Kinetics and Mechanism of the Oxidation of the Glutathione Dimer

Time-resolved spectra were fitted using SPECFIT 3.0.36 global analysis software (Spectrum Associates, Chapel Hill, NC). All data points represent the ...
12 downloads 19 Views 155KB Size
Download PDF
Chem. Res. Toxicol. 2007, 20, 79-87

79

Kinetics and Mechanism of the Oxidation of the Glutathione Dimer by Hypochlorous Acid and Catalytic Reduction of the Chloroamine Product by Glutathione Reductase Pe´ter Nagy and Michael T. Ashby* Department of Chemistry and Biochemistry, UniVersity of Oklahoma, Norman, Oklahoma 73019 ReceiVed August 9, 2006

Oxidized glutathione (GSSG) reacts with two molar equivalents of HOCl/OCl- (a neutrophil-derived oxidant and a common biocide) to form the dichloro (bis-N-chloro-γ-L-glutamyl) derivative (NDG). The reaction of less than two molar equivalents of HOCl with GSSG does not yield the unsymmetrical monochloro derivative (NCG) but rather a stoichiometric amount of NDG and GSSG. This result is explained by a faster reaction of the second equivalent of HOCl with NCG than that of the first equivalent of HOCl with GSSG. The rates of reaction of GSSG2-, GSSG3-, and GSSG4- (successive deprotonation of the ammonium groups) have been investigated, and it is clear that GSSG2- is unreactive, whereas GSSG4- is about twice as reactive as GSSG3-. Accordingly, the following mechanism is proposed (constants for 5 °C): H+ + OCl- ) HOCl, pK1 ) -7.47; GSSG2- ) GSSG3- + H+, pK2 ) 8.5; GSSG3) GSSG4- + H+, pK3 ) 9.5; GSSG3- + HOCl f NCG3- + H2O, k4 ) 2.7(2) × 106 M-1 s-1; GSSG4+ HOCl f NCG4- + H2O, k5 ) 3.5(3) × 107 M-1 s-1; NCG3- f NDG4- + H+, k6 ) fast; and NCG4+ HOCl f NDG4- + H2O, k7 ) fast. At physiologic pH, the k4 pathway dominates. NDG decomposes at pH 7.4 in a first-order process with kdec ) 4.22(1) × 10-4 s-1 (t1/2 ) 27 min). Glutathione reductase (EC 1.6.4.2) is capable of catalyzing the reduction of NDG by NADPH. The only NDG-derived product that is observed (by NMR) after the reduction by NADPH is GSH. Thus, in the presence of the GOR/ NADPH system, GSH is capable of redox buffering a 3/2 mol equiv of HOCl rather than a 1/2 mol equiv as previously assumed. Introduction Glutathione plays an essential role in the health of many types of organisms, particularly those that survive in an aerobic atmosphere. Glutathione is the predominant non-protein thiol in eukaryotes (1) and some prokaryotes (2, 3), and millimolar concentrations are found in some cells. The term glutathione is collectively used to refer to the tripeptide L-γ-glutamyl-Lcysteinylglycine in both its reduced (GSH)1 and oxidized dimeric (GSSG) forms. The biological activity of GSH is generally identified with the sulfhydryl group of the cysteine (CySH) residue, although this contribution will draw the reader’s attention to the terminal R-amino moiety of the glutamate residue. In contrast to the benevolent properties of GSH, the parent amino acid CySH is known to be cytotoxic in high concentrations, in part because it drives Fenton chemistry (4-6). Furthermore, the oxidized disulfide form, cystine, is insoluble at physiological pH, which is a property that accounts for a significant percentage of all urinary stones that are caused by genetic diseases (7). Thus, millimolar concentrations of GSH provide the desirable reaction * To whom correspondence should be addressed. Fax: 405-325-6111. E-mail: [email protected]. 1 1Abbreviations: GSH, reduced glutathione; CySH, cysteine; GSSG, oxidized glutathione; ROS, reactive oxygen species; NCC, N-chlorocystine; NDC, N,N′-dichlorocystine; NCG, the N-chloro-γ-L-glutamyl derivative of GSSG; NDG, the bis-N-chloro-γ-L-glutamyl derivative of GSSG; NADPH, β-nicotinamide-adenine-dinucleotide-phosphoric acid (reduced form); GOR, glutathione reductase (EC 1.6.4.2); NMR, nuclear magnetic resonance; SVD, singular value decomposition; EDTA, ethylenediamine-N,N,N′,N′-tetraacetic acid; WT, wild-type; LD50, median dose of a toxic substance required to kill half the members of a tested population; MIC, minimum inhibitory concentration.

properties of the sulfhydryl moiety (Vide infra) while avoiding the deleterious properties of CySH. Although GSH also plays critical roles in catalysis (8), metabolism (9), signal transduction (10), gene expression (11), apoptosis (12), and the detoxification of xenobiotic compounds (13), we focus here on its antioxidant properties (14). GSH is the principal intracellular non-protein thiol in most cells, and as such, it serves to maintenance intracellular redox states. It is a nucleophilic scavenger and an electron donor via the sulfhydryl group of its CySH residue. GSH is a cofactor for glutathione reductase, which is an antioxidant enzyme that maintains glutathione in its reduced state (15). The redox buffering capability of GSH is achieved via direct reaction with reactive oxygen species (ROS) as well as through the indirect maintenance of other antioxidants. For example, GSH is an efficient scavenger of many radical and non-radical ROS (16). Additionally, GSH is involved in the regeneration of the antioxidant

10.1021/tx060184g CCC: $37.00 © 2007 American Chemical Society Published on Web 12/22/2006

