ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 26, No. 1 Copyright © 1996, Institute for Clinical Science, Inc.
Interaction of Carcinogenic Metal Compounds with Deoxyribonucleic Acid Repair Processes ANDREA HARTWIG, Ph.D., REGINA SCHLEPEGRELL, HEIKE DALLY, and MAIKE HARTMANN Department of Biology and Chemistry, University of Bremen, 28334 Bremen, Germany
ABSTRACT The potentials of nickel(II) and cadmium(II) to interfere with the repair of different types of deoxyribonucleic acid (DNA) lesions was investigated. Concerning the nucleotide excision repair pathway, nickel(II) has been shown to reduce the incision and the ligation frequency after ultraviolet (UV)-irradiation. When applying a gel mobility shift assay and HeLa nuclear cell free extracts, nickel(II) diminishes the specific binding of a protein to UV-damaged DNA, suggesting that nickel(II) interferes with the DNA-protein interactions involved in the damage recognition after UV-irradiation. Similarly, the incision frequency is reduced in the pres ence of low concentrations of cadmium(II). Concerning the repair of oxi dative DNA damage induced by visible light, non-cytotoxic concentrations of nickel(II) caused a complete repair inhibition of DNA base modifica tions like 7,8-dihydro-8-oxoguanine (8-hydroxyguanine) and of DNA strand breaks. Since the repair of DNA damage is essential for the prevention of cancer, its inhibition may account for the carcinogenic action of the respec tive metal compounds. Introduction Compounds of nickel, cadmium, cobalt and arsenic are well established carcino gens to humans and to experimental ani mals. However, the mechanisms leading to tumor formation are still not under stood, since the induction of DNA damSend reprint requests to: Dr. Andrea Hartwig, University of Bremen, Department of Biology and Chemistry, Postfach 330440, 28334 Bremen, Ger many.
age and the mutagenic potentials are rather weak and mainly restricted to cytotoxic concentrations of the metals. In contrast, cytotoxicity and genotoxicity enhancing effects in combination with other DNA damaging agents are more pronounced and observed at lower concentrations.1 Based on these findings, the potential of carcinogenic metal compounds to interfere with the repair of different types of DNA lesions was investigated. Possible interactions were examined
31 0091-7370/96/0100-0031 $01.20 © Institute for Clinical Science, Inc.
32
HARTWIG, SCHLEPEGRELL, DALLY, AND HARTMANN
with (1) the nucleotide excision repair system, which represents the major pathway eliminating a broad spectrum of DNA lesions induced by many environ mental mutagens from the genome, and (2) the repair of oxidative DNA lesions, since they are induced continuously owing to cellular oxygen metabolism; if not repaired, they are implicated in muta tions, cancer and aging.2 To investigate the effects of metal com pounds on nucleotide excision repair, UVC light (254 nm) was used as a well characterized DNA damaging agent, since both major DNA photoproducts, the cyclobutane pyrimidine dimer and the pyrimidine-(6-4)-pyrimidone-photoproduct ((6-4)-photoproduct), are removed by this repair pathway.3 In pre vious studies, nickel(II), as w ell as cobalt(II), arsenic(III) and cadmium(II), have been shown to dim inish the removal of UV-induced cyclobutane pyrimidine dimers.1 The present study was undertaken to elucidate the effects on distinct steps of the repair process and the potentially underlying mechanisms. Concerning the repair of oxidative DNA damage and its potential inhibition, visible light was applied, which induces predominantly 8-hydroxyguanine and, to a lesser extent, DNA strand breaks.4 The induction and removal of the DNA base modifications were monitored by their sensitivity towards the bacterial formamidopyrimidine-DNA glycosylase (Fpg protein). Special emphasis will be given to the results obtained with nickel(II), and the effects observed in comparative studies w ith cadm ium (II), cobalt(II) and arsenic(III) will be summarized. Materials and Methods Cell Culture
containing 5 percent fetal bovine serum. The cultures were incubated at 37°C with 5 percent C 0 2 in air and 100 per cent humidification. UV-lRRADIATION
Ultraviolet irradiation of cells was per formed with a General Electric germi cidal lamp.* I r r a d ia t io n
w it h
V is ib l e L ig h t
Irradiation was carried out essentially as described by Pflaum et al.4 The cells were illuminated as monolayers in cell culture dishes covered with phosphate buffered saline (140 mM NaCl, 3 mM KC1, 8 mM Na2H P0 4 1 mM KH2P 04, 1 mM CaCl2, 0.5 mM MgCl2, and 0.1 per cent glucose) on ice with a 1000 W halo gen lamp for 20 min. in a distance of 33 cm corresponding to 450 kj/m2. D e t e c t io n o f D N A S t r a n d b y A l k a l in e U n w in d in g
Breaks
The strand breaks were detected as described previously. Briefly, 2 x 10® HeLa cells were allowed to attach for at least 24 h and treated as described for the respective experiments. Afterwards, the medium was removed and an alkaline solution was added containing 0.03 M NaOH, 0.02 M Na 2H P 0 4 and 0.9 M NaCl. Separation of single- and double stranded DNA was perform ed on hydroxyapatite columns. The DNA con tent of both fractions was determined fluorimetrically by adding the dye bisbenzimidetrihydrochloride.t The num ber of DNA strand breaks was calculated from calibration with X-rays.
