Kinetic Glutathione Chemoassay To Quantify Thiol Reactivity of

742

Chem. Res. Toxicol. 2009, 22, 742–750

Kinetic Glutathione Chemoassay To Quantify Thiol Reactivity of Organic ElectrophilessApplication to r,β-Unsaturated Ketones, Acrylates, and Propiolates Alexander Bo¨hme,†,‡ Diana Thaens,†,‡ Albrecht Paschke,† and Gerrit Schuü¨rmann*,†,‡ UFZ Department of Ecological Chemistry, Helmholtz Centre for EnVironmental Research, Permoserstrasse 15, 04318 Leipzig, Germany, and Institute for Organic Chemistry, Technical UniVersity Bergakademie Freiberg, Leipziger Strasse 29, 09596 Freiberg, Germany ReceiVed December 27, 2008

Glutathione (GSH) is a soft nucleophile and, as such, can be used to sense the reactivity of electrophilic agents toward the thiol group and other electron-rich sites of molecular structures. A new kinetic GSH chemoassay is introduced that employs a photometric method to quantify GSH loss and enables an efficient determination of second-order rate constants, kGSH, of the reaction between electrophilic substances and GSH. Comparison with an existing 2 h static assay shows that the new kinetic variant is superior with respect to the detectable range of electrophilic reactivity and to confounding factors such as additional GSH loss due to oxidation. Analysis of the chemoassay degradation kinetics provides insight into the characteristic reaction times and the contributions of GSH-electrophile Michael addition and GSH oxidation to the overall GSH loss. For 15 R,β-unsaturated ketones, nine acrylates, and two propiolates acting as Michael acceptors, the measured kGSH values span ca. 5 orders of magnitude. Moreover, log kGSH correlates with the compounds’ toxicity toward the ciliates Tetrahymena pyriformis in terms of 48 h log EC50 (50% growth inhibition) values, yielding a squared correlation coefficient (r2) of 0.91 and a root-mean-square error of 0.30 log units. It shows that for these and related compounds, aquatic toxicity is driven by electrophilic reactivity. The findings demonstrate that the kinetic GSH chemoassay can be used as an efficient tool to analyze, interpret, and predict correspondingly reactive toxicity in the context of qualitative and quantitative structure-activity relationship studies and as a nonanimal tool of integrated testing strategies for REACH to screen compounds for excess toxicity. Introduction In aquatic toxicology, it is a widely accepted paradigm that baseline narcosis is the minimum toxicity associated with any neutral organic compound, which in turn can be estimated from simple quantitative structure-activity relationships (QSARs) employing the logarithmic octanol-water partition coefficient to quantify the compound’s hydrophobicity (1). Accordingly, baseline narcosis can be taken as a pragmatic reference to identify a specific or reactive toxicity mechanism, provided that these exert excess toxicity (relative to the narcosis regression line) (2). While narcotic toxicity can be predicted quantitatively for those species for which respective QSARs have been calibrated, a correspondingly general predictive approach is lacking for excess toxic compounds. Accordingly, an effective way to reduce the need for experimental testing is to discriminate, directly from molecular structure, narcosis-level compounds from substances exerting excess toxicity (3). The former could then be evaluated employing existing species-specific narcosis QSARs, and only the latter would require a more detailed investigation of their actual toxic potency. In the context of the new European chemical legislation REACH, corresponding classification methods could serve as a component of integrated testing strategies (ITS) (4), enabling an early decision about * To whom correspondence should be addressed. Tel: +49-341-2351262. Fax: +49-341-235-1785. E-mail: [email protected]. † UFZ Helmholtz Centre for Environmental Research. ‡ Technical University Bergakademie Freiberg.

the possible reduction or replacement of aquatic toxicity testing for narcotic compounds. Reactive toxicity results from covalent bonding to proteins or endogenous biochemical substrates and is usually expected to result in excess toxicity. Because in this case toxicity is directly related to chemical reactivity, an appropriate assessment of the latter provides an alternative route to its nonanimal screening. Thus, chemical reactivity testing is a further option as an ITS component, complementing nontest information gained through chemical category analysis, chemical and biological read-across (extrapolation across compounds or species), and QSARs if available for the reactive compound class of interest. Electrophiles form an important class of reactive, excesstoxic compounds and have been studied with respect to their acute aquatic toxicity since some time (2). Early investigations include the use of chemoassays to quantify the electrophilic reactivity of organic halides (5), aldehydes (6), and epoxides (7). The electrophilic impact on aquatic toxicity has also been studied employing quantum chemical parameters for various compound classes such as organophosphorus compounds (8), aldehydes, SN2-reactive halogenated compounds and R,βunsaturated carbonyl compounds (9), redox cyclers (10), and phenol derivatives (11). Moreover, electrophilic toxicity was investigated with respect to skin sensitization (12, 13) and chromosomal aberration (14). Chemoassays employing glutathione (GSH) as a model nucleophile have been used for some time to quantify the electrophilic reactivity of R,β-unsaturated carbonyls acting as

10.1021/tx800492x CCC: $40.75  2009 American Chemical Society Published on Web 03/25/2009

Thiol ReactiVity of Organic Electrophiles

Michael acceptors and of SN2 electrophiles such as aliphatic, allylic, and benzylic halides as well as R-halogenated carbonyls. Earlier work focused on reaction rate constants (15-18), and more recently, a 2 h photometric assay was developed to determine 50% reaction concentration (RC50) values as a quantification of the electrophilic reactivity toward GSH (19, 20). Interestingly, static (fixed-time) assays to quantify the chemical reactivity of toxicologically relevant compounds have been introduced ca. 60 years ago (21) and are inferior to (typically but not necessarily more time-consuming) explicit determinations of reaction rate constants with respect to the detectable range of reactivity. In view of the increased interest in the use of chemoassays as an ITS tool to screen compounds for their potential to exert reactive toxicity, it appeared useful to investigate and possibly extend the actual range of applicability of the GSH chemoassay. Building on the photometric approach (19, 20), we have developed a variant of the GSH chemoassay that provides an efficient determination of the second-order reaction rate constants for a wide range of electrophilic reactivity and can also be used to determine the amount of GSSG (oxidized GSH) formation as additional GSH loss process. As shown in more detail below, the latter is mediated through both DMSO (cosolvent dimethyl sulfoxide) and solution-phase oxygen (22) and may govern the overall GSH loss process for low-reactive electrophiles. First results with this kinetic GSH chemoassay in terms of second-order and pseudo-first-order kGSH values are presented for 15 R,β-unsaturated ketones and 11 R,β-unsaturated carboxylates (acrylates, methacrylates, crotonates, and propiolates) and are discussed in the context of reactive toxicity. Furthermore, for all compounds, RC50 values are also reported and compared to corresponding literature data (19, 20) (as far as available) and to kGSH. Moreover, detailed analysis of the chemoassay degradation kinetics provides insight into the characteristic reaction times and their impact on the relative contributions of GSH-electrophile Michael addition and GSSG formation to the overall GSH loss. The findings demonstrate the capability of the kinetic GSH chemoassay to quantify electrophilic reactivity over a large value range and to serve as a respective ITS tool for the analysis and prediction of the reactive toxicity of electrophiles.

