- Email: [email protected]
A Critical Evaluation of the Effect of Sorbitol on the Ferric–Xylenol Orange Hydroperoxide Assay
Analytical Biochemistry 284, 217–220 (2000) doi:10.1006/abio.2000.4696, available online at http://www.idealibrary.com on
A Critical Evaluation of the Effect of Sorbitol on the Ferric–Xylenol Orange Hydroperoxide Assay Craig Gay and Janusz M. Gebicki 1 Department of Biological Sciences, Macquarie University, Sydney 2109, Australia
Received January 4, 2000
Measurement of hydroperoxide concentration by the ferric–xylenol orange assay has the advantages of simplicity, convenience, and inertness to oxygen (Gay, C., et al. (1999) Anal. Biochem. 273, 149 –155). However, its sensitivity is limited by the molar absorption coefficients, which are less than 50,000 M ⴚ1 cm ⴚ1 for most hydroperoxides. An earlier report showed that this could be significantly enhanced by the inclusion of 100 mM sorbitol in the assay solution, resulting in an increase of the apparent absorption coefficient for H 2O 2 from 4.46 ⴛ 10 4 to 2.24 ⴛ 10 5 M ⴚ1 cm ⴚ1 (Wolff, S. P. (1994) Methods Enzymol. 233, 182–189). It was claimed that the technique was also valid for other hydroperoxides, such as butyl and cumyl. In an attempt to extend this modification to a wider range of hydroperoxides, we have confirmed the enhancement of the assay of H 2O 2 by sorbitol. However, the sensitivity of the measurements of butyl, cumyl, amino acid, protein, and human blood serum hydroperoxides was only approximately doubled by the inclusion of sorbitol. A mechanism explaining the difference in the assay of H 2O 2 and the organic hydroperoxides is proposed. © 2000 Academic Press
The measurement of hydroperoxide concentrations based on the oxidation of ferrous iron in acid solution and the formation of a colored complex between the ferric iron generated and xylenol orange has the advantages of ease of application, reasonable sensitivity, and independence of the presence of oxygen. Unfortunately, one of the earlier and most widely used protocols (FOX 2) (1) proved to have significant limitations. These were largely overcome by a series of modifications recommended in a recent study, which gave new procedures able to produce reliable estimates of the concentrations of a wide range of hydroperoxides in aqueous or alcohol solutions (2, 3). 1
Fax: 61 02 9850 8245. E-mail: [email protected].
0003-2697/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
However, the lower limit of measurement remained at about 1 nmol of hydroperoxide. The finding that the FOX 2 procedure had to be modified prompted us to examine its promising variant, FOX 1 (1). In this, the amount of ferric iron generated by H 2O 2 and measured by the xylenol orange complex was enhanced about 5-fold by the inclusion of sorbitol (D-glucitol) in the reaction mixture (1). It was claimed that the technique could also be applied to other water-soluble hydroperoxides, such as butyl and cumyl. We now confirm the enhancing action of sorbitol in the assay of H 2O 2 but find that it only approximately doubles the sensitivity of measurement of a range of organic hydroperoxides. Two reaction sequences are proposed, which can account for these differences. MATERIALS AND METHODS
All chemicals were of at least analytical grade and used without further purification. Glassware was cleaned in hot concentrated nitric acid and thoroughly rinsed with water passed through a four-stage purification system with a 0.2-m final filter (Millipore-Waters, Sydney). This procedure was found to be essential for the removal of traces of iron. Hydrogen peroxide and sulfuric acid were from Aldrich Chemical Co., Sydney, and the ferrous and ferric ammonium sulfates from BDH Chemicals Pty. Ltd., Sydney. tert-Butyl and cumyl hydroperoxides were from Koch-Light (Colnbrook, UK), and xylenol orange [o-cresosulfonphthalein-3,3-bis(sodium methyliminodiacetate)] and BSA 2 (fatty acid free) from Sigma (St. Louis, MO). Sorbitol (99%⫹) was supplied by the Aldrich Chemical Co. (Milwaukee, WI) and used directly. 2 Abbreviations used: XO, xylenol orange; Fe–XO, ferric–xylenol orange complex; –OOH, the hydroperoxide group; ROOH, organic hydroperoxide; t-Bu–OOH, tert-butyl hydroperoxide; Cu–OOH, cumene hydroperoxide; BSA, bovine serum albumin.
