Mutation Research, 250 11991) 95-101 (¢ 1991 Elsevier Science Publishers B.V. All rights reserved 0027-5107/91/$03.50 ADONIS 002751079100168W
95
MUT 02514
Role of antioxidants in protecting cellular D N A from damage by oxidative stress Elizabeth A.L. Martins, Leda S. Chubatsu and Rog6rio Meneghini Department of Bi¢nrhemistry, Unit,ersity of S~o Paulo, CP 20780, CEP 01498, S~o Paulo (Brazil) (Accepted 5 April 1991)
Keywords: DNA damage; Oxidative stress; Antioxidants
Summary We have previously derived 2 V79 clones resistant to menadione (Mdl cells) and cadmium (Cdl cells), respectively. They both were shown to be cross-resistant to hydrogen peroxide. There was a modification in the antioxidant repertoire in these ceils as compared to the parental cells. Mdl presented an increase in catalase and glutathione peroxidase activities whereas Cdl cells exhibited an increase in metallothionein and glutathione contents. The susceptibility of the DNA of these cells to the damaging effect of H 2 0 2 was tested using the DNA precipitation assay. Both Mdl and Cdl DNAs were more resistant to the peroxide action. In the case of Mdl cells it seems clear that the extra resistance is provided by the increase in the two H 2 0 2 scavenger enzymes, catalase and glutathione peroxidase. In the case of Cdl ceils the activities of these enzymes as well as of superoxide dismutases ( C u / Z n and Mn) are unaltered as compared to the parental ceils. The facts that parental cells exposed to 100/zM Zn 2÷ in the medium exhibit an increase in metallothionein but not in glutathione and that these ceils become more resistant to the DNA-damaging effect of H 2 0 2 suggest that this protein might play a protective role in vivo against the O H radical attack on DNA.
An interesting possibility to modify the biochemical repertoire of cells in a relatively specific way is by deriving cell variants resistant to toxic agents. Quite often an alteration in the amount of one or a few cell components is involved in the
Correspondence: Dr. R. Meneghini, Department of Biochemistry, University of Silo Paulo, CP 20780, CE 01498, Silo Paulo (Brazil).
Abbreciations: SOD, superoxide dismutase; GPX, glutathione peroxidase; MT, metallothionein; PBS, phosphate-buffered saline (8.1 mM Na2HPO 4, 1.47 mM KH2PO4, 1.68 mM KCI, 137 mM NaCI, pH 7.0).
increase in resistance and the variants may be useful for investigation of biochemical mechanisms implicated in a given phenomenon. Recently we develop cells resistant to menadione, a vitamin K-derived quinone (Martins and Meneghini, 1990) and to cadmium (Mello-Filho et al., 19881. In both cases cross-resistance to the lethal action of hydrogen peroxide was observed. In menadione-resistant cells significant increases in catalase activity and in glutathione content were observed (Martins and Meneghini, 1990) whereas in cadmium-resistant cells there was a dramatic increase in the amount of the protein metallothionein (Mello-Filho et al., 1988) and a signifi-
96
cant increase in glutathione (Chubatsu ct al., 19911. No studies had bccn carried out in these cells to determine the genotoxic effect of hydrogen peroxide, a common end product of many types of oxidative stresses. The production of DNAstrand breaks by this peroxide has been investigated in several cell lines and the results led us to propose a model according to which DNA-bound Fell reacts with H , O 2 by the Fenton reaction (Fc2~+ H 2 0 _, ---, Fe 3" + O H + O H ") to generate O H radical; this reactive site-generated radical will then attack D