80 Chem. Res. Toxicol., Vol. 20, No. 1, 2007

ascorbate from its oxidized form, which is accomplished via GSH-dependent dehydroascorbate reductase (17). Because of the connection that exists between oxidative stress and inflammatory disease, there is considerable interest in the pecking order of antioxidants (18). Until recently, most of the attention has focused on the relative reactivity of antioxidants toward 1e-oxidants (18), but there is mounting anecdotal evidence that 2e-oxidants such as the neutrophil-derived ROS hypochlorous acid (HOCl) may play significant roles in inflammatory disease (19). Both inorganic and organic reductants may be involved in reactions with HOCl in ViVo. Pattison and Davies have investigated the rates of the reactions of HOCl with the organic components of proteins and lipids (20-22). For the organic moieties that have been studied thus far, the sulfurcontaining amino acids CySH and methionine (Met) have been found to be the most reactive: Met > Cys . cystine ∼ His ∼ R-amino > Trp > Lys . Tyr ∼ Arg > backbone amides > Gln ∼ Asn (20). In addition to these organic compounds, we have noted that the endogenous inorganic ion thiocyanate (SCN-) is as reactive toward HOCl as CySH and Met (23, 24). The reaction between HOCl and SCN- produces hypothiocyanite (OSCN-), a species that is not lethal to mammalian cells (23). Although some attention has been focused on measuring the relative rates of the reactions of ROS and antioxidants, it is not clear for most of these reactions if long-lived reactive intermediates are generated (secondary ROS), and in some cases, the products are not known. Furthermore, ambiguities exist for the rate data that in some cases preclude a prediction of the site(s) of reaction. For example, the disulfide and R-amino moieties of model compounds have exhibited similar rate constants for their reactions with HOCl, but we recently reported that the reaction of cystine with HOCl cleanly produces N,N′dichlorocystine (NDC) as the first observable product and that NDC exhibits a substantial lifetime at physiologic pH (25). Other largely unexplored variables that may affect absolute rates are the steric and environmental effects that are introduced as polypeptide bonds form between the amino acids. In the present study, we have examined the kinetics and mechanism of the reaction of GSSG with HOCl. Such a reaction could prove to be important during near-lethal oxidative stress by HOCl, where the redox buffering capacity of a cell is depleted as GSH is oxidized to GSSG. In consideration of our previous observation for cystine (25), it is perhaps not surprising that we find that the dichloro (bis-N-chloro-γ-L-glutamyl) derivative (NDG) is formed when GSSG is reacted with HOCl. What is surprising is the fact that NDG is long-lived at physiologic pH (t1/2 ≈ 27 min) and that it is catalytically reduced by glutathione reductase (GOR, EC 1.6.4.2), in the presence of NADPH, to produce GSH. Following a discussion of the kinetics and mechanisms of the aforementioned reactions, we will address the possible relevance of NDG in the context of cell defense during oxidative stress by HOCl.

Materials and Methods Materials. Water was doubly distilled in glass. GSSG, deuterium chloride (35 wt %) solution in D2O, NaClO4, and K3PO4 were used as received from Sigma-Aldrich. NaH2PO4‚H2O, Na2HPO4, and Na3PO4‚12 H2O were used as received from Mallinckrodt. Deuterium oxide (99.9%) was obtained from Cambridge Isotope Laboratories. The tetrasodium salt of β-nicotinamide adenine dinucleotide phosphate (NADPH, reduced form) and the glutathione reductase from baker’s yeast (as a suspension in 3.6 M (NH4)2SO4 at pH 7.0, containing 0.1 mM dithiothreitol) were obtained from SigmaAldrich. The enzyme was stored at T ) 4 °C, and NADPH was stored at T ) -20 °C until needed. The concentrations of the stock

Nagy and Ashby NADPH solutions were spectrophotometrically determined under alkaline conditions (pH > 11) using (NADPH)340 nm ) 6.22 mM-1 cm-1. The buffer solutions were prepared from the solids NaH2PO4‚H2O, Na2HPO4, and Na3PO4‚12 H2O; the ionic strength was adjusted with NaClO4, and the pH was adjusted with NaOH or HClO4. Stock solutions of NaOCl were prepared by sparging Cl2 into a 0.3 M solution of NaOH. The sparging was stopped when [OCl-] was ca. 100 mM, as determined spectrophometrically ( (OCl-)292 nm ) 350 M-1 cm-1). The concentrations of stock solutions of OCl- were determined iodometrically. The concentration of OCl- in solutions that were prepared from the titrated stock solutions were confirmed spectrophotometrically ((OCl-)292 nm ) 350 M-1 cm-1). pH/pD Measurements. pH measurements were made with an Orion Ion Analyzer EA920 using a Ag/AgCl combination pH electrode. All pH measurements were corrected for the “Irving factor” of the working medium (26), and a temperature correction was applied. The electrode response was corrected for measurements under alkaline conditions with a gradient of standard solutions that were individually titrated. A value of pKw of 13.79 was used for the [OH-] or [H+] calculations according to Martell and Smith (27). pD measurements in D2O were made using the same pH electrode by adding 0.4 units to the measurement. Stopped-Flow Studies. Kinetic measurements were made with a HI-TECH SF-61 DX2 stopped-flow spectrophotometer using a Xe arc lamp. Temperature control of the observation cell was achieved with a Lauda RC-20 circulator. The monochromatic kinetic traces were fitted with HI-TECH KinetAsyst 3.14 software (HiTech, U.K.). Time-resolved spectra were fitted using SPECFIT 3.0.36 global analysis software (Spectrum Associates, Chapel Hill, NC). All data points represent the average of at least nine mixing cycles. The GSSG and H+ dependencies of the rate law were determined under pseudo-first-order conditions at λ ) 250 and 300 nm. Time-resolved UV-vis spectra were assembled from individual kinetic traces that were measured for the stoichiometric reaction of HOCl and GSSG in the λ ) 230-350 nm range in 5 nm increments. For the latter conditions ([HOCl] ) 2[GSSG]), the kinetic traces exhibited second-order kinetics. NMR Studies. 1H NMR spectra were recorded with a Varian XL-300 spectrometer at 20 °C. Deuterated buffers were prepared from D2O solutions of anhydrous K3PO4 by adding DCl. The chemical shifts (ppm) were referenced to sodium 3-(trimethylsilyl)1-propanesulfonate (δ ) 0.015 ppm). Turbulent mixing of the reagents was necessary to ensure homogeneity of reaction mixtures in the time frame of the chemical reaction, and this was achieved for the NMR studies by employing a hand mixer comprising two Hamilton syringes and a T-mixer. Failure to quickly mix solutions of HOCl and GSSG produced different, irreproducible results. UV-vis Spectroscopy. Electronic spectra were measured using a HP 8452A diode array spectrophotometer or the monochromator of the HI-TECH SF-61 DX2 stopped-flow instrument with a Xe arc lamp. Assay for the GOR Reduction of GSSG. The activity of glutathione reductase was assayed using a published procedure (28). The depletion of NADPH was followed at λ ) 340 nm with and without added EDTA (0.5 mM). Conditions for a typical assay solution: [iP] ) 0.1 M, [NADPH] ) 0.1 mM, [GOR] ) 0.025 U/mL ∼ 1.3 nM, [GSSG] ) 1 mM, [EDTA] ) 0.5 mM, pH 7.56, and T ) 25 °C. The reaction was initiated by the addition of GSSG. Unit definition: One unit will reduce 1.0 µmol of oxidized glutathione per min at pH 7.6 at 25 °C. Assay for the GOR Reduction of NDG. The same method that was used to assay for the GOR reduction of GSSG was also used to assay for the GOR reduction of NDG. Conditions for a typical assay solution: [iP] ) 0.1M, [NADPH] ) 0.1 mM, [GOR] ) 0.025 U/mL ∼ 1.3 nM, [NDG] ) 0.5 mM, pH 7.56, and T ) 25 °C. The reaction was initiated by the addition of NDG. Reaction Simulations. Computational simulations of the in ViVo reactions of HOCl (e.g., Figure 10) were achieved with computer programs that were coded in Mathematica. The Euler method (29) was employed with the following second-order (M-1 s-1) rate

Glutathione Dimer Oxidation by Hypochlorous Acid

Figure 1. Time-resolved UV-visible spectra illustrating the consumption of OCl- at 292 nm and the appearance of NDG at 252 nm. Conditions: [GSSG]0 ) 1 mM, [OCl-]0 ) 2 mM, [iP] ) 0.1 M, pH 10.4, and T ) 18 °C. These spectra were assembled from monochromatic kinetic traces that were spaced 5 nm apart. The individual spectra illustrated in this Figure are spaced 23 ms apart (but the actual absorption-time traces were spaced 4 ms apart). Note the isosbestic point.