217
218
GAY AND GEBICKI
Aerobic Hydroperoxide Assay Solutions containing the hydroperoxide were mixed with the appropriate volumes of reagents to give final concentrations of 25 mM H2SO 4, 100 –150 M xylenol orange, and 100 –250 M ferrous iron in a volume of 2 ml. The final pH was 1.8 ⫾ 0.5. After 30 min in the dark, the absorbance was read at 560 nm, with XO/Fe2⫹ as blank. We found that even the most carefully prepared 5 mM ferrous ammonium sulfate solutions contained about 1 M ferric iron, present in the solid salt. RESULTS AND DISCUSSION FIG. 1. Effect of added sorbitol on the net absorbance of the ferric– xylenol orange complex generated by H 2O 2. The color was produced in the standard aerobic assay (Materials and Methods) and in the presence of the final concentrations of sorbitol shown. The results shown were averaged from three measurements, with standard deviations falling within the data points.
Hydroperoxide Solutions Stock solutions of H2O 2, tert-butyl hydroperoxide, and cumene hydroperoxide were prepared by dissolving 20 l of the reagent, as supplied, in 100 ml of water. The solutions were stable for at least 2 weeks at 4°C. Serum samples were prepared from blood obtained from healthy volunteers by venipuncture and collected in Vacutainer serum separation tubes (Bioclone Australia, Sydney). Blood cells were removed by 10-min centrifugation at 600g and the serum collected as clear fluid above the Vacutainer gel plug. If not used immediately, the serum was stored in 1.5-ml aliquots at ⫺80°C. Hydroperoxide groups were generated in amino acids by exposure of 20 mM aqueous solutions to 60Co radiation at a dose rate of 20 Gy min⫺1, with occasional stirring to ensure air saturation. Undiluted serum and solutions containing 2 mg ml ⫺1 of BSA were oxidized similarly, but with slow oxygen bubbling. The H2O 2 produced by the radiation was removed by 260 U of catalase per milliliter of liquid.
Measurement of the effect of concentration of sorbitol on the absorbance of the Fe–XO complex shows that maximum and stable color developed between 50 and 400 mM sorbitol (Fig. 1). Higher concentrations were not tested, and in subsequent experiments 100 mM concentration was used, as recommended in the FOX 1 protocol (1). The full absorbance developed in 15 min at room temperature (data not shown). Tests with increasing concentrations of H 2O 2 in the standard assay solutions supplemented with 100 mM sorbitol (Fig. 2) showed a spectacular increase in the yield of the Fe–XO complex in the presence of sorbitol. Since the reaction with Fe2⫹ under conditions of this assay gives rise to 2.2 Fe3⫹ ions per H 2O 2 (3), in the presence of sorbitol this yield increased to 20.6 Fe 3⫹ per H 2O 2, or almost 10-fold. The discrepancy between this result and the value of 5-fold increase reported for FOX 1 (1) may be due to differences in the lengths of chain reactions producing these high yields of Fe 3⫹, caused by terminating impurities, or to a problem in measurement of the concentration of H2O 2 in the earlier work. The determination of the precise yield of Fe3⫹ under different conditions may need further study, but it is clear that the presence of sorbitol greatly enhanced the
Anaerobic Hydroperoxide Assay The method was based on a published procedure (4). Quartz spectrophotometer cuvettes equipped with closefitting Teflon stoppers were flushed with oxygen-free nitrogen and filled with 2.0 ml of thoroughly degassed 1:2 acetic acid:water solution containing 10% KI (w/v). The initial absorbance was measured at 358 nm. Up to 500 l of the hydroperoxide solution was then added, and the cuvette was flushed again with nitrogen for 10 s, restoppered, and incubated in the dark at 50°C for 20 min. After cooling, the absorbance was measured at 358 nm and the hydroperoxide concentration calculated from the difference between the initial and final readings, using ⑀ ROOH ⫽ 2.9 ⫻ 10 4 M ⫺1 cm ⫺1.
FIG. 2. Net absorbance of solutions of H 2O 2 in the standard Fe–XO assay in the absence (open circles) and presence (closed circles) of 100 mM sorbitol. The correct concentration of the H 2O 2 was determined by the anaerobic assay (Materials and Methods). The standard deviations of four measurements fall within the data points.