N A to produce damage (Meneghini, 19881. This model has bccn supported by others (Cochrane et al., 1988); but alternative proposals appeared according to which DNA-strand breaks produced by H , O ~ arc brought about by Ca-induced nuclease (Cantoni et al., 1989) or by lipoperoxidcs (Ochi and Ccrutti, 1987). To discriminate among these possibilities is not a simple task and requires multiple approaches. One of them is the use of variant cells that exhibit different susceptibility to H 20 2 in terms of DNA-strand breaks, which may then be compared to their biochemical properties in terms of antioxidant repertoire, in the present study wc determined the production of DNA-strand breaks by H 2 0 2 in cadmium-resistant and menadioncresistant ceils. Both were shown to bc less susceptible than the parental cells, indicating that these variants might be useful for the study of mechanisms of DNA-strand breaks by oxidative stress. Material and methods
Materials Cell culture and derication of cell cariants M8 is a clone of V79 Chinese hamster lung fibroblast. The cells wcrc routinely grown in Dulbccco's modified Eagle's medium, pH 7.0, supplemcntcd with 10% ( v / v ) fetal calf serum, 472 units of pcnicillin/ml and 94 ~zg of streptomycin/ml. The cells wcrc kept in humidified C O 2 / a i r (1 : 19) at 37 "C. Mycoplasma-free ceils were used in the experiments, as determined by staining with Hoechst 33528 (Fox, 1981). Mdl and Cdl are clones derived from M8 by selection for resistance to menadione and cadmium, respectively,
after increasing stresses produced by these compounds, in a chronic fashion. I)ctails of the procedure have been previously described (MelloFilho et al., 1988; Martins and Mencghini, lt,~90). Mdl and Cdl were routinely grown as M8 cells, except that they were supplemented with 1.3× 10 s M menadione or 220 # M CdSO,~/II(I u M ZnSO~, respectively.
Irradiation A cobalt radioactive source of y-irradiation was used. at a dose rate of 7.62 × 104 r a d / h . Cells were irradiated in 3-cm pctri dishes, containing PBS. l lydrogen peroxide treatment Cells were labeled for 24 h in medium containing 1 izCi/ml 3H-methylthymidinc (50 C i / m m o l c , Amersham), after which they were trypsinized and seeded into multi-wells (Costar, 2 cm2/16 mm diameter well). Each well received I(XI.(I(X) cells, which were allowed to grow for 15 h in radioactive-free medium. At this point the medium was removed, the ceils were washed with PBS and supplied with I ml of PBS containing the indicated H 2 0 2 concentration, determined just beforc the experiment (Cotton and Dunford, 19731. Incubation proceeded for 30 min at 37 ° C, in the dark. It is important to note that the indicated H : O 2 concentrations correspond to those at the beginning of the incubation, since cellular enzymes rapidly consume H_,O 2 from the medium (Hoffmann ct al., 1984). For this reason, any consideration of the H ~O~ dose effect should take into account the number of cells, which will greatly influence the kinetics of H ~O~ consumption. DNA precipitation assay This is a modified version of the assay described by Olive (19881. After cell treatment 300 /11 of lysis buffer (10 mM Tris, 10 mM EDTA, 0.05 M N a O H , 2% SDS, pH 12.4) and 3(X1/.~1 of 120 mM KCI are added to the well. Incubation is at 6 5 ° C for 10 min in a water bath, after which 400 p,l of the lysate is transferred into Eppcndorf tubes and kept for 5 rain in an ice bath. A D N A - p r o t e i n K - S D S complex is formed in this process, which precipitates at low temperature.