Figure 2. 1H NMR spectrum of NDG at 300 MHz. Conditions: [GSSG]0 ) 5 mM, [OCl-]0 ) 10 mM, [iP] ) 0.1 M in D2O, pD 7.2, and T ) 20 °C.

constants and initial concentrations (mM): kGSH ) 3.0 × 107, kGSSG ) 1.0 × 105, ksulfhydryl ) 3.0 × 107, kdisulfide ) 1.6 × 105, kMet ) 3.8 × 107, kHis ) 1.0 × 105, kamino ) 1.0 × 105, [HOCl]0 + [OCl-]0 ) 48, [GSH]0 ) 0-50, [GSSG]0 ) 0, [sulfhydryl]0 ) 13, [disulfide]0 ) 2, [Met]0 ) 26, [His]0 ) 16, and [amino]0 ) 3. The [HOCl]0 + [OCl-]0 was chosen to be four times the average concentration of cytoplasmic GSH in WT E. coli cells to simulate the experimental conditions that were employed by Chesney et al. (Vide infra). Our reasons for selecting these initial conditions are explained in detail in the Discussion.

Results Figure 1 illustrates time-resolved UV-vis spectra for the reaction of GSSG with HOCl/OCl- at pH 10.4 with the consumption of OCl- (λmax ) 292 nm) and the formation of a product with λmax ) 252 nm and  ) 1070 M-1 cm-1 (assuming a single molecular product) (30). The product was further characterized by 1H NMR spectroscopy (Figure 2). The stoichiometry of the reaction was also investigated by 1H NMR spectroscopy. The addition of two molar equivalents of HOCl/ OCl- to GSSG produced one molar equivalent of the initial product (Figure 2). The addition of one molar equivalent of HOCl/OCl- to GSSG produced one-half molar equivalent of the initial product and one-half molar equivalent of unreacted GSSG (data not shown). Similar stoichiometry is indicated when UV-visible spectrophotometry is used to follow the reaction (data not shown). The kinetics of the reaction of GSSG and HOCl/OCl- were investigated at pH 10.4. Under pseudo-first-order conditions (excess GSSG), the observed exponential decay of the absorption-time traces demonstrate that the reaction rate is first-order with respect to [HOCl]T ) [HOCl] + [OCl-], and the resulting pseudo-first-order rate constants exhibit a first-order depend-

Chem. Res. Toxicol., Vol. 20, No. 1, 2007 81

Figure 3. Effect of [GSSG]0 on the reaction rate. Conditions: [GSSG]0 ) 1-10 mM, [OCl-]0 ) 0.15 mM, [iP] ) 0.1 M, I ) 1 M (NaClO4 + Na2HPO4 + Na3PO4) pH 10.35, and T ) 18 °C. Kinetic traces were evaluated at 300 nm.

Figure 4. Effect of [OH-] on the reaction rate. Conditions: [GSSG]0 ) 1 mM, [OCl-]0 ) 0.2 mM, [iP] ) 0.1 M, I ) 1 M (NaClO4 + Na2HPO4 + Na3PO4) pH 10.31-11.61, and T ) 18 °C. Kinetic traces were evaluated at 300 nm. Note: the first two data points were fit to a double-exponential function (see text).

ency with respect to [GSSG]T0 ) [GSSG2-] + [GSSG3-] + [GSSG4-] (Figure 3). The reaction kinetics are also inversely dependent upon the [OH-] (Figure 4). Note that the linear fits of the data in Figures 3 and 4 pass through the origins. The reaction rates of HOCl/OCl- with GSSG near pH 8.5 are too fast to follow by stopped-flow at 18 °C. However, the reaction can be followed at 5 °C for all [H+] (Figure 5). Most of the monochromatic kinetic traces were well described by a singleexponential (first-order) model, but two exponentials were required for the kinetic traces at high pH (pH > 11) (Figure 6). In the absence of a catalyst, the kinetics of the decomposition of NDG at neutral pH exhibit first-order behavior (Figure 7). The products of the decomposition of NDG have not been determined. NDG reacts slowly with the reductant NADPH with pseudo-first-order kinetics to produce GSSG (confirmed by 1H NMR). The reaction between NDG and NADPH is catalyzed by GOR, with the eventual production of GSH (confirmed by 1H NMR). A standard procedure that has been used to assay for the GOR-catalyzed reduction of GSSG (Figure 8) was also used to analyze the reduction of NDG (Figure 9) (28). As reported previously, EDTA has a profound influence on the assay for the reduction of GSSG (Figure 8), but EDTA has no effect on the kinetics of the reduction of NDG (data not shown) (28).

Discussion Characterization of NDG. We have previously demonstrated that cystine reacts with two molar equivalents of HOCl/OClto produce N,N′-dichlorocystine (NDC) (25). NDC was characterized by its electronic spectrum, which exhibited an absorption band at λmax ) 255 nm ( ) 1150 M-1 cm-1) that is characteristic of a chloroamine (30). Aliphatic chloroamines

82 Chem. Res. Toxicol., Vol. 20, No. 1, 2007

Figure 5. (Top) Computed keff for the k4 pathway (dot-dashed), k5 pathway (dashed), and combined k4 and k5 pathways (solid) at 5 °C. The observed keff (open circles) with standard deviations (as error bars) are also plotted. (Bottom) Computed speciation of HOCl (solid), OCl(dot dashed), GSSG2- (long dashed), GSSG3- (dotted), and GSSG4(short dashed) at 5 °C. Conditions: [GSSG]0 ) 0.6 mM, [OCl-]0 ) 0.06 mM, [iP] ) 0.1 M, pH 7-11, I ) 1 M (NaClO4 + Na2HPO4 + Na3PO4), and T ) 5 °C.