SORBITOL EFFECT ON HYDROPEROXIDE ASSAY
FIG. 3. Net absorbance of the Fe–XO complexes generated by tertbutyl and cumene hydroperoxides in the absence (lower line) and presence (upper line) of 100 mM sorbitol. Data are represented by tert-butyl hydroperoxide without sorbitol (open circles), cumene hydroperoxide without sorbitol (open triangles), tert-butyl hydroperoxide with sorbitol (closed circles), and cumene hydroperoxide with sorbitol (closed triangles). Hydroperoxide concentrations were measured by the anaerobic assay. The results are averaged from three experiments, with standard deviations falling within the data points.
sensitivity of the measurement of H2O 2 by the Fe–XO assay. This was not the case for the other hydroperoxides studied. The results shown in Fig. 3 and Table 1, obtained with cumene and tert-butyl hydroperoxides, show approximately a doubling of the amount of Fe3⫹ normally generated by the inclusion of 100 mM sorbitol in the assay solution. Virtually the same enhancement was obtained with BSA–OOH and a somewhat higher one with the amino acids (Table 1, graphs not shown). Finally, the presence of sorbitol led to a similar increase in the levels of hydroperoxide residues measured in oxidized serum (Fig. 4). In the case of serum, it is not possible to express the results in terms of yield of Fe3⫹ per hydroperoxide group, since no independent assay of the latter was possible. However, the enhancement in the formation of the
TABLE 1
Enhancement of the Apparent Molar Absorption Coefficients of Hydroperoxides by Sorbitol Hydroperoxide
⫺Sorbitol
H 2O 2 44,400 ⫾ 1,400 tert-Butyl–OOH 96,900 ⫾ 4,000 Cumene–OOH 103,800 ⫾ 4,000 Valine–OOH 51,600 ⫾ 850 Lysine–OOH 52,100 ⫾ 360 BSA–OOH 35,900 ⫾ 700
⫹Sorbitol
Enhancement factor
415,600 ⫾ 1,300 209,300 ⫾ 4,200 196,100 ⫾ 4,600 141,200 ⫾ 4,300 152,300 ⫾ 850 81,000 ⫾ 500
9.36 1.90 2.16 2.74 2.92 2.26
Note. All coefficients are in units M ⫺1 cm ⫺1 and represent the average of at least three determinations and their standard deviations. Those measured in the absence of sorbitol lie within the ranges of corresponding values determined in an earlier study (3).
219
FIG. 4. Net absorbance produced in the Fe–XO assay by oxidized human blood serum in the absence (open circles) and presence (closed circles) of 100 mM sorbitol. The serum was irradiated in a ␥ source (Materials and Methods) and the volume added to the Fe–XO assay solution shown on the x axis. The absolute peroxide concentrations could not be measured because the iodide assay is affected by the high serum protein concentration. The results are from a single experiment, representative of several, each producing slightly different slopes because each serum sample contained different amounts of hydroperoxides after irradiation.
Fe–XO complex in serum by the sorbitol was 2.29, similar to that of BSA–OOH, which agrees with our findings that protein hydroperoxides are an early product generated in irradiated serum (5). The mechanism previously suggested for the formation of the high yield of Fe 3⫹ in the presence of sorbitol postulated a chain reaction (1). The statement that the FOX 1 assay could be used not only for H2O 2 but also for other water-soluble hydroperoxides (1) implies that all would yield similar amounts of Fe3⫹ per –OOH group and that the same mechanism should be valid generally. Our results show that this is incorrect (Table 1) and that the reactions initiated by reduction of H2O 2 by Fe 2⫹ in the presence of sorbitol are uniquely different from those of the organic hydroperoxides. This is caused by differences in the reactivities of free radicals formed in the initial reduction of the different hydroperoxides. In the case of H 2O 2, the following sequence of reactions can account for the experimental results: H2O2 ⫹ Fe 2⫹ 3 HO • ⫹ OH ⫺ ⫹ Fe 3⫹ ,
[1]
HO • ⫹ R2CHOH 3 H2O ⫹ R2C •OH,
[2]
R2C •OH ⫹ O2 3 R2C共OO • 兲OH,
[3]
R2C共OO • 兲OH 3 R2CAO ⫹ HO 2• ,
[4]
2HO 2• 3 H2O2 ⫹ O2 ,
[5]
where R 2CHOH is the sorbitol molecule, CH 2OH (CHOH) 4CH 2OH. The C represents any one of the six
220
GAY AND GEBICKI