Materials and Methods Chemical Substances. The 26 test chemicals used in this study were provided by Merck (Darmstadt, Germany), Alfa Aesar (Karlsruhe, Germany), Sigma (Missouri, United States), Acros Organics (Geel, Belgium), and Fluka (Buchs, Switzerland). The purity of all compounds was g95%. Reduced GSH, DMSO, 2,2′dinitro-5,5′-dithiodibenzoic acid (DTNB), disodium hydrogen phosphate (anhydrous), and potassium dihydrogen phosphate (anhydrous) were obtained from Merck, all being at p.a. grade. Kinetic GSH Chemoassay. A buffer solution containing 0.0648 mol/L disodium hydrogen phosphate and 0.0153 mol/L potassium dihydrogen phosphate was prepared and stored at 25 °C in a climate chamber. The DTNB solution was obtained by suspending 3.98 g of DTNB in 200 mL of buffer solution and adding 1 N sodium hydroxide until the pH value of 7.4 was reached. The GSH solution was freshly prepared for every experiment by diluting 0.0215 g of 0 reduced GSH in 50 mL of buffer solution (cGSH ) 1.4 mmol/L ) initial GSH concentration at reaction time t ) 0). The stock solution of the test compound was generated by filling a defined amount of DMSO into a 50 mL volumetric flask. Note that the DMSO volume varied (3.3 or 10 mL per 50 mL stock solution), depending on the amount of stock solution required for the experiment (600 or 200 µL). Afterward, the test compound was added gravimetrically, and

Chem. Res. Toxicol., Vol. 22, No. 4, 2009 743 the flask was filled with buffer solution. The resulting compound concentrations were in the range from 0.1 to 25 mmol/L. Reaction rate constants were determined using glass-coated deep well plates (96 wells) as reaction batches with a total reaction volume of 2000 µL. The first column of the plate was used for the blank (2000 µL of buffer solution and 40 µL of DTNB). Columns 2 and 3 were used to determine the initial GSH concentration (200 µL of GSH stock solution, 1200 or 1600 µL of buffer solution, 40 µL of DTNB, and 600 or 200 µL of test compound stock solution). To determine the absorbances, 300 µL of each well was transferred to a 370 µL 96 well plate and analyzed via UV-vis spectrometry at 412 nm (SpectraMax 384 Plus, Molecular Devices Corporation, United States). The remaining wells were prepared with the required amount of buffer solution, 200 µL of GSH stock solution, and the corresponding volume of the stock solution of the test compound. The reactions proceeded under isothermal conditions in a climate chamber with plate agitation. After certain reaction times (g5 min), 40 µL of DTNB solution was given into all wells of a column to stop the GSH degradation, and 300 µL was transferred to the 370 µL well plate for UV-vis analysis at 412 nm. Modification for Highly Reactive Compounds. For highly reactive contaminants (in our case, 1-penten-3-one, 1-hexen-3-one, and 1-octene-3-one), the total reaction volume was reduced to 1 mL, with small magnetic stirrers in each well to achieve an immediate mixing with the DTNB solution, allowing us to use smaller time slices (e2 min). Determination of GSH Concentration. The quantification of the GSH concentration was based on external calibration. To this end, the UV-vis absorbance was measured at different GSH concentrations ranging from 0.00152 to 0.239 mmol/L, resulting in a highly significant correlation between absorbance and GSH concentration (r2 ) 0.9996, s ) 2 × 10-6). Determination of the GSH concentration was based on eight replicates obtained from every column. The associated standard deviations were e0.05, demonstrating a very good repeatability. For every test compound, the reaction rate constant was determined from at least three replicates (separate experiments), taking 6-8 data points per rate plot for high- and low-reactive compounds, and 6-11 data points for compounds with intermediate reactivities. The associated r2 values were at least 0.99 in most cases and, for few compounds, 0.96-0.98. Reaction Rates. Rate constants of reaction between the GSH and the electrophilic test compounds, kGSH, were determined employing pseudo-first-order kinetics for less reactive electrophiles and second-order kinetics for the more reactive compounds. The -1 overall pseudo-first-order rate constant kpseudo GSH (min ) was quantified as the slope of the linear regression of ln cGSH - ln c0GSH on reaction time t. Then, the second-order rate constant, kGSH (L mol-1 min-1), was determined from kpseudo GSH and the initial test compound concentration, c0, according to pseudo 0 kGSH ) kGSH /c

(1)

For the more reactive compounds, regression of ln cGSH - ln c on reaction time t yielded slope k′, which was converted to the secondorder rate constant kGSH according to 0 - c0) kGSH ) k'/ (cGSH

(2)

The associated standard deviations of the means were calculated using Microsoft Excel. For all 26 test compounds, degradation half-lives due to their reaction with GSH were calculated according to second-order kinetics:

t1/2 )

kGSH

( )

cA0 1 ln 2 - 0 × (cB0 - cA0) cB

(3)

744

Chem. Res. Toxicol., Vol. 22, No. 4, 2009

Böhme et al.

0 where cA0 is the smaller value of cGSH (initial GSH concentration) and c0 (initial test compound concentration) and c0B is the respective larger value. GSSG Formation. The experimentally determined GSH degradation rate can be confounded by the concurrent GSH loss process through oxidative reaction with DMSO and diluted solution-phase oxygen (22, 23) forming GSSG, which becomes relevant for slow reactions (see below). Therefore, the pseudo-first-order rate constant of GSH loss through oxidation to GSSG was determined in the absence of a test compound by following the described experimental procedure. Ample replicates were performed to quantify the standard error of the oxidative GSH loss () standard error of the kinetic GSH assay). To investigate the dependence of GSSG formation on the chemoassay conditions in terms of DMSO, oxygen (that is assumed to partition from air into the solution according to its Henry’s law constant), and a possible cosolvency effect due to the presence of test compounds, four additional experiments were carried out. First, all solutions and vessels were flushed with argon to eliminate essentially all solution-phase oxygen, yielding the pseudo-first-order DMSO rate constant of the reaction of GSH and DMSO (kGSSG ). Second, DMSO was replaced by a respective amount of buffer solution, thus enabling us to quantify the reaction rate constant of GSSG 2 formation mediated solely through oxygen (kOGSSG ). Third, GSH loss was analyzed in the absence of both DMSO and solution-phase oxygen to confirm that there are no further side reactions of GSH. Fourth, GSSG formation was studied in the presence of additional nonreactive polar organics, covering the typical range of polarity of the electrophiles of interest. To this end, pyridine, 3-pyridinol, 3-propyne-1-ol, and 1-propanol were selected as respective test substances to mimic the possible cosolvency effect of the (more reactive) electrophilic test compounds. The GSH degradation half-life due to oxidation was calculated from the pseudo-first-order oxidation rate constant, kpseudo GSSG , according to

t1/2 )

ln 2 pseudo kGSSG

(4)

Statistical Parameters. To evaluate the statistical performance of linear regression equations, the following parameters have been employed: squared correlation coefficient, r2; predictive squared correlation coefficient evaluated through leave-one-out cross2 validation, qcv ; root-mean-square error of calibration, rms; crossvalidated root-mean-square error of prediction, rmscv; and F test value, F1,n-2 (with n ) number of compounds).

