Free Radical Biology & Medicine. Vol. 10, pp. 69-77, 1991
0891-5849/91 $3.00 + .00 Copyright¢ 1991PergamonPress plc
Printedin the USA. All rightsreserved.
- Original • "Contribution USE OF CYANIDE AND DIETHYLDITHIOCARBAMATE IN THE ASSAY OF SUPEROXIDE DISMUTASES
JAWAID IQBAL* and PHILIP WHITNEYt Calvin and Flavia Oak Asthma Research and TreatmentFacility, PulmonaryResearch Laboratories (R-120), Departmentof Medicine, Universityof Miami School of Medicine, P.O. Box 016960 Miami, FL 33136 and VA Medical Center, Miami, FL 33125 (Received 27 April 1990; Revised 24 August 1990; Accepted 16 October 1990)
Abstract--Eucaryotes have two major forms of superoxide dismutase (SOD). Cu,ZnSOD and MnSOD; in most tissues Cu,ZnSOD is present in higher amountsthan MnSOD. To assay MnSOD, Cu,ZnSODcan be inhibitedselectivelyby millimolarconcentrations of cyanide ion. However, calculationof MnSOD activity from the differentialcyanide inhibitionassay is complex and small experimental errors can cause large errors in the calculated MnSOD activity. We have assessed how interactionof cyanide and hydrogen peroxide with cytochrome c can lead to further errors in the xanthineoxidase--cytochrome c assay for SOD. Alternatively, Cu,ZnSOD can be completely inactivatedby 50 mM diethyldithiocarbamate(DDC) at 30°(?for 1 h without affecting the activityof MnSOD. Since DDC reduces cytochrome c, the treated samples must be thoroughly dialyzed or desalted before assay. In the case of lung homogenates, dialysis is not an extra step since fresh, untreated samples must also be dialyzed or desalted before assaying by the cytochrome c method. Cu,ZnSOD activity is equal to the activity in the untreated sample minus the activity in the DDCtreated portion of the sample. Another copper chelator, triethylenetetramine,did not inactivate Cu,ZnSOD and could not be used instead of DDC. For accurate measurementof both enzymes in samples where MnSODcontributesonly a small fractionof the total SOD activity, the DDC method has the advantage that it provides a direct measure of the MnSOD activity without interferenceby Cu,ZnSOD. Keywords--Enzyme assay, Superoxide dismutase, Cyanide, Diethyldithiocarbamate,Cytochrome c, Hydrogen peroxide
the solution in question is valid in high cyanide. Cyanide is not an innocuous component of the assay mixture. Cyanide complexes with cytochrome c, 3'4 and inhibits cytochrome oxidase and cytochrome c peroxidase; 5 it also catalyzes the oxidation of alpha-ketoaldehydes and alpha-ketoalcohols in a superoxide-mediated mechanism. 6 In this paper, we report several ways cyanide can interfere with the SOD assay. In view of these complexities introduced by the use of cyanide, we considered other methods to measure MnSOD activity. Sodium dodecyl sulfate has been used to inactivate MnSOD selectively, 7 but this assay is better suited for samples with a low ratio of Cu,ZnSOD to MnSOD because MnSOD activity is calculated from the small difference in two large numbers representing total SOD and Cu,ZnSOD activities. The differential pH assay is based on the observation that Cu,ZnSOD appears to be l0 times more active at pH 10 than at pH 7.8 whereas MnSOD appears to be only twice as active at the higher pH. 5 The differential pH method is also ill suited for measuring MnSOD because the fraction of the total SOD activity contributed by MnSOD will be even lower at pH 10 than at pH 7.8.