carbon positions in sorbitol, which the hydroxyl radical attacks randomly at close to diffusion-controlled rate. Since sorbitol is the most abundant molecule present, it will be the main target. Reaction [1] is the wellknown Fenton process. Reaction [3] has a rate constant of about 2 ⫻ 10 9 M ⫺1 s ⫺1 (6), while that of [4] is 2700 s ⫺1 in acid solution (7). The H 2O 2 regenerated in reaction [5] can propagate a chain reaction through [1]. The reduction of the organic hydroperoxides does not proceed by a chain mechanism. Unlike the random reaction of the HO • with sorbitol (reaction [2]), the RO • reacts primarily at the terminal carbons of sorbitol (C in the formula RCH 2OH), producing the following reaction sequence: ROOH ⫹ Fe 2⫹ 3 RO • ⫹ HO ⫺ ⫹ Fe 3⫹ ,
[6]
RO • ⫹ RCH2OH 3 ROH ⫹ RC •HOH,
[7]
RC •HOH ⫹ O2 3 RCH共OO • 兲OH,
[8]
RCH共OO • 兲OH ⫹ Fe 2⫹ ⫹ H ⫹ 3 RCH共OOH兲OH ⫹ Fe 3⫹ ,
[9]
RCH共OOH兲OH ⫹ Fe 2⫹ 3 RCH共O • 兲OH ⫹ Fe 3⫹ ⫹ OH ⫺ ,
[10]
RCH共O • 兲OH ⫹ Fe 2⫹ ⫹ H ⫹ 3 RCH共OH兲OH ⫹ Fe 3⫹ ,
report demonstrated that Cu–OOH reduction by Fe 2⫹ gives a methyl radical, which enters a series of 19 reactions, resulting in the formation of several Fe 3⫹ ions for every –OOH group present (9). A similar process can be assumed for the t-Bu–OOH, since it also generates the methyl radical on reaction with Fe 2⫹. Extensive search of the literature did not reveal any reaction mechanism that could explain the doubling of the yield of Fe 3⫹ from these hydroperoxides by sorbitol (Table 1). In the case of BSA, amino acids, and serum, the higher than expected doubling of the yield of Fe 3⫹ by sorbitol (reactions [6]–[11]) was probably also caused by side reactions of free radicals known to be generated in the reduction of protein hydroperoxides (10). All these reactions are difficult to control and are likely to give different sorbitol enhancement factors under different conditions. Separate experiments also showed that sorbitol had no effect on the molar absorption coefficient of the Fe–XO complex. The main results of this study are the confirmation of the ability of sorbitol to enhance significantly the sensitivity of the Fe–XO assay of H 2O 2 and the finding of a much smaller enhancement for organic hydroperoxides. It seems likely that, in most practical applications, the possible complications resulting from inclusion of a high concentration of sorbitol in solutions of hydroperoxides other than H 2O 2 will not make this modification worthwhile. ACKNOWLEDGMENTS
[11]
This study was supported by Macquarie University grants. C.G. is holder of an Australian Postgraduate Award.
ROOH ⫹ 4Fe 2⫹ ⫹ 2H ⫹ ⫹ O2 ⫹ RCH2OH 3 ROH ⫹ 4Fe 3⫹ ⫹ 2OH ⫺ ⫹ RCH共OH兲OH. This gives an overall stoichiometry of 4 Fe 3⫹ ions for every –OOH reacting. Since each ROOH group gives normally close to 2 Fe 3⫹ ions (3), the enhancement by sorbitol of the yield of Fe 3⫹ should be about 2 (Table 1). The elimination of HO 2• from the ␣ and positions in sorbitol is slower than from the other sites in the molecule (8), allowing reaction [9] to compete effectively with [4]. At present, both these mechanisms are speculative, but have the potential to explain the experimental observations. There is little doubt that, especially as some very reactive species are generated in the presence of Fe 2⫹, reactions not considered here contribute to the overall yield of Fe 3⫹, accounting for the occurrence of sorbitol enhancement factors greater than 2. In the case of cumene hydroperoxide, an earlier
REFERENCES 1. Wolff, S. P. (1994) Methods Enzymol. 233, 182–189. 2. Gay, C., Collins, J., and Gebicki, J. M. (1999) Anal. Biochem. 273, 143–148. 3. Gay, C., Collins, J., and Gebicki, J. M. (1999) Anal. Biochem. 273, 149 –155. 4. Hicks, M., and Gebicki, J. M. (1979) Anal. Biochem. 99, 249 –253. 5. Collins, J., Gay, C., and Gebicki, J. M. (1999) Redox Rep. 4, 327–328. 6. Willson, R. L. (1970) Int. J. Radiat. Biol. 17, 349 –358. 7. Bothe, E., Schulte-Frohlinde, D., and von Sonntag, C. (1978) J. Chem. Soc., Perkin Trans. 2 416 – 420. 8. Bothe, E., Schuchman, M. N., Schulte-Frohlinde, D., and von Sonntag, C. (1978) Photochem. Photobiol. 28, 639 – 644. 9. Fordham, J. W. L., and Williams, H. L. (1950) J. Am. Chem. Soc. 72, 4465– 4469. 10. Davies, M. J., Fu, S., and Dean, R. T. (1995) Biochem. J. 305, 643– 649.
