- Email: [email protected]
Critical evaluation of the use of hydroethidine as a measure of superoxide anion radical
Free Radical Biology & Medicine, Vol. 25, No. 7, pp. 826 – 831, 1998 Copyright © 1998 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/98 $19.00 1 .00
PII S0891-5849(98)00163-4
Original Contribution CRITICAL EVALUATION OF THE USE OF HYDROETHIDINE AS A MEASURE OF SUPEROXIDE ANION RADICAL LUDMIL BENOV*, LAURA SZTEJNBERG*,
and IRWIN
FRIDOVICH*
*Department of Biochemistry, Duke University Medical Center, Durham, NC, USA (Received 15 May 1998; Revised 19 June 1998; Accepted 22 June 1998)
Abstract—The fluorogenic oxidation of hydroethidine (HE) to ethidium (E1) has been used as a measure of O22. Evaluation of this method confirms that O22, but not O2 or H2O2, rapidly oxidizes HE to E1. However the ratio of E1 produced per O22 introduced decreased as the flux of O22 was increased. This suggested that HE can catalyze the dismutation of O22 and this was affirmed. HE was oxidized to a red product, distinct from E1 by ferricytochrome c and a similar oxidation may occur within Escherichia coli. HE inhibited the growth and killed a SOD-null strain to a greater extent than the SOD-replete parental strain and these effects were much diminished under anaerobic conditions. This indicated that E1 was responsible for the toxicity of HE and indeed E1 was seen to be toxic under both aerobic and anaerobic conditions. In view of the data presented HE can be recommended as a qualitative but not as a quantitative measure of O22. © 1998 Elsevier Science Inc. Keywords—Hydroethidine; Dihydroethidium; Ethidium, reduction of; Superoxide, detection of; Free radical
importantly that HE can catalyze the dismutation of O22. This latter process interferes with the fluorimetric measurement of O22 with HE.
INTRODUCTION
Hydroethidine (HE)1 is oxidized to the fluorescent ethidium (E1) by O22 whether added as KO2, or produced by activated leukocytes [1]. Moreover, the oxidation of HE to E1 was shown to be rapid when the oxidant was O22; but not when O2, H2O2, HOCl or ONOO2 were used [2]. Consequently HE has been applied as a detector of intracellular O22 [3]. We have investigated chromogenic and lumigenic detectors of O22, such as the tetrazoliums [4] and lucigenin [5], and have found them to be prone to artifact in that they can cause O22 production even in systems not producing this radical in the absence of these compounds. We wondered whether in hydroethidine we might finally have a convenient measure of O22, free from such problems. In the experiments reported herein we verify earlier reports [1] that the oxidation of HE to E1 is rapidly caused by O22 but not by H2O2 [2]. However we note that HE can be oxidized by cytochrome c and more
MATERIALS AND METHODS
HE was purchased from Polysciences Inc. and was dissolved, to 63.5 mM, in either N,N-dimethyl formamide or dimethyl sulfoxide. This stock solution was stored under N2 at 280°C. Working solutions were prepared by diluting this stock into 50 mM potassium phosphate at pH 7.8. This was done immediately prior to use. DNA and E1 were from Sigma. The oxidation of HE to E1 was monitored fluorimetrically by exciting at 470 nm and following emission at 590 nm. The sodA sodB strains of Escherichia coli were JI132 [6], KK135 and QC2510 [7,8] and the corresponding parental strains were AB1157 [6], AN387 [9], and QC 2509 [7,8], respectively. Cultures were grown to A600 nm5 0.5– 0.8 in Luria Bertani (LB) medium at 37°C, with shaking at 200 rpm. When extracts were needed, the cells were collected by centrifugation, washed with 50 mM phosphate at pH 7.8, and then resuspended in this buffer before disruption with a
Address correspondence to: Dr. Irwin Fridovich, Department of Biochemistry, Duke University Medical Center, Durham, NC 27710, USA; Tel: (919) 684-5122; Fax (919) 684-8885. 1 This strain was obtained from Dr. James Imlay. 826
Hydroethidine to measure O22
1
Fig. 1. Oxidation of HE to E by the XO reaction. Reaction mixtures contained 63.5 mM HE, 375 mM X 1 5 mg/ml of XO in 50 mM potassium phosphate 1 0.1 mM EDTA at pH 7.8 and 25° C. Additional components were: line 1, none; line 2, 3 mg/ml SOD; line 3, 10 mg/ml SOD. Line 4 was as in line 1 but with XO omitted.
French Press. After clarification by centrifugation, protein concentration was determined colorimetrically [10]. When intact cells were to be treated with HE, the washed cells were suspended in the phosphate buffer plus 0.2% glucose and 127 mM HE, to A600 nm 5 0.5– 0.8, and were then shaken at 37°C for 30 min. The cells were then washed twice in phosphate buffer and were then extracted with the French Press and the clarified extracts were examined for E1 fluorescence. When luminescence was to be measured in cell suspensions, lucigenin at 0.2 mM was used in place of HE but the cells were not disrupted with a French press. The cells were washed, resuspended to A600 nm 5 0.5 in phosphate buffer containing 0.2% glucose, 0.2 mM lucigenin and luminescence was measured. RESULTS
Oxidation of HE to E1 by O22 Line 1, Fig. 1 presents the rate of oxidation of HE to E1 by the flux of O22 generated by the xanthine oxidase reaction. That this oxidation was due to O22, and not to H2O2 or to O2, is shown by the dose-dependent inhibition by SOD (lines 2, 3 and 4). Catalase, in contrast was without effect (not shown). These results are in full accord with previous reports [1–3]. It should be noted that xanthine oxidase plus xanthine did not cause the reduction of E1 to HE, even under anaerobic conditions (data not shown). Reduction of E1 by glucose oxidase This enzyme does not produce O22; yet in the presence of NBT11 it catalyzes an SOD-inhibitable forma-
827
Fig. 2. E1 mediates the reduction of cyt c. Reaction mixtures contained 20 mM cyt c, 1.0 mM glucose and 1.0 mg/ml of glucose oxidase in 50 mM TRIS buffer at pH 7.4 and 25°C. Where indicated E1 was added to 100 mM and SOD to 10 mg/ml.