Eucaryotes possess two major forms of superoxide dismutase (SOD), Cu,ZnSOD and MnSOD; Cu,ZnSOD is mainly cytosolic and MnSOD is mitochondrial. ~ Methods to measure the activities of these two forms selectively are based on differential responses of the enzymes to inhibitors, denaturants, or pH. The use of cyanide is based on the inhibition of Cu,ZnSOD but not MnSOD by the cyanide ion. 2 In tissues with high ratios of Cu,ZnSOD to MnSOD, errors in the assay in high concentrations of cyanide will not have a great influence on the calculated activity for Cu,ZnSOD because the activity in high cyanide is small compared to the uninhibited SOD activity. However, errors in the measurements in high cyanide will directly affect the calculated activity for MnSOD. Furthermore, before the differential cyanide inhibition method can be used to quantitate MnSOD activity, one must demonstrate that the assay of *Current address: PreclinicalScience Building, Georgetown University Medical Center, 3900 Reservoir Road N.W., Washington, DC 20007. tAuthor to whom correspondenceshould be addressed. 69
J. IQBAt_and P. WHITNEY
The best method for assaying samples with high ratios of Cu,ZnSOD to MnSOD would inactivate the Cu,ZnSOD while maintaining the full activity of MnSOD and the validity of the assay system. Inactivation of Cu,ZnSOD can be achieved by treatment with diethyldithiocarbamate (DDC), a reagent that complexes and removes copper from Cu,ZnSOD. 8'9 In this report, we described conditions that permit the use of DDC to assay MnSOD without the use of high concentrations of cyanide. Some of these results have appeared in an abstract. ~o
Male Sprague-Dawley rats (200-250 g) from Charles River Breeding Laboratories were maintained in the Animal Care Facility at the University of Miami. We purchased bovine xanthine oxidase from Boehringer Mannheim (Indianapolis, IN), bovine Cu,Zn superoxide dismutase from Diagnostic Data Co. (Mountain View, CA), hydroxyapatite and DEAE Bio Gel A from Bio Rad (Richmond, CA), and Sephadex G 100SF, xanthine, bathocuproinedisulfonic acid, diethylenetriaminepentaacetic acid, triethylenetetramine tetrahydrochloride, diethyldithiocarbamate (DDC) and horse heart cytochrome c (Type VI) from Sigma Chemical Co. (St. Louis, MO) In some experiments, we used cytochrome c that had been passed through a 1.5 × 90 cm column of Sephadex G-100SF in 50 mM potassium phosphate (KPi) buffer (pH 7.8). Properties of the chromatographed cytochrome c were not distinguishably different than the unfractionated protein.
Tissue preparation Tissue samples for SOD assays were obtained from rats that were killed by cutting the great vessels of the abdomen after anesthetization with sodium pentobarbitol ( - 5 0 mg/kg). Lungs were perfused with 0.9% NaC1. Each lung was homogenized in 10 mL 5 mM KPi buffer (pH 7.8) for 50 s with a Polytron Pl10/35 (Brinkman Instruments (Westbury, NY)) operated at highest speed. The homogenate was centrifuged at 27,000 g for 45 min. Rat liver homogenates were processed similarly, but centrifuged at 100,000 g for 60 min.
Purification of superoxide dismutases Rat lung Cu,ZnSOD was purified as described by Crapo et al. 5 except DEAE Bio Gel A was substituted for DE-52. MnSOD was purified 20-fold and separated from Cu,ZnSOD by chromatography through a column of hydroxyapatite. ~
Assay of SOD activi~, The xanthine oxidase-cytochrome c method was used for all assays of SOD activity. 5 A low concentration of sodium cyanide (0.015 mM) was used to inhibit cytochrome oxidase. A unit of SOD activity is the amount that halves the rate of reduction of cytochrome c. We only used 1 mL assay volume, so our unit of activity is three times that of Crapo et al. 5 Based on published equations 1213 and adding a correction for a blank measured in the absence of xanthine oxidase, a unit of activity can be calculated from eq. 1 where vo is the rate of reduction of cytochrome c (A55o - 0 . 0 2 5 m i n - l ) in the presence of xanthine oxidase but without SOD, v is the rate with xanthine oxidase and SOD, and v' is the rate with SOD but without xanthine oxidase. U=v
The value of v' was negligible for purified SOD samples or for total SOD in dialyzed or desalted lung homogenates. The procedure utilizing diethyldithiocarbamate (DDC) to inactivate Cu,ZnSOD required two portions of each sample if they were to be assayed for both Cu,ZnSOD and MnSOD. One was untreated control and the other was incubated at pH 7.8 with 50 mM DDC at 30°C for 1 h. Both samples were then desalted or dialyzed. Desalting was accomplished using Bio Rad Econo-Pac 10DG columns equilibrated and eluted with 50 mM KPj buffer (pH 7.8) -0.1 mM EDTA. For dialysis, we used three changes of 400 volumes of 5 mM KP i buffer (pH 7.8) -0.1 mM EDTA. The SOD activity in the control was total SOD activity; the DDC-treated portion had only MnSOD activity. Cu,ZnSOD activity was obtained by subtracting the activity of the treated sample from that of the control. To assess recovery from desalting or dialysis, we measured protein concentrations by the method of Bradford. 14 (Unless the samples were first precipitated with trichloroacetic acid, the bichinchonic acid protein assay ~5 was not appropriate for this purpose because small molecular weight components in lung homogenates contributed about 20% of the color that developed with the bichinchonic acid (BCA) reagent; as a result, the apparent recoveries determined with the BCA reagent were 20% lower than the actual protein recovery.) Since volume changes during dialysis were small, satisfactory corrections could also be made from the ratio of the total volume (or weight) of each sample before and after dialysis.