Results and Discussion Static GSH Chemoassay. Using the previously introduced experimental procedure (19, 20), RC50 (reaction concentration 50%) values have been determined for 15 R,β-usaturated ketones, nine acrylates, and two propiolates and compared with respective data from the literature as far as available. The results are summarized in Table 1 and representsto our best knowledgesthe first data independent from the original laboratory (19). As can be seen from Table 1, the thiol reactivity of the test compounds ranges from high (1-pentene-3-one, RC50 ) 0.09 mM) to low (methyl methacrylate, RC50 ) 74.1 mM). For ethyl methacrylate and methyl tiglate, no RC50 value could be determined under the 2 h static assay conditions. While agreement with the original data (19, 20) is good for most of the substances, some compounds such as 3-methyl-3-penten2-one, n-propyl acrylate, and methyl tiglate show larger deviations. In the context of interpreting toxicological findings, the presently determined interlaboratory RC50 accuracy of around 0.3 log units could be considered acceptable. Apart from that, however, the 2 h RC50 assay sensitivity appears to be limited

Table 1. RC50 (50% Reaction Concentration) Data for the Reaction of 26 Compounds with GSH within 120 Min (Static Chemoassay) Together with Associated Standard Errors (() and Corresponding Data from Literature RC50 (mM) this RC50 work ((mM) (mM) (20)

compound

CAS

1-pentene-3-one 1-hexene-3-one 1-octene-3-one methyl propiolate ethyl propiolate 3-hexyne-2-one 3-pentene-2-one 2-octene-4-one 2-cyclopentene-1-one 4-hexene-3-one 3-heptene-2-one 3-octene-2-one methyl acrylate 3-nonene-2-one ethyl acrylate n-propyl acrylate n-butyl acrylate 3-methyl-3-pentene-2-one 4-methyl-3-pentene-2-one 2-methyl-2cyclopentene-1-one methyl crotonate ethyl crotonate 3-methyl-3cyclopentene-1-one methyl methacrylate ethyl methacrylate methyl tiglate

1629-58-9 1629-60-3 4312-99-6 922-67-8 623-47-2 1679-36-3 625-33-2 4643-27-0 930-30-3 2497-21-4 1119-44-4 1669-44-9 96-33-3 14309-57-0 140-88-5 925-60-0 141-32-2 565-62-8 141-79-7 1120-73-6

0.09 0.08 0.11 0.14 0.13 0.16 0.21 0.22 0.58 0.28 0.43 0.63 0.42 0.44 0.47 0.49 0.55 5.17 31.4 8.40

0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.009 0.006 0.007 0.80 4.8 0.01

0.053 0.052 0.051 0.095 0.091 0.12 0.11

623-43-8 623-70-1 2758-18-1

32.6 34.8 26.9

0.5 0.2 0.5

25 23

80-62-6 97-63-2 6622-76-0

74.1 a a

1.1 a a

75 NRASb 6.1

0.34 0.46 0.55 0.52 0.80 0.80 10 28

a Within 120 min, 50% GSH degradation could not be achieved at the saturation limit of the test compound. b Reported in the literature as not reactive at saturation (20).

to an intermediate range of electrophilic reactivity for the following reasons: First, the low end of the detectable reactivity is determined by the water solubility (Sw) of the test chemical, which in turn limits the maximum amount of compound offered for reaction within 2 h. An example is ethyl methacrylate (Sw ) 0.048 mol/L), which was considered nonreactive previously (20). While its water solubility (0.048 mol/L) is 3-fold lower than the one of methyl methacrylate (0.150 mol/L), both compounds yield similar second-order rate constants (see Table 2). Another issue in this context is the reported RC50 value of 6.1 mM for methyl tiglate (20), a crotonate (and thus a 3-methyl acrylate) with log Kow ) 1.69 and Sw ) 0.033 mol/L. This value is very unlikely for the following reasons: The water solubility of methyl tiglate is still lower than the one of ethyl methacrylate (s.a.), and the additional methyl group at the β-carbon makes methyl tiglate significantly less susceptible for the Michael 1,4addition. Indeed, we could not obtain an RC50 value for this compound (see Table 1). Second, the RC50 assay is also less sensitive for differences in electrophilic reactivity at the high reactivity end. Examples of compounds with similar RC50 values but rate constants differing by a factor of ca. 10 are discussed in the next section. Third, some of the literature RC50 data (19, 20) are even lower than the minimum concentration required to degrade 50% of GSH, which is equal to 50% of the initial GSH concentration (0.07 mM). It demonstrates the problem associated with a statistical (rather than actual) RC50 determination, keeping in mind that the RC50 assay as published does not foresee to replenish the electrophile as it is consumed (which would in principle be possible but at the cost of making the experimental procedure significantly more complex).


Chem. Res. Toxicol., Vol. 22, No. 4, 2009 745

Table 2. Second-Order Reaction Rate Constants (kGSH) for the Reaction of 26 Carbonyl and Carboxyl Compounds with GSH in the Kinetic Chemoassay Together with Associated Standard Errors ((), log Kow, and Toxicity to Tetrahymena pyriformis in Terms of log EC50 (Effective Concentration Yielding 50% Growth Inhibition) compound

CAS

kGSH (L mol-1 min-1)

((L mol-1 min-1)

1-pentene-3-one 1-hexene-3-one 1-octene-3-one methyl propiolate ethyl propiolate 3-hexyne-2-one 3-pentene-2-one 2-octene-4-one 2-cyclopentene-1-one 4-hexene-3-one 3-heptene-2-one 3-octene-2-one methyl acrylate 3-nonene-2-one ethyl acrylate n-propyl acrylate n-butyl acrylate

1629-58-9 1629-60-3 4312-99-6 922-67-8 623-47-2 1679-36-3 625-33-2 4643-27-0 930-30-3 2497-21-4 1119-44-4 1669-44-9 96-33-3 14309-57-0 140-88-5 925-60-0 141-32-2

second-order kinetics (eq 2) 1261 1173 1074 117 105 80.0 26.7 26.1 25.6 24.2 12.5 11.4 11.4 10.8 10.6 10.2 8.54

3-methyl-3-pentene-2-one 4-methyl-3-pentene-2-one 2-methyl-2-cyclopentene-1-one methyl crotonate ethyl crotonate 3-methyl-2-cyclopentene-1-one methyl methacrylate ethyl methacrylate methyl tiglate

565-62-8 141-79-7 1120-73-6 623-43-8 623-70-1 2758-18-1 80-62-6 97-63-2 6622-76-0

pseudo-first-order kinetics (eq 1) 0.779d 0.208d 0.200d 0.164d 0.161d 0.074d 0.072d 0.058d 0.007d

63 30 13 8 5 1.3 0.9 0.5 0.1 0.2 0.3 0.2 0.3 0.2 0.1 0.3 0.20 0.018 0.007 0.001 0.005 0.002 0.002 0.004 0.005 0.001

log Kowa

log EC50b (mol/L)

0.90 1.39 2.37 0.09 0.58 0.52 0.82 2.29 0.71 1.31 1.80 2.29 0.73 2.79 1.22 1.71 2.20

-4.52 -4.66 -4.92 -4.77 -4.70 -4.32 -3.54 -4.01c -3.64 -3.93 -3.70 -3.74 -3.55 -3.98 -3.52 -3.53 -3.52

1.37 1.37 1.26 1.44 1.63 1.26 1.28 1.77 1.69

-2.66 -2.36 -2.17 -2.08 -2.24 -1.68 -1.78 -2.07 -2.21

a Log Kow calculated using KOWWIN v 1.67 (24). b Taken from the literature (25-27) and converted to mol/L. c EC50 (effective concentration yielding 50% growth inhibition of T. pyriformis) determined by Franziska Schramm (UFZ). d Corrected by the pseudo-first-order rate constant of pseudo oxidative GSH loss through GSSG formation due to reactions with DMSO and oxygen, kGSSG (see eq 5).