A Critical Evaluation of the Effect of Sorbitol on the Ferric–Xylenol Orange Hydroperoxide Assay Craig Gay and Janusz M. Gebicki 1 Department of Biological Sciences, Macquarie University, Sydney 2109, Australia
Received January 4, 2000
Measurement of hydroperoxide concentration by the ferric–xylenol orange assay has the advantages of simplicity, convenience, and inertness to oxygen (Gay, C., et al. (1999) Anal. Biochem. 273, 149 –155). However, its sensitivity is limited by the molar absorption coefficients, which are less than 50,000 M ⴚ1 cm ⴚ1 for most hydroperoxides. An earlier report showed that this could be significantly enhanced by the inclusion of 100 mM sorbitol in the assay solution, resulting in an increase of the apparent absorption coefficient for H 2O 2 from 4.46 ⴛ 10 4 to 2.24 ⴛ 10 5 M ⴚ1 cm ⴚ1 (Wolff, S. P. (1994) Methods Enzymol. 233, 182–189). It was claimed that the technique was also valid for other hydroperoxides, such as butyl and cumyl. In an attempt to extend this modification to a wider range of hydroperoxides, we have confirmed the enhancement of the assay of H 2O 2 by sorbitol. However, the sensitivity of the measurements of butyl, cumyl, amino acid, protein, and human blood serum hydroperoxides was only approximately doubled by the inclusion of sorbitol. A mechanism explaining the difference in the assay of H 2O 2 and the organic hydroperoxides is proposed. © 2000 Academic Press
The measurement of hydroperoxide concentrations based on the oxidation of ferrous iron in acid solution and the formation of a colored complex between the ferric iron generated and xylenol orange has the advantages of ease of application, reasonable sensitivity, and independence of the presence of oxygen. Unfortunately, one of the earlier and most widely used protocols (FOX 2) (1) proved to have significant limitations. These were largely overcome by a series of modifications recommended in a recent study, which gave new procedures able to produce reliable estimates of the concentrations of a wide range of hydroperoxides in aqueous or alcohol solutions (2, 3). 1
Fax: 61 02 9850 8245. E-mail: [email protected].
0003-2697/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
However, the lower limit of measurement remained at about 1 nmol of hydroperoxide. The finding that the FOX 2 procedure had to be modified prompted us to examine its promising variant, FOX 1 (1). In this, the amount of ferric iron generated by H 2O 2 and measured by the xylenol orange complex was enhanced about 5-fold by the inclusion of sorbitol (D-glucitol) in the reaction mixture (1). It was claimed that the technique could also be applied to other water-soluble hydroperoxides, such as butyl and cumyl. We now confirm the enhancing action of sorbitol in the assay of H 2O 2 but find that it only approximately doubles the sensitivity of measurement of a range of organic hydroperoxides. Two reaction sequences are proposed, which can account for these differences. MATERIALS AND METHODS
All chemicals were of at least analytical grade and used without further purification. Glassware was cleaned in hot concentrated nitric acid and thoroughly rinsed with water passed through a four-stage purification system with a 0.2-m final filter (Millipore-Waters, Sydney). This procedure was found to be essential for the removal of traces of iron. Hydrogen peroxide and sulfuric acid were from Aldrich Chemical Co., Sydney, and the ferrous and ferric ammonium sulfates from BDH Chemicals Pty. Ltd., Sydney. tert-Butyl and cumyl hydroperoxides were from Koch-Light (Colnbrook, UK), and xylenol orange [o-cresosulfonphthalein-3,3-bis(sodium methyliminodiacetate)] and BSA 2 (fatty acid free) from Sigma (St. Louis, MO). Sorbitol (99%⫹) was supplied by the Aldrich Chemical Co. (Milwaukee, WI) and used directly. 2 Abbreviations used: XO, xylenol orange; Fe–XO, ferric–xylenol orange complex; –OOH, the hydroperoxide group; ROOH, organic hydroperoxide; t-Bu–OOH, tert-butyl hydroperoxide; Cu–OOH, cumene hydroperoxide; BSA, bovine serum albumin.