Other assays Xanthine oxidase activity was assayed by measuring uric acid production as determined from the rate of in-
Pitfalls in assaying superoxide dismutases
crease in A293in 50 mM KP i buffer (pH 7.8)-0.1 mM EDTA - 0.1 mM xanthine at 25°C. Catalase activity was assayed at 240 nm.16 Protein was analyzed by the method of Bradford 14 using bovine serum albumin as standard. Spectra were recorded with a Cary 14 spectrophotometer. Cytochrome c concentration and state of oxidation were determined from absorbances at 550 nm before and after reduction with - 1 mg sodium dithionite per 2.5 mL by using molar extinction coefficients given by Massay. 17
u.I (J Z
uric acid + 02 + H202 Cytc 2~ + 02
> > >
02 + H202 Cyt c 3+ q- H202 Cyt c 3+ CN
eq. 4 eq. 5 eq. 6
Equations 5 and 6 are reactions that can interfere with the assay under suboptimal conditions.
Influence of cyanide on the xanthine oxidase-cytochrome c system We will use the term "cyanide" to include both its acid and salt forms. Hydrocyanic acid is a weak acid with a pK a above 9; addition of 1.5 mM NaCN to 50 mM KPi buffer at pH 7.8 resulted in almost complete conversion to HCN and required the addition of 1.4 mM HCI to maintain the proper pH. Without HC1, addition of 1.5 or 6 mM NaCN increased the pH of the 50 mM buffer to 7.94 or 8.6, respectively. Since even small changes in pH will alter the specific activity measured in the SOD assay, maintenance of a constant pH is important. Cyanide has been widely employed in the xanthine oxidase-cytochrome c assay for SOD. Low concentrations of cyanide (5-50 ~M) have been used to inhibit peroxidases and cytochrome oxidase in crude samples. High concentrations of cyanide (1-2 mM) have been employed in a differential cyanide inhibition assay to inhibit Cu,ZnSOD and permit the quantitation of both Cu,ZnSOD and MnSOD. 2"18 These uses assume that cyanide does not influence the basic assay system in which superoxide, formed by xanthine oxidase during oxidation of xanthine to uric acid, is assayed by the rate it reduces ferricytochrome c. We have explored two interactions of cyanide with cytochrome c and assessed
WAVELENGTH (nm) Fig. 1. Difference spectrum of cyanide-ferricytochrome c versus ferricytochrome c. Ferricytochrome c (20 p.M) in 50 mM KP i buffer (pH 7.8)-0.1 mM EDTA was in sample and reference cuvettes for the baseline. A second spectrum was recorded 1.5 h after adding 1.5 mM NaCN + 1.4 mM HCI to the sample cuvette and an equal volume of buffer to the reference cuvette.