HeLa cells were grown as monolayers* Bioblock scientific, 254 nm. in minimal essential medium (MEM) t Hoechst 33258
INTERACTION OF CARCINOGENIC METAL COMPOUNDS WITH DNA REPAIR PROCESSES
Q u a n t if ic a t io n o f F p g -S e n s it iv e S it e s
The procedure applied to detect enzym e sen sitiv e sites has b een described in detail previously.6 Briefly, cells were lysed as monolayers (0.006 M Na2H P04, 0.001 M KH2P 04) 0.137 M NaCl, 0.003 M KC1, 0.1% octyl phenoxy polyethoxy ethanol (Triton X 100), and treated with a high salt solution (2 M NaCl, 0.01 M ethylene diaminotetraacetic acid [EDTA] and 0.002 M Tris[hydroxymethyl]aminomethane [Tris], pH 8.0). The nucleoids were then incu bated with the Fpg protein (1 (xg/ml) in buffer consisting of 0.05 M sodium phos phate, pH 7.5, 0.01 M EDTA, and 100 mM NaCl. At the end of incubation, an alkaline solution was added yielding a final concentration of 0.07 N NaOH, 0.013 M EDTA, and 0.37 M NaCl, pH 12.3, and the DNA was allowed to unwind for 30 min in the dark. The fur ther steps of neutralization, separation of single- and double-stranded DNA and quantification of lesion frequency were performed as described above. G e l M o b i l i t y S h i f t A ssa y
Nuclear protein extracts were prepared from HeLa cells essentially as described by Schreiber et al.7 To detect damagespecific DNA-protein interactions, a digoxygenin-labeled synthetic oligonu cleotide (48 bp) was applied which was either unirradiated or irradiated with 18 kj/m 2 UVC. For the binding reaction, 2 |xg protein extract were pretreated with 300 fmol of the unlabeled oligonucleo tide in a gel shift buffer (13.3 mM Hepes, 9.6% glycerin, 100 mM KC1, 5 mM MgCl2, 1 mM EDTA, 1.2 jxg bovine serum albumin, 0.4 mM dithiothreitol, 75 (xM phenylmethylsulfonyl fluoride, and 4 ng leupeptin, pH 7.9) for 10 min at room temperature. In addition, 0.5 |xg poly[d(I-C)] as well as 60 fmol of the digoxy-
33
genin-labeled oligonucleotide were added for another 35 min. Afterwards, the binding mixture was loaded on a poly acrylamide gel (6 %; 45 mM Tris, 45 mM boric acid, 1 mM EDTA), and electropho resis was conducted at 110 V for 2.5 h. Southern blot was done in a semi-dry electro-blotting apparature applying a positively charged nylon membrane. The detection of the digoxygenin-labeled oli gonucleotide was performed calorimetrically by the alkaline phosphatase conju gated to an Anti-digoxygenin antibody using nitroblue tetrazolium salt (NBT) and 5-bromo-4-chloro-3-indoyl phos phate (X-phosphate) as substrates. Results and Discussion I n t e r a c t io n s w it h N u c l e o t id e E x c is io n R e p a ir b y N ic k e l (II)
Nickel(II) has been shown to diminish the removal of UV-induced cyclobutane pyrimidine dimers8’9; however, the mechanism leading to repair inhibition remains to be elucidated. Nucleotide excision repair can be roughly divided into four different steps, namely: (1) dam age recognition, (2) incision on both sites of the lesion with the subsequent dis placement of the damaged oligonucleo tide, (3) polymerization, and (4) ligation of the repair patches.10 During the repair process, DNA strand breaks are gener ated owing to incisions, which are ligated again at the end of the repair event. In a previous study by us ,9 the alkaline unwinding technique was applied to fol low the transient appearance of DNA strand breaks after UV-irradiation. In the presence of nickel(II), less DNA strand breaks were detected, suggesting an interference with the damage recogni tion/incision step in excision repair. However, once DNA strand breaks are generated, they remain open for a pro longed period of time in the presence of nickel(II), indicating that the ligation