217
218
GAY AND GEBICKI
Aerobic Hydroperoxide Assay Solutions containing the hydroperoxide were mixed with the appropriate volumes of reagents to give final concentrations of 25 mM H2SO 4, 100 –150 M xylenol orange, and 100 –250 M ferrous iron in a volume of 2 ml. The final pH was 1.8 ⫾ 0.5. After 30 min in the dark, the absorbance was read at 560 nm, with XO/Fe2⫹ as blank. We found that even the most carefully prepared 5 mM ferrous ammonium sulfate solutions contained about 1 M ferric iron, present in the solid salt. RESULTS AND DISCUSSION FIG. 1. Effect of added sorbitol on the net absorbance of the ferric– xylenol orange complex generated by H 2O 2. The color was produced in the standard aerobic assay (Materials and Methods) and in the presence of the final concentrations of sorbitol shown. The results shown were averaged from three measurements, with standard deviations falling within the data points.
Hydroperoxide Solutions Stock solutions of H2O 2, tert-butyl hydroperoxide, and cumene hydroperoxide were prepared by dissolving 20 l of the reagent, as supplied, in 100 ml of water. The solutions were stable for at least 2 weeks at 4°C. Serum samples were prepared from blood obtained from healthy volunteers by venipuncture and collected in Vacutainer serum separation tubes (Bioclone Australia, Sydney). Blood cells were removed by 10-min centrifugation at 600g and the serum collected as clear fluid above the Vacutainer gel plug. If not used immediately, the serum was stored in 1.5-ml aliquots at ⫺80°C. Hydroperoxide groups were generated in amino acids by exposure of 20 mM aqueous solutions to 60Co radiation at a dose rate of 20 Gy min⫺1, with occasional stirring to ensure air saturation. Undiluted serum and solutions containing 2 mg ml ⫺1 of BSA were oxidized similarly, but with slow oxygen bubbling. The H2O 2 produced by the radiation was removed by 260 U of catalase per milliliter of liquid.
Measurement of the effect of concentration of sorbitol on the absorbance of the Fe–XO complex shows that maximum and stable color developed between 50 and 400 mM sorbitol (Fig. 1). Higher concentrations were not tested, and in subsequent experiments 100 mM concentration was used, as recommended in the FOX 1 protocol (1). The full absorbance developed in 15 min at room temperature (data not shown). Tests with increasing concentrations of H 2O 2 in the standard assay solutions supplemented with 100 mM sorbitol (Fig. 2) showed a spectacular increase in the yield of the Fe–XO complex in the presence of sorbitol. Since the reaction with Fe2⫹ under conditions of this assay gives rise to 2.2 Fe3⫹ ions per H 2O 2 (3), in the presence of sorbitol this yield increased to 20.6 Fe 3⫹ per H 2O 2, or almost 10-fold. The discrepancy between this result and the value of 5-fold increase reported for FOX 1 (1) may be due to differences in the lengths of chain reactions producing these high yields of Fe 3⫹, caused by terminating impurities, or to a problem in measurement of the concentration of H2O 2 in the earlier work. The determination of the precise yield of Fe3⫹ under different conditions may need further study, but it is clear that the presence of sorbitol greatly enhanced the
Anaerobic Hydroperoxide Assay The method was based on a published procedure (4). Quartz spectrophotometer cuvettes equipped with closefitting Teflon stoppers were flushed with oxygen-free nitrogen and filled with 2.0 ml of thoroughly degassed 1:2 acetic acid:water solution containing 10% KI (w/v). The initial absorbance was measured at 358 nm. Up to 500 l of the hydroperoxide solution was then added, and the cuvette was flushed again with nitrogen for 10 s, restoppered, and incubated in the dark at 50°C for 20 min. After cooling, the absorbance was measured at 358 nm and the hydroperoxide concentration calculated from the difference between the initial and final readings, using ⑀ ROOH ⫽ 2.9 ⫻ 10 4 M ⫺1 cm ⫺1.
FIG. 2. Net absorbance of solutions of H 2O 2 in the standard Fe–XO assay in the absence (open circles) and presence (closed circles) of 100 mM sorbitol. The correct concentration of the H 2O 2 was determined by the anaerobic assay (Materials and Methods). The standard deviations of four measurements fall within the data points.