tion of the purple formazan [4], and in the presence of lucigenin it causes an SOD-inhibitable luminescence [5]. This is the case because glucose oxidase can reduce these compounds to their corresponding radicals; which can, in turn, reduce O2 to O22. It was therefore of interest to see whether E1 could give rise to a comparable artifactual production of O22. Figure 2 shows that addition of E1 to a reaction mixture containing glucose oxidase, glucose, and cytochrome c, caused a reduction of the cytochrome c which was not inhibitable by SOD. It follows that glucose oxidase is able to catalyze the reduction of E1 to a form which directly reduces the cytochrome. Moreover that reduced form of E1 did not detectably autoxidize with production of O22. In an attempt to favor such autoxidation this experiment was repeated under 100% O2 and results identical to those shown in Fig. 2 were obtained (not shown). These results suggested that glucose oxidase was reducing E1 i.e. to HE; which is known not to readily autoxidize. In that case HE should rapidly reduce cyt c and the results in Fig. 3 show that it did so. Since the oxidation of HE to E1 provides two electrons while the reduction of cyt c is a univalent process, one molecule of HE should reduce two of cyt c. In fact, as the ratio of cyt c to HE in the reaction mixture was increased the ratio of cyt c reduced to HE oxidized rose to 4. This is presented in Table 1 and suggested that cyt c could oxidize HE to products beyond E1. Figure 4 presents the spectra of HE (line 3), E1 (line 1) and of the product of the oxidation of HE by excess cyt c (line 2). Clearly a product or products other than E1 were obtained when excess cyt c oxidized HE. Since E1 did not itself reduce cyt c it must be supposed that an intermediate in the oxidation of HE by cyt c, probably E• led to products more oxidized than E1. The product obtained by the oxidation of HE by a four-fold molar excess of cytochrome c, followed by
828
L. BENOV et al.
Fig. 3. The reduction of cyt c by HE. The reaction mixture contained 10 mM cyt c and 6.35 mM HE in 50 mM potassium phosphate at pH 7.8 and at 25°C.
ultrafiltration to remove the cytochrome, was reddish. The sodA sodB E. coli (JI132) incubated aerobically with HE gave a distinctly red pellet when sedimented. This indicates that a similar oxidation of HE to a product other than E1 occurred in the cells. The parental strain AB1157 showed much less coloration under identical conditions. The reduction of E1 to HE by glucose oxidase plus glucose was directly demonstrated by following the absorbance of E1 at 460 nm; where HE does not significantly absorb (Fig. 4). When 10 mM E1 was incubated with 5.0 mM glucose and 1.0 mg/ml glucose oxidase in 50 mM Tris at pH 7.4, the absorbance at 460 nm decreases as a function of time. If this was really due to reduction of E1 to HE then subsequent aerobic addition of 375 mM xanthine 1 5 mg/ml of xanthine oxidase as a source of O22 should cause an increase in A460 nm. It did so (data not shown).
Fig. 4. Absorption spectra of E1, HE, and the product of oxidation of HE by excess cyt c. HE was at 63.5 mM in 50 mM potassium phosphate at pH 7.8 (line 3). E1 at 63.5 mM in this buffer (line 1). HE (63.5 mM), oxidized by a 4-fold molar excess of cyt c for 1 h, was freed of cyt c by ultrafiltration through a Centricon 3000 (from Amicon) (line 2).
cytc to a plateau [11]. At that plateau the rate of cyt c reduction may be equated to the rate of O22 production; because the cyt c is then outcompeting the spontaneous dismutation of O22. It was of interest to determine the number of HE molecules oxidized to E1 per O22 consumed. This was done at different fluxes of O22, achieved at different concentrations of xanthine oxidase. Figure 5 presents the amount of O22 produced (line 1) and of the HE oxidized to E1 (line 2) as a function of the xanthine oxidase used. As expected the O22 produced per minute was nearly a
HE catalyzes the dismutation of O22 The rate of reduction of cyt c by the xanthine oxidase reaction increases with the concentration of Table 1. Cyt c Reduction by HE HE (mM)
Cyt c (mM)
Cyt c Reduced (mM)
Cyt cred/HE
20 10 6 2 1
20 20 20 20 20
19 6 1.0 15 6 1.0 12 6 1.5 6 6 1.0 4.0 6 0.5
0.95 1.5 2 3 4
The reaction mixture contained 20 mM cyt c and the indicated concentration of HE in 50 mM potassium phosphate, pH 7.8. Cyt c reduction was followed at 550 nm using « 5 21,000 M21 cm21.
Fig. 5. Reduction of cyt c and oxidation of HE as a function of [XO]. Reaction mixtures contained 10 mM cyt c, 375 mM xanthine, and the indicated amounts of XO in 50 mM potassium phosphate at pH 7.8 and 25°C. Additional components were: line 1, 10 mM cyt c; or line 2, 63.5 mM HE. Panel A, the cyt c reduced (line 1), or the HE oxidized to E1 (line 2) per minute; Panel B the ratio of cyt c reduced per HE oxidized.