Nagy and Ashby

Figure 7. Kinetic trace at λ ) 252 nm (50% of the data illustrated) depicting the decomposition of NDG at pH 7.11 together with an exponential fit (4.30((0.01) × 10-4 s-1). (Inset) Time-resolved spectra. Conditions: [GSSG]0 ) 0.75 mM, [OCl-]0 ) 1.5 mM, [iP] ) 0.1 M, pH 7.11, and T ) 25 °C. The spectra are 24 min apart.

Figure 8. Kinetic trace at λ ) 340 nm depicting the GOR-catalyzed reduction of GSSG by NADPH, with and without EDTA, with linear fits (dashed lines with reciprocal slopes 1.8 × 10-3 and 4.3 × 10-4 s-1, respectively). Conditions: [GSSG]0 ) 1 mM, [NADPH]0 ) 0.1 mM, [GOR] ) 0.025 U/mL ≈ 1.3 nM, [iP] ) 0.1 M, pH 7.56, and T ) 25 °C.

Figure 6. One (dashed line) and two (solid line) exponential fits of a kinetic trace (30% of the data points illustrated) at pH 11.6. This is the first data point in Figure 4. Conditions: [GSSG]0 ) 1 mM, [OCl-]0 ) 0.2 mM, [iP] ) 0.1 M, I ) 1 M (NaClO4 + Na2HPO4 + Na3PO4), pH 11.6, and T ) 18 °C. The residuals of the one- and two-exponential fits evidence that the latter is a better model. In contrast, the last three data points of Figure 4 exhibited clean single-exponential decay. (Inset) Time-resolved UV-visible spectra that illustrate that the fact that the isosbestic point apparent at pH 10.4 (Figure 1) is absent at pH 11.6. Conditions: [GSSG]0 ) 1 mM, [OCl-]0 ) 2 mM, [iP] ) 0.1 M, pH 11.6, and T ) 18 °C.

typically exhibit λmax ≈ 245-260 nm and  ≈ 300-420 M-1 s-1 (31). The molar absorptivity of NDC was consistent with the formation of one chloroamine moiety per equivalent of HOCl/OCl-. The stoichiometry observed in the NMR experiments and the symmetry of the 1H NMR spectra of NDC were consistent with the formation of a dichloro derivative. Analogously, the electronic spectrum (Figure 1, λmax ) 252 nm and  ) 1070 M-1 cm-1) and 1H NMR spectrum (Figure 2) we

have measured for the initial product of the reaction of GSSG and HOCl are consistent with their assignment to NDG, the symmetrical dichloroamine (bis-N-chloro-γ-L-glutamyl) derivative of GSSG. A comparison of the 1H NMR spectra of GSSG and NDG demonstrates that the most substantial differences in the chemical shifts (Table 1) and coupling constants (Table 2) are associated with the glutamate residue. Kinetics and Mechanism of the Formation of NDG. Given our characterization of the initial GSSG-derived product of the reaction of GSSG and HOCl/OCl- as NDG, we have interpreted the observed kinetics according to the reaction mechanism and rate law of Scheme 1. We have previously derived a rate law for an analogous mechanism (25). The mechanism of Scheme 1 is consistent with the observed 2:1 HOCl/OCl-:GSSG stoichiometry as well as the observed first-order dependency of the concentrations of HOCl/OCl- and GSSG (Figure 3). Figure 4 illustrates an inverse dependency on [OH-] under alkaline conditions (i.e., a first-order dependency on [H+]), which suggests that HOCl is the reacting oxidant. The intercept of zero in Figure 4 demonstrates that OCl- does not exhibit a

Glutathione Dimer Oxidation by Hypochlorous Acid

Chem. Res. Toxicol., Vol. 20, No. 1, 2007 83 Scheme 1. Proposed Mechanism, Eqilibrium, and Rate Constants (at 5 °C) and the Rate Law for the Reaction of GSSG and HOCl for Pseudo-First-Order Conditions (Excess GSSG)

Figure 9. Kinetic trace at λ ) 340 nm depicting the reduction of NDG by NADPH, with and without GOR. A linear fit of the pseudo-zeroorder part of the enzyme-catalyzed reaction is illustrated as a dashed line. Conditions: [NDG]0 ) 0.5 mM, [NADPH]0 ) 0.1 mM, [GOR] ) zero or 0.025 U/mL ≈ 1.3 nM, [iP] ) 0.1 M, pH 7.56, and T ) 25 °C. Table 1. 1H NMR Chemical Shifts (ppm) for GSSG and NDG in D2O (100 mM Phosphate Buffer at pD 7.2) at 20 °C Gly R-Ha Gly R-Hb Cys R-H Cys β-Ha Cys β-Hb Glu R-H Glu β-Ha Glu β-Hb Glu γ-Ha Glu γ-Hb

GSSGa

NDG

3.78 3.75 4.74 3.30 2.97 3.77 2.15 2.15 2.55 2.51

3.79 3.79 4.79 3.33 2.96 3.48 1.95 1.95 2.46 2.41

a

Rabenstein, D. L., and Keire, D. A. (1989) Nuclear magnetic resonance spectroscopy of glutathione. Coenzymes Cofactors 3, 67-101. Table 2. 1H NMR Coupling Constants (Hz)a for GSSG and NDG in D2O (100 mM Phosphate Buffer at pD 7.2) at 20 °C 3J

Gly R-Ha,Gly R-Hb

3J

Cys R-H,Cys β-Ha

3J

Cys R-H,Cys β-Hb

3J

Cys β-Ha,Cys β-Hb

3J

Glu R-H,Glu β-Ha

3J

Glu R-H,Glu β-Hb

2J

Glu β-Ha,Glu β-Hb

3J

Glu β-Ha,Glu γ-Ha

3J

Glu β-Hb,Glu γ-Ha

3J

Glu β-Ha,Glu γ-Hb

3J

Glu β-Hb,Glu γ-Hb

3J

Glu γ-Ha,Glu γ-Hb

GSSGb

NDG

17.4 4.1 9.7 14.3 6.2 6.3

17.7 3.9 9.6 14.1 6.9 6.9

7.7 7.8 7.6 7.7 15.4

8.1 8.1 7.2 7.2 13.5

a

Absolute coupling constants. The signs of the coupling constants were not determined. b Rabenstein, D. L., and Keire, D. A. (1989) Nuclear magnetic resonance spectroscopy of glutathione. Coenzymes Cofactors 3, 67-101.

measurable rate of reaction with GSSG. The rates of reaction of HOCl with nucleophiles are typically 103-105 faster than those with OCl- (32-34). Although it was possible to measure the kinetics of the reaction of HOCl/OCl- with GSSG at 18 °C under very basic conditions, the reaction became too fast for such measurements near pH 8.5. However, it proved possible to measure the rate between pH 6 and 11 at 5 °C (Figure 5). The reaction kinetics exhibit an inflection point with a maximum rate at about pH 8.5, which is indicative of a competitive acid-base equilibrium. The rate decreases above pH 8.5, which is consistent with the reaction of HOCl and not OCl- because the pKa of HOCl is

