|Home | About | Journals | Submit | Contact Us | Français|
Oxidative stress-related damage to the DNA macromolecule produces lesions that are implicated in various diseases. To understand damage to DNA, it is important to study the free radical reactions causing the damage. Measurement of DNA damage has been a matter of debate as most of the available methods measure the end product of a sequence of events and provide limited information on the initial free radical formation. We report a measurement of free radical damage in DNA induced by a Cu(II)-H2O2 oxidizing system using immuno-spin trapping supplemented with electron paramagnetic resonance. In this investigation, the short-lived radical generated is trapped by the spin trap 5,5-dimethyl-1-pyrroline N-oxide (DMPO) immediately upon formation. The DMPO adduct formed is initially electron paramagnetic resonance active, but is subsequently oxidized to the stable nitrone adduct, which can be detected and visualized by immuno-spin trapping and has the potential to be further characterized by other analytical techniques. The radical was found to be located on the 2′-deoxyadenosine (dAdo) moiety of DNA. The nitrone adduct was repaired on a time scale consistent with DNA repair. In vivo experiments for the purpose of detecting DMPO–DNA nitrone adducts should be conducted over a range of time in order to avoid missing adducts due to the repair processes.
DNA-centered radicals are implicated as a root cause of carcinogenesis and aging through their initiation of genomic damage. Organisms must maintain the integrity of their DNA in order to remain healthy and propagate. DNA is continuously exposed to exogenous and endogenous mutagens including reactive oxygen and nitrogen species that can alter its integrity (1–3). Indeed, oxidatively generated DNA damage is an inevitable consequence of cellular metabolism. Injury to this macromolecule can have severe biological consequences including mutation, carcinogenic transformation, cell death and reproductive death (4). Considerable attention has been focused on the formation, stability and detection of various DNA damage modifications, which have been widely studied in terms of formation, mutagenesis and repair. Oxidatively generated damage of cellular DNA by free radicals may be a significant factor in human carcinogenesis (5). Extensive mechanistic studies have provided a large amount of information that allows for the identification of the chemical pathways leading to the existence of oxidative stress. The detection and characterization of radical transients are important in understanding the mechanisms involved.
The hydroxyl radical is the most reactive of the several reactive oxygen species and, almost uniquely among free radicals, reacts rapidly with DNA (3). When DNA damage occurs to the extent that it can no longer be repaired, senescence, cell death and carcinogenesis may occur.
Several immunological assays have also been designed for the measurement of oxidatively generated DNA damage (6,7). However, the immunoassays developed to detect DNA damage have been plagued by a lack of specificity. Several attempts have been made to use antibodies raised against 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodGuo, one of the most popular markers of oxidative DNA damage), but it has been difficult to obtain a sufficiently specific antibody. 8-oxodGuo differs from dGuo by a single oxygen atom, which is a challenge for the specificity of the antibody. The applications of this method have been limited by cross reactivity of the antibodies with normal DNA bases and other abundant biological constituents (8). In addition, antibodies to a specific nucleotide oxidation product will not recognize the numerous other oxidation products.
While it is critical to detect and identify the primary DNA radicals in order to understand DNA damage, detection of DNA radicals is difficult due to their short lifetimes. The spin-trapping technique in electron spin resonance (ESR) relies on the addition of the free radical across the double bond of a spin trap to form a more stable adduct than was originally detected by ESR. Although more stable than the original free radical, the radical adduct still has a comparatively short lifetime and can be oxidized to a stable end product, a nitrone, which preserves the chemical bond at the site of high spin density formed by spin trapping. The base, or sugar, where the spin trap is attached can then be determined by MS or by other analytical techniques such as NMR. A method developed in our laboratory, called immuno-spin trapping (IST), detects the stable nitrone after oxidation to combine the specificity of spin trapping with the sensitivity of an antigen–antibody-based assay (9–11).
IST has been very successful in detecting radical adducts in proteins (11–17), DNA (18–21), cells (11,22,23) and tissue (24,25). Its major limitation is that the chemical structure of the free radical is not identified by the IST procedure. Since detection of 5,5-dimethyl-1-pyrroline N-oxide (DMPO)–DNA nitrone adducts is done by dot blots or enzyme-linked immunosorbent assay (ELISA), the identification of a radical as DNA-derived and not a contaminant (protein-derived) depends on the absolute purity of the DNA as achieved by traditional DNA purification methods. The structural identification of DNA-derived radicals detected by ELISA in previous works (18–21,24) is further compromised by the limited investigations of DNA radical adducts by ESR or the corresponding DNA nitrone adducts by mass spectrometry (MS). We have recently been able to combine IST with ESR, MS and MS/MS to detect and identify a DNA radical adduct (26).