Hydroethidine to measure O22
Fig. 6. Comparison of HE and lucigenin assays. Cells grown in LB to A600 nm 5 0.5– 0.8 were washed and resuspended to A600 nm 5 0.5 in 50 mM potassium phosphate buffer pH 7.8 1 0.2% glucose and 0.2 mM lucigenin or 63.5 mM HE. Lucigenin luminescence was measured as previously described (14) exactly 4 min later. E1 fluorescence was measured in clarified extracts of cells incubated 30 min at 37°C in a shaking water bath, washed and French pressed, data presented were corrected for the fluorescence of extracts prepared from anaerobic controls treated in the same way.
829
the oxidation of 7.0 mM NH2OH to NO22 per 20 min and this NO22 production was 50% inhibited by 3.15 mM HE. Thus 3.15 mM HE intercepted 14.3 mM O22 or ;4.6 O22/HE. Since the divalent oxidation of HE to E1 requires only 2.0 O22 this indicates that HE can catalyze the dismutation of O22. This was further explored by comparing the rates of cyt c reduction and HE oxidation by a given flux of O22 produced by the XO reaction. A flux of O22 which caused the reduction of 12 mM cyt c/min was found to oxidize only 2.5 mM HE to E1/min. This was the case when the concentrations of cyt c and of HE were great enough to achieve maximal rates of cyt c reduction and of HE oxidation, respectively. In this case the ratio of O22/HE oxidized to E1 was 4.8, which again supports the idea of dismutation of O22 by HE.
linear function of [xanthine oxidase] (line 1), but the rate of oxidation of HE to E1 followed fluorimetrically was not (line 2). Thus the yield of HE oxidized to E1 decreased as the rate of production of O22 increased. This suggested that HE might catalyze the dismutation of O22 or alternately that O22 was being lost to the spontaneous dismutation [12] at higher fluxes of O22. To examine whether HE could catalyze the dismutation of O22, an indicating scavenger for O22 other than cytochrome c was necessary in order to avoid the problem of the direct reduction of cytochrome c by HE. Hydroxylamine, which is oxidized to NO22 by O22 [13], was chosen and it was ascertained that NH2OH did not directly react with E1 or with HE. A xanthine oxidase reaction which produced 28.6 mM O22 per 20 min, on the basis of cytochrome c reduction, was seen to cause
Table 2. Lucigenin Luminescence and HE Fluorescence in sodA sodB and Parental Strainsa Ratio sodA sodB/parental Strain Luminescence Fluorescence Luminescence/ Fluorescence a
JI132/AB1157
QC2510/QC2509
KK135/AN387
13.5 5.9 2.3
5.0 2.2 2.3
9.2 3.0 3.1
Data are from Fig. 6.
Fig. 7. Growth of JI132 and AB1157 in the presence of HE and E1. 50 ml overnight JI132 and AB1157 cultures in LB medium were transferred into 10 ml M9CA medium containing 63.5 mM HE or E1. Cells were grown aerobically, or anaerobically, in a Coy chamber, and growth was monitored by measuring A600 nm at intervals. Panel A, aerobic; panel B, anaerobic. Line 1 — JI132 1 E1; line 2 — JI132 1 HE; line 3 — JI132; line 4 — AB1157 1 E1; line 5 — AB1157 1 HE; line 6 — AB1157.
830
L. BENOV et al. DISCUSSION
Measuring itracellular [SOD] Lucigenin luminescence can be used as a measure of [SOD] in E. coli [14]. The conversion of HE to E1 should be similarly useful. Figure 6 presents the application of these two methods to three sets of sodA sodB, and the corresponding parental, strains. In all three cases the sodA sodB strains showed substantially more lucigenin luminescence and E1 fluorescence than did the corresponding parental strains. Yet the ratios of lucigenin luminescence in the SOD-null strains compared to the SOD-replete strains was greater by a factor of 2.3–3.1 than the corresponding ratios of E1 fluorescence. This is shown in Table 2 and may reflect the ability of HE to catalyze the dismutation of O22 thus muting the difference between SOD-null and SOD-replete strains.
The toxicity of HE Since E1 can intercalate into double stranded nucleic acids and since HE can be oxidized to E1 by O22; one would expect that the toxicity of HE should reflect intracellular levels of O22. That, in turn, suggests that HE should be more toxic to a sodA sodB than to a SODcompetent parental strain and further that its toxicity to the sodA sodB strain should be dioxygen-dependent. Comparison of lines 2 and 3 in Fig. 7A illustrates the toxicity of 63.5 mM HE to the sodA sodB E. coli; while lines 5 and 6 demonstrate the much smaller effect exerted on the parental strain. The greater toxicity of E1 on the two strains is presented by lines 1 and 4. When the same experiment was repeated under anaerobic conditions the toxicity of HE to the sodA sodB strain was much diminished, as shown by lines 2 and 3 in Fig. 7B and, as anticipated, E1 exerted its toxicity even in the absence of dioxygen (line 1). It is likely that E1 can be reduced within the cells and that the net redox balance is shifted towards oxidation in the SOD-null strains. That could explain the much greater toxicity of E1 to the sodA sodB (lines 1) than to the parental (lines 4) strain. The expectation is thus that there would be more E1 reduction in the parental strain and more HE oxidation in the SOD-null strain. The partial escape from E1 toxicity, after 4 h of anaerobic incubation with the SOD-null (lines 1 panel 7B), could thus be due to reduction of E1 to HE unopposed by its reoxidation by O22. It should be noted that the HE used contained some E1 due to autoxidation prior to the imposition of anaerobiosis. This explains the weak toxicity of HE to the SOD-null strain even under anaerobic conditions (compare lines 2 and 3 in panel B).