34
HARTWIG, SCHLEPEGRELL, DALLY, AND HARTMANN
step is inhibited as well.9,11 Since the ini tiation of repair events is very complex in eucaryotic cells, involving the coordi nated action of at least 15 to 18 polypep tides,12 the interference with this process by nickel(II) was investigated in more detail. One important prerequisite for the initiation of repair events is the recogni tion of the respective DNA lesions. To investigate w hether or not nickel(II) interferes with the damage recogni tion step, its effect was studied on the binding of a protein exerting high affinity for UV-damaged DNA. A HeLa nuclear cell free extract was prepared and, by applying a gel mobility shift assay, a pro tein was identified which binds with
Lane UV Ni(II) [jiM]
1
0
higher affinity to a UV-irradiated syn thetic oligonucleotide as compared to the same unirradiated oligonucleotide. Most likely, this protein is the UV-damaged DNA-binding protein, described previ ously by other authors,13 which is miss ing in some cell extracts derived from patients with the DNA repair disorder Xeroderma pigmentosum group E. When the cells were incubated with nickel(II) for 24 h before the preparation of the cell extracts, the binding capacity of the protein decreased dose-dependent at concentrations as low as 50 n-M nickel(II), indicating that nickel(II) disrupts DNA-protein interactions involved in damage recognition (figure 1). This dis-
2 + 0
3 + 50
4 5 + + 250 600
F i g u r e 1. Effect of nickel(II) on the binding of a UV-damage specific protein, detected in a gel mobility shift assay. HeLa cells were incubated for 24 h with the respective concentrations of NiCl2 before the preparation of nuclear protein extracts. The binding reaction was conducted as described in Materials and Methods, applying either an unirradiated digoxigenin-labeled oligonucleotide (Lane 1) or an UV-irradiated digoxigenin-labeled oligonucleotide (Lanes 2-5).
INTERACTION OF CARCINOGENIC METAL COMPOUNDS WITH DNA REPAIR PROCESSES Intact DNA
Ligation
Exposure to DNA damaging agents
As(III), Ni(II), Cd(II)
H
35
TTTT
tttttY
Damage recognition/ incision
Repair polymerisation
Co(II) I
I I
I L -J
I
Ni(II), Cd(II),Co(II), As(III)
I I
I I I I I I I I
im
F ig u r e 2. Inhibition of nucleotide excision repair by compounds of nickel(II), cadmium(II), cobalt(II)
and arsenic(III): proposed sites of action.
turbance of DNA-protein interactions by nickel(II) was reversible by the addition of 10 mM magnesium(II) to the gel shift buffer (data not shown), indicating that the competition between these two metal ions might play a decisive role in repair inhibition. Furthermore, an enhance ment of cytotoxicity by nickel(II) was also observed in combination with the cytostatic drug dis-diamminedichloroplatinum(II).14 as well as a repair inhibi tion of the respective DNA adducts induced by this agent,* suggesting that the nucleotide excision repair system is affected in general by nickel(II).