97
Low-molecular single-strand DNA is released from the bulk of the DNA during this procedure and is separated from the precipitate by centrifugation at 3500 rpm for 10 min at 10°C. The supernatant is transferred to 1.5 × 2.0 cm pieces of thick filter paper; the sediment is dissolved in 100 >1 of water at 6 5 ° C and transferred to another piece of filter paper. These are serially washed in 5% TCA, ethanol (twice) and acetone, dried and their radioactivity counted in a scintillation counter. For plotting, the percentage of precipitated DNA is calculated for each sample and normalized to the value of precipitated DNA in the untreated controls, taken as 100%. The real percentage of precipitated DNA in untreated controls averaged 90% throughout the experiments. In all plots the points represent the mean of 3 independent determinations and the vertical bars represent the standard deviation. Results
The DNA precipitation assay is very sensitive and convenient to detect single-strand breaks (Olive, 1988). In our hands this assay was comparable in sensitivity to the alkaline elution assay (Kohn et ai., 1972) and had the advantages of being faster and producing more reproducible results. To estimate the number of single-strand breaks, an experiment was run in which V79 M8 cells were y-irradiated and processed by the DNA precipitation assay. The results (Fig. 1) show that the percentage of precipitated DNA correlates PRECIPITATED
DNA (%)
exponentially with the radiation dose according to the equation: log P = 2.01 - 2.039 × 10 -.4 (dose)
(1)
where P is the percentage of precipitated DNA. Because of the linear relationship between strand breaks and radiation dose (Lehmann and Ormerod, 1970) it is possible to establish an exponential relationship between percentage of precipitated DNA and strand breaks. In order to do that we used the assay of sedimentation in alkaline sucrose gradient ( H o f f m a n n and Meneghini, 1979), which directly gives the number of strand breaks produced by the dose of ",/-irradiation. The results (not shown) revealed a factor of 2.73× 10 -'2 b r e a k s / d a l t o n per rad, very close to figures determined by others (Kohn et ai., 1972). In this experiment the dose range was 5-25 krad. Combining this factor with expression 1 gives: no. strand breaks/dalton = 10 -8 ( 2 . 6 9 - 1.34 x log P )
(2)
Fig. 2 shows how hydrogen peroxide causes a loss in the ability of V79 M8 DNA to precipitate. The effect is linear up to 4 0 / z M , above which it tends to level off. We previously determined this tendency for DNA-strand breaks to reach saturation (Hoffmann et al., 1984) indicating that some factor is limiting, possibly an iron-reducing agent or the iron ion itself. Using equation 2 on the linear portion of the curve gives the yield of 1.42 × 10 -m strand breaks per dalton per ~ M H202.
801 6o~ 50 i__ o
(~
2
-
r
400
-%-- . . . . .
600 RADS
~
.....
800
1(500
1200
Fig. 1. DNA precipitation assay to determine single-strand breaks produced by y-radiation.
We previously derived a cell line resistant to menadione (Mdl) which turned out to be more resistant to the lethal action of H 2 0 2 than the M8 parental cells (Martins and Meneghini, 1990). Fig. 3 shows that Mdl cells are also more resistant to the DNA-damaging effect of H 2 0 2 than M8 cells. The same concentration of this peroxide produces 11 times more strand breaks in M8 cells than in Mdl cells. Cdl cells were derived to be resistant to cadmium and they also showed cross-resistance to the lethal effect of H 2 0 2 , as compared to the
98 120 ~
PRECIPITATED D N A (°/o) .
.
.
.
.
.
.
.
.
.
.
.
.
PRECIt'tTA--D
DNA(°/o
.
+OOor i
+°!
\ -'10. 30
30[ ..... o
,o
so o PEROXIDE (t~M)
so
HYDROGEN
7o
80
Fig. 2. DNA precipitation assay to determine single-stramt breaks in cells CXlm)sed to I1 ,O,.
parental M8 cells (Mello-Filho ct al., 1988). Fig. 4 shows that Cdl cells are also more resistant to the DNA-damaging action of H z O 2 than M8 cells. In this case the difference is only 2-fold and seems to reflect the narrower variation in survival to H z O ~ exhibited by Cdl cells, as compared to Mdl cells (Martins and Mencghini, 1990; MelloFilho et al., 1988). A pertinent question is the nature of the agent(s) responsible for the extra resistance of D N A to H z O 2 in these cell variants. For Mdl cells the SOD activity (total) remained unaltered as compared to the parental M8 cells, whereas a 2.8-fold increase in catalase activity was observed (Martins and Meneghini, 1990). On the other
[
PRECIPITATED
DNA ( % )
~
i
90~ ~
~
6O
401 . o
. o
. 2o
1
.