SORBITOL EFFECT ON HYDROPEROXIDE ASSAY
FIG. 3. Net absorbance of the Fe–XO complexes generated by tertbutyl and cumene hydroperoxides in the absence (lower line) and presence (upper line) of 100 mM sorbitol. Data are represented by tert-butyl hydroperoxide without sorbitol (open circles), cumene hydroperoxide without sorbitol (open triangles), tert-butyl hydroperoxide with sorbitol (closed circles), and cumene hydroperoxide with sorbitol (closed triangles). Hydroperoxide concentrations were measured by the anaerobic assay. The results are averaged from three experiments, with standard deviations falling within the data points.
sensitivity of the measurement of H2O 2 by the Fe–XO assay. This was not the case for the other hydroperoxides studied. The results shown in Fig. 3 and Table 1, obtained with cumene and tert-butyl hydroperoxides, show approximately a doubling of the amount of Fe3⫹ normally generated by the inclusion of 100 mM sorbitol in the assay solution. Virtually the same enhancement was obtained with BSA–OOH and a somewhat higher one with the amino acids (Table 1, graphs not shown). Finally, the presence of sorbitol led to a similar increase in the levels of hydroperoxide residues measured in oxidized serum (Fig. 4). In the case of serum, it is not possible to express the results in terms of yield of Fe3⫹ per hydroperoxide group, since no independent assay of the latter was possible. However, the enhancement in the formation of the
TABLE 1
Enhancement of the Apparent Molar Absorption Coefficients of Hydroperoxides by Sorbitol Hydroperoxide
⫺Sorbitol
H 2O 2 44,400 ⫾ 1,400 tert-Butyl–OOH 96,900 ⫾ 4,000 Cumene–OOH 103,800 ⫾ 4,000 Valine–OOH 51,600 ⫾ 850 Lysine–OOH 52,100 ⫾ 360 BSA–OOH 35,900 ⫾ 700
⫹Sorbitol
Enhancement factor
415,600 ⫾ 1,300 209,300 ⫾ 4,200 196,100 ⫾ 4,600 141,200 ⫾ 4,300 152,300 ⫾ 850 81,000 ⫾ 500
9.36 1.90 2.16 2.74 2.92 2.26
Note. All coefficients are in units M ⫺1 cm ⫺1 and represent the average of at least three determinations and their standard deviations. Those measured in the absence of sorbitol lie within the ranges of corresponding values determined in an earlier study (3).
219
FIG. 4. Net absorbance produced in the Fe–XO assay by oxidized human blood serum in the absence (open circles) and presence (closed circles) of 100 mM sorbitol. The serum was irradiated in a ␥ source (Materials and Methods) and the volume added to the Fe–XO assay solution shown on the x axis. The absolute peroxide concentrations could not be measured because the iodide assay is affected by the high serum protein concentration. The results are from a single experiment, representative of several, each producing slightly different slopes because each serum sample contained different amounts of hydroperoxides after irradiation.
Fe–XO complex in serum by the sorbitol was 2.29, similar to that of BSA–OOH, which agrees with our findings that protein hydroperoxides are an early product generated in irradiated serum (5). The mechanism previously suggested for the formation of the high yield of Fe 3⫹ in the presence of sorbitol postulated a chain reaction (1). The statement that the FOX 1 assay could be used not only for H2O 2 but also for other water-soluble hydroperoxides (1) implies that all would yield similar amounts of Fe3⫹ per –OOH group and that the same mechanism should be valid generally. Our results show that this is incorrect (Table 1) and that the reactions initiated by reduction of H2O 2 by Fe 2⫹ in the presence of sorbitol are uniquely different from those of the organic hydroperoxides. This is caused by differences in the reactivities of free radicals formed in the initial reduction of the different hydroperoxides. In the case of H 2O 2, the following sequence of reactions can account for the experimental results: H2O2 ⫹ Fe 2⫹ 3 HO • ⫹ OH ⫺ ⫹ Fe 3⫹ ,
[1]
HO • ⫹ R2CHOH 3 H2O ⫹ R2C •OH,
[2]
R2C •OH ⫹ O2 3 R2C共OO • 兲OH,
[3]
R2C共OO • 兲OH 3 R2CAO ⫹ HO 2• ,
[4]
2HO 2• 3 H2O2 ⫹ O2 ,
[5]
where R 2CHOH is the sorbitol molecule, CH 2OH (CHOH) 4CH 2OH. The C represents any one of the six
220
GAY AND GEBICKI