In this work we used the sensitive IST technique to detect, visualize and identify primary DNA radical damage free of artifacts because additional nitrone adduct formation is impossible once the DMPO is diluted by DNA extraction, and subsequent nitrone adduct formation is precluded. Dilution of DMPO to 1mM is sufficient to prevent artifactual nitrone formation because, at DMPO concentrations below this, DNA radicals will decay before they can be trapped in detectable amounts. Here, we have chosen a Cu(II)-H2O2 oxidizing system to induce hydroxyl-free radical damage to DNA. Copper-induced oxidative damage by H2O2 is attributed to the formation of hydroxyl radical by a mechanism analogous to the iron-catalyzed Haber–Weiss cycle, and a Cu(I) intermediate is formed (27). Copper-mediated damage is significant because it can be toxic as in Wilson’s disease (28), which is a copper overload disorder. Furthermore, there is increasing evidence of elevated levels of copper in tumor growth and angiogenesis (29–31). The hydroxyl radical generated reacts at diffusion-controlled rates with virtually any macromolecule, including DNA. Cu(II) bound to DNA undergoes cyclic reduction. The reducing agent in our system is hydrogen peroxide or superoxide radical anion formed by the Haber–Weiss cycle, which reduces the metallic complex yielding Cu(I)-bound DNA. Subsequently, this reduced state can react with H2O2 in the Fenton reaction, yielding the hydroxyl radical which, in turn, causes site-specific damage to nuclear DNA. We examined this damage using ESR and IST and detected and identified a nitrone adduct on the 2′-deoxyadenosine moiety in cellular DNA (26). With confocal microscopy, we were also able to visualize these DNA radical adducts at the nucleus and see that the free radical formation was significantly greater in the nuclear fraction than in the cytosolic fraction. Finally, we examined the effect of DNA damage repair on the DMPO–DNA nitrone adduct and investigated the inhibition of this repair mechanism by using known inhibitors of excision repair pathways.
2′-Deoxycytidine (dCyd), 2′-deoxyguanosine (dGuo), thymidine (Thd) and 2′-deoxyadenosine (dAdo) were purchased from MP Biomedicals (Irvine, CA, USA). Calf thymus DNA was obtained from Sigma Aldrich (St Louis, MO, USA). Methoxyamine (TRC 102) was obtained from Supelco (Bellefonte, PA, USA) and nickel chloride was from Alfa Aesar (Ward Hill, MA, USA). The spin trap DMPO was purchased from Dojindo (Rockville, MD, USA). The DMPO concentration was measured at 228nm, assuming a molar absorption co-efficient of 7800M−1cm−1. Hydrogen peroxide was obtained from Fisher Scientific Company (Fairlawn, NJ, USA). The hydrogen peroxide concentration was verified using UV absorption at 240nm (ε240nm=43.6M−1cm−1). All buffers used were treated with Chelex 100 ion exchange resin (Bio-Rad Laboratories, Hercules, CA, USA) to avoid transition metal-catalyzed reactions.
Typically, reaction mixtures contained 2.5–5mM nucleoside, 300μM CuCl2, 100μM H2O2 and 100mM DMPO in 100mM Chelex-treated phosphate buffer (pH 7.4) and were incubated for 1h at 37°C. For 2′-deoxyguanosine experiments, a saturated solution of the nucleoside was used.
The mouse macrophage cell line RAW 264.7 was grown and maintained in Dulbecco’s modified eagle medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum with penicillin and streptomycin (10000 units penicillin and 10mg streptomycin per ml in 0.9% NaCl). Upon reaching 80% confluence, the cells were counted and seeded at the needed density. Exponentially growing RAW cells (1–3million/well) were seeded into 12-well plates and treated with 100µM CuCl2/100µM H2O2 in the presence of DMPO for 12–15h at 37°C in a humidified 5% CO2 environment. After this treatment, cell viability was determined using the tryphan blue exclusion assay. The cells were pelleted and frozen at −80°C until DNA extraction and analysis of the DMPO nitrone adducts.
Exponentially growing RAW cells (1–3million/well) were seeded into 12-well plates and treated with 100µM CuCl2/100µM H2O2 in the presence of DMPO. After treatment, the number of viable cells was determined by tryphan blue dye exclusion and counted manually with a hemocytometer.
Nuclear/cytosolic fractionation was carried out with treated and untreated cells using a nuclear/cytosolic fractionation kit from BioVision (Mountain View, CA, USA).