3o
95
MUT 02514
Role of antioxidants in protecting cellular D N A from damage by oxidative stress Elizabeth A.L. Martins, Leda S. Chubatsu and Rog6rio Meneghini Department of Bi¢nrhemistry, Unit,ersity of S~o Paulo, CP 20780, CEP 01498, S~o Paulo (Brazil) (Accepted 5 April 1991)
Keywords: DNA damage; Oxidative stress; Antioxidants
Summary We have previously derived 2 V79 clones resistant to menadione (Mdl cells) and cadmium (Cdl cells), respectively. They both were shown to be cross-resistant to hydrogen peroxide. There was a modification in the antioxidant repertoire in these ceils as compared to the parental cells. Mdl presented an increase in catalase and glutathione peroxidase activities whereas Cdl cells exhibited an increase in metallothionein and glutathione contents. The susceptibility of the DNA of these cells to the damaging effect of H 2 0 2 was tested using the DNA precipitation assay. Both Mdl and Cdl DNAs were more resistant to the peroxide action. In the case of Mdl cells it seems clear that the extra resistance is provided by the increase in the two H 2 0 2 scavenger enzymes, catalase and glutathione peroxidase. In the case of Cdl ceils the activities of these enzymes as well as of superoxide dismutases ( C u / Z n and Mn) are unaltered as compared to the parental ceils. The facts that parental cells exposed to 100/zM Zn 2÷ in the medium exhibit an increase in metallothionein but not in glutathione and that these ceils become more resistant to the DNA-damaging effect of H 2 0 2 suggest that this protein might play a protective role in vivo against the O H radical attack on DNA.
An interesting possibility to modify the biochemical repertoire of cells in a relatively specific way is by deriving cell variants resistant to toxic agents. Quite often an alteration in the amount of one or a few cell components is involved in the
Correspondence: Dr. R. Meneghini, Department of Biochemistry, University of Silo Paulo, CP 20780, CE 01498, Silo Paulo (Brazil).
Abbreciations: SOD, superoxide dismutase; GPX, glutathione peroxidase; MT, metallothionein; PBS, phosphate-buffered saline (8.1 mM Na2HPO 4, 1.47 mM KH2PO4, 1.68 mM KCI, 137 mM NaCI, pH 7.0).
increase in resistance and the variants may be useful for investigation of biochemical mechanisms implicated in a given phenomenon. Recently we develop cells resistant to menadione, a vitamin K-derived quinone (Martins and Meneghini, 1990) and to cadmium (Mello-Filho et al., 19881. In both cases cross-resistance to the lethal action of hydrogen peroxide was observed. In menadione-resistant cells significant increases in catalase activity and in glutathione content were observed (Martins and Meneghini, 1990) whereas in cadmium-resistant cells there was a dramatic increase in the amount of the protein metallothionein (Mello-Filho et al., 1988) and a signifi-
96
cant increase in glutathione (Chubatsu ct al., 19911. No studies had bccn carried out in these cells to determine the genotoxic effect of hydrogen peroxide, a common end product of many types of oxidative stresses. The production of DNAstrand breaks by this peroxide has been investigated in several cell lines and the results led us to propose a model according to which DNA-bound Fell reacts with H , O 2 by the Fenton reaction (Fc2~+ H 2 0 _, ---, Fe 3" + O H + O H ") to generate O H radical; this reactive site-generated radical will then attack D N A to produce damage (Meneghini, 19881. This model has bccn supported