a change of the Trx fluorescence intensity upon going from the reduced to the oxidized forms during the reduction of the Cys-117/Cys-63 MsrB disulfide bond as already described for MsrA and MsrB from N. meningitidis (16,17). When the excitation was done at 310 nm, the contribution of the MsrB to the quenching of the fluorescence signal is indeed small compared with that of Trx (spectra not shown). From the curve of k obs versus Trx red concentration, assuming binding of Trx red to MsrB ox is in rapid equilibrium and using Equation 3, a K s value of 40 M for Trx red and k f and k r values of 1.9 ϫ 10 Ϫ3 s Ϫ1 and ϳ0 s Ϫ1 can be determined, respectively (Table II). The value of k f is in the same range as the k cat value. Therefore, these results strongly support a rate-limiting step in C31S/T63C MsrB that either precedes or is concomitant to formation of Trx ox .
Properties of S31C/C63S MsrB from N. meningitidis-Catalytic constants of S31C/C63S N. meningitidis MsrB determined under steady-state conditions were obtained with Ac-L-Met-R,S-SO-NHMe as a substrate. For Ac-L-Met-R,S-SO-NHMe K m , Trx red , and k cat values were 0.5 mM, 13 M, and 0.12 s Ϫ1 , respectively, at pH 8.0 (Table II).
Kinetic characterizations of the rate of formation of Ac-L-Met-NHMe and of the reduction of the Cys-117/Cys-31 disulfide bond by Trx red were then done by using the same methods as those described for the wild type. For the rate of Ac-L-Met-NHMe formation at pH 8.0, a burst of Ac-L-Met-NHMe formation was observed with a stoichiometry of 0.9 mol of Ac-L-Met-NHMe per mole of MsrB red . However, the rate was too fast to be determined with the apparatus adapted for quenched-flow experiments. In contrast, at pH 5.5, a k obs value of 12 s Ϫ1 was attainable with a stoichiometry of 0.9 mol of Ac-L-Met-NHMe per mole of enzyme (Table III). Therefore, the data demonstrated that the rate-limiting step in the S31C/C63S MsrB takes place after formation of the sulfenic acid intermediate and is probably associated with the Trx-recycling process as in the wild type. The kinetics of this process was then studied under single turnover conditions using the Trx ox fluorescence message as already described for the wild type. When the excitation was done at 310 nm, the contribution of MsrB to the quenching of the fluorescence signal was indeed small compared with that of Trx (spectra not shown). Assuming binding of Trx red to MsrB ox is in rapid equilibrium and using Equation 3, a K s value of 390 M for Trx red and k f and k r values of 40 s Ϫ1 and ϳ 0 s Ϫ1 can be determined from the curve k obs versus Trx red concentration, respectively. The fact that the values of k f and of K s for Trx red are 330-fold and 30-fold higher than those determined under steady-state conditions, respectively, strongly suggests that the rate-limiting step is associated with release of Trx ox similar to that proposed for the wild-type enzyme (17). DISCUSSION In a previous study carried out on the N. meningitidis MsrB, the recycling reductase process was shown to be Trx-dependent via formation of a disulfide bond between Cys-117 and Cys-63 (13). In the present study, we show that Cys-31 in the X. campestris MsrB plays a role similar to that of the Cys-63 in the N. meningitidis MsrB. Several data argue in favor of this interpretation. First, Trx-dependent recycling activity was observed for the X. campestris MsrB, which does not possess a cysteine at position 63. Second, a loss of two cysteines was observed upon the reductase step in the absence of Trx. Third, when Ser was substituted for Cys-31, X. campestris C31S MsrB was as efficient in the reductase step as the wild type but displayed no recycling activity with Trx red . Fourth, S31C/C63S MsrB from N. meningitidis showed recycling Trx-dependent activity similar to the wild type, in contrast to the C63S MsrB, which did not show any recycling activity with Trx red (13).
As for the N. meningitidis MsrB, the rate-limiting step in the X. campestris enzyme was associated with the Trx-recycling process. The fact that Cys-31 can efficiently substitute for Cys-63 in the recycling process raises the question of how the Cys-117/Cys-31 disulfide bond of the X. campestris MsrB is efficiently formed and reduced by Trx red .