briefly summarized and updated (figure 2). Regarding cadmium(II), an accumula tion of DNA strand breaks after UV-irradiation has been reported by Nocentini et al 15 indicating an inhibition of the polymerization/ligation step of the repair process. However, when investigating the total incision frequency during the repair of UV-induced DNA damage determined in the presence of the repair inhibitors aphidicolin and hydroxyurea, a dose-dependent inhibition of the inci sion frequency was observed by us at concentrations as low as 1 jxM cadmium(II), which is well below cytotoxic concentrations (table I). This indicates that the process of DNA damage recogni I n t e r a c t io n s w it h N u c l e o t id e tion/incision is disturbed by cadmium(II) E x c is io n R e p a ir b y C a d m iu m (II), as well. Similar effects were observed C o b a l t (II) a n d A r s e n ic (III) with cobalt(II). While an inhibition of the Besides nickel(II), other carcinogenic polymerisation step has been suggested metal compounds have been shown to previously ,16 more detailed investiga inhibit the repair of UV-induced DNA tions revealed also a reduction of the damage as w ell. The results have incision frequency after UV-irradiation.17 recently been reviewed 1 and are only Finally, arsenic has been shown repeat edly to impair the ligation capacity in * Unpublished observation by I. Krüger and A. nucleotide excision repair, possibly by the inactivation of essential SH-groups.18 Hartwig.
HARTWIG, SCHLEPEGRELL, DALLY, AND HARTMANN
36
G to T transversions.20 The frequency of enzyme-sensitive sites was quantified Effect of Cadmium (II) on the Incision Frequency by the determination of DNA strand After Ultraviolet Irradiation breaks measured by the alkaline unwind ing method. Regarding the ability of Treatment Incisions/Cell % of Control nickel(II) to induce oxidative DNA dam age, the spontaneously occurring fre 100.0 UV (5 J/m2) 17652 ± 1247 quency of Fpg-sensitive sites is only sig U V + 1)xM Cd(ll) 14958 ±1234 84.7 ± 7.0 nificantly enhanced at the cytotoxic U V + 5|iMCd(ll) 9874 ±1260 55.9 ±7.1 concentration of 0.75 mM (data not UV + 10jiM Cd(ll) 10007 ±1384 56.7 ± 7.8 shown). To investigate the effect of nickel(II) on the repair of oxidative DNA HeLa cells were pretreated with hydroxyurea damage, HeLa cells were irradiated with (10mM) and aphidicolin (15 | .M) for 1 hour, visible light, which induces predomi irradiated with ultraviolet, and postincubated in the nantly 8 -hydroxyguanine and to a lesser presence of the inhibitors for 30 minutes. When extent DNA strand breaks.4 Within 4 h treated with cadmium (II), CdChwas added 1 hour before the addition of aphidicolin and hydroxyurea after irradiation, the Fpg-sensitive sites and was present during all subsequent were repaired to about 63 percent, and incubations. The DNA strand breaks were the DNA strand breaks were ligated com pletely. In the presence of nickel(II), determined by alkaline unwinding. however, the repair of both types of was impaired; it caused a repair Our observations demonstrate addition lesions inhibition of DNA strand breaks and ally an inhibition of the incision step at Fpg-sensitive sites at concentrations as low concentrations.t low as 50 |xM (data not shown). At 250 (jlM, the damage removal was blocked I n t e r a c t io n w it h t h e completely (figure 3). R e p a ir o f O x id a t iv e D N A TABLE I
0
D am age
by
N ic k e l (II)
As an indicator of oxidative DNA dam age, the frequency of Fpg-sensitive sites was determined by a very sensitive method established in our labora tory originally for the detection of UV-induced cyclobutane pyrimidine dimers.6 The Fpg protein isolated from Escherichia coli specifically removes 8-hydroxyguanine and, to a lesser extent, imidazol ring-opened forms of guanine and adenine, and the resulting abasic sites are converted into DNA single strand breaks by the associated AP endo nuclease activity.19 Among these DNA adducts, 8-hydroxyguanine is the most relevant, since it is mutagenic by causing t U npublished observations by A. Hartwig, L.H.F. Mullenders, et al.