7.47 at 5 °C and I ) 1. Our estimated pKa value of HOCl at 5 °C is based upon previously measured values between 15 and 50 °C (35). The observed decrease in rate below pH 8.5 suggests that the amines, and not the ammonium moieties of GSSG, are responsible for reactivity. We estimate the two pKa values of the amines of GSSG to be 8.5 and 9.5 at 5 °C and I ) 1 on the basis of previously measured values at 5-25 °C at I ) 0.15 and I ) 0.15-1 at 25 °C (27). We have previously concluded that the different conjugate bases of cystine contribute to the rate law for its reaction with HOCl to produce NDC (25). Using similar logic, we conclude that the magnitudes of the observed rate and the width-at-halfheight of the keff versus pH (Figure 5, top) cannot be explained by the k4 or k5 pathway alone. Using [H+], k4, and k5 as variables in the equation of Scheme 1, we subjected the experimental data of Figure 5 to nonlinear least-squares analysis. That analysis yielded the rate constants k4 ) 2.7(2) × 106 M-1 s-1 and k5 ) 3.5(3) × 107 M-1 s-1 at 5 °C. These rate constants are consistent with the values for k4 and k5 that were determined at the peripheries of Scheme 1, low pH, and high pH, where the respective k4 and k5 pathways would dominate. Furthermore, the relative magnitudes of the rate constants, about 1:13, are consistent with the expectation that the tetraanion GSSG4should be a better nucleophile. However, despite the smaller rate constant for the reaction of GSSG3-, the k4 pathway yields a competitive (almost equal) maximum rate because the pKa of GSSG3- is better matched with the pKa of HOCl compared with the pKa of GSSG4-. The k4 pathway dominates at physiological pH for the same reason (Figure 5). At high pH, a doubleexponential equation was required to model the kinetic traces (Figure 6), which is consistent with the sequential reaction model of Scheme 1. The exponent of the first exponential was found to correspond to k5 and the exponent of the second exponential

84 Chem. Res. Toxicol., Vol. 20, No. 1, 2007

Nagy and Ashby

was assigned to k7. Under very basic conditions, k7 ≈ 10 × k5. To our knowledge, this is the first report of an absolute rate constant for the reaction of HOCl with GSSG. The values we have determined for k4 and k5 for GSSG are comparable to the related rate constants we have previously reported for cystine (4.3(2) × 106 M-1 s-1 and 1.6(2) × 107 M-1 s-1, respectively, at 5 °C) (25). However, because of better pKa matching between HOCl with cystine (versus GSSG), cystine is about twice as reactive at physiologic pH as GSSG. The magnitudes of the rate constants we have measured for the formation of the chloroamine derivatives of cystine and GSSG at pH 7 (ca. 105 M-1 s-1) are comparable to the rates that have been reported for model reactions between HOCl and R-amino compounds (ca. 105 M-1 s-1) (20). Decomposition of NDG. In the absence of other reactants, NDG eventually spontaneously decomposes via a first-order process that is consistent with hydrolysis or possibly with an intramolecular redox reaction (Figure 7). Because this decomposition process is unimolecular (or pseudo-first-order if the co-reactant is water), the half-life for decomposition is independent of the initial concentration of NDG. From the rate constant that was determined from the kinetic data of Figure 7 (4.30(1) × 10-4 s-1 by SVD analysis of polychromatic kinetic traces), we compute a half-life for NDG at pH 7.1 of approximately 27 min. This observed rate is comparable to those that have been attributed to hydrolysis of other chloroamines, ca. 10-5 to 10-6 s-1 (36, 37). However, other reaction pathways are feasible, including Grob fragmentation to give aldehydes (38) and internal oxidation of the disulfide moiety (25). In addition, Winterbourn et al. have identified a sulfonamide derivative when GSH is reacted with HOCl (39). It is not clear what role, if any, NDG plays in the formation of the latter species. GOR-Catalyzed Reduction of NDG. GOR catalyzes the reduction of GSSG by NADPH (Figure 8). However, it has been previously noted that the thiol product of the reaction as well as other thiols can have a marked effect on the efficacy of the enzyme (28). The inhibitory effects of thiols, which have been attributed to the formation of disulfides at the cysteine active site of GOR, can be ameliorated by the addition of EDTA (Figure 8). NDG reacts with NADPH via a slow second-order process (Figure 9, fit not shown). However, the reaction is accelerated in the presence of GOR (Figure 9). The pseudozero-order kinetics that are observed in the presence of substoichiometric (catalytic) amounts of GOR are expected for enzyme catalysis (Figure 9). The only glutathione-derived product that is observed by 1H NMR following the GORcatalyzed reduction of NDG by NADPH is GSH. Because the enzyme-catalyzed reaction is faster than the rate of decomposition of NDG (Figure 7) and because it is also faster than the uncatalyzed reaction of NDG and NADPH (Figure 9), we conclude that NDG is capable of serving as a substrate for GOR. Note that it was not necessary to add EDTA to the assay to reduce NDG (data not shown), presumably because the thiol that is produced reacts rapidly with the remaining chloroamine to yield disulfide:

RSH + R′NHCl f RSCl + R′NH2

(1)

RSCl + RSH f RSSR + HCl

(2)

Regarding the mechanism of the GOR-catalyzed reduction of NDG by NADPH, by analogy, it is conceivable that the Cys58/Cys-63 active site of GOR in its reduced state reacts with the chloroamine moieties of NDG to produce GSSG. The GSSG