A fourth and more subtle problem is caused by the fact that to determine RC50, a set of reaction batches with increasing compound concentrations is required (19, 20). The latter implies a gradual variation in reaction medium through correspondingly changed cosolvent (DMSO) concentrations that affects the reactivity of both GSH and the electrophile. On the one hand, the rate of GSSG formation as confounding side reaction increases with increasing DMSO concentration (see below). On the other hand, varying DMSO concentrations translate into a varying solution polarity, which in turn is likely to affect the rates of both GSSG formation and GSH-electrophile reaction (because transition state formation is increasingly stabilized by increasing polarity). It follows further that with this assay, correction for GSSG formation as mentioned above (23) would require separate respective measurements for each cosolvent concentration of the reaction medium. Without this correction, GSH degradation through electrophilic attack by the test compound tends to be overestimated, which is increasingly important with decreasing electrophilic reactivity. This problem leads to a further issue that could also affect the reproducibility of 2 h RC50 measurements. For a given test compound, the initial concentration can be obtained either from a relatively large volume of a low-concentration stock solution or from a relatively small volume of a respective highconcentration stock solution. It follows that the same initial test compound concentration may be accompanied by significantly different DMSO concentrations, resulting in correspondingly different GSH loss rates after 2 h due to the different rates of GSSG formation through the GSH-DMSO reaction pathway (see below). Kinetic GSH Chemoassay. For the test set of 15 R,βunsaturated ketones and 11 R,β-unsaturated carboxylic acid esters (acrylates, methacrylates, crotonates, and propiolates), the

second-order rate constants for their reaction with GSH, kGSH (L mol-1 min-1), are listed in Table 2. Associated log Kow (logarithmic octanol/water partition coefficient) values have been calculated with KOWWIN (24), and data for the toxicity toward the ciliates Tetrahymena pyrifomis in terms of 48 h log EC50 (effective concentration to inhibit population growth by 50%) values have been collected from literature (25-27). The decision between pseudo-first-order kinetics (eq 1) and second-order kinetics (eq 2) was made as follows: For compounds with a change in concentration below 5%, kGSH was determined through application of eq 1, while for all other compounds, eq 2 was applied. As can be seen from Table 2, the latter group covers 17 compounds with kGSH larger than 8 L mol-1 min-1, and for the former group of nine compounds, kGSH is smaller than 1 L mol-1 min-1. Note further that for the latter subset of low-reactive electrophiles, kGSH was corrected for GSSG formation as outlined in the next section. For the total compound set, chemical reactivity in terms of kGSH spans more than 5 orders of magnitude, ranging from 1261 (1-pentene-3-one) to 0.007 L mol-1 min-1 (methyl tiglate). Note that kGSH represents a total rate constant (corrected for GSSG formation if relevant, see below), reflecting the total GSH loss due to 1,4-addition (Michael addition) and 1,2-addition of the electrophiles as outlined in Figure 1. Because the thiol group is a soft nucleophile, however, we assume that kGSH is in fact dominated by the 1,4-addition involving the β-carbon of the electrophile (that is significantly softer than the carbonyl carbon). In contrast to the RC50 assay (see above and Table 1), the electrophilic reactivity of ethyl methacrylate (kGSH ) 0.058 L mol-1 min-1) and methyl tiglate (kGSH ) 0.007 L mol-1 min-1) can now be quantified (see also Table 2). The former turns out to be close to the one of methyl methacrylate (kGSH ) 0.072 L mol-1 min-1). The latter is still a factor of 10 less reactive and

746


Böhme et al. DMSO resulted in the pseudo-first-order rate constant kGSSG ) 0.0741 -2 -1 × 10 min . It suggests that this route of DMSO-mediated GSSG formation proceeds according to

GSH

GSH + DMSO f GS• + DMS•OH 98 GSSG + DMS + H2O (6) Figure 1. Michael addition (conjugated 1,4-addition, top) and 1,2addition (bottom) of a nucleophile (Nu-H) to R,β-unsaturated ketones (Y d R d alkyl; R, H at R- and β-carbon), acrylates (Y d OR; in our study, H at R- and β-carbon), methacrylates (Y d OR; methyl at R-carbon; in our study, H at β-carbon), crotonates (Y d OR; H at R-carbon; methyl at β-carbon), tiglate (Y d OR; methyl at R- and β-carbon), and propiolates (CtC; Y d OR, H at R- and β-carbon). The reaction with GSH as nucleophile is assumed to proceed predominantly via 1,4-addition, thus setting kGSH ≈ k1,4 for the associated second-order rate constant.

may be close to the sensitivity limit of the kinetic assay at the low end of thiol reactivity. A further difference in reactivity order between the static and the kinetic GSH assay concerns methyl propiolate, ethyl propiolate, and 3-hexyne-2-one as compared to 1-octene-3-one. For these four compounds, RC50 ranges from 0.11 to 0.16 mM (see Table 1). By contrast, 1-octene-3-one is more reactive in terms of kGSH by a factor of at least 10 (kGSH, 1074 vs 80-117 L mol-1 min-1; see Table 2) and, in fact, turned out to be the third most reactive compound of the 26 electrophiles under present investigation. A similarly large difference between both assays is observed when comparing 1-pentene-3-one with 4-hexene-3-one. The former is more reactive toward GSH than the latter by a factor of 48 (kGSH, 1261 vs 24.2 L mol-1 min-1), as opposed to a factor of only 3.5 when employing RC50 values (0.08 vs 0.28 mM; Table 1). Correction for GSSG Formation. The extent of GSH oxidation by DMSO and oxygen (22, 23) was determined under kinetic assay conditions, using a blank variant (no electrophilic test compound) augmented by the typical DMSO/buffer mixture. For the experimental analysis, assumptions have been made as follows: First, the concentrations of DMSO (cDMSO ) 0.279 mol/ L) and diluted oxygen in solution (that is governed the ambient O2 concentration according to its Henry’s law constant) were taken as constant during the reaction time. Second, GSSG formation was considered to proceedsat least mainlysthrough reaction of GSH with either DMSO or oxygen (which was at least indirectly checked through additional experiments as outlined below). The kinetics of GSH oxidation can then be described according to eq 5:

-

dcGSH DMSO O2 × cGSH + kGSSG × cGSH ) kGSSG dt DMSO O2 ) (kGSSG + kGSSG ) × cGSH pseudo ) kGSSG × cGSH

(5)

The experimentally determined pseudo-first-order rate constant of the overall oxidative GSH loss through GSSG formation is pseudo ) 0.115 ((0.009) × 10-2 min-1, yielding a GSH halfkGSSG life of t1/2 ) 603 min (see eq 4) when employing (as usual for both the static and the kinetic assays) c0GSH ) 1.40 × 10-4 mol/L as the initial GSH concentration. Separate analysis of the GSH loss through reaction with DMSO at a fixed concentration (cDMSO ) 0.279 mol/L) without oxygen (achieved through flushing with argon as outlined above)

with the first reaction step being rate-determining for the overall process, yielding pseudo-first-order kinetics governed by kDMSO GSSG (see also the left-hand side of eq 5); in eq 6, DMS denotes dimethyl sulfide. Replacement of DMSO through a corresponding amount of buffer with oxygen remaining as the only oxidizing agent for GSH resulted again in pseudo-first-order kinetics for the GSH O2 loss and the associated rate constant kGSSG ) 0.0436 × 10-2 -1 min . Here, pseudo-first-order kinetics is compatible with a GSSG formation according to GSH

GSH + O2 f GS• + HOO• 98 GSSG + H2O2 (7) where again the first reaction step is assumed to be rateO2 (see also the left-hand determining and thus governed by kGSSG side of eq 5). DMSO O2 + kGSSG ) Summing up both individual rate constants (kGSSG -2 -1 gives 0.117 × 10 min , which is in excellent agreement with pseudo value in the presence of the experimentally determined kGSSG both DMSO and diluted oxygen (s.a.). This suggests further that under the chemoassay conditions, the oxidative GSSG formation is indeed governed by the reactions of both DMSO and solution-phase oxygen with GSH and that a further GSSG formation route is unlikely to play an important role. In relevant cases (see below and Table 2), kpseudo GSSG (eq 5) was subtracted from the initially determined overall pseudo-first-order rate constant kpseudo GSSG (that includes oxidative GSH loss) to derive kGSH according to eq 1. A final issue concerns the possible difference in solution polarity between the blank variant of the assay and its condition after addition of the test compound to run the actual measurement of thiol reactivity. As mentioned above, increasing polarity of the solution is expected to increasingly stabilize the transition state formation of the oxidative GSH degradation. In this context, it should be noted that GSH with a pKa of about 9.2 (28) is prevalent in aqueous solution as a neutral (undissociated) reactant at pH 7.4 (assay condition) to more than 90% and thus likely less polar than the transition-state structure formed during its oxidation. Accordingly, addition of a polar test compound could in principle accelerate (or at least modify) the GSSG formation rate, the latter of which has not yet been addressed so far. To investigate this aspect, an additional set of essentially nonreactive test compounds (pyridine, 3-pyridinol, 3-propyne1-ol, and 1-propanol) was selected such that it covers the relevant range of substrate polarity. The concentrations were chosen to mimic the experimental conditions with the electrophilic substances (see Table 3) and range from about 3.70 × 10-3 to about 4.10 × 10-2 mol/L. In the presence of these pseudo cosolvents, the individual kGSSG values (determined without replicates) ranged from 0.109 × 10-2 to 0.123 × 10-2 min-1, yielding a corresponding mean value of kpseudo GSSG of 0.114 ((0.006) × 10-2 min-1 in good agreement with the associated reference value of 0.115 ((0.009) × 10-2 min-1. These findings



