The goal of this protocol is the detection of the DNA oxidation marker, 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxo-dGuo) by HPLC-ED, in DNA from cultured cells or animal tissues.
Oxidative stress is associated with many physiological and pathological processes, as well as xenobiotic metabolism, leading to the oxidation of biomacromolecules, including DNA. Therefore, efficient detection of DNA oxidation is important for a variety of research disciplines, including medicine and toxicology. A common biomarker of oxidatively damaged DNA is 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxo-dGuo; often erroneously referred to as 8-hydroxy-2'-deoxyguanosine (8-OH-dGuo or 8-oxo-dG)). Several protocols for 8-oxo-dGuo measurement by high pressure liquid chromatography with electrochemical detection (HPLC-ED) have been described. However, these were mainly applied to purified DNA treated with pro-oxidants. In addition, due to methodological differences between laboratories, mainly due to differences in analytical equipment, the adoption of published methods for detection of 8-oxo-dGuo by HPLC-ED requires careful optimization by each laboratory. A comprehensive protocol, describing such an optimization process, is lacking. Here, a detailed protocol is described for the detection of 8-oxo-dGuo by HPLC-ED, in DNA from cultured cells or animal tissues. It illustrates how DNA sample preparation can be easily and rapidly optimized to minimize undesirable DNA oxidation that can occur during sample preparation. This protocol shows how to detect 8-oxo-dGuo in cultured human alveolar adenocarcinoma cells (i.e., A549 cells) treated with the oxidizing agent KBrO3, and from the spleen of mice exposed to the polycyclic aromatic hydrocarbon dibenzo(def,p)chrysene (DBC, formerly known as dibenzo(a,l)pyrene, DalP). Overall, this work illustrates how an HPLC-ED methodology can be readily optimized for the detection of 8-oxo-dGuo in biological samples.
Reactive oxygen species (ROS), whose steady-state levels can increase during many pathological conditions and xenotoxic metabolism, contribute to an increased frequency of oxidative DNA damage. Among several possible nucleobases oxidation products, oxidative DNA damage can readily be measured using the stable marker 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxo-dGuo), which is one of the oxidized forms of 2’-deoxyguanosine (dGuo)1. 8-oxo-dGuo is the most abundant DNA lesion2 and, therefore, has been studied to greater detail as a DNA oxidation biomarker despite the existence of multiple DNA oxidation products 3. In humans, this damage can be repaired via base excision repair by 8-oxoguanine glycosylase 1 (hOGG1)4. If left unrepaired, 8-oxo-dGuo can contribute to the formation of base pair-substitution mutations (i.e., G to T transversions)4. Importantly, 8-oxo-dGuo is an established marker for DNA damage in relation to the initiation and promotion of carcinogenesis2. Therefore, accurate quantification of 8-oxo-dGuo is a useful and desirable biomarker of oxidative DNA damage5.
There is widespread confusion in the literature regarding the correct names for oxidatively-damaged forms of 2-deoxyguanosine and, moreover, the correct name of the compound(s) routinely measured as a biomarker of oxidative DNA damage6. The 6,8-diketo and 6-enol,8-keto tautomeric forms of 8-oxo-dGuo (shown in Figure 1) are the two most prominent tautomers discussed in the literature5,7. The 6,8-diketo form is the most prominent form at physiological pH of 7.4, and is the most prominent DNA oxidation product7. Therefore, 8-oxo-dGuo, rather than 8-hydroxy-dGuo is the most appropriate name for this oxidation product6. It is also important to note that 2-deoxyguanosine (dGuo), rather than nucleobase guanine (Gua) or ribonucleoside guanosine (Guo), respectively, is detected by most methods6.
Accurate detection and quantification of 8-oxo-dGuo is challenging due to: i) variability in the digestion of the DNA sample, ii) adventitious oxidation of dGuo to 8-oxo-dGuo that can occur during sample preparation, and iii) the need for effective validation of the analytical HPLC-ED method8. In this protocol, we aimed at achieving i) by providing conditions, favorable for complete DNA digestion and ii) by the inclusion metal chelator and chelator-treated solutions and a special DNA-isolating reagent, while iii) was only partially addressed by inclusion of positive controls and thus providing that the method is capable of detecting 8-oxo-dGuo in biological samples. Further validation is beyond the scope of this paper. However, we are confident that this protocol will help the prospective users determine the extent to which they need to formally validate the protocol, depending on their purposes. A list of steps required for the formal validation of the method is provided further. During the development and deployment of a method for 8-oxo-dGuo detection, published methods were reviewed and consolidated. Thus, this method eliminates the need to gather information from several published sources that often lack important experimental details while also providing rapid and straightforward means of testing if the method for the detection and quantification of 8-oxo-dGuo has been adopted successfully. This adapted method was employed to successfully analyze DNA samples from cultured cells and murine tissue. This video article will assist other groups in establishing an effective method for reliable detection and quantification of 8-oxo-dGuo by HPLC-ED.