that would be produced in such a reaction would in turn be reduced by the GOR/NADPH system via the conventional mechanism. Alternatively, it is possible that NDG serves as a substrate for GOR as GSSG does, albeit to produce a chloroamine derivative of GSH that would react via reactions 1 and 2 to produce the same end products. The mechanism of the GOR-catalyzed reduction of NDG by NADPH is currently under investigation. Possible Physiological Relevance of NDG. As noted in the introduction, GSH is the principal intracellular non-protein thiol, and its sulfhydryl group presumably affords the first defense against oxidative stress by HOCl. But, what happens when that defensive barrier is overwhelmed? To ascertain whether the amino groups of GSSG are capable of competing effectively with respect to the other functional groups within a cell, it is necessary to determine the relative reactivities of all of the potential targets of HOCl. This is a daunting task because the rate constants for the reaction of HOCl with actual proteins have never been measured. Indeed, there is no consensus on the precise rate constants for free residues, such as cysteine (20, 38, 40) and cystine (20, 25). Nonetheless, it appears clear that the relevant rate constants for the free residues can be grouped into two categories: those that are ca. 107 M-1 s-1 (sulfhydryl groups and thioethers) and those that are ca. 105 M-1 s-1 (disulfides, amino groups, and imidazole rings). Competition for HOCl by these functional groups is generally proportional to the products of their rate constants and their respective concentrations. Because it is recognized that there are some residues in proteins that are not accessible by oxidative insults, these rate constants can be assumed to be the upper limits for the proteinaceous derivatives of these functional groups (20). In addition to the limited information that is available regarding the rate constants, it is clear that estimations of the concentrations of the reactants in ViVo will also have large uncertainties. Thus, in attempting to model the chemistry that occurs in ViVo, we will concern ourselves here with the trends and not with absolute outcomes. We have selected Escherichia coli as a model because there is a substantial amount of information available concerning the cellular content of this bacterium (41). Although HOCl will certainly cause damage to the periplasm and the membranes, we will restrict our discussion to the role of GSH within the cytoplasm. Furthermore, we will focus here on peptide residues and ignore possible nuclear damage entirely by HOCl (6). To postpone a discussion of the roles of repair mechanisms (e.g., GOR) (42), we will first address the issue of whether GSSG could play a role in protecting proteins from a bolus of HOCl. Because the intracellular concentrations of GSH/GSSG are 5001000 times that of NADPH and other intracellular redox systems (14), it can be expected that the rate of the oxidation of GSH by HOCl will far exceed the rate that GOR will be able to reduce GSSG. Our approach will be to employ a kinetic model that comprises a parallel competition between a reaction of HOCl with GSH and the average protein in E. coli. The concentration of glutathione in E. coli has been determined to be ca. 27 µmol/g dry weight, when the bacterium is cultured aerobically (3). Under typical conditions, roughly 99.5% of the glutathione is maintained in the reduced state (2). Because the dry weight of an E. coli cell is ca. 3 × 10-13 g and the volume of the cytoplasm is ca. 6.7 × 10-16 L (41), we calculate [GSH] to be ∼12 mM in the cytoplasm. With respect to the protein model, we will focus only on the CySH, Met, cystine, His, and the terminal R-amino groups. The other proteinaceous functional groups are expected to be considerably less reactive (20). Free

Glutathione Dimer Oxidation by Hypochlorous Acid

Figure 10. Simulation of the in ViVo reaction of HOCl in the cytoplasm of E. coli as a function of the amount of GSH (see text). The rightmost ordinate axis refers to HOCl/GSH stoichiometry (bold line). The arrows mark the concentration of GSH expected in WT E. coli. The amount of HOCl that was employed in the simulation corresponds to four times this amount of GSH. After the point on the line marked with an X, GSH has been depleted, and GSSG is overoxidized to NDG. By this point in the model, all of the oxidant-accessible protein sulfhydryl and thioether groups have been oxidized. The percent of remaining protein disulfide is plotted for models that include (solid line) and exclude (dashed line) the overoxidation of GSSG. Similarly, the percent of remaining His and terminal amino groups are plotted for models that include (dot-dashed line) and exclude (dotted line) the overoxidation of GSSG.

amino acids were not included in the model. Assuming 106 cytoplasmic proteins (41), an average size of 360 residues (41), and an average distribution of amino acids in the cytoplasmic proteins of prokaryotes (43, 44), we estimate [thiol] + 2 × [disulfide] ) 15 mM, [Met] ) 26 mM, and [His] ) 16 mM. Using the average thio/disulfide redox status for proteins in the cytoplasm (45, 46), we compute [thiol] ) 13 mM and [disulfide] ) 1 mM. Assuming that each of the 106 cytoplasmic proteins bear one terminal R-amino group (41), we estimate [terminal R-amino groups] ) 3 mM. It is important to point out that the absolute concentration of each functional group is irrelevant because it is their relative abundances that will dictate the outcome. Thus, we will abandon the absolute concentrations for the remainder of the discussion, in favor of relative concentrations, with the objective of determining the trend of varying the redox buffering capacity versus the protein residues that require protection (i.e., the ratio of the number of molar equivalents; nHOCl /nGSH o o ). We have chosen the amount of HOCl reacting HOCl (no , corresponding to the initial number of moles of HOCl in ViVo for the bolus model)2 to be four times the estimated initial number of moles of GSH in the cytoplasm of WT E. coli (nGSH o ) for reasons that will become apparent later. Using these estimated concentrations and the rate constants that have been published for the reaction of HOCl/OCl- with model compounds at neutral pH (20), together with the rate constants that are reported herein, one can model the depletion of the targets as a function of the ratio nHOCl /nGSH (which can o o also be related to GSH/protein for our model) using Euler’s is held constant, and nGSH is method (29). In Figure 10, nHOCl o o varied. It is important to point out that the use of initial rates (i.e., the rate constants times the initial concentrations of the 2 We point out that a model that comprises the rate-limiting diffusion of HOCl across a cytoplasmic membrane, followed by partitioning of the HOCl via the aforementioned rate constants and intracellular concentrations, is expected to yield precisely the same result as that of a model that consists of a bolus of HOCl.

Chem. Res. Toxicol., Vol. 20, No. 1, 2007 85

reaction partners) will not provide a realistic picture of this model because the reaction in question is one in which the concentrations of the reactants will change as they are depleted by the bolus2 of HOCl. For example, sulfhydryl groups are converted to disulfide groups that continue to participate in the reaction. The results of our simulation are summarized in Figure 10. Despite the primitivity of the model, there are some general conclusions that can be drawn concerning the possible role of GSSG as an antioxidant. It is clear that GSSG will not likely play a role (vis-a`-vis NCG or NDG) in protecting oxidantaccessible proteinaceous sulfhydryl and thioether moieties from oxidation to disulfides and sulfoxides, respectively (see the caption for Figure 10). Thus, all of the sulfhydryl and thioether moieties are oxidized before overoxidation is predicted by our model. However, we reiterate that this model is based upon the assumption that all of the proteinateous sulfhydryl and thioether moieties are equally reactive, whereas in reality, these groups are expected to exhibit a continuum of reactivity. Accordingly, the model in Figure 10 probably reflects the minimum role of GSSG overoxidation. Although the reaction of GSSG with HOCl is not predicted to protect the more reactive groups (e.g., those that react with rate constants of ca. 107), GSSG may play a role in protecting proteinaceous imidazole and terminal amino groups, and it may also serve to protect the disulfide and sulfoxide groups in proteins from overoxidation (47). Furthermore, we note that there are protective enzyme systems available in the cytoplasm to reduce disulfides (48) and sulfoxides (49), and the production of disulfides and sulfoxides (as well as higher oxidation states) may facilitate the signaling of oxidative stress (50, 51). Although Figure 10 provides an enticing possibility for the involvement of GSSG in the redox buffering of HOCl, we point out once more that the model we have employed is very crude. We have noted previously that one of the significant challenges associated with the development of kinetic models for in ViVo chemical reactions is that the outcomes of such models can often be significantly altered by the inclusion of one or more additional variables (24). However, there is some experimental evidence that GSSG is capable of buffering HOCl ex ViVo and in ViVo. Chesney et al. have observed an LD50 for a GSH-deficient strain of E. coli that is about two times lower than that of the WT (52). Because the strains are otherwise isogenic, this difference in survival is attributed to GSH stasis. Remarkably, exogenous GSSG fully restored the resistance of the GSH-deficient strain toward hypochlorous acid (53), which we attribute to the formation of extracellular NDG upon insult by HOCl. We have independently determined that exogenous NDG is not cytotoxic to E. coli (unpublished results). Furthermore, because it was determined by Chesney et al. that 50% of the cells survived when they were challenged by 50% more HOCl than would be expected to oxidize the cytoplasmic GSH to GSSG, this raises the issue of whether NDG plays a role in ViVo.3 Chesney et al. observed that the stoichiometry between the amount of HOCl added and the amount of bacterial GSH that was oxidized was 4:1, and they suggested that the GSH is being oxidized to a higher oxidation state than the disulfide (52). That point is marked with arrows in Figure 10. Given the results we have presented herein, we believe this hypothesis has merit. 3 We have determined the minimum inhibitory concentration (MIC) of HOCl toward 107 E. coli cells/mL (MG1655) in 100 mM phosphate buffer (i.e., no organic load) to be 30 µM (data not shown). Assuming that all of the extracellular HOCl eventually enters the cytoplasms of these cells, the MIC represents about 100 times the amount of HOCl that we have employed in our model calculations. Of course, the HOCl will in reality find targets other than those that have been included in our computational model.