Table 3. Kinetics of the GSH Loss in the Static Chemoassay of 24 Compounds, Accounting for Both the GSH-Electrophile Reaction and Oxidative GSSG Formationa compound 1-pentene-3-one 1-hexene-3-one 1-octene-3-one methyl propiolate ethyl propiolate 3-hexyne-2-one 3-pentene-2-one 2-octene-4-one 2-cyclopentene-1-one 4-hexene-3-one 3-heptene-2-one 3-octene-2-one methyl acrylate ethyl acrylate 3-nonene-2-one n-propyl acrylate n-butyl acrylate

kGSH × c0 × pseudo app c0 (mol/L) 10-2 (min-1) kGSSG /kGSH t1/2 (min)b second-order kinetics (eqs 2 and 8) 0.09 × 10-3 11.3 0.08 × 10-3 9.38 0.11 × 10-3 11.8 0.14 × 10-3 1.65 1.37 0.13 × 10-3 -3 0.16 × 10 1.28 -3 0.21 × 10 0.561 0.22 × 10-3 0.574 0.58 × 10-3 1.48 0.28 × 10-3 0.680 0.43 × 10-3 0.538 0.63 × 10-3 0.718 -3 0.42 × 10 0.483 -3 0.47 × 10 0.503 -3 0.44 × 10 0.475 -3 0.49 × 10 0.500 0.468 0.55 × 10-3

0.01 0.01 0.01 0.07 0.08 0.09 0.21 0.20 0.08 0.17 0.21 0.16 0.24 0.23 0.24 0.23 0.25

pseudo-first-order kinetics (eqs 1 and 10) 0.349 0.33 3-methyl-3-pentene-2-one 5.17 × 10-3 4-methyl-3-pentene-2-one 31.4 × 10-3 0.738 0.16 -3 methyl crotonate 32.6 × 10 0.535 0.22 -3 ethyl crotonate 34.8 × 10 0.560 0.21 -3 2-methyl-28.40 × 10 0.168 0.68 cyclopentene-1-one -3 3-methyl-226.9 × 10 0.202 0.57 cyclopentene-1-one -3 methyl methacrylate 74.1 × 10 0.496 0.23

12.9 19.7 9.88 69.4 76.2 73.6 154 149 50 119 142 103 159 151 149 151 160 200 94.1 130 124 415

(9)

as pseudo-first-order rate constant for the observed GSH degradation (see eq 1). The overall GSH loss kinetics accounting for both the pseudo-first-order GSH-electrophile reaction and the pseudo-first-order reactions of GSH with DMSO and with solution-phase oxygen can then be written as

dcGSH pseudo pseudo × cGSH + kGSSG × cGSH ) kGSH dt

140

For ethyl methacrylate and methyl tiglate, RC50 values could not be determined under the static chemoassay conditions within 120 min (see the last two rows of Table 1). The pseudo-first-order rate constant of pseudo GSH oxidation to GSSG (see eq 5) is kGSSG ) 0.115 ((0.009) × 10-2 0 min-1, and the initial GSH concentration at reaction time t ) 0 is cGSH ) 1.40 × 10-4 mol/L in both assays; c0 ) initial electrophile concentration (at t ) 0) ) RC50; kGSH × c0 ) pseudo-first-order rate constant of the loss of GSH due to its reaction with the electrophile at t pseudo app ) 0 (see eq 9) with kGSH being taken from Table 2; kGSSG /kGSH ) app contribution of GSSG formation to the overall GSH loss, where kGSH ) pseudo kGSH × c0 + kGSSG is the apparent overall rate constant of GSH loss due to both GSH-electrophile reaction and GSSG formation at t ) 0 (which, in the case of pseudo-first-order kinetics, holds also for t > 0 in app pseudo pseudo terms of kGSH ) kGSH + kGSSG ; see eq 10). Because the extrapolation procedure (19, 20) to determine RC50 implies variation of cDMSO that pseudo would make a precise determination of kGSSG difficult, cDMSO ) 0.279 mol/L of the kinetic assay conditions was applied as a rough approximation for the kinetic analysis of the static assay, keeping in pseudo mind that increasing cDMSO increases kGSSG (see text). b Half-lives t1/2 were calculated according to second-order kinetics (see eq 3), neglecting the concurrent GSH oxidation to GSSG.

demonstrate that for the assay conditions, polar cosolvency effects are unlikely to play an important role. Chemoassay Degradation Kinetics. In case of second-order kinetics for the overall GSH loss, the actual contributions of GSSG formation and GSH-electrophile reaction are triggered by the time-dependent concentrations of GSH (cGSH) and electrophile (c). This condition applies to the more reactive electrophiles, and here, the relevant differential equation can be written as

dcGSH pseudo × cGSH ) kGSH × c × cGSH + kGSSG dt pseudo ) (kGSH × c + kGSSG ) × cGSH

pseudo kGSH × c(t) ≈ kGSH × c0 ) kGSH

-

344

a

-

to quantify the contributions of both loss processes to the overall GSH degradation rate at a given time t, provided that c(t) is known. In this case of actual second-order kinetics, we will confine ourselves to comparing the initial rate of GSHelectrophile reaction with the concurrent GSSG formation. To this end, kGSH × c is quantified at t ) 0 [where c(0) ≡ c0 is the known initial concentration of the test compound]. Introducing app pseudo ≡ kGSH × c0 + kGSSG , the terms kGSH × the abbreviation kGSH 0 app pseudo app c /kGSH and kGSSG /kGSH then quantify the relative contributions of both the GSH-electrophile reaction and the GSSG formation to the overall GSH loss rate at the beginning (t ) 0) of the GSH degradation. For increasingly less reactive electrophiles, their correspondingly slow decrease in concentration with time can be increasingly well ignored, yielding