In accord with previous reports [1,2] we find HE to be oxidized to E1 by O22, whereas the rates of oxidation by O2 or by H2O2 are much slower. The reduction of E1 to HE, by glucose oxidase plus glucose, does increase cyt c reduction, but that is due to a direct effect of HE and not to HE-mediated O22 production; since this cyt c reduction was not SOD-inhibitable. The oxidations of HE by cyt c consumed more than the theoretical 2 cyt c/HE and the product was not E1. This indicates that cyt c can oxidize HE to species more oxidized than E1. At the same time it must be noted that E1 did not reduce cyt c. Hence the pathway of oxidation of EH by cyt c does not go through E1. It can be supposed that the univalent oxidation of EH by cyt c yields a radical E. which may dimerize before further oxidation by cyt c. O22, unlike cyt c, does oxidize HE to E1; but the yield of E1 was less than 1.0 per 2.0 O22. This suggested that, in addition to stoichiometrically scavenging O22, EH was capable of catalyzing the dismutation of O22. The following reactions may be proposed: HE 1 O 2 2 1 H 1
3
E 1 H 2O 2
(1)
E 1 O 2 2 1 2H 1
3
E 1 1 H 2O 2
(2)
3
HE 1 E
3
E 2 E
(4)
3
HE 1 O 2
(5)
•
E 1 E 1 H •
•
E 1 E •
1
•
E 1 O 22 1 H 1 •
•
1
(3)
Reactions 1–3 account for the oxidation of HE to E1 while reactions 1 plus 5 account for the dismutation of O22 catalyzed by HE. This ability of HE to catalyze the dismutation of O22 can explain the exponential appearance of line 1 in Fig. 1. Thus the rate of oxidation of HE to E1 increased with time because, as HE was consumed, it caused less dismutation of O22; leaving a larger proportion of the O22 to oxidize HE. We conclude that the HE conversion to E1 may be useful as a qualitative indicator of O22, but should not be used as a quantitative measure of this radical because it can catalyze the dismutation of O22 and therefore the yield of E1 produced per O22 will be less than stoichiometric and may be variable. Moreover, it is likely that the oxidation of HE by cytochrome c has some counterpart in the cell. Acknowledgements — This work was supported by grants from the Council for Tobacco Research, U.SA., Inc. (2871BR1) and the National Institutes of Health (HL56025-03).
Hydroethidine to measure O22 REFERENCES [1] Rothe, G.; Valet, G. Flow cytometric analysis of respiratory burst activity in phagocytes with hydroethidine and 29,79-dichlorofluorescin. J. Leukocyte Biol. 47:440 – 448; 1989. [2] Bindokas, V. P.; Jordan, J.; Lee, C. C.; Miller, R. J. Superoxide production in rat hippocampal neurons: selective imaging with hydroethidine J. Neurosci. 16:1324 –1326; 1989. [3] Carter, W. O.; Narayanan, P. K.; Robinson, J. P. Intracellular hydrogen peroxide and superoxide anion detection in endothelial cells. J. Leukocyte Biol. 55:253–258; 1994. [4] Liochev, S. I.; Fridovich, I. Superoxide from glucose oxidase or from nitroblue tetrazolium? Arch. Biochem. Biophys. 318:408 – 410; 1995. [5] Liochev, S. I.; Fridovich, I. Lucigenin (bis-N-methylacridinium) as a mediator of superoxide anion production. Arch. Biochem Biophys. 337:115–120; 1997. [6] Imlay, J.; Linn, S. Mutagenesis and stress responses induced in Escherichia coli by hydrogen peroxide. J. Bacteriol. 169:2967– 2976; 1987. [7] Carlioz, A.; Touati, D. Isolation of superoxide dismutase mutants in Escherichia coli: is superoxide dismutase necessary for aerobic life? EMBO J. 5:623– 630; 1986. [8] Pianzzola, M. J.; Soubes, M.; Touati, D. Overproduction of the rbo gene product from desulfovibrio species suppresses all deleterious effects of lack of superoxide dismutase in Escherichia coli. J. Bacteriol. 178:6736 – 6742; 1996
831
[9] Imlay, J. A metabolic enzyme that rapidly produces superoxide, fumarate reductase of Escherichia coli. J. Biol. Chem. 270:19767–19777; 1995. [10] Lowry, O. H.; Rosebrough, N. F.; Farr, A. L.; Randall, R. J. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265–275; 1951. [11] McCord, J. M.; Fridovich, I. The reduction of cytochrome c by milk xanthine oxidase. J. Biol. Chem. 243:5753–5760; 1968. [12] Bielski, B. H.; Cabelli, D. E.; Arudi, R. L. Reactivity of HO2/O22 radicals in aqueous solution. J. Phys. Chem. 14:1041–1097; 1985. [13] Elstner, E. F.; Heupel, A. Inhibition of nitrite formation from hydroxylammonium chloride: a simple assay for superoxide dismutase. Anal. Biochem. 70:616 – 620; 1976. [14] Liochev, S. I.; Fridovich, I. Lucigenin luminescence as a measure of intracellular superoxide dismutase activity in Escherichia coli. Proc. Nat. Acad. Sci. USA 94:2891–2896; 1997.