how they can influence the correlation of superoxide concentration with the observed rate of reduction of cytochrome c. The first interaction is the complex with ferricytochrome c in which cyanide displaces the methionine ligand to the heme iron and strongly inhibits the reduction of cytochrome C. 3'19 Binding of cyanide causes a shift in the visible spectrum leading to a large increase in absorbance at 417 nm, a large decrease at 403 nm, and several smaller changes including an increase at 550 nm (Fig. 1). The complex forms relatively slowly. We obtained a half-time of 10 min in 1.5 mM cyanide (Fig. 2); this is in good agreement with a half-time of 13 min calculated from the data of George and Tsou 3 that was obtained at a higher buffer concentration. For 1.5 mM cyanide plus 10 txM ferricytochrome c, the initial rate of increase in absorbance at 550 nm was less than 0.0005 m i n - ~, so it makes only a small contribution to the rate of change in absorbance as free ferricytochrome c is being reduced by superoxide in the SOD assay and would cause the SOD activity to be 4% lower than the true value. To correct for it, the 0.0005 min-1 should be subtracted from v0 in eq. 1. A more important consideration for the SOD assay is that the cyanide-cytochrome c complex is not reduced by superoxide, so complex formation decreases the ef-
J. IQBALand P. WHn'NEY
zw z "~ 30 LM
20 TIME (min)
Fig. 2. Rate of cyanide inhibition of reduction of ferricytochromec. Cyanide (1.5 raM) binding to 14 0.M cytochrome c in 50 mM KP, buffer (pH 8.0)~0.1 mM EDTA was followed at 417 nm (closed circles). Cyanide inhibition of the SOD assay was measured by adding 1.5 rnM NaCN to 10 ~M ferricytochromec in 50 mM KP, buffer (pH 8.0) -0.1 mM EDTA - 100 ~M xanthine at designated times before initiating the assay with xanthine oxidase and measuring the initial rate of reduction of cytochrome c at 550 nm (open circles): each point in this curve is from a separate experiment with a different preincubation period. fective concentration of cytochrome c. The other curve (open circles) in Fig. 2 shows how the initial rate of reduction of cytochrome c in SOD assay cocktail progressively decreases with increasing length of time for preincubation of cyanide with cytochrome c before initiating the assay with xanthine oxidase. The decrease in rate of reduction of cytochrome c in the assay system lags behind the rate of complex formation (Fig. 2) because cytochrome c in the assay medium is present at nearly saturating concentration. ]2 Consequently, a 5-min preincubation was required before there was a significant decrease in initial rate of reduction of cytochrome c in the SOD assay, and a relatively large fraction of cytochrome c had to be inactivated by cyanide before falling into the more linear region of the curve for the rate of reduction vs concentration of free cytochrome c. Inactivation of ferricytochrome c by 1.5 mM cyanide was not a problem if the cyanide was added immediately before initiating the reaction with xanthine oxidase. Inactivation is more rapid and becomes a more serious problem at higher concentrations of cyanide. An-
other consideration is that the rate of complex formation is very pH dependent with a peak at pH 9. 3 At pH 10 the rate was more than twice that at pH 7.83 The problem of cyanide inactivation of ferricytochrome c has been alluded to in the past 2"2° and some incorrect information has also appeared; 5 to our knowledge, this is the first published evaluation of its effect on the SOD assay. A second effect of cyanide is to inhibit oxidation of ferrocytochrome c by H202. In a solution of partially reduced cytochrome c (8 ~M ferricytochrome c and 2 txM ferrocytochrome c), 2 ~M H202 completely oxidized the ferrocytochrome c within 3 rain. The oxidation was inhibited 50% by 4.5 p.M cyanide, 73% by 15 txM cyanide, and 97% by 150 ~zM cyanide. Sodium azide was a much weaker inhibitor 05o = 3 raM). Catalase (1500 units/mL assay medium) was nearly as effective as 150 I~M cyanide. Catalase is inhibited by cyanide (K i = 8 p.M at pH 7), so catalase should not be used along with cyanide. Addition of xanthine oxidase (in the absence of xanthine) did not influence the rate, so lactoperoxidase was not an impurity in xanthine oxidase. The oxidation of ferrocytochrome c by H202 has a marked influence on the xanthine oxidase-cytochrome c system if the cytochrome c is not all in the oxidized form. When 20% of the cytochrome c was initially in the reduced form, the addition of xanthine oxidase brought about a brief phase of reduction of ferricytochrome c followed by a reversal that completely oxidized the cytochrome c (trace 1, Fig. 3) even though the xanthine oxidase was continuing to produce O ~ - . If cyanide was included in the assay medium, the oxidation reaction was inhibited and a linear trace was obtained for reduction of ferricytochrome c (trace 2, Fig. 3). If cyanide was added later in the assay, the oxidation was stopped at that point and a linear reduction trace obtained (trace 3, Fig. 3). An explanation of these results is that xanthine oxidase produces more H202 than O 2- at this pH 21 and 0 2- spontaneously dismutes to H202 and 0 2, thereby depleting 0 2 - and increasing H202. In the presence of cyanide, ferrocytochrome c is protected against oxidation by H202. The means of this protection is not known but is is likely due to a complex of cyanide with ferrocytochrome c. If such a complex is formed, it probably does not involve the heme, since no change in visible spectrum was noted when 100 txM NaCN was added to 40 p.M ferrocytochrome c. In relating these results to the SOD assay, it should be noted that the initial slope in trace 1, Fig. 3 is only about 70% of that in trace 2 and it is very transient. When the assay was begun with ferricytochrome c that had no ferrocytochrome c, then a linear initial trace was obtained that was not enhanced by adding cyanide. The rate of uric acid production by xanthine oxidase was found to be independent of catalase or 3 mM cyanide.