Exponentially growing RAW cells (1–3million/well) were seeded into 12-well plates and treated with 100µM CuCl2 and 100µM H2O2 in the presence of DMPO. At designated time points, treated cells were harvested and stored at −80°C until DNA extraction and analysis of the DMPO nitrone adducts.
Exponentially growing RAW cells at 80–90% confluence were harvested in ice-cold lysis buffer to release soluble cellular proteins. The cells were scraped from the plates, transferred to centrifuge tubes and placed on ice to ensure efficient lysis. Lysates were centrifuged at 10000rpm for 5min. The supernatant was pooled and added to a prepared DMPO–DNA nitrone adduct reaction mixture.
Two known inhibitors, methoxyamine (TRC 102) of the base excision repair (BER) pathway and NiCl2 of the nucleotide excision repair (NER) pathway, were each added to exponentially growing cells and incubated for 20–24h. The cells were then harvested and lysed as was done for the repair study above.
Cell pellets stored at −80°C were re-suspended in digestion buffer with proteinase K, and DNA extraction was carried out as described previously (18,19). The extraction procedure preserves the nitrone adduct covalently bound to the DNA (18,19). DNA concentration and purity were measured from the absorbance at 260 and 280nm. Pure DNA will exhibit an A260/A280 ratio between 1.8 and 2.
A standard ELISA procedure in 96-well plates (Greiner Labortechnik, Frickenhausen, Germany) was used as described in detail (18). Briefly, DNA solutions were diluted to 5μgml−1 in 1× PBS. A volume of 25μl of the DNA solution and 25μl of Reacti-Bind DNA coating solution were placed in each well of the plate and the plate incubated for 2–4h at 37°C. The plates were washed once with washing buffer (1× PBS containing 0.05% non-fat dry milk and 0.1% Tween-20) and blocked with blocking buffer (1× PBS containing 3% non-fat dry milk) for 2h at 37°C or overnight at 4°C. After washing once with the washing buffer, the rabbit anti-DMPO serum (diluted 1:10000 in the wash buffer) was added and incubated at 37°C for 60min. After three more washes, 100μl of the secondary antibody (HRP-conjugated), diluted 1:10000 in the wash buffer, was added and incubated for 60min. After three washes, the antigen–antibody complexes were developed using the Immobilon chemiluminescence substrate (Millipore Corp., Billerica, MA, USA) and the light emitted was recorded as arbitrary light units using X-Fluor software (Tecan US, Research Triangle Park, NC, USA).
Upon mixing, the reaction mixture was immediately transferred to a flat cell and the ESR scan initiated within 1min after starting the reaction. ESR spectra were obtained with an ELEXSYS E500 ESR spectrometer (Bruker Biospin, Billerica, MA, USA) equipped with an ER4122SHQ cavity operating at 9.76GHz and at room temperature. The ESR spectrometer settings were as follows: scan range, 100G; modulation frequency, 100kHz; modulation amplitude, 1.0G; microwave power, 20mW; receiver gain, 2×104; time constant, 327ms; and conversion time, 327ms.
For confocal microscopy studies, treated RAW cells were allowed to adhere to polylysine-coated plates (MatTek) for 30min on ice. The cells were fixed with 0.4% paraformaldehyde and permeabilized with 0.01% Surfact-Amps X-100 (Thermo Scientific). After blocking with 0.1% BSA in PBS, DMPO nitrone adducts were stained using anti-DMPO (raised in rabbit) antisera. After 1h incubation, the cells were washed and the secondary Alexaflour anti-rabbit antibody conjugated with fluorescein (Molecular Probes Invitrogen) was added.
ESR studies were carried out with each of the 2′-deoxyribonucleosides (dGuo, dAdo, Thd and dCyt). When 5mM (or saturated, for dGuo) solutions of the nucleosides reacted with CuCl2 (300μM), H2O2 (100μM) and DMPO (100mM), a radical adduct was detected only with 2′-deoxyadenosine (26).
When calf-thymus DNA reacted with CuCl2 (300μM), H2O2 (100μM) and 100mM DMPO, no assignable electron paramagnetic resonance (EPR) spectra were detected (26). Our attempts to digest the DNA to obtain motionally narrowed, sharper isotropic spectra were not successful. Therefore, in order to try to mimic DNA, we treated an equimolar mixture of adenosine, thymidine, cytidine and guanosine with CuCl2 (300μM), H2O2 (100μM) and 100mM DMPO. A spectrum was obtained of a dAdo-derived, nitrogen-centered DMPO radical adduct (Figure 1A) nearly identical to that obtained from the reaction of only dAdo under similar circumstances (26). The presence of dAdo and all other reaction components (CuCl2, H2O2 and DMPO) was necessary for radical adduct formation. Figure 1 shows the comparison of the adduct obtained from the four-nucleoside mixture with the simulated spectra obtained from the dAdo-only reaction mixture (Figure 1B).