86 Chem. Res. Toxicol., Vol. 20, No. 1, 2007

Acknowledgment. We are grateful to the American Heart Association (0555677Z), the Petroleum Research Fund (42850AC4), and the National Science Foundation (CHE-0503984) for their financial support of various aspects of this project. Supporting Information Available: Mathematica file used in the simulation of Figure 10. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Meister, A. (1988) Glutathione metabolism and its selective modification. J. Biol. Chem. 263, 17205-17208. (2) Smirnova, G. V., and Oktyabrsky, O. N. (2005) Glutathione in bacteria. Biochemistry (Moscow) 70, 1199-1211. (3) Fahey, R. C., Brown, W. C., Adams, W. B., and Worsham, M. B. (1978) Occurrence of glutathione in bacteria. J. Bacteriol. 133, 11261129. (4) Berglin, E. H., Edlund, M. B. K., Nyberg, G. K., and Carlsson, J. (1982) Potentiation by L-cysteine of the bactericidal effect of hydrogen peroxide in Escherichia coli. J. Bacteriol. 152, 81-88. (5) Carlsson, J., Granberg, G. P. D., Nyberg, G. K., and Edlund, M. B. K. (1979) Bactericidal effect of cysteine exposed to atmospheric oxygen. Appl. EnViron. Microbiol. 37, 383-390. (6) Park, S., and Imlay, J. A. (2003) High levels of intracellular cysteine promote oxidative DNA damage by driving the Fenton reaction. J. Bacteriol. 185, 1942-1950. (7) Konigsberger, E., Wang, Z., and Konigsberger, L.-C. (2000) Solubility of L-Cystine in NaCl and artificial urine solutions. Monatsh. Chem. 131, 39-45. (8) Dolphin, D., Avramovic, O., and Paulson, R.,Eds. (1989) Coenzymes and Cofactors: Glutathione: Chemical, Biochemical, and Medical Aspects, Vol. 3, parts A and B, John Wiley and Sons, New York. (9) Wu, G., Fang, Y.-Z., Yang, S., Lupton, J. R., and Turner, N. D. (2004) Glutathione metabolism and its implications for health. J. Nutr. 134, 489-492. (10) Shelton, M. D., Chock, P. B., and Mieyal, J. J. (2005) Glutaredoxin: Role in reversible protein S-glutathionylation and regulation of redox signal transduction and protein translocation. Antioxid. Redox Signaling 7, 348-366. (11) Pineda-Molina, E., and Lamas, S. (2003) Redox-regulated glutathionylation of transcription factors: a regulatory mode for gene expression. NATO Sci. Ser., Ser. I 344, 89-101. (12) Coppola, S., and Ghibelli, L. (2000) GSH extrusion and the mitochondrial pathway of apoptotic signalling. Biochem. Soc. Trans. 28, 56-61. (13) Armstrong, R. N. (1997) Structure, catalytic mechanism, and evolution of the glutathione transferases. Chem. Res. Toxicol. 10, 2-18. (14) Schafer, F. Q., and Buettner, G. R. (2001) Redox environment of the cell as viewed through the redox state of the glutathione disulfide/ glutathione couple. Free Radical Biol. Med. 30, 1191-1212. (15) Schirmer, R. H., Krauth-Siegel, R. L., and Schulz, G. E. (1989) Glutathione Reductase, in Coenzymes and Cofactors: Glutathione: Chemical, Biochemical, and Medical Aspects (Dolphin, D., Avramovic, O., and Paulson, R., Eds.), Vol. 3, part A, pp 553-596, John Wiley and Sons, New York. (16) Ishikawa, T., and Sies, H. (1989) Glutathione as an Antioxidant: Toxicological Aspects, in Coenzymes and Cofactors: Glutathione: Chemical, Biochemical, and Medical Aspects (Dolphin, D., Avramovic, O., and Paulson, R., Eds.) Vol. 3, part B, pp 85-109, John Wiley and Sons, New York. (17) Wells, W. W., and Xu, D. P. (1994) Dehydroascorbate reduction. J. Bioenerg. Biomembr. 26, 369-377. (18) Buettner, G. R. (1993) The pecking order of free radicals and antioxidants: Lipid peroxidation, a-tocopherol, and ascorbate. Arch. Biochem. Biophys. 300, 535-543. (19) Heinecke, J. W. (2003) Oxidative stress: New approaches to diagnosis and prognosis in atherosclerosis. Am. J. Cardiol. 91, 12A-16A. (20) Pattison, D. I., and Davies, M. J. (2001) Absolute rate constants for the reaction of hypochlorous acid with protein side chains and peptide bonds. Chem. Res. Toxicol. 14, 1453-1464. (21) Pattison, D. I., Hawkins, C. L., and Davies, M. J. (2003) Hypochlorous acid-mediated oxidation of lipid components and antioxidants present in low-density lipoproteins: Absolute rate constants, product analysis, and computational modeling. Chem. Res. Toxicol. 16, 439449. (22) Hawkins, C. L., Pattison, D. I., and Davies, M. J. (2003) Hypochloriteinduced oxidation of amino acids, peptides and proteins. Amino Acids 25, 259-274.