(8)

pseudo DMSO O2 ) kGSSG + kGSSG , the rate constant of GSSG (where kGSSG formation, has been defined through eq 5). Evaluation of the pseudo would enable us product term kGSH × c as compared to kGSSG

pseudo pseudo ) (kGSH + kGSSG ) × cGSH app ) kGSH × cGSH

(10)

app denoting the associated apparent first-order rate with kGSH pseudo app pseudo app /kGSH and kGSSG /kGSH quantify the constant. In this case, kGSH relative contributions of the GSH-electrophile reaction and GSSG formation to the overall GSH loss. In Table 3, the following details of the GSH loss kinetics under static chemoassay conditions are given for 24 compounds (omitting ethyl methacrylate and methyl tiglate where RC50 values could not be obtained within 120 min; see Table 1): kGSH × c0 (that equals kpseudo GSH in the case of pseudo-first-order kinetics pseudo app /kGSH (the for the GSH-electrophile reaction; see eq 9), kGSSG relative contribution of GSSG formation to the overall GSH loss rate at t ) 0, which remains constant for t > 0 in case of overall pseudo-first-order kinetics; see eq 10), and the half-life t1/2 calculated according to eq 3 (without taking into account the GSSG formation). In Table 4, the corresponding data are given for all 26 compounds under kinetic chemoassay conditions. As can be seen from Table 3, the relative contribution of oxidative GSSG formation to the overall GSH loss rate at t ) 0 is below 10% for only seven compounds, up to 25% (n-butyl acrylate) for the remaining 10 more reactive compounds with second-order kinetics (upper part of Table 3), and up to 68% (2-methyl-2-cyclopentene-1-one) for the seven less reactive compounds where pseudo-first-order kinetics could be applied pseudo app /kGSH directly (lower part of Table 3). In this latter case, kGSSG quantifies the overall error due to neglection of GSSG formation app [because kapp GSH(t) ) kGSH(0) under pseudo-first-order conditions]. Note further that the GSH half-lives resulting only from the GSH-electrophile reaction are above 120 min for 14 of the 24 compounds. Keeping in mind that the reaction is 1:1 stoichiometric, it follows that within the static assay duration of 2 h, the 50% degradation of GSH could not have been achieved solely through attack of the electrophile. It follows that for these substances and compounds with similar reactivities, RC50 values

748


Böhme et al.

Table 4. Kinetics of the GSH Loss in the Kinetic Chemoassay for All 26 Compounds, Accounting for Both the GSH-Electrophile Reaction and Oxidative GSSG Formationa compound

c0 (mol/L)

kGSH × c0 × pseudo app t1/2 k /k 10-2 (min-1) GSSG GSH (min)b

second-order kinetics (eqs 2 and 8) 1-pentene-3-one 1.96 × 10-4 24.7 1-hexene-3-one 1.93 × 10-4 22.6 1-octene-3-one 1.11 × 10-4 11.9 methyl propiolate 3.26 × 10-4 3.81 -4 2.70 ethyl propiolate 2.57 × 10 -4 3-hexyne-2-one 2.66 × 10 2.13 -4 3-pentene-2-one 1.15 × 10 3.07 2-octene-4-one 1.23 × 10-3 3.21 2-cyclopentene-1-one 1.44 × 10-3 3.67 4-hexene-3-one 1.41 × 10-3 3.41 3-heptene-2-one 1.29 × 10-3 1.61 methyl acrylate 2.04 × 10-3 2.33 -3 3-octene-2-one 1.78 × 10 2.03 -3 ethyl acrylate 1.38 × 10 1.46 -3 3-nonene-2-one 1.15 × 10 1.24 -3 n-propyl acrylate 1.86 × 10 1.90 2.31 n-butyl acrylate 2.71 × 10-3

0.01 0.01 0.01 0.03 0.04 0.05 0.04 0.04 0.03 0.03 0.07 0.05 0.06 0.08 0.09 0.06 0.05

pseudo-first-order kinetics (eqs 1 and 10) 1.15 0.09 3-methyl-3-pentene-2-one 1.48 × 10-2 4-methyl-3-pentene-2-one 4.08 × 10-2 0.849 0.12 -2 methyl crotonate 3.27 × 10 0.536 0.18 -2 ethyl crotonate 1.80 × 10 0.209 0.28 -2 2-methyl-24.17 × 10 0.834 0.12 cyclopentene-1-one -2 3-methyl-23.91 × 10 0.293 0.28 cyclopentene-1-one -2 methyl methacrylate 3.87 × 10 0.279 0.29 ethyl methacrylate 1.17 × 10-2 0.0679 0.63 methyl tiglate 1.62 × 10-2 0.0113 0.91

3.56 3.90 9.72 20.7 30.5 38.5 23.4 22.3 19.4 20.9 44.3 30.4 34.9 48.8 57.8 37.3 30.4 54.8 72.0 106 171 73.1 170 176 379 548

a The pseudo-first-order rate constant of GSH oxidation to GSSG (see pseudo eq 5) is kGSSG ) 0.115 ((0.009) × 10-2 min-1, and the initial GSH 0 concentration at reaction time t ) 0 is cGSH ) 1.40 × 10-4 mol/L in both assays; c0 ) initial electrophile concentration at t ) 0; kGSH × c0 ) pseudo-first-order rate constant of the loss of GSH due to its reaction with the electrophile at t ) 0 with kGSH being taken from Table 2; pseudo app kGSSG /kGSH ) contribution of GSSG formation to the overall GSH loss, app pseudo where kGSH ) kGSH × c0 + kGSSG is the apparent overall rate constant of GSH loss due to both GSH-electrophile reaction and GSSG formation at t ) 0 (which, in the case of pseudo-first-order kinetics, holds also for app pseudo pseudo t > 0 in terms of kGSH ) kGSH + kGSSG ; see eq 10). b The half-life was calculated according to second-order kinetics (see eq 3), neglecting the concurrent GSH oxidation to GSSG.

are likely to be underestimated if the concurrent oxidative GSSG formation is not taken into account. As mentioned above, the literature value of 6.1 mM reported (20) for methyl tiglate (last column of Table 1) appears to be unrealistic. In fact, our kinetic analysis yields t1/2 ) 16233 min for this compound, demonstrating that a proper RC50 cannot be attained within 120 min. For the three most reactive compounds 1-pentene-3-one (kGSH ) 1261 M-1 min-1; see Table 2), 1-hexene-3-one (kGSH ) 1173 M-1 min-1) and 1-octene-3-one (kGSH ) 1074 M-1 min-1), the initial loss due to GSSG formation provides only 1% of the pseudo app /kGSH ) 0.01, fourth total GSH degradation rate at t ) 0 (kGSSG column of Table 3). Here, the half-lives of the GSH loss due to the GSH-electrophile reaction (neglecting GSSG formation) are below 20 min (last column of Table 3). Thus, highly reactive compounds will have been degraded well before 2 h, leaving an appreciable time for oxidative GSSG formation after completion of the GSH-electrophile reaction, provided the initial electrophile concentration has been lower than the initial GSH concentration (which is the case for thesstatistically derived (19, 20)sRC50 values of these three compounds; see Table 1). Note, however, that under the chemoassay conditions including the c0GSH value, the maximum GSH degradation due to oxidation

alone is 1.73 × 10-5 mol/L (corresponding to 12% of the initial GSH concentration of 1.40 × 10-4 mol/L) after 2 h. Taking 1-hexene-3-one as an example with an initial test concentration of 0.8 × 10-4 mol/L () RC50; see Table 1) that 0 , 4 × t1/2 (calculated through iterative application is below cGSH of eq 3) yields 30 min as time where the compound concentration would have been lowered by 0.75 × 10-4 mol/L, corresponding to ca. 95% degradation. Correspondingly, cGSH would have been reduced to 0.65 × 10-4 mol/L at t ) 30 min when neglecting GSSG formation (and thus already below the final 2 h RC50, which is 0.7 × 10-4 mol/L by definition). For the remaining 90 min, GSH oxidation would yield an additional GSH decrease by about 0.06 × 10-4 mol/L (calculated from pseudo-first-order kinetics of GSH oxidation), corresponding to ca. 8% of the final RC50 value. These calculations that approximate the actual concurrent processes through a consecutive consideration of GSH-electrophile reaction and GSH oxidation suggest that for highly reactive compounds, the systematic error in RC50 caused by neglecting GSSG formation would be around 8%. Note further that the RC50 values of 1-pentene-3-one, 1-hexene-3-one, and 1-octene-3-one as reported in literature (19, 20) (see last column of Table 1) are even lower than 0.7 × 10-4 mol/L that would theoretically be required to degrade GSH by 50%, keeping in mind that the static assay does not include a procedure to replenish the electrophile as it is consumed (see the Materials and Methods). From this point of view, the RC50 error was in fact at least 24-27% (0.27 ) 1 - 0.051/0.07) for these compounds and is probably caused by both the statistical nature of the RC50 (extrapolation) and the additional GSH degradation due to oxidation. In Table 4, details of the degradation kinetics are summarized for the kinetic GSH chemoassay. Because c0 (selected individually to achieve good degradation kinetics) differs from RC50 (that by definition was c0 in the static assay), different values app ≡ for kGSH × c0 (with kGSH taken from Table 2) and for kGSH 0 pseudo kGSH × c + kGSSG were obtained, resulting in correspondingly different oxidation contributions to the overall GSH loss rate pseudo app /kGSH (4th column of Table 4). The at t ) 0 in terms of kGSSG latter are now below 10% for all 17 compounds evaluated with second-order kinetics and up to 91% (methyl tiglate) for the nine compounds evaluated with pseudo-first order kinetics. The differences in c0 between the static and the kinetic chemoassay yield also correspondingly different degradation half-lives t1/2 (see eq 3), which now are below 10 min for the three most reactive compounds, below 60 min for the group of 17 more reactive compounds (upper part of Table 4), and up to 548 min (methyl tiglate) for the remaining nine compounds (lower part of Table 4). As outlined above, the kinetic chemoassay includes a procedure to correct kGSH for oxidative GSSG formation. On the basis of our preliminary experience, we suggest to apply app this correction to the initially measured value kGSH when at t ) 0 the condition pseudo kGSSG app kGSH