Conclusions The data presented in this study pro vide further evidence that the inhibition of DNA repair processes represents a common mechanism in the genotoxicity of nickel(II), cadmium(II), cobalt(II), and arsenic(III). The reasons leading to repair inhibition, however, might be quite different for the respective metal compounds, depending on their ability to compete with essential metal ions like Mg2+, as demonstrated for nickel(II), but also with Zn2+ as an essential cofactor in some repair enzymes, or with Ca2+, which might be involved in the regula tion of DNA repair processes. Since the DNA is continuously damaged by exoge nous and endogenous sources, an impaired repair capacity might not only be relevant for enhancing effects in com
INTERACTION OF CARCINOGENIC METAL COMPOUNDS WITH DNA REPAIR PROCESSES
15000
ID C“cO
01 0®
Interaction of Carcinogenic Metal Compounds with Deoxyribonucleic Acid Repair Processes ANDREA HARTWIG, Ph.D., REGINA SCHLEPEGRELL, HEIKE DALLY, and MAIKE HARTMANN Department of Biology and Chemistry, University of Bremen, 28334 Bremen, Germany
ABSTRACT The potentials of nickel(II) and cadmium(II) to interfere with the repair of different types of deoxyribonucleic acid (DNA) lesions was investigated. Concerning the nucleotide excision repair pathway, nickel(II) has been shown to reduce the incision and the ligation frequency after ultraviolet (UV)-irradiation. When applying a gel mobility shift assay and HeLa nuclear cell free extracts, nickel(II) diminishes the specific binding of a protein to UV-damaged DNA, suggesting that nickel(II) interferes with the DNA-protein interactions involved in the damage recognition after UV-irradiation. Similarly, the incision frequency is reduced in the pres ence of low concentrations of cadmium(II). Concerning the repair of oxi dative DNA damage induced by visible light, non-cytotoxic concentrations of nickel(II) caused a complete repair inhibition of DNA base modifica tions like 7,8-dihydro-8-oxoguanine (8-hydroxyguanine) and of DNA strand breaks. Since the repair of DNA damage is essential for the prevention of cancer, its inhibition may account for the carcinogenic action of the respec tive metal compounds. Introduction Compounds of nickel, cadmium, cobalt and arsenic are well established carcino gens to humans and to experimental ani mals. However, the mechanisms leading to tumor formation are still not under stood, since the induction of DNA damSend reprint requests to: Dr. Andrea Hartwig, University of Bremen, Department of Biology and Chemistry, Postfach 330440, 28334 Bremen, Ger many.
age and the mutagenic potentials are rather weak and mainly restricted to cytotoxic concentrations of the metals. In contrast, cytotoxicity and genotoxicity enhancing effects in combination with other DNA damaging agents are more pronounced and observed at lower concentrations.1 Based on these findings, the potential of carcinogenic metal compounds to interfere with the repair of different types of DNA lesions was investigated. Possible interactions were examined
31 0091-7370/96/0100-0031 $01.20 © Institute for Clinical Science, Inc.
32
HARTWIG, SCHLEPEGRELL, DALLY, AND HARTMANN
with (1) the nucleotide excision repair system, which represents the major pathway eliminating a broad spectrum of DNA lesions induced by many environ mental mutagens from the genome, and (2) the repair of oxidative DNA lesions, since they are induced continuously owing to cellular oxygen metabolism; if not repaired, they are implicated in muta tions, cancer and aging.2 To investigate the effects of metal com pounds on nucleotide excision repair, UVC light (254 nm) was used as a well characterized DNA damaging agent, since both major DNA photoproducts, the cyclobutane pyrimidine dimer and the pyrimidine-(6-4)-pyrimidone-photoproduct ((6-4)-photoproduct), are removed by this repair pathway.3 In pre vious studies, nickel(II), as w ell as cobalt(II), arsenic(III) and cadmium(II), have been shown to dim inish the removal of UV-induced cyclobutane pyrimidine dimers.1 The present study was undertaken to elucidate the effects on distinct steps of the repair process and the potentially underlying mechanisms. Concerning the repair of oxidative DNA damage and its potential inhibition, visible light was applied, which induces predominantly 8-hydroxyguanine and, to a lesser extent, DNA strand breaks.4 The induction and removal of the DNA base modifications were monitored by their sensitivity towards the bacterial formamidopyrimidine-DNA glycosylase (Fpg protein). Special emphasis will be given to the results obtained with nickel(II), and the effects observed in comparative studies w ith cadm ium (II), cobalt(II) and arsenic(III) will be summarized. Materials and Methods Cell Culture
containing 5 percent fetal bovine serum. The cultures were incubated at 37°C with 5 percent C 0 2 in air and 100 per cent humidification. UV-lRRADIATION
Ultraviolet irradiation of cells was per formed with a General Electric germi cidal lamp.* I r r a d ia t io n
w it h
V is ib l e L ig h t
Irradiation was carried out essentially as described by Pflaum et al.4 The cells were illuminated as monolayers in cell culture dishes covered with phosphate buffered saline (140 mM NaCl, 3 mM KC1, 8 mM Na2H P0 4 1 mM KH2P 04, 1 mM CaCl2, 0.5 mM MgCl2, and 0.1 per cent glucose) on ice with a 1000 W halo gen lamp for 20 min. in a distance of 33 cm corresponding to 450 kj/m2. D e t e c t io n o f D N A S t r a n d b y A l k a l in e U n w in d in g
Breaks
The strand breaks were detected as described previously. Briefly, 2 x 10® HeLa cells were allowed to attach for at least 24 h and treated as described for the respective experiments. Afterwards, the medium was removed and an alkaline solution was added containing 0.03 M NaOH, 0.02 M Na 2H P 0 4 and 0.9 M NaCl. Separation of single- and double stranded DNA was perform ed on hydroxyapatite columns. The DNA con tent of both fractions was determined fluorimetrically by adding the dye bisbenzimidetrihydrochloride.t The num ber of DNA strand breaks was calculated from calibration with X-rays.