Ensure that all animal husbandry, housing, handling and experimentation adhere to local rules and regulations and that experimentation protocols are approved prior to commencing any study. For the described experiments, animal care, handling, and treatment were approved by the Health Canada Animal Care Committee. See the “Reagents table” for the suppliers’ information.
1. Collecting Biological Samples
2. DNA Extraction, Precipitation and Wash (For Tissues Proceed Directly to 2.2)
3. Enzymatic Digestion
4. HPLC Run: Preparation of Mobile Phase, Instrumentation Setup and Maintenance
5. Preparation of Standards
6. Agarose Gel Electrophoresis
NOTE: Agarose gel electrophoresis may be performed to verify the completeness of DNA digestion.
dGuo was observed to have a retention time of 4.7 min whereas 8-oxo-dGuo had a retention time of approximately 6.4 min (Figure 2A and B). There is about 1,000-fold difference in the peak heights between the two analytes, as seen in Figure 2C. Voltammograms for 8-oxo-dGuo and dGuo were obtained by running standards at a working potential in the range of +0.2 to +1.1 V. The optimum working potential for 8-oxo-dGuo was determined to be +0.5 V, and +0.9 V for dGuo (Figure 3). These potentials are in agreement with other glassy carbon electrodes described in the literature14,15. The limit of detection and dynamic range of 8-oxo-dGuo and dGuo electrochemical detection is reported to be in the femtomole and nanomole range, respectively8,9. Standard curves for dGuo and 8-oxo-dGuo should be run daily to ensure linear detection across a suitable concentration range and proper performance of the instrument (Figure 4). Standard curves are constructed by plotting the peak area as a function of known analyte concentration (Supplementary Figure 2) that allows one to calculate the analyte concentration in the samples from the equation generated from the standard curves.
Complete DNA digestion is required for accurate measurements of 8-oxo-dGua frequency and several genomic DNA digestion methods have been suggested8,15,16. Where necessary, the efficacy of DNA extraction and digestion can be verified using agarose gel electrophoresis (Figure 5A) and HPLC-ED (Figure 5B). Lanes with obvious DNA smear in Figure 5A (e.g., lanes 2, 4, 6, 9, 11, and 13) indicate the presence of undigested DNA while samples with digested DNA fragments run off the gel and were thus undetected. The plateau in Figure 5B indicates that further incubation of DNA with digestion enzymes beyond 1.5 hr does not result in greater amount of dGuo detected, suggesting nearly complete digestion after 1.5 hr. In addition, both agarose gel electrophoresis and HPLC-ED were used to test the utility of the phosphate-based DNA digestion buffer employed here. This buffer is well matched to the HPLC mobile phase and does not give rise to undesirable detector noise attributable to mixing of dissimilar solutions (data not shown). Therefore, the switch from the buffer suggested in the literature (i.e., Tris-HCl8) to sodium phosphate buffer is recommended for this protocol (left versus right side of Figure 5A). Please note that the underlying assumptions for the optimization of DNA digestion were that enzyme inhibition and sample inhomogeneity were negligible and digestion time was the only limiting factor, as stated above. A possibility that incomplete DNA digestion resulted in fragments that are too small to be detected on the agarose gel remains. Even though such a systematic error would affect all samples equally, further experiments (e.g., treatment of cultured cells with a radiolabelled pro-oxidant10) could be conducted to completely rule out such a possibility.
Unlike standard solutions, the digested DNA, whether from isolated cells or mouse tissues, produced several additional peaks (Figure 6A and 6B). These were distant from the 8-oxo-dGuo and dGuo peaks, but spiking experiments as a part of validation exercises outlined in the Discussion are needed to examine if additional peaks (i.e., sample impurity due to matrix effects, for example) interfere with quantification. The 8-oxo-dGuo peak was confirmed by correspondence of the retention time with the standard; and moreover, increase in the 8-oxo-dGuo peak for samples from cells or animals that underwent pro-oxidant treatment (Figure 7). Pro-oxidant KBrO3 participates in one-electron abstraction from guanine that leads to 8-oxo-dGuo formation17. It is noteworthy that in the absence of the pro-oxidant stress, 2.4 molecules of 8-oxo-dGuo per 107 dGuo were detected in A549 cells (Figure 7A). This is comparable to 4.5 molecules of 8-oxo-dGuo/107 dGuo detected in A549 cells observed using an alternative, mass spectrometry-based approach designed to address potential shortcomings of the HPLC-ED method9. DBC treatment (i.e., daily 20.0 mg DBC/kg bw per day for three days) increased 8-oxo-dGuo levels in mouse spleen DNA. This is also an expected result, since ROS is known to be produced during the metabolism of polycyclic aromatic hydrocarbons in vivo8. The relative levels of 8-oxo-dGuo in the spleen of unstressed animals were 8.3-9.4 8-oxo-dGuo/106 dGuo (Figure 7B), which is in excellent agreement with published values of four 8-oxo-dG molecules/106 dGuo measured in murine splenic cells under standard conditions by HPLC-ED18. Figures 6 and 7 show that the levels of 8-oxo-dGuo above the baseline are very small. Thus if DNA is oxidized to lower levels than observed here, the levels of 8-oxo-dGuo may be indistinguishable from baseline, which should be kept in mind by prospective protocol users.