ABBREVIATIONS
HE— hydroethidine E1— ethidium GO— glucose oxidase SOD—superoxide dismutase
PII S0891-5849(98)00163-4
Original Contribution CRITICAL EVALUATION OF THE USE OF HYDROETHIDINE AS A MEASURE OF SUPEROXIDE ANION RADICAL LUDMIL BENOV*, LAURA SZTEJNBERG*,
and IRWIN
FRIDOVICH*
*Department of Biochemistry, Duke University Medical Center, Durham, NC, USA (Received 15 May 1998; Revised 19 June 1998; Accepted 22 June 1998)
Abstract—The fluorogenic oxidation of hydroethidine (HE) to ethidium (E1) has been used as a measure of O22. Evaluation of this method confirms that O22, but not O2 or H2O2, rapidly oxidizes HE to E1. However the ratio of E1 produced per O22 introduced decreased as the flux of O22 was increased. This suggested that HE can catalyze the dismutation of O22 and this was affirmed. HE was oxidized to a red product, distinct from E1 by ferricytochrome c and a similar oxidation may occur within Escherichia coli. HE inhibited the growth and killed a SOD-null strain to a greater extent than the SOD-replete parental strain and these effects were much diminished under anaerobic conditions. This indicated that E1 was responsible for the toxicity of HE and indeed E1 was seen to be toxic under both aerobic and anaerobic conditions. In view of the data presented HE can be recommended as a qualitative but not as a quantitative measure of O22. © 1998 Elsevier Science Inc. Keywords—Hydroethidine; Dihydroethidium; Ethidium, reduction of; Superoxide, detection of; Free radical
importantly that HE can catalyze the dismutation of O22. This latter process interferes with the fluorimetric measurement of O22 with HE.
INTRODUCTION
Hydroethidine (HE)1 is oxidized to the fluorescent ethidium (E1) by O22 whether added as KO2, or produced by activated leukocytes [1]. Moreover, the oxidation of HE to E1 was shown to be rapid when the oxidant was O22; but not when O2, H2O2, HOCl or ONOO2 were used [2]. Consequently HE has been applied as a detector of intracellular O22 [3]. We have investigated chromogenic and lumigenic detectors of O22, such as the tetrazoliums [4] and lucigenin [5], and have found them to be prone to artifact in that they can cause O22 production even in systems not producing this radical in the absence of these compounds. We wondered whether in hydroethidine we might finally have a convenient measure of O22, free from such problems. In the experiments reported herein we verify earlier reports [1] that the oxidation of HE to E1 is rapidly caused by O22 but not by H2O2 [2]. However we note that HE can be oxidized by cytochrome c and more
MATERIALS AND METHODS
HE was purchased from Polysciences Inc. and was dissolved, to 63.5 mM, in either N,N-dimethyl formamide or dimethyl sulfoxide. This stock solution was stored under N2 at 280°C. Working solutions were prepared by diluting this stock into 50 mM potassium phosphate at pH 7.8. This was done immediately prior to use. DNA and E1 were from Sigma. The oxidation of HE to E1 was monitored fluorimetrically by exciting at 470 nm and following emission at 590 nm. The sodA sodB strains of Escherichia coli were JI132 [6], KK135 and QC2510 [7,8] and the corresponding parental strains were AB1157 [6], AN387 [9], and QC 2509 [7,8], respectively. Cultures were grown to A600 nm5 0.5– 0.8 in Luria Bertani (LB) medium at 37°C, with shaking at 200 rpm. When extracts were needed, the cells were collected by centrifugation, washed with 50 mM phosphate at pH 7.8, and then resuspended in this buffer before disruption with a
Address correspondence to: Dr. Irwin Fridovich, Department of Biochemistry, Duke University Medical Center, Durham, NC 27710, USA; Tel: (919) 684-5122; Fax (919) 684-8885. 1 This strain was obtained from Dr. James Imlay. 826
Hydroethidine to measure O22
1
Fig. 1. Oxidation of HE to E by the XO reaction. Reaction mixtures contained 63.5 mM HE, 375 mM X 1 5 mg/ml of XO in 50 mM potassium phosphate 1 0.1 mM EDTA at pH 7.8 and 25° C. Additional components were: line 1, none; line 2, 3 mg/ml SOD; line 3, 10 mg/ml SOD. Line 4 was as in line 1 but with XO omitted.
French Press. After clarification by centrifugation, protein concentration was determined colorimetrically [10]. When intact cells were to be treated with HE, the washed cells were suspended in the phosphate buffer plus 0.2% glucose and 127 mM HE, to A600 nm 5 0.5– 0.8, and were then shaken at 37°C for 30 min. The cells were then washed twice in phosphate buffer and were then extracted with the French Press and the clarified extracts were examined for E1 fluorescence. When luminescence was to be measured in cell suspensions, lucigenin at 0.2 mM was used in place of HE but the cells were not disrupted with a French press. The cells were washed, resuspended to A600 nm 5 0.5 in phosphate buffer containing 0.2% glucose, 0.2 mM lucigenin and luminescence was measured. RESULTS
Oxidation of HE to E1 by O22 Line 1, Fig. 1 presents the rate of oxidation of HE to E1 by the flux of O22 generated by the xanthine oxidase reaction. That this oxidation was due to O22, and not to H2O2 or to O2, is shown by the dose-dependent inhibition by SOD (lines 2, 3 and 4). Catalase, in contrast was without effect (not shown). These results are in full accord with previous reports [1–3]. It should be noted that xanthine oxidase plus xanthine did not cause the reduction of E1 to HE, even under anaerobic conditions (data not shown). Reduction of E1 by glucose oxidase This enzyme does not produce O22; yet in the presence of NBT11 it catalyzes an SOD-inhibitable forma-
827
Fig. 2. E1 mediates the reduction of cyt c. Reaction mixtures contained 20 mM cyt c, 1.0 mM glucose and 1.0 mg/ml of glucose oxidase in 50 mM TRIS buffer at pH 7.4 and 25°C. Where indicated E1 was added to 100 mM and SOD to 10 mg/ml.