HeLa cells were grown as monolayers* Bioblock scientific, 254 nm. in minimal essential medium (MEM) t Hoechst 33258
INTERACTION OF CARCINOGENIC METAL COMPOUNDS WITH DNA REPAIR PROCESSES
Q u a n t if ic a t io n o f F p g -S e n s it iv e S it e s
The procedure applied to detect enzym e sen sitiv e sites has b een described in detail previously.6 Briefly, cells were lysed as monolayers (0.006 M Na2H P04, 0.001 M KH2P 04) 0.137 M NaCl, 0.003 M KC1, 0.1% octyl phenoxy polyethoxy ethanol (Triton X 100), and treated with a high salt solution (2 M NaCl, 0.01 M ethylene diaminotetraacetic acid [EDTA] and 0.002 M Tris[hydroxymethyl]aminomethane [Tris], pH 8.0). The nucleoids were then incu bated with the Fpg protein (1 (xg/ml) in buffer consisting of 0.05 M sodium phos phate, pH 7.5, 0.01 M EDTA, and 100 mM NaCl. At the end of incubation, an alkaline solution was added yielding a final concentration of 0.07 N NaOH, 0.013 M EDTA, and 0.37 M NaCl, pH 12.3, and the DNA was allowed to unwind for 30 min in the dark. The fur ther steps of neutralization, separation of single- and double-stranded DNA and quantification of lesion frequency were performed as described above. G e l M o b i l i t y S h i f t A ssa y
Nuclear protein extracts were prepared from HeLa cells essentially as described by Schreiber et al.7 To detect damagespecific DNA-protein interactions, a digoxygenin-labeled synthetic oligonu cleotide (48 bp) was applied which was either unirradiated or irradiated with 18 kj/m 2 UVC. For the binding reaction, 2 |xg protein extract were pretreated with 300 fmol of the unlabeled oligonucleo tide in a gel shift buffer (13.3 mM Hepes, 9.6% glycerin, 100 mM KC1, 5 mM MgCl2, 1 mM EDTA, 1.2 jxg bovine serum albumin, 0.4 mM dithiothreitol, 75 (xM phenylmethylsulfonyl fluoride, and 4 ng leupeptin, pH 7.9) for 10 min at room temperature. In addition, 0.5 |xg poly[d(I-C)] as well as 60 fmol of the digoxy-
33
genin-labeled oligonucleotide were added for another 35 min. Afterwards, the binding mixture was loaded on a poly acrylamide gel (6 %; 45 mM Tris, 45 mM boric acid, 1 mM EDTA), and electropho resis was conducted at 110 V for 2.5 h. Southern blot was done in a semi-dry electro-blotting apparature applying a positively charged nylon membrane. The detection of the digoxygenin-labeled oli gonucleotide was performed calorimetrically by the alkaline phosphatase conju gated to an Anti-digoxygenin antibody using nitroblue tetrazolium salt (NBT) and 5-bromo-4-chloro-3-indoyl phos phate (X-phosphate) as substrates. Results and Discussion I n t e r a c t io n s w it h N u c l e o t id e E x c is io n R e p a ir b y N ic k e l (II)
Nickel(II) has been shown to diminish the removal of UV-induced cyclobutane pyrimidine dimers8’9; however, the mechanism leading to repair inhibition remains to be elucidated. Nucleotide excision repair can be roughly divided into four different steps, namely: (1) dam age recognition, (2) incision on both sites of the lesion with the subsequent dis placement of the damaged oligonucleo tide, (3) polymerization, and (4) ligation of the repair patches.10 During the repair process, DNA strand breaks are gener ated owing to incisions, which are ligated again at the end of the repair event. In a previous study by us ,9 the alkaline unwinding technique was applied to fol low the transient appearance of DNA strand breaks after UV-irradiation. In the presence of nickel(II), less DNA strand breaks were detected, suggesting an interference with the damage recogni tion/incision step in excision repair. However, once DNA strand breaks are generated, they remain open for a pro longed period of time in the presence of nickel(II), indicating that the ligation