g 0.10

(11)

-2 min-1) is met. Accordingly, GSSG (with kpseudo GSSG ) 0.115 × 10 formation was neglected for the kGSH determination of all 17 compounds evaluated with second-order kinetics and applied to kGSH of the nine less reactive compounds with pseudo-firstorder kinetics listed in the lower part of Table 4 (see also Table 2).


Figure 2. Log Kow vs log kGSH for the test set of 26 compounds as listed in Table 2. The subset of 14 compounds with log kGSH values in the range 1.00-2.20 is plotted with star symbols, and the (partly overlapping) subset of 22 compounds within 2.5 log Kow units is plotted with filled triangles.

The present findings demonstrate that without correction for GSSG formation, both RC50 and kGSH could represent apparent values. With the static assay, RC50 tends to be underestimated for both low-reactive compounds and highly reactive substances. We estimate the associated errors (only due to neglecting GSSG formation) to be around 10-30% for the former and around 8% for the latter. With the kinetic assay, a convenient procedure to correct kGSH for GSSG formation is available (exploiting also the fact that cDMSO can easily be hold constant), complementing its larger sensitivity with respect to the detectable range of thiol reactivity. Relationship of kGSH with Reactive Toxicity. For the 26 test compounds, log Kow spans 2.7 units (from 0.09 to 2.79) and log EC50 (48-h Tetrahymena pyriformis growth inhibition) (26-28) spans 3.2 units (from -1.68 to -4.92; Table 2). However, regression of log EC50 on log Kow yields poor statistics (r2 < 0.01). It indicates that for this compound set, the variation in aquatic toxicity is not driven by the variation in membrane affinity. In Figure 2, log Kow is plotted against log kGSH. While the overall correlation is low (r2 ) 0.02), two partly overlapping subsets can be identified: For 14 compounds (star symbols in Figure 2), log kGSH varies only between 1.00 and 2.20 as opposed to a log Kow variation over the whole value range. The partly overlapping other subset (filled triangles in Figure 2) contains 22 compounds that cover the full log kGSH range of ca. 5 units but only 2.5 log Kow units. Interestingly, most of the compounds of the former subset (except 3-hexyne-2-one, methyl propiolate, and ethyl propiolate) are associated with a quite narrow toxicity range (-3.5 g log EC50 g -4.0), while for the latter, log EC50 varies over the whole range observed for the total set. Regression of log EC50 on log kGSH yields

Log EC50 (M) ) -0.673 ((0.042) ×


Figure 3. Toxicity toward T. pyriformis in terms of log EC50 (48 h growth inhibition 50%) vs electrophilic reactivity quantified through log kGSH for the test set of 26 compounds listed in Table 2. The associated linear regression relationship is given in eq 12 and yields r2 ) 0.91 and rms ) 0.30 log units.

optimistic characterization of the correlation between log EC50 and log kGSH. Nevertheless, the analysis demonstrates that for the present set of 26 R,β-unsaturated ketones, acrylates, and propiolates, the overall variation in toxicity is mainly driven by differences in their electrophilic reactivity as quantified through log kGSH. In a previous study with 41 R,β-unsaturated carbonyls, a corresponding relationship has been reported for the correlation between log EC50 and log RC50 (r2 ) 0.85) (20). Note further that when exploring eq 12 for the predictive screening of compounds containing the relevant structural moiety, its mechanistic domain is limited to the Michael addition. Relationship between RC50 and kGSH. In the intermediate reactivity range, RC50 is directly related to kGSH under pseudofirst-order conditions. Indeed, omission of seven compounds at the high and low reactivity end (0.21 mM e RC50 e 34.8 mM) as well as of three further outliers (2-cyclopentene-1-one, 2-methyl-2-cyclopentene-1-one, and 3-methyl-2-cyclopentene1-one) yields the following regression equations (employing currently determined RC50 values):

Log RC50 (mM) ) -0.996 ((0.023) × log kGSH (M-1 min-1) + 0.729 ((0.024) (13) 2 ) 0.99, rms ) 0.07, rmscv ) where n ) 14, r2 ) 0.99, qcv 0.08, and F1,12 ) 1803.

Log kGSH (M-1 min-1) ) -0.998 ((0.023) × log RC50 (mM) + 0.732 ((0.019) (14) 2 ) 0.99, rms ) 0.07, rmscv ) where n ) 14, r2 ) 0.99, qcv 0.08, and F1,12 ) 1803. In the case of second-order kinetics, introduction of RC50 ) c0 (according to the definition of RC50) in eq 3 leads to

log kGSH (M-1 min-1) - 2.877 ((0.067) (12) 2 ) 0.89, rms ) 0.30, rmscv ) where n ) 26, r2 ) 0.91, qcv 0.34, and F1,24 ) 257. The associated data in Figure 3 reveal a somewhat clustered distribution, indicating that the formally derived calibration and leave-one-out cross-validation statistics tend to provide a too

kGSH )

t1/2

(

)

0 cGSH 1 RC50 ln 2 ≡ kGSH 0 RC × (RC50 - cGSH) 50

(15)

with t1/2 ) 120 min (static assay condition). For the 14 compounds, the resultant correlation between experimental log

750


RC50 kGSH (Table 2) and log kGSH predicted from (currently determined) experimental RC50 (Table 1) and the initial GSH concentration through eq 15 yields

Log kGSH (M-1 min-1) ) 0.998 ((0.024) × RC50 (M-1 min-1) + 0.028 ((0.025) (16) log kGSH 2 ) 0.99, rms ) 0.07, rmscv ) where n ) 14, r2 ) 0.99, qcv 0.08, and F1,12 ) 1802, and with both slope and intercept being close to their theoretically expected values of 1 by 0, respectively. Thus, eqs 13-16 can be used to convert RC50 into kGSH (and vice versa), provided that the above-mentioned underlying assumptions apply.