tion of the purple formazan [4], and in the presence of lucigenin it causes an SOD-inhibitable luminescence [5]. This is the case because glucose oxidase can reduce these compounds to their corresponding radicals; which can, in turn, reduce O2 to O22. It was therefore of interest to see whether E1 could give rise to a comparable artifactual production of O22. Figure 2 shows that addition of E1 to a reaction mixture containing glucose oxidase, glucose, and cytochrome c, caused a reduction of the cytochrome c which was not inhibitable by SOD. It follows that glucose oxidase is able to catalyze the reduction of E1 to a form which directly reduces the cytochrome. Moreover that reduced form of E1 did not detectably autoxidize with production of O22. In an attempt to favor such autoxidation this experiment was repeated under 100% O2 and results identical to those shown in Fig. 2 were obtained (not shown). These results suggested that glucose oxidase was reducing E1 i.e. to HE; which is known not to readily autoxidize. In that case HE should rapidly reduce cyt c and the results in Fig. 3 show that it did so. Since the oxidation of HE to E1 provides two electrons while the reduction of cyt c is a univalent process, one molecule of HE should reduce two of cyt c. In fact, as the ratio of cyt c to HE in the reaction mixture was increased the ratio of cyt c reduced to HE oxidized rose to 4. This is presented in Table 1 and suggested that cyt c could oxidize HE to products beyond E1. Figure 4 presents the spectra of HE (line 3), E1 (line 1) and of the product of the oxidation of HE by excess cyt c (line 2). Clearly a product or products other than E1 were obtained when excess cyt c oxidized HE. Since E1 did not itself reduce cyt c it must be supposed that an intermediate in the oxidation of HE by cyt c, probably E• led to products more oxidized than E1. The product obtained by the oxidation of HE by a four-fold molar excess of cytochrome c, followed by
828
L. BENOV et al.
Fig. 3. The reduction of cyt c by HE. The reaction mixture contained 10 mM cyt c and 6.35 mM HE in 50 mM potassium phosphate at pH 7.8 and at 25°C.
ultrafiltration to remove the cytochrome, was reddish. The sodA sodB E. coli (JI132) incubated aerobically with HE gave a distinctly red pellet when sedimented. This indicates that a similar oxidation of HE to a product other than E1 occurred in the cells. The parental strain AB1157 showed much less coloration under identical conditions. The reduction of E1 to HE by glucose oxidase plus glucose was directly demonstrated by following the absorbance of E1 at 460 nm; where HE does not significantly absorb (Fig. 4). When 10 mM E1 was incubated with 5.0 mM glucose and 1.0 mg/ml glucose oxidase in 50 mM Tris at pH 7.4, the absorbance at 460 nm decreases as a function of time. If this was really due to reduction of E1 to HE then subsequent aerobic addition of 375 mM xanthine 1 5 mg/ml of xanthine oxidase as a source of O22 should cause an increase in A460 nm. It did so (data not shown).
Fig. 4. Absorption spectra of E1, HE, and the product of oxidation of HE by excess cyt c. HE was at 63.5 mM in 50 mM potassium phosphate at pH 7.8 (line 3). E1 at 63.5 mM in this buffer (line 1). HE (63.5 mM), oxidized by a 4-fold molar excess of cyt c for 1 h, was freed of cyt c by ultrafiltration through a Centricon 3000 (from Amicon) (line 2).
cytc to a plateau [11]. At that plateau the rate of cyt c reduction may be equated to the rate of O22 production; because the cyt c is then outcompeting the spontaneous dismutation of O22. It was of interest to determine the number of HE molecules oxidized to E1 per O22 consumed. This was done at different fluxes of O22, achieved at different concentrations of xanthine oxidase. Figure 5 presents the amount of O22 produced (line 1) and of the HE oxidized to E1 (line 2) as a function of the xanthine oxidase used. As expected the O22 produced per minute was nearly a
HE catalyzes the dismutation of O22 The rate of reduction of cyt c by the xanthine oxidase reaction increases with the concentration of Table 1. Cyt c Reduction by HE HE (mM)
Cyt c (mM)
Cyt c Reduced (mM)
Cyt cred/HE
20 10 6 2 1
20 20 20 20 20
19 6 1.0 15 6 1.0 12 6 1.5 6 6 1.0 4.0 6 0.5
0.95 1.5 2 3 4
The reaction mixture contained 20 mM cyt c and the indicated concentration of HE in 50 mM potassium phosphate, pH 7.8. Cyt c reduction was followed at 550 nm using « 5 21,000 M21 cm21.
Fig. 5. Reduction of cyt c and oxidation of HE as a function of [XO]. Reaction mixtures contained 10 mM cyt c, 375 mM xanthine, and the indicated amounts of XO in 50 mM potassium phosphate at pH 7.8 and 25°C. Additional components were: line 1, 10 mM cyt c; or line 2, 63.5 mM HE. Panel A, the cyt c reduced (line 1), or the HE oxidized to E1 (line 2) per minute; Panel B the ratio of cyt c reduced per HE oxidized.