34
HARTWIG, SCHLEPEGRELL, DALLY, AND HARTMANN
step is inhibited as well.9,11 Since the ini tiation of repair events is very complex in eucaryotic cells, involving the coordi nated action of at least 15 to 18 polypep tides,12 the interference with this process by nickel(II) was investigated in more detail. One important prerequisite for the initiation of repair events is the recogni tion of the respective DNA lesions. To investigate w hether or not nickel(II) interferes with the damage recogni tion step, its effect was studied on the binding of a protein exerting high affinity for UV-damaged DNA. A HeLa nuclear cell free extract was prepared and, by applying a gel mobility shift assay, a pro tein was identified which binds with
Lane UV Ni(II) [jiM]
1
0
higher affinity to a UV-irradiated syn thetic oligonucleotide as compared to the same unirradiated oligonucleotide. Most likely, this protein is the UV-damaged DNA-binding protein, described previ ously by other authors,13 which is miss ing in some cell extracts derived from patients with the DNA repair disorder Xeroderma pigmentosum group E. When the cells were incubated with nickel(II) for 24 h before the preparation of the cell extracts, the binding capacity of the protein decreased dose-dependent at concentrations as low as 50 n-M nickel(II), indicating that nickel(II) disrupts DNA-protein interactions involved in damage recognition (figure 1). This dis-
2 + 0
3 + 50
4 5 + + 250 600
F i g u r e 1. Effect of nickel(II) on the binding of a UV-damage specific protein, detected in a gel mobility shift assay. HeLa cells were incubated for 24 h with the respective concentrations of NiCl2 before the preparation of nuclear protein extracts. The binding reaction was conducted as described in Materials and Methods, applying either an unirradiated digoxigenin-labeled oligonucleotide (Lane 1) or an UV-irradiated digoxigenin-labeled oligonucleotide (Lanes 2-5).
INTERACTION OF CARCINOGENIC METAL COMPOUNDS WITH DNA REPAIR PROCESSES Intact DNA
Ligation
Exposure to DNA damaging agents
As(III), Ni(II), Cd(II)
H
35
TTTT
tttttY
Damage recognition/ incision
Repair polymerisation
Co(II) I
I I
I L -J
I
Ni(II), Cd(II),Co(II), As(III)
I I
I I I I I I I I
im
F ig u r e 2. Inhibition of nucleotide excision repair by compounds of nickel(II), cadmium(II), cobalt(II)
and arsenic(III): proposed sites of action.
turbance of DNA-protein interactions by nickel(II) was reversible by the addition of 10 mM magnesium(II) to the gel shift buffer (data not shown), indicating that the competition between these two metal ions might play a decisive role in repair inhibition. Furthermore, an enhance ment of cytotoxicity by nickel(II) was also observed in combination with the cytostatic drug dis-diamminedichloroplatinum(II).14 as well as a repair inhibi tion of the respective DNA adducts induced by this agent,* suggesting that the nucleotide excision repair system is affected in general by nickel(II).