carbon positions in sorbitol, which the hydroxyl radical attacks randomly at close to diffusion-controlled rate. Since sorbitol is the most abundant molecule present, it will be the main target. Reaction [1] is the wellknown Fenton process. Reaction [3] has a rate constant of about 2 ⫻ 10 9 M ⫺1 s ⫺1 (6), while that of [4] is 2700 s ⫺1 in acid solution (7). The H 2O 2 regenerated in reaction [5] can propagate a chain reaction through [1]. The reduction of the organic hydroperoxides does not proceed by a chain mechanism. Unlike the random reaction of the HO • with sorbitol (reaction [2]), the RO • reacts primarily at the terminal carbons of sorbitol (C in the formula RCH 2OH), producing the following reaction sequence: ROOH ⫹ Fe 2⫹ 3 RO • ⫹ HO ⫺ ⫹ Fe 3⫹ ,
[6]
RO • ⫹ RCH2OH 3 ROH ⫹ RC •HOH,
[7]
RC •HOH ⫹ O2 3 RCH共OO • 兲OH,
[8]
RCH共OO • 兲OH ⫹ Fe 2⫹ ⫹ H ⫹ 3 RCH共OOH兲OH ⫹ Fe 3⫹ ,
[9]
RCH共OOH兲OH ⫹ Fe 2⫹ 3 RCH共O • 兲OH ⫹ Fe 3⫹ ⫹ OH ⫺ ,
[10]
RCH共O • 兲OH ⫹ Fe 2⫹ ⫹ H ⫹ 3 RCH共OH兲OH ⫹ Fe 3⫹ ,
report demonstrated that Cu–OOH reduction by Fe 2⫹ gives a methyl radical, which enters a series of 19 reactions, resulting in the formation of several Fe 3⫹ ions for every –OOH group present (9). A similar process can be assumed for the t-Bu–OOH, since it also generates the methyl radical on reaction with Fe 2⫹. Extensive search of the literature did not reveal any reaction mechanism that could explain the doubling of the yield of Fe 3⫹ from these hydroperoxides by sorbitol (Table 1). In the case of BSA, amino acids, and serum, the higher than expected doubling of the yield of Fe 3⫹ by sorbitol (reactions [6]–[11]) was probably also caused by side reactions of free radicals known to be generated in the reduction of protein hydroperoxides (10). All these reactions are difficult to control and are likely to give different sorbitol enhancement factors under different conditions. Separate experiments also showed that sorbitol had no effect on the molar absorption coefficient of the Fe–XO complex. The main results of this study are the confirmation of the ability of sorbitol to enhance significantly the sensitivity of the Fe–XO assay of H 2O 2 and the finding of a much smaller enhancement for organic hydroperoxides. It seems likely that, in most practical applications, the possible complications resulting from inclusion of a high concentration of sorbitol in solutions of hydroperoxides other than H 2O 2 will not make this modification worthwhile. ACKNOWLEDGMENTS
[11]
This study was supported by Macquarie University grants. C.G. is holder of an Australian Postgraduate Award.
ROOH ⫹ 4Fe 2⫹ ⫹ 2H ⫹ ⫹ O2 ⫹ RCH2OH 3 ROH ⫹ 4Fe 3⫹ ⫹ 2OH ⫺ ⫹ RCH共OH兲OH. This gives an overall stoichiometry of 4 Fe 3⫹ ions for every –OOH reacting. Since each ROOH group gives normally close to 2 Fe 3⫹ ions (3), the enhancement by sorbitol of the yield of Fe 3⫹ should be about 2 (Table 1). The elimination of HO 2• from the ␣ and positions in sorbitol is slower than from the other sites in the molecule (8), allowing reaction [9] to compete effectively with [4]. At present, both these mechanisms are speculative, but have the potential to explain the experimental observations. There is little doubt that, especially as some very reactive species are generated in the presence of Fe 2⫹, reactions not considered here contribute to the overall yield of Fe 3⫹, accounting for the occurrence of sorbitol enhancement factors greater than 2. In the case of cumene hydroperoxide, an earlier
REFERENCES 1. Wolff, S. P. (1994) Methods Enzymol. 233, 182–189. 2. Gay, C., Collins, J., and Gebicki, J. M. (1999) Anal. Biochem. 273, 143–148. 3. Gay, C., Collins, J., and Gebicki, J. M. (1999) Anal. Biochem. 273, 149 –155. 4. Hicks, M., and Gebicki, J. M. (1979) Anal. Biochem. 99, 249 –253. 5. Collins, J., Gay, C., and Gebicki, J. M. (1999) Redox Rep. 4, 327–328. 6. Willson, R. L. (1970) Int. J. Radiat. Biol. 17, 349 –358. 7. Bothe, E., Schulte-Frohlinde, D., and von Sonntag, C. (1978) J. Chem. Soc., Perkin Trans. 2 416 – 420. 8. Bothe, E., Schuchman, M. N., Schulte-Frohlinde, D., and von Sonntag, C. (1978) Photochem. Photobiol. 28, 639 – 644. 9. Fordham, J. W. L., and Williams, H. L. (1950) J. Am. Chem. Soc. 72, 4465– 4469. 10. Davies, M. J., Fu, S., and Dean, R. T. (1995) Biochem. J. 305, 643– 649.