RAW 264.7 macrophage cells were treated with CuCl2 and H2O2 and DNA extracted as outlined earlier (18,19). Viability assays showed that the cells were ~90% viable even at 22h post-treatment (Figure 2A). It is known that DNA radicals can be detected in DNA-extracted cells using an anti-DMPO antibody in ELISA experiments (18,22). Here, production of the DNA nitrone adducts required the presence of all components of the reaction mixture (Figure 2B). In order to determine the localization of the DMPO nitrone adducts, we separated the cytosolic and the nuclear fractions using a nuclear/cytosolic fractionation kit from BioVision and analyzed them by ELISA (Figure 2C). Clearly, the nuclear fraction had a higher level of DMPO nitrone adducts. In order to determine whether the observed chemiluminescence was due to DNA adducts and not protein (specifically histone in the nuclear fraction), DNA extraction was carried out from both fractions and concentration determined. The resulting ELISA showed a substantial increase in nitrone adduct in the nuclear fraction as compared to the cytosolic fraction (Figure 2D). No protein–DMPO adducts were detected in Western blots of either the cytosolic or the nuclear fractions (data not shown).
Parallel confocal fluorescence microscopy experiments were also performed to visually determine the localization of the nitrone adducts. After fixation and permeabilization, cells were stained with rabbit anti-DMPO serum followed by Alexa 567 conjugated anti-rabbit antisera to stain the DMPO nitrone adducts. The data showed that the cells exposed to CuCl2/H2O2 had a fluorescent signal localized primarily in the nucleus (Figure 3A). In the absence of CuCl2, H2O2 and/or DMPO (Figure 3B), no such fluorescence was detected. These data indicate that the radical formation is largely localized to the nucleus and is generated by the reaction of copper with H2O2.
The formation and repair of DMPO–DNA nitrone adducts were assessed over a 24-h time course. Figure 4A shows a time profile of the formation of the adduct. The DMPO–DNA nitrone adduct formed within 30min, then increased slowly over the next several hours before reaching an adduct maximum at ~12h. Thereafter, apparent repair of the DMPO–DNA nitrone adduct occurred, and the adduct formation then dropped to a value of ~50% of the maximum value. As there exists the possibility of cell replication, causing dilution of the DNA–DMPO due to DNA synthesis during this time frame, we further tested the direct repair of the DMPO–DNA nitrone adduct formation. Preformed DNA–DMPO adduct from calf thymus treated with 300µM CuCl2 and100µM H2O2 in the presence of 100mM DMPO was prepared, and cell lysate from the RAW 267.4 cells added. The DMPO–DNA nitrone adduct was analyzed over a 24-h time frame. The DNA–DMPO adduct formed was clearly repaired over a relatively short period of time (Figure 4B).
Cell lysates containing inhibitors of the BER and NER pathways were added to preformed DNA adducts (from calf thymus treated with 300µM CuCl2 and100µM H2O2 in the presence of DMPO), and repair observed in the above experiment was inhibited when the methoxyamine was added to cell lysate (Figure 4C). NiCl2 added to cell lysate did not have this effect, indicating that the repair of this damage is predominantly by the BER pathway although the repair pathways are overlapping.
The measurement of oxidized bases and nucleosides in DNA may be used to gain insights into the nature and importance of chemical reactions generated in DNA by oxidizing agents. Significant improvement for the measurement of oxidatively generated damage to DNA has been obtained by the use of HPLC coupled to MS and MS/MS (3). This method has been gaining prominence and has overcome many of the limitations that previously hindered the measurement of oxidatively generated DNA damage (3). It now represents the usual method of choice for identifying modified DNA bases (1,2,32).
IST, which can be combined with MS and MS/MS, is a relatively new method for studying DNA radical damage (18–21). Our recent investigations with EPR and with IST with MS and MS/MS have allowed the characterization of nitrone adducts formed in Fenton systems by the hydroxyl radical (26). The objective of the present work was to determine the time course of formation and repair of free radical-derived DNA nitrone adduct, and to use confocal microscopy to visualize the damage to the DNA in the nucleus. Separation of cytosolic and nuclear fractions and subsequent DNA purification demonstrated the presence of DNA–DMPO adduct, presumably of the known DMPO/N6-centered adenosine nitrone adduct in the nucleus (26). MS and MS/MS confirmed the DMPO/N6-centered adenosine in the nuclear fraction (data not shown).