Hydroethidine to measure O22
Fig. 6. Comparison of HE and lucigenin assays. Cells grown in LB to A600 nm 5 0.5– 0.8 were washed and resuspended to A600 nm 5 0.5 in 50 mM potassium phosphate buffer pH 7.8 1 0.2% glucose and 0.2 mM lucigenin or 63.5 mM HE. Lucigenin luminescence was measured as previously described (14) exactly 4 min later. E1 fluorescence was measured in clarified extracts of cells incubated 30 min at 37°C in a shaking water bath, washed and French pressed, data presented were corrected for the fluorescence of extracts prepared from anaerobic controls treated in the same way.
829
the oxidation of 7.0 mM NH2OH to NO22 per 20 min and this NO22 production was 50% inhibited by 3.15 mM HE. Thus 3.15 mM HE intercepted 14.3 mM O22 or ;4.6 O22/HE. Since the divalent oxidation of HE to E1 requires only 2.0 O22 this indicates that HE can catalyze the dismutation of O22. This was further explored by comparing the rates of cyt c reduction and HE oxidation by a given flux of O22 produced by the XO reaction. A flux of O22 which caused the reduction of 12 mM cyt c/min was found to oxidize only 2.5 mM HE to E1/min. This was the case when the concentrations of cyt c and of HE were great enough to achieve maximal rates of cyt c reduction and of HE oxidation, respectively. In this case the ratio of O22/HE oxidized to E1 was 4.8, which again supports the idea of dismutation of O22 by HE.
linear function of [xanthine oxidase] (line 1), but the rate of oxidation of HE to E1 followed fluorimetrically was not (line 2). Thus the yield of HE oxidized to E1 decreased as the rate of production of O22 increased. This suggested that HE might catalyze the dismutation of O22 or alternately that O22 was being lost to the spontaneous dismutation [12] at higher fluxes of O22. To examine whether HE could catalyze the dismutation of O22, an indicating scavenger for O22 other than cytochrome c was necessary in order to avoid the problem of the direct reduction of cytochrome c by HE. Hydroxylamine, which is oxidized to NO22 by O22 [13], was chosen and it was ascertained that NH2OH did not directly react with E1 or with HE. A xanthine oxidase reaction which produced 28.6 mM O22 per 20 min, on the basis of cytochrome c reduction, was seen to cause
Table 2. Lucigenin Luminescence and HE Fluorescence in sodA sodB and Parental Strainsa Ratio sodA sodB/parental Strain Luminescence Fluorescence Luminescence/ Fluorescence a
JI132/AB1157
QC2510/QC2509
KK135/AN387
13.5 5.9 2.3
5.0 2.2 2.3
9.2 3.0 3.1
Data are from Fig. 6.
Fig. 7. Growth of JI132 and AB1157 in the presence of HE and E1. 50 ml overnight JI132 and AB1157 cultures in LB medium were transferred into 10 ml M9CA medium containing 63.5 mM HE or E1. Cells were grown aerobically, or anaerobically, in a Coy chamber, and growth was monitored by measuring A600 nm at intervals. Panel A, aerobic; panel B, anaerobic. Line 1 — JI132 1 E1; line 2 — JI132 1 HE; line 3 — JI132; line 4 — AB1157 1 E1; line 5 — AB1157 1 HE; line 6 — AB1157.
830
L. BENOV et al. DISCUSSION
Measuring itracellular [SOD] Lucigenin luminescence can be used as a measure of [SOD] in E. coli [14]. The conversion of HE to E1 should be similarly useful. Figure 6 presents the application of these two methods to three sets of sodA sodB, and the corresponding parental, strains. In all three cases the sodA sodB strains showed substantially more lucigenin luminescence and E1 fluorescence than did the corresponding parental strains. Yet the ratios of lucigenin luminescence in the SOD-null strains compared to the SOD-replete strains was greater by a factor of 2.3–3.1 than the corresponding ratios of E1 fluorescence. This is shown in Table 2 and may reflect the ability of HE to catalyze the dismutation of O22 thus muting the difference between SOD-null and SOD-replete strains.
The toxicity of HE Since E1 can intercalate into double stranded nucleic acids and since HE can be oxidized to E1 by O22; one would expect that the toxicity of HE should reflect intracellular levels of O22. That, in turn, suggests that HE should be more toxic to a sodA sodB than to a SODcompetent parental strain and further that its toxicity to the sodA sodB strain should be dioxygen-dependent. Comparison of lines 2 and 3 in Fig. 7A illustrates the toxicity of 63.5 mM HE to the sodA sodB E. coli; while lines 5 and 6 demonstrate the much smaller effect exerted on the parental strain. The greater toxicity of E1 on the two strains is presented by lines 1 and 4. When the same experiment was repeated under anaerobic conditions the toxicity of HE to the sodA sodB strain was much diminished, as shown by lines 2 and 3 in Fig. 7B and, as anticipated, E1 exerted its toxicity even in the absence of dioxygen (line 1). It is likely that E1 can be reduced within the cells and that the net redox balance is shifted towards oxidation in the SOD-null strains. That could explain the much greater toxicity of E1 to the sodA sodB (lines 1) than to the parental (lines 4) strain. The expectation is thus that there would be more E1 reduction in the parental strain and more HE oxidation in the SOD-null strain. The partial escape from E1 toxicity, after 4 h of anaerobic incubation with the SOD-null (lines 1 panel 7B), could thus be due to reduction of E1 to HE unopposed by its reoxidation by O22. It should be noted that the HE used contained some E1 due to autoxidation prior to the imposition of anaerobiosis. This explains the weak toxicity of HE to the SOD-null strain even under anaerobic conditions (compare lines 2 and 3 in panel B).