Pitfalls in assaying superoxide dismutases
halve the measured rate of reduction of cytochrome c. We conclude that it is best to use fresh solutions with cytochrome c almost completely in the oxidized form.
Differential cyanide inhibition assay
Fig. 3. Cyanide inhibition of oxidation of ferricytochrome c by H202. Trace 1. The assay was initiated by adding xanthine oxidase to 8 i.tM ferricytochrome c/2 I.tM ferricytochrome c/100 p.M xanthine in 50 mM KP, buffer (pH 7.8)-0.1 mM EDTA. The absorbance at 550 am was recorded with time. Trace 2 is the same as trace l except 0.15 mM NaCN was included. Trace 3 is the same as trace 1 except 0.15 mM NaCN was added 2 min after the assay was begun.
Therefore, if cyanide alters the apparent rate of reduction of ferricytochrome c it is probably due to an effect on cytochrome c. Our lot of Sigma type VI ferricytochrome c did not contain detectable amounts of ferrocytochrome c, but it has been reported that commercial ferricytochrome c may contain 5-20% ferrocytochrome c. 22 A sample that does contain significant levels of ferrocytochrome c can be used after it is oxidized with ferricyanide and then dialyzed or desalted. A second aspect of the problem is maintaining the cytochrome c in the oxidized form; significant levels of ferrocytochrome c were found in our assay cocktail of ferricytochrome c plus xanthine when it was kept at 25°C for several hours or at 4°C for several days. A small fraction of ferrocytochrome c can be tolerated in the assay of total SOD activity in the presence of 0.015 mM cyanide. However, if the objective is to compare initial rates at low and high concentrations of cyanide, the presence of ferrocytochrome c will complicate the assays because part of the effect of high cyanide will be more complete inhibition of oxidation by H20 2. In this case, the measured net rate of reduction of cytochrome c at lower cyanide concentration would be less than the actual rate of reduction, the measured rate would underestimate the concentration of 0 2 - , and a higher concentration of SOD would be required to
The differential cyanide inhibition assay is based on the lack of inhibition of MnSOD by cyanide at concentrations that cause nearly complete inhibition of Cu,ZnSOD. 2 Cyanide inhibition of rat Cu,ZnSOD was rapid and completely reversible; after incubation of Cu,ZnSOD with 5 mM cyanide at pH 7.8 for 30 min, inhibition was reversed within the time it took ( - 5 s) to get the cuvette into the spectrophotometer. Since there was no irreversible inhibition, the inhibition was characterized by a K i. We found a Ki of 130 IxM for inhibition of rat lung Cu,ZnSOD by cyanide. This Ki is the same as that reported for human Cu,ZnSOD, 23 twice as high as the K i for the bovine enzyme, 23 and less than half that reported for the equine enzyme. 24 The xanthine oxidase-cytochrome c assay generally includes a low concentration of cyanide (0.015 mM) to inhibit cytochrome oxidase. That low concentration of cyanide inhibited Cu,ZnSOD only 7%. The unit of activity of Cu,ZnSOD was defined as that activity expressed in 0.015 mM cyanide. At high cyanide concentration (1.5 mM), purified rat Cu,ZnSOD was 92% inhibited compared to its activity in 0.015 mM cyanide. Defining I as the fraction of Cu,ZnSOD inhibited at high cyanide, then the activities of Cu,ZnSOD (C) and MnSOD (M) were calculated from activities in 0.015 mM cyanide (L) and 1.5 mM cyanide (H) with the following equations:
H - (1 - I)L I
Measurement of I requires a purified preparation of the same Cu,ZnSOD as is in the mixture to be assayed. If most of the SOD activity is due to Cu,ZnSOD, the accuracy of the assay for Cu,ZnSOD will not be greatly influenced by errors in H. However, the activity of MnSOD is strongly dependent on accurate values of H and I, both of which are measured in the presence of a high concentration of cyanide. A 1% error in I would produce more than a 10% error in MnSOD activity. Measurement of I may be further complicated because crude extracts of tissues may reduce the effective cyanide concentrations, s Although Cu,ZnSOD inhibition will be more complete at higher concentrations of cyanide, the problem of residual Cu,ZnSOD activity must still be addressed. For instance, if Cu,ZnSOD accounts for 90%
J. IQBALand P. WHITNEY
~ j n