Inspection of the x-ray structures of the N. meningitidis 2 and N. gonorrhoeae (9) MsrBs shows a core domain composed of two anti-parallel ␤ sheets from strands ␤1, ␤2, ␤9, and ␤3-␤7 decorated by three ␣-helices. The active site that contains Cys-117 and Cys-63 is on the surface of the protein. Cys-117 is situated on the ␤8 strand. Its sulfur atom is positioned at least 11.3 Å apart the oxygen atom of Ser-31, whereas it is at 3.2 Å from the sulfur atom of Cys-63 and thus is in a favorable position to form the Cys-117/Cys-63 disulfide bond. The two positions 63 and 31 are situated into two distinct loops. Therefore, two hypotheses could be postulated to explain the formation of the Cys-117/Cys-31 disulfide bond. In the first one, the two Msrs belong to two distinct structural sub-classes. In that case, each loop would be located differently in the two MsrBs such that the distance between Cys-117 and the cysteine involved in the recycling process is compatible with formation of the corresponding disulfide bond. The fact that the loop bearing Cys-63 in the N. gonorrhoeae and N. meningitidis MsrBs is well positioned in the crystal structure to form a Cys-117/Cys-63 disulfide bond rather favors this interpretation. An alternative explanation is that the flexibility of the MsrB structures generates multiple conformers of the two loops bearing Cys-63 or Cys-31 such that Cys-63 or Cys-31 can a priori approach Cys-117 and then form with a similar efficiency a disulfide bond with Cys-117. The fact that S31C/C63S N. meningitidis MsrB exhibits a k cat value with Trx red similar to that of the wild type rather favors the second hypothesis. In that context, the knowledge of the x-ray structure of the X. campestris MsrB, which is under determination, will be of particular interest. Character-ization of the disulfide bond formed, i.e. Cys-117/Cys-63 or Cys-117/Cys-31 or both, in a N. meningitidis MsrB in which a cysteine would be introduced at position 31 will also be informative.
The facts that the Trx-recycling process is efficient in the N. meningitidis MsrB and X. campestris MsrB as well and is likely rate-limited by Trx ox release support productive interactions between MsrB ox and Trx red , thus permitting efficient two-electron chemical exchange. In the case of the C31S/T63C X. campestris MsrB, the rate of formation of Trx ox is 7000-fold decreased compared with the wild type and becomes ratelimiting in the Trx-recycling process. Although the K s value for Trx red is similar to that of the wild type, it is possible that the drastic decrease in k cat is due to nonproductive interactions between MsrB ox and Trx red , thus preventing an efficient twoelectron chemical exchange. This can be due to differences in the amino acid environment near the introduced Cys-63 compared with that near Cys-31. Another cause could have been the presence of the Zn 2ϩ binding site in the X. campestris MsrB. About 50% of the MsrBs, including the X. campestris, E. coli, and Drosophila (22) MsrBs, contain the two CXXC motifs involved in the binding of Zn 2ϩ , whereas the others such as the N. meningitidis and N. gonorrhoeae MsrBs do not possess Cys-45, Cys-48, Cys-94, and Cys-97 and thereby have no Zn 2ϩ atom bound. This is, however, unlikely because recent studies done on the E. coli MsrB, which also possesses Cys-63 and a metal-binding site, have shown that T31C/C63S E. coli MsrB exhibits an efficient Trx-recycling activity similar to the wild type. 3 As deduced from Fig. 1, 32% of the MsrB amino acid sequences neither possess Cys-63 nor Cys-31. Therefore, this raises the question as to how these MsrBs regenerate their Cys-117 when it is oxidized in the sulfenic acid form. Sequence alignments show that some MsrBs have a cysteine at position 60 ( Fig. 1). This position is situated on the same loop within which Cys-63 is located, and thus Cys-60 could eventually play a role similar to that of Cys-63. This has been confirmed by recent studies that show that introducing Cys-60 in the C63S N. meningitidis MsrB also generates a MsrB with Trx-recycling activity, with a k cat value in the same range as that of the k cat of the wild type. 3 Again, this argues for a greater flexibility of the MsrB structures. Anyway, 25% of the putative MsrBs have no cysteine that can substitute for Cys-63, Cys-60, or Cys-31. This is the case for Rhodobacter capsulatus MsrB, for which the sulfenic acid intermediate formed on the Cys-117 is not reduced in vitro by Trx red . 3 Assuming that the other putative MsrBs behave similarly to the R. capsulatus MsrB and are expressed and active in vivo, this supports other recycling processes in which Trx is not involved. Their biochemical characterizations remain to be done.