Figure 1. Structure of 8-oxo-dGuo and its tautomer 8-OH-dGuo. Note that 8-oxo-dGuo rather than 8-OH-dGuo is the major tautomer at pH 7.4.
Figure 2. Retention times for dGuo (A) and 8-oxo-dGuo (B). HPLC–ED chromatograms were obtained from a 10.0 μl injection of either dGuo, (1.0 nmol total) or 8-oxo-dGuo (1,000 fmol total) and detected at 0.9 V (A) or 0.5 V (B). The retention peak for dGuo was approximately 4.7 min (A) and that for 8-oxo-dGuo was approximately 6.4 min (B). The broad peak at 2-3 min is likely caused by injection disturbances and is unaffected (i.e., always present at similar intensity) by the concentration of 8-oxo dGuo. (C) Superimposed chromatograms from two separate injections described in (A) and (B) detected by UV (260 nm). The standards were run on the same day using identical run conditions. Please click here to view a larger version of this figure.
Figure 3. Optimizing detector voltage. (A) Voltammogram of dGuo (final concentration 1 nmol) detected by HPLC-ED analysis at a range of 0.7-1.1 V. (B) Voltammogram of 8-oxo-dGuo (final concentration 500 fmol) detected by HPLC-ED analysis at a range of 0.3-0.7 V. Means of three independent experiments (N=3) ± standard deviation shown. In some cases error bars are smaller than the plotting symbols.
Figure 4. Standard curve for dGuo (A) and 8-oxo-dGuo (B). Ten microliters of either dGuo or 8-oxo-dGuo was injected and measured by HPLC-ED at 0.9 V (A) or 0.6 V (B). Means of three independent experiments (N=3) ± standard deviation shown. In some cases error bars are smaller than the plotting symbols.
Figure 5. Optimization of DNA digestion. (A) Eighty microgram of salmon testes DNA was digested as described8 in Tris-HCl, glycine-acetate buffer (left) or sodium phosphate buffer (right). Agarose gel electrophoresis (1.0%) was run at 150 V for 45 min. Lane 1 and 8: DNA ladder; lanes 2, 4, and 6: undigested DNA; lanes 3, 5, and 7: digested DNA; lanes 9, 11, and 13: undigested DNA (phosphate buffer); lanes 10, 12, and 14: digested DNA (phosphate buffer). (B) Salmon testes DNA (80 µg) was digested herein by incubating with DNase I, PDE I, and AP for 0.5-2.5 hr per each step, in 50 mM phosphate buffer containing magnesium chloride and DFO. Fifty microliters of each sample, containing 7.3 µg of digested DNA, was analyzed by HPLC (i.e., 0.9 V and 1 ml/min flow) to detect dGuo. Figure shows means of three independent experiments (N=3) ± standard error of the mean. Please click here to view a larger version of this figure.
Figure 6. Chromatograms for 8-oxo-dGuo detection in cultured cells (A) or animal tissue (B). (A) Digested DNA from A549 cells, treated or untreated with KBrO3. (B) Digested DNA from the spleen of DBC-treated mice. Peaks shifted to the left due to the removal of guard column (due to pressure build-up in the apparatus; their position was confirmed with standards). Standards and samples were run on the same day using identical run conditions and 10 ml injection volumes, at 1 ml/min flow rate and 0.5 and 0.9 V for dGuo and 8-oxo-dGuo, respectively. Please click here for a larger version of panel A; here for panel B.
Figure 7. Quantification of 8-oxo-dGuo in biological samples. Peaks were integrated and compared to standard curves (see Supplementary Figure 2 for more details) to yield 8-oxo-dGuo levels in DNA from A549 cells, treated with KBrO3 (A), or 8-oxo-dGuo levels in DNA from the spleen of DBC-treated mice (B). Stars (*) indicate statistical significance (p < 0.05, one-way analysis of variance (ANOVA) (A), Student’s t-test (B)). Means of three (A) or five (B) independent experiments ± standard error of the mean are shown. Samples were collected 4 or 24 hr after the 3-day treatment.