briefly summarized and updated (figure 2). Regarding cadmium(II), an accumula tion of DNA strand breaks after UV-irradiation has been reported by Nocentini et al 15 indicating an inhibition of the polymerization/ligation step of the repair process. However, when investigating the total incision frequency during the repair of UV-induced DNA damage determined in the presence of the repair inhibitors aphidicolin and hydroxyurea, a dose-dependent inhibition of the inci sion frequency was observed by us at concentrations as low as 1 jxM cadmium(II), which is well below cytotoxic concentrations (table I). This indicates that the process of DNA damage recogni I n t e r a c t io n s w it h N u c l e o t id e tion/incision is disturbed by cadmium(II) E x c is io n R e p a ir b y C a d m iu m (II), as well. Similar effects were observed C o b a l t (II) a n d A r s e n ic (III) with cobalt(II). While an inhibition of the Besides nickel(II), other carcinogenic polymerisation step has been suggested metal compounds have been shown to previously ,16 more detailed investiga inhibit the repair of UV-induced DNA tions revealed also a reduction of the damage as w ell. The results have incision frequency after UV-irradiation.17 recently been reviewed 1 and are only Finally, arsenic has been shown repeat edly to impair the ligation capacity in * Unpublished observation by I. Krüger and A. nucleotide excision repair, possibly by the inactivation of essential SH-groups.18 Hartwig.
HARTWIG, SCHLEPEGRELL, DALLY, AND HARTMANN
36
G to T transversions.20 The frequency of enzyme-sensitive sites was quantified Effect of Cadmium (II) on the Incision Frequency by the determination of DNA strand After Ultraviolet Irradiation breaks measured by the alkaline unwind ing method. Regarding the ability of Treatment Incisions/Cell % of Control nickel(II) to induce oxidative DNA dam age, the spontaneously occurring fre 100.0 UV (5 J/m2) 17652 ± 1247 quency of Fpg-sensitive sites is only sig U V + 1)xM Cd(ll) 14958 ±1234 84.7 ± 7.0 nificantly enhanced at the cytotoxic U V + 5|iMCd(ll) 9874 ±1260 55.9 ±7.1 concentration of 0.75 mM (data not UV + 10jiM Cd(ll) 10007 ±1384 56.7 ± 7.8 shown). To investigate the effect of nickel(II) on the repair of oxidative DNA HeLa cells were pretreated with hydroxyurea damage, HeLa cells were irradiated with (10mM) and aphidicolin (15 | .M) for 1 hour, visible light, which induces predomi irradiated with ultraviolet, and postincubated in the nantly 8 -hydroxyguanine and to a lesser presence of the inhibitors for 30 minutes. When extent DNA strand breaks.4 Within 4 h treated with cadmium (II), CdChwas added 1 hour before the addition of aphidicolin and hydroxyurea after irradiation, the Fpg-sensitive sites and was present during all subsequent were repaired to about 63 percent, and incubations. The DNA strand breaks were the DNA strand breaks were ligated com pletely. In the presence of nickel(II), determined by alkaline unwinding. however, the repair of both types of was impaired; it caused a repair Our observations demonstrate addition lesions inhibition of DNA strand breaks and ally an inhibition of the incision step at Fpg-sensitive sites at concentrations as low concentrations.t low as 50 |xM (data not shown). At 250 (jlM, the damage removal was blocked I n t e r a c t io n w it h t h e completely (figure 3). R e p a ir o f O x id a t iv e D N A TABLE I
0
D am age
by
N ic k e l (II)
As an indicator of oxidative DNA dam age, the frequency of Fpg-sensitive sites was determined by a very sensitive method established in our labora tory originally for the detection of UV-induced cyclobutane pyrimidine dimers.6 The Fpg protein isolated from Escherichia coli specifically removes 8-hydroxyguanine and, to a lesser extent, imidazol ring-opened forms of guanine and adenine, and the resulting abasic sites are converted into DNA single strand breaks by the associated AP endo nuclease activity.19 Among these DNA adducts, 8-hydroxyguanine is the most relevant, since it is mutagenic by causing t U npublished observations by A. Hartwig, L.H.F. Mullenders, et al.
Conclusions The data presented in this study pro vide further evidence that the inhibition of DNA repair processes represents a common mechanism in the genotoxicity of nickel(II), cadmium(II), cobalt(II), and arsenic(III). The reasons leading to repair inhibition, however, might be quite different for the respective metal compounds, depending on their ability to compete with essential metal ions like Mg2+, as demonstrated for nickel(II), but also with Zn2+ as an essential cofactor in some repair enzymes, or with Ca2+, which might be involved in the regula tion of DNA repair processes. Since the DNA is continuously damaged by exoge nous and endogenous sources, an impaired repair capacity might not only be relevant for enhancing effects in com
INTERACTION OF CARCINOGENIC METAL COMPOUNDS WITH DNA REPAIR PROCESSES
15000
ID C“cO
01 0®