by others (Cochrane et al., 1988); but alternative proposals appeared according to which DNA-strand breaks produced by H , O ~ arc brought about by Ca-induced nuclease (Cantoni et al., 1989) or by lipoperoxidcs (Ochi and Ccrutti, 1987). To discriminate among these possibilities is not a simple task and requires multiple approaches. One of them is the use of variant cells that exhibit different susceptibility to H 20 2 in terms of DNA-strand breaks, which may then be compared to their biochemical properties in terms of antioxidant repertoire, in the present study wc determined the production of DNA-strand breaks by H 2 0 2 in cadmium-resistant and menadioncresistant ceils. Both were shown to bc less susceptible than the parental cells, indicating that these variants might be useful for the study of mechanisms of DNA-strand breaks by oxidative stress. Material and methods
Materials Cell culture and derication of cell cariants M8 is a clone of V79 Chinese hamster lung fibroblast. The cells wcrc routinely grown in Dulbccco's modified Eagle's medium, pH 7.0, supplemcntcd with 10% ( v / v ) fetal calf serum, 472 units of pcnicillin/ml and 94 ~zg of streptomycin/ml. The cells wcrc kept in humidified C O 2 / a i r (1 : 19) at 37 "C. Mycoplasma-free ceils were used in the experiments, as determined by staining with Hoechst 33528 (Fox, 1981). Mdl and Cdl are clones derived from M8 by selection for resistance to menadione and cadmium, respectively,
after increasing stresses produced by these compounds, in a chronic fashion. I)ctails of the procedure have been previously described (MelloFilho et al., 1988; Martins and Mencghini, lt,~90). Mdl and Cdl were routinely grown as M8 cells, except that they were supplemented with 1.3× 10 s M menadione or 220 # M CdSO,~/II(I u M ZnSO~, respectively.
Irradiation A cobalt radioactive source of y-irradiation was used. at a dose rate of 7.62 × 104 r a d / h . Cells were irradiated in 3-cm pctri dishes, containing PBS. l lydrogen peroxide treatment Cells were labeled for 24 h in medium containing 1 izCi/ml 3H-methylthymidinc (50 C i / m m o l c , Amersham), after which they were trypsinized and seeded into multi-wells (Costar, 2 cm2/16 mm diameter well). Each well received I(XI.(I(X) cells, which were allowed to grow for 15 h in radioactive-free medium. At this point the medium was removed, the ceils were washed with PBS and supplied with I ml of PBS containing the indicated H 2 0 2 concentration, determined just beforc the experiment (Cotton and Dunford, 19731. Incubation proceeded for 30 min at 37 ° C, in the dark. It is important to note that the indicated H : O 2 concentrations correspond to those at the beginning of the incubation, since cellular enzymes rapidly consume H_,O 2 from the medium (Hoffmann ct al., 1984). For this reason, any consideration of the H ~O~ dose effect should take into account the number of cells, which will greatly influence the kinetics of H ~O~ consumption. DNA precipitation assay This is a modified version of the assay described by Olive (19881. After cell treatment 300 /11 of lysis buffer (10 mM Tris, 10 mM EDTA, 0.05 M N a O H , 2% SDS, pH 12.4) and 3(X1/.~1 of 120 mM KCI are added to the well. Incubation is at 6 5 ° C for 10 min in a water bath, after which 400 p,l of the lysate is transferred into Eppcndorf tubes and kept for 5 rain in an ice bath. A D N A - p r o t e i n K - S D S complex is formed in this process, which precipitates at low temperature.