Mobile phase | 50.0 mM Na phosphate buffer, pH 6.2, containing 6% MeOH and 2.0 mM KCl |
Mobile phase flow rate | 1.0 ml/min |
Column type | YMC-BASIC with bonded spherical silica |
Column length | 15 cm |
Column inner diameter | 3.5 mm |
Column temperature | 35 °C |
Detector temperature | 29 °C |
Voltage setting for 8-oxo-dGuo | +0.5 V |
Voltage setting for dGuo | +0.9 V |
Injection volume | 10.0 ml |
Table 1. Instrument parameters for detection and quantification of 8-oxo-dGuo by HPLC-ED.
nmoles | 100 nM standard, µl | HPLC Mobile phase, µl | Dilution, fold |
5 | 300 | 0 | 1 |
2.5 | 150 | 150 | 2 |
1 | 60 | 240 | 5 |
0.5 | 30 | 270 | 10 |
0.25 | 15 | 285 | 20 |
0.1 | 6 | 294 | 50 |
Table 2. Preparation of standard curve for dGuo.
fmoles | 250 nM standard, µl | HPLC Mobile phase, µl | Dilution, fold |
2,500 | 300 | 0 | 1 |
1,000 | 120 | 180 | 2.5 |
500 | 150 | 150 | 5 |
250 | 150 | 150 | 10 |
100 | 120 | 180 | 25 |
50 | 150 | 150 | 50 |
Table 3. Preparation of standard curve for 8-oxo-dGuo.
Supplementary Figure 1. Comparison of electrochemical (EC) and UV-based detection of dGuo. Standard curve was created for dGuo detection by electrochemical detection (EC) or UV detection (260 nm). Peak areas were integrated for dGuo peak at approximately 4.7 min by both methods. N=3, means ± standard deviation are shown.
Supplementary Figure 2. Quantification of dGuo from its standard curve. An example of the use of standard curve to determine the concentration of dGuo is shown. (A) Peak area can be determined by integration using standard software. (B) This information is used to construct a standard curve in order to use the equation of the line to calculate the unknown analyte concentration in the sample. Please click here to view a larger version of this figure.
Although 8-oxo-dGuo has been reported as a useful biomarker of DNA oxidation, its reliable quantification can pose a challenge. Although several published methods exist, there is a need for a comprehensive, descriptive overview of protocol to permit researchers to deploy the method in their laboratories. Here we present a detailed overview of an HPLC-based protocol that will permit new users to establish an effective method for 8-oxo-dGuo detection and quantification.
Three major methods that have been described for quantification of 8-oxo-dGuo. These include enzyme-linked immunosorbent assay (ELISA)-based commercial kits19, liquid chromatography/mass spectrometry (MS) methods (e.g., as described elsewhere9 and HPLC-ED methods, such as that described herein. ELISA methods can suffer from interferences with certain compounds in biological samples19. In general, results obtained using chromatographic techniques are considered more reliable for 8-oxo-dGuo quantification in comparison with ELISA-based methods20. In comparing the HPLC-ED methods with the MS-based methods, the advantage of the former is relatively inexpensive equipment (i.e., approximately an order of magnitude less costly than equipment for the MS-based methods). In addition, HPLC-MS are superior to HPLC-ED methods as the former are not prone to potential interferences with electroactive compounds and provide unambiguous, quantitative measurements using isotopes21.
Concerns have been raised regarding the use of traditional phenol extraction methods for the extraction of genomic DNA to be used for 8-oxo-dGuo quantification22. Several researchers have noted that the conditions associated with the standard phenol extraction can introduce unwanted dGuo oxidation thus increasing baseline levels1,9. To address this problem, cell and animal genomic DNA can be extracted and purified using a modified DNAzol protocol22. DNAzol contains high concentrations of guanidine thiocyanate, and previous studies9 have noted lower basal level 8-oxo-dGuo formation when DNAzol is used compared to traditional phenol- and NaI-based extraction methods. The iron chelating compound DFO was used to minimize ROS production and DNA oxidation during sample preparation. As noted above, the 8-oxo-dGuo levels in unstressed A549 cells (Figure 7A) were highly similar to what was measured by an MS-based method13. This suggests that incorporation of precautions described for the MS-based methods13 permitted minimization of undesirable DNA oxidation during sample preparation and analysis to achieve accurate 8-oxo-dGuo quantification by HPLC-ED. Timing of necropsy or harvesting culture cells after pro-oxidant treatment may be another important factor, as seen here by declining 8-oxo-dGuo levels in the spleen at 24 hr vs. 4 hr after the last treatment (Figure 7B). This may be due to the activation of the base excision repair to eliminate DNA damage.
The two obvious limitations of the described method are: (1) the requirement for rather large quantities of DNA (about 80 μg per sample to achieve three technical replicates), and (2) potential interference from co-eluting compounds. It was found that only a very small (i.e., about 15 mg) piece of spleen (about 15% of total tissue weight) was sufficient to obtain 80 μg of DNA, leaving a sufficient amount of tissue for