Conclusions Reactivity can determine the mechanism of toxic action of chemical substances, and govern the level of their toxicity. In this context, R,β-unsaturated carbonyl and carboxylic acid ester compounds form a prominent class of Michael acceptors that may impair the functioning of endogenous macromolecules through covalent binding at electron-rich sites. The newly introduced kinetic GSH chemoassay offers an efficient way to quantify electrophilic reactivity and, in this way, to identify compounds that have substantial potential for excess toxicity. Because of the kinetic approach with determining reaction rate constants rather than concentration values yielding a 50% reaction within a fixed exposure time, previous limitations with respect to solubility, range of reactivity, and the possibility to correct for GSH oxidation by DMSO and diluted oxygen as additional loss processes could be removed. The resultant second-order rate constants, kGSH, provide a quantitative measure of electrophilic reactivity based on the thiol scale and can be used as screen for excess toxicity prior to biological testing. As such, the kinetic GSH chemoassay may prove useful as a component of ITS for chemical substances, both for industrial in-house screening and under the framework of the new European chemical legislation REACH. Acknowledgment. Thanks are due to our colleagues Franziska Schramm, Elke Bu¨ttner, and Uwe Schro¨ter for technical support and assistance with some experiments. This study was financially supported by the European Union through the projects CAESAR (contract no. SSPI-022674-CAESAR) and OSIRIS (contract no. GOCE-CT-2007-037017).

References (1) Lipnick, R. L. (1995) Structure-activity relationships. In Fundamentals of Aquatic Toxicology, 2nd ed. (Rand, G. M., Ed.) pp 609-655, Taylor & Francis, Washington, DC. (2) Lipnick, R. L. (1991) Outliers: Their origin and use in the classification of molecular mechanisms of toxicity. Sci. Total EnViron. 109/110, 131–153. (3) Von der Ohe, P. C., Ku¨hne, R., Ebert, R.-U., Altenburger, R., Liess, M., and Schuü¨rmann, G. (2005) Structural alertssA new classification model to discriminate excess toxicity from narcotic effect levels of organic compounds in the acute daphnid assay. Chem. Res. Toxicol. 18, 536–555. (4) Van Leeuwen, K., Vermeire, T., Eds. Risk Assessment of Chemicals: An Introduction, 2nd ed., 686 pp, Springer, Dordrecht, The Netherlands. (5) Hermens, J. L. M., Busser, F., Leeuwangh, P., and Musch, A. (1985) Quantitative correlation studies between the acute lethal toxicity of 15 organic halides to the guppy (Poecilia reticulata) and chemical reactivity towards 4-nitrobenzylpyridine. Toxicol. EnViron. Chem. 9, 219–236.

Böhme et al. (6) Deneer, J. W., Seinen, W., and Hermens, J. L. M. (1988) The acute toxicity of aldehydes to the guppy. Aquat. Toxicol. 12, 185–192. (7) Deneer, J. W., Sinnige, T. L., Seinen, W., and Hermens, J. L. M. (1988) A quantitative structure-activity relationship for the acute toxicity of some epoxy compounds to the guppy. Aquat. Toxicol. 13, 195–204. (8) Schuü¨rmann, G. (1990) QSAR analysis of the acute toxicity of organic phosphorothionates using theoretically derived molecular descriptors. EnViron. Toxicol. Chem. 9, 417–428. (9) Karabunarliev, S., Mekenyan, O. G., Karcher, W., Russom, C. L., and Bradbury, S. P. (1996) Quantum-chemical descriptors for estimating the acute toxicity of electrophiles to the fathead minnow (Pimephales promelas): An analysis on molecular mechanisms. Quant. Struct.-Act. Relat. 15, 302–310. (10) Schmitt, H., Altenburger, R., Jastorff, B., and Schuü¨rmann, G. (2000) Quantitative structure-activity analyses of the algae toxicity of nitroaromatic compounds. Chem. Res. Toxicol. 13, 441–450. (11) Schuü¨rmann, G., Aptula, A. O., Ku¨hne, R., and Ebert, R.-U. (2003) Stepwise discrimination between four modes of toxic action of phenols in the Tetrahymena pyriformis assay. Chem. Res. Toxicol. 16, 974– 987. (12) Patlewicz, G., Roberts, D. W., and Uriarte, E. (2008) A comparison of reactivity schemes for the prediction skin sensitization potential. Chem. Res. Toxicol. 21, 521–541. (13) Karlberg, A.-T., Bergstro¨m, M. A., Bo¨rje, A., Luthman, K., and Nilsson, J. L. G. (2008) Allergic contact dermatitissFormation, structural requirements, and reactivity of skin sensitizers. Chem. Res. Toxicol. 21, 53–69. (14) Mekenyan, O. G., Todorov, M., Seafimova, R., Stoeva, S., Aptula, A., Finking, R., and Jacob, E. (2007) Identifying the structural requirements for chromosomal aberration by incorporating molecular flexibility and metabolic activation of chemicals. Chem. Res. Toxicol. 20, 1927–1941. (15) Esterbauer, H., Zoolner, H., and Scholz, N. (1975) Reaction of glutathione with conjugated carbonyls. Z. Naturforsch. 30c, 466–473. (16) McCarthy, T. J., Hayes, E. P., Schwartz, C. S., and Witz, G. (1994) The reactivity of selected acrylate esters toward glutathione and deoxyribonucleosides in Vitro: Structure-activity relationships. Fundam. Appl. Toxicol. 22, 543–548. (17) Clark, E. D., Greenshow, D. T., and Adams, D. (1998) Metabolism related assays and their application to agrochemical research: Reactivity of pesticides with glutathione and glutathione transferase. Pestic. Sci. 54, 385–393. (18) Freidig, A. P., Verhaar, H. J. M., and Hermens, J. L. M. (1999) Quantitative structure-property relationships for chemical reactivity of acrylates and methacrylates. EnViron. Toxicol. Chem. 18, 1133– 1139. (19) Schultz, T. W., Yarbrough, J. W., and Johnson, E. L. (2005) Structureactivity relationships for reactivity of carbonyl-containing compounds with glutathione. SAR QSAR EnViron. Res. 16, 313–322. (20) Yarbrough, J. W., and Schultz, T. W. (2007) Abiotic sulfhydryl reactivity: A predictor of aquatic toxicity for carbonyl-containing R,βunsaturated compounds. Chem. Res. Toxicol. 20, 558–562. (21) Everett, J. L., and Ross, W. C. J. (1949) Aryl-2-halogenoalkylamines. Part II. J. Chem. Soc. 1972–1983. (22) Homer, N. Z. M., Reglinski, J., Sowden, R., Spickett, C. M., Wilson, R., and Walker, J. J. (2005) Dimethylsulfoxide oxidizes glutathione in vitro and in human erythrocytes: Kinetic analysis by 1H NMR. Cryobiology 50, 317–324. (23) Misra, H. P. (1974) Generation of superoxide free radical during the autooxidation of thiols. J. Biol. Chem. 249, 2151–2155. (24) Meylan, W. M. (2004) KOWWIN 1.67, Syracuse Research Corporation, Syracuse, NY. (25) Schultz, T. W., and Yarbrough, J. W. (2004) Trends in structuretoxicity relationships for carbonyl-containing R,β-unsaturated compounds. SAR QSAR EnViron. Res. 15, 139–146. (26) Schultz, T. W., Netzeva, T. I., Roberts, D. W., and Cronin, M. T. D. (2005) Structure-toxicity relationships for the effect to Tetrahymena pyriformis of aliphatic, carbonyl-containing, R,β-unsaturated chemicals. Chem. Res. Toxicol. 18, 330–341. (27) Netzeva, T. I., Schultz, T. W., Aptula, A. O., and Cronin, M. T. D. (2003) Partial least square modelling of the acute toxicity of aliphatic compounds to Tetrahymena pyriformis. SAR QSAR EnViron. Res. 14, 265–283. (28) Jung, G., Breitmaier, E., and Voelter, W. (1972) Dissoziationsgleichgewichte von Glutathion. Eine Fourier-Transform-13C-NMR spektroskopische Untersuchung der pH-Abha¨ngigkeit der Ladungsverteilung. Eur. J. Biochem. 24, 438–445.

TX800492X

Kinetic Glutathione Chemoassay To Quantify Thiol Reactivity of

Recommend Documents