In accord with previous reports [1,2] we find HE to be oxidized to E1 by O22, whereas the rates of oxidation by O2 or by H2O2 are much slower. The reduction of E1 to HE, by glucose oxidase plus glucose, does increase cyt c reduction, but that is due to a direct effect of HE and not to HE-mediated O22 production; since this cyt c reduction was not SOD-inhibitable. The oxidations of HE by cyt c consumed more than the theoretical 2 cyt c/HE and the product was not E1. This indicates that cyt c can oxidize HE to species more oxidized than E1. At the same time it must be noted that E1 did not reduce cyt c. Hence the pathway of oxidation of EH by cyt c does not go through E1. It can be supposed that the univalent oxidation of EH by cyt c yields a radical E. which may dimerize before further oxidation by cyt c. O22, unlike cyt c, does oxidize HE to E1; but the yield of E1 was less than 1.0 per 2.0 O22. This suggested that, in addition to stoichiometrically scavenging O22, EH was capable of catalyzing the dismutation of O22. The following reactions may be proposed: HE 1 O 2 2 1 H 1
3
E 1 H 2O 2
(1)
E 1 O 2 2 1 2H 1
3
E 1 1 H 2O 2
(2)
3
HE 1 E
3
E 2 E
(4)
3
HE 1 O 2
(5)
•
E 1 E 1 H •
•
E 1 E •
1
•
E 1 O 22 1 H 1 •
•
1
(3)
Reactions 1–3 account for the oxidation of HE to E1 while reactions 1 plus 5 account for the dismutation of O22 catalyzed by HE. This ability of HE to catalyze the dismutation of O22 can explain the exponential appearance of line 1 in Fig. 1. Thus the rate of oxidation of HE to E1 increased with time because, as HE was consumed, it caused less dismutation of O22; leaving a larger proportion of the O22 to oxidize HE. We conclude that the HE conversion to E1 may be useful as a qualitative indicator of O22, but should not be used as a quantitative measure of this radical because it can catalyze the dismutation of O22 and therefore the yield of E1 produced per O22 will be less than stoichiometric and may be variable. Moreover, it is likely that the oxidation of HE by cytochrome c has some counterpart in the cell. Acknowledgements — This work was supported by grants from the Council for Tobacco Research, U.SA., Inc. (2871BR1) and the National Institutes of Health (HL56025-03).
Hydroethidine to measure O22 REFERENCES [1] Rothe, G.; Valet, G. Flow cytometric analysis of respiratory burst activity in phagocytes with hydroethidine and 29,79-dichlorofluorescin. J. Leukocyte Biol. 47:440 – 448; 1989. [2] Bindokas, V. P.; Jordan, J.; Lee, C. C.; Miller, R. J. Superoxide production in rat hippocampal neurons: selective imaging with hydroethidine J. Neurosci. 16:1324 –1326; 1989. [3] Carter, W. O.; Narayanan, P. K.; Robinson, J. P. Intracellular hydrogen peroxide and superoxide anion detection in endothelial cells. J. Leukocyte Biol. 55:253–258; 1994. [4] Liochev, S. I.; Fridovich, I. Superoxide from glucose oxidase or from nitroblue tetrazolium? Arch. Biochem. Biophys. 318:408 – 410; 1995. [5] Liochev, S. I.; Fridovich, I. Lucigenin (bis-N-methylacridinium) as a mediator of superoxide anion production. Arch. Biochem Biophys. 337:115–120; 1997. [6] Imlay, J.; Linn, S. Mutagenesis and stress responses induced in Escherichia coli by hydrogen peroxide. J. Bacteriol. 169:2967– 2976; 1987. [7] Carlioz, A.; Touati, D. Isolation of superoxide dismutase mutants in Escherichia coli: is superoxide dismutase necessary for aerobic life? EMBO J. 5:623– 630; 1986. [8] Pianzzola, M. J.; Soubes, M.; Touati, D. Overproduction of the rbo gene product from desulfovibrio species suppresses all deleterious effects of lack of superoxide dismutase in Escherichia coli. J. Bacteriol. 178:6736 – 6742; 1996
831
[9] Imlay, J. A metabolic enzyme that rapidly produces superoxide, fumarate reductase of Escherichia coli. J. Biol. Chem. 270:19767–19777; 1995. [10] Lowry, O. H.; Rosebrough, N. F.; Farr, A. L.; Randall, R. J. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265–275; 1951. [11] McCord, J. M.; Fridovich, I. The reduction of cytochrome c by milk xanthine oxidase. J. Biol. Chem. 243:5753–5760; 1968. [12] Bielski, B. H.; Cabelli, D. E.; Arudi, R. L. Reactivity of HO2/O22 radicals in aqueous solution. J. Phys. Chem. 14:1041–1097; 1985. [13] Elstner, E. F.; Heupel, A. Inhibition of nitrite formation from hydroxylammonium chloride: a simple assay for superoxide dismutase. Anal. Biochem. 70:616 – 620; 1976. [14] Liochev, S. I.; Fridovich, I. Lucigenin luminescence as a measure of intracellular superoxide dismutase activity in Escherichia coli. Proc. Nat. Acad. Sci. USA 94:2891–2896; 1997.
ABBREVIATIONS
HE— hydroethidine E1— ethidium GO— glucose oxidase SOD—superoxide dismutase