Nagy and Ashby (23) Ashby, M. T., Carlson, A. C., and Scott, M. J. (2004) Redox buffering of hypochlorous acid by thiocyanate in physiologic fluids. J. Am. Chem. Soc. 126, 15976-15977. (24) Nagy, P., Beal, J. L., and Ashby, M. T. (2006) Thiocyanate is an efficient endogenous scavenger of the phagocytic killing agent hypobromous acid. Chem. Res. Toxicol. 19, 587-593. (25) Nagy, P., and Ashby, M. T. (2005) Reactive sulfur species: Kinetics and mechanism of the oxidation of cystine by hypochlorous acid to give N,N′-dichlorocystine. Chem. Res. Toxicol. 18, 919-923. (26) Irving, H. M. N. H., Miles, M. G., and Pettit, L. D. (1967) A study of some problems in determining the stoichiometric proton dissociation constants of complexes by potentiometric titrations using a glass electrode. Anal. Chim. Acta 38, 475-488. (27) Martell, A. E., and Smith, R. M. (1989) Critical Stability Constants, Vol. 6, Plenum Press, New York. (28) Smith, I. K., Vierheller, T. L., and Thorne, C. A. (1988) Assay of glutathione reductase in crude tissue homogenates using 5,5′dithiobis(2-nitrobenzoic acid). Anal. Biochem. 175, 408-413. (29) Coe, D. A. (1987) Numerical solutions of kinetic equations on a spreadsheet. J. Chem. Educ. 64, 496-497. (30) Thomas, E. L., Grisham, M. B., and Jefferson, M. M. (1986) Preparation and characterization of chloramines. Methods Enzymol. 132, 569-585. (31) Kovacic, P., Lowery, M. K., and Field, K. W. (1970) Chemistry of N-bromamines and N-chloramines. Chem. ReV. 70, 639-665. (32) Gerritsen, C. M., and Margerum, D. W. (1990) Non-metal redox kinetics: hypochlorite and hypochlorous acid reactions with cyanide. Inorg. Chem. 29, 2757-2762. (33) Fogelman, K. D., Walker, D. M., and Margerum, D. W. (1989) Nonmetal redox kinetics: hypochlorite and hypochlorous acid reactions with sulfite. Inorg. Chem. 28, 986-993. (34) Kumar, K., and Margerum, D. W. (1987) Kinetics and mechanism of general-acid-assisted oxidation of bromide by hypochlorite and hypochlorous acid. Inorg. Chem. 26, 2706-2711. (35) Adam, L. C., Fabian, I., Suzuki, K., and Gordon, G. (1992) Hypochlorous acid decomposition in the pH 5-8. region. Inorg. Chem. 31, 3534-3541. (36) Gray, E. T., Jr., Margerum, D. W., and Huffman, R. P. (1978) Chloramine equilibriums and the kinetics of disproportionation in aqueous solution. ACS Symp. Ser. 82, 264-277. (37) Margerum, D. W., Gray, E. T., Jr., and Huffman, R. P. (1978) Chlorination and the formation of N-chloro compounds in water treatment. ACS Symp. Ser. 82, 278-291. (38) Armesto, X. L., Canle, L, M., Garcia, M. V., and Santaballa, J. A. (1998) Aqueous chemistry of N-halo-compounds. Chem. Soc. ReV. 27, 453-460. (39) Pullar, J. M., Vissers, M. C. M., and Winterbourn, C. C. (2001) Glutathione oxidation by hypochlorous acid in endothelial cells produces glutathione sulfonamide as a major product but not glutathione disulfide. J. Biol. Chem. 276, 22120-22125. (40) Winterbourn, C. C. (1985) Comparative reactivities of various biological compounds with myeloperoxidase-hydrogen peroxidechloride, and similarity of the oxidant to hypochlorite. Biochem. Biophys. Acta 840, 204-210. (41) Sundararaj, S., Guo, A., Habibi-Nazhad, B., Rouani, M., Stothard, P., Ellison, M., and Wishart, D. S. (2004) The CyberCell Database (CCDB): a comprehensive, self-updating, relational database to coordinate and facilitate in silico modeling of Escherichia coli. Nucleic Acids Res. 32, D293-D295. (42) Carmel-Harel, O., and Storz, G. (2000) Roles of the glutathione- and thioredoxin-dependent reduction systems in the Escherichia coli and Saccharomyces cerevisiae responses to oxidative stress. Annu. ReV. Microbiol. 54, 439-461. (43) Pe’er, I., Felder, C. E., Man, O., Silman, I., Sussman, J. L., and Beckmann, J. S. (2003) Proteomic signatures: Amino acid and oligopeptide compositions differentiate among phyla. Proteins: Struct., Funct., Bioinf. 54, 20-40. (44) Neidhardt, F. C. (1987) Chemical Composition of Escherichia coli, in Escherichia coli Salmonella typhimurium pp 3-6, ASM Press, Washington D.C. (45) Cumming, R. C., Andon, N. L., Haynes, P. A., Park, M., Fischer, W. H., and Schubert, D. (2004) Protein disulfide bond formation in the cytoplasm during oxidative stress. J. Biol. Chem. 279, 21749-21758. (46) Rietsch, A., and Beckwith, J. (1998) The genetics of disulfide bond metabolism. Annu. ReV. Genet. 32, 163-184. (47) Georgiou, G., and Masip, L. (2003) Biochemistry: An overoxidation journey with a return ticket. Science 300, 592-594. (48) Aslund, F., and Beckwith, J. (1999) The thioredoxin superfamily: redundancy, specificity, and gray-area genomics. J. Bacteriol. 181, 1375-1379. (49) Weissbach, H., Resnick, L., and Brot, N. (2005) Methionine sulfoxide reductases: history and cellular role in protecting against oxidative damage. Biochim. Biophys. Acta 1703, 203-212.

Glutathione Dimer Oxidation by Hypochlorous Acid (50) Watson, W. H., Yang, X., Choi, Y. E., Jones, D. P., and Kehrer, J. P. (2004) Thioredoxin and its role in toxicology. Toxicol. Sci. 78, 3-14. (51) Hoshi, T., and Heinemann, S. H. (2001) Regulation of cell function by methionine oxidation and reduction. J. Physiol. 531, 1-11. (52) Chesney, J. A., Eaton, J. W., and Mahoney, J. R., Jr. (1996) Bacterial glutathione: a sacrificial defense against chlorine compounds. J. Bacteriol. 178, 2131-35.

Chem. Res. Toxicol., Vol. 20, No. 1, 2007 87 (53) Carr, A. C., Decker, E. A., Park, Y., and Frei, B. (2001) Comparison of low-density lipoprotein modification by myeloperoxidase-derived hypochlorous and hypobromous acids. Free Radical Biol. Med. 31, 62-72.

TX060184G