Using ESR, MS and MS/MS, we determined that the radical trapped by DMPO was located at N6 of the adenosine (26). Hydroxyl radical can add to the double bond at C4, C5 and C8 of the adenosine (33–37), of which C4 constitutes the most abundant addition site (81%) (11). This adduct undergoes proton or hydrogen transfer, followed by dehydration, to form an N6-dehydrogenated radical. N6 has also been shown elsewhere to be the most likely site of hydrogen abstraction by hydroxyl radical (38,39). However, the absence of detection of other DNA radicals is not proof of their absence. It should be noted that this type of site-specific hydroxyl radical generation should not form as many DNA-derived radicals as ionizing radiation where freely diffusing hydroxyl radical is formed.
When the RAW 267.4 cells were treated with CuCl2 and H2O2, the radical generated was trapped by DMPO immediately upon formation to form a DMPO nitrone adduct. The DMPO radical adduct formed is initially EPR active but subsequently decays to the more stable nitrone adduct, which can be very conveniently detected by the IST method (20,26,30) and confirmed and identified by MS and MS/MS. The radical is trapped spontaneously by DMPO, and artifactual generation of DMPO adducts during the extraction and digestion steps is minimized, if not eliminated. Furthermore, as it is a DMPO adduct that is formed, its identity can be confirmed by MS/MS from its fragmentation to DMPO and the parent radical ion, thereby eliminating the requirement of a pure product for identification.
It should be recognized that the oxidation event does not have to take place inside the cell nucleus, but may occur either within the cytosol or even in the extracellular compartment. However, the damage inflicted by CuCl2/H2O2 Fenton chemistry to the RAW 267.4 macrophages was indeed found to be largely localized to the nuclear DNA. This localization was evident from the confocal data as well as from the immunochemical detection of the cytosolic and nuclear fractions.
We also examined the persistence of the DMPO–DNA adducts formed. When the adducts were monitored over a period of time, they formed detectable amounts within 30min, then increased slowly over the next several hours before reaching a maximum at ~12h. Thereafter, repair of the DMPO–DNA nitrone adduct apparently dominates, accompanied by a drop in the adduct level to a value of ~50% of the maximum. However, because this effect could also be caused by turnover due to DNA synthesis, we also studied the repair of the preformed DNA–DMPO adduct added to cell lysate. Preformed DMPO–DNA nitrone repair by treatment with cell lysate was largely complete within an hour. Clearly, the DMPO–DNA nitrone adduct concentrations detected were the result of continuous hydroxyl radical attack on DNA balanced against DNA repair processes. DNA repair enzymes modify DNA damage including removal of DNA adducts, and there are multiple and overlapping DNA repair pathways (40). Inhibition of the repair pathways will block this mechanism. We used two known inhibitors, namely methoxyamine (TRC 102) of the BER pathway (41) and NiCl2 of the NER pathway (42). Methoxyamine is a small molecule repair inhibitor (41) and inhibited the repair of the single-base lesions in preformed DMPO–DNA nitrones.
In summary, with our method of combining IST with ESR, we were able to confirm that DMPO traps a radical adduct from DNA in accordance with published ELISA studies (18–21). This methodology provides an unequivocal assignment of the radical formed in DNA damage induced by the hydroxyl radical in a copper/hydrogen peroxide Fenton system (26). The radical produced when cellular DNA was treated with a Fenton-like system was found to be at the adenosine moiety, which may be an intermediate in a sequence of events leading to the formation of 8-oxo-dGuo (43) and other products such as 8-oxo-7,8-dihydro-2′-deoxyadenosine (8-oxodAdo). Although speculative for us, it has been predicted by hole-trapping researchers that the adenine radical cation contributes to the hole-transfer process through A/T sequences and exists as a real chemical intermediate (44–48). One-electron oxidation of DNA results in migration and localization of the electron loss center to guanine (43).
The damage was found to be localized in the nuclear DNA, but was repaired over a period of time, probably through the BER pathway. The repair study is preliminary in nature and makes clear the interfering effect of DNA repair in a biological system. It is necessary that in vivo experiments studying DMPO–DNA nitrone adducts be conducted over a range of times as repair of the DMPO–DNA nitrone adduct does occur.
Funding for open access charge: Intramural Research Program of the National Institutes of Health and the National Institute of Environmental Health Sciences.
Conflict of interest statement. None declared.
We gratefully acknowledge Jean Corbett for her unconditional help during all stages of the project. We are very thankful to Dr. Marilyn Ehrenshaft and Dr. Fiona Summers for reviewing the manuscript. We thank Dr. Ann Motten and Mary Mason for their invaluable help in preparing this manuscript.