97
Low-molecular single-strand DNA is released from the bulk of the DNA during this procedure and is separated from the precipitate by centrifugation at 3500 rpm for 10 min at 10°C. The supernatant is transferred to 1.5 × 2.0 cm pieces of thick filter paper; the sediment is dissolved in 100 >1 of water at 6 5 ° C and transferred to another piece of filter paper. These are serially washed in 5% TCA, ethanol (twice) and acetone, dried and their radioactivity counted in a scintillation counter. For plotting, the percentage of precipitated DNA is calculated for each sample and normalized to the value of precipitated DNA in the untreated controls, taken as 100%. The real percentage of precipitated DNA in untreated controls averaged 90% throughout the experiments. In all plots the points represent the mean of 3 independent determinations and the vertical bars represent the standard deviation. Results
The DNA precipitation assay is very sensitive and convenient to detect single-strand breaks (Olive, 1988). In our hands this assay was comparable in sensitivity to the alkaline elution assay (Kohn et ai., 1972) and had the advantages of being faster and producing more reproducible results. To estimate the number of single-strand breaks, an experiment was run in which V79 M8 cells were y-irradiated and processed by the DNA precipitation assay. The results (Fig. 1) show that the percentage of precipitated DNA correlates PRECIPITATED
DNA (%)
exponentially with the radiation dose according to the equation: log P = 2.01 - 2.039 × 10 -.4 (dose)
(1)
where P is the percentage of precipitated DNA. Because of the linear relationship between strand breaks and radiation dose (Lehmann and Ormerod, 1970) it is possible to establish an exponential relationship between percentage of precipitated DNA and strand breaks. In order to do that we used the assay of sedimentation in alkaline sucrose gradient ( H o f f m a n n and Meneghini, 1979), which directly gives the number of strand breaks produced by the dose of ",/-irradiation. The results (not shown) revealed a factor of 2.73× 10 -'2 b r e a k s / d a l t o n per rad, very close to figures determined by others (Kohn et ai., 1972). In this experiment the dose range was 5-25 krad. Combining this factor with expression 1 gives: no. strand breaks/dalton = 10 -8 ( 2 . 6 9 - 1.34 x log P )
(2)
Fig. 2 shows how hydrogen peroxide causes a loss in the ability of V79 M8 DNA to precipitate. The effect is linear up to 4 0 / z M , above which it tends to level off. We previously determined this tendency for DNA-strand breaks to reach saturation (Hoffmann et al., 1984) indicating that some factor is limiting, possibly an iron-reducing agent or the iron ion itself. Using equation 2 on the linear portion of the curve gives the yield of 1.42 × 10 -m strand breaks per dalton per ~ M H202.
801 6o~ 50 i__ o
(~
2
-
r
400
-%-- . . . . .
600 RADS
~
.....
800
1(500
1200
Fig. 1. DNA precipitation assay to determine single-strand breaks produced by y-radiation.
We previously derived a cell line resistant to menadione (Mdl) which turned out to be more resistant to the lethal action of H 2 0 2 than the M8 parental cells (Martins and Meneghini, 1990). Fig. 3 shows that Mdl cells are also more resistant to the DNA-damaging effect of H 2 0 2 than M8 cells. The same concentration of this peroxide produces 11 times more strand breaks in M8 cells than in Mdl cells. Cdl cells were derived to be resistant to cadmium and they also showed cross-resistance to the lethal effect of H 2 0 2 , as compared to the
98 120 ~
PRECIPITATED D N A (°/o) .
.
.
.
.
.
.
.
.
.
.
.
.
PRECIt'tTA--D
DNA(°/o
.
+OOor i
+°!
\ -'10. 30
30[ ..... o
,o
so o PEROXIDE (t~M)
so
HYDROGEN
7o
80
Fig. 2. DNA precipitation assay to determine single-stramt breaks in cells CXlm)sed to I1 ,O,.
parental M8 cells (Mello-Filho ct al., 1988). Fig. 4 shows that Cdl cells are also more resistant to the DNA-damaging action of H z O 2 than M8 cells. In this case the difference is only 2-fold and seems to reflect the narrower variation in survival to H z O ~ exhibited by Cdl cells, as compared to Mdl cells (Martins and Mencghini, 1990; MelloFilho et al., 1988). A pertinent question is the nature of the agent(s) responsible for the extra resistance of D N A to H z O 2 in these cell variants. For Mdl cells the SOD activity (total) remained unaltered as compared to the parental M8 cells, whereas a 2.8-fold increase in catalase activity was observed (Martins and Meneghini, 1990). On the other
[
PRECIPITATED
DNA ( % )
~
i
90~ ~
~
6O
401 . o
. o
. 2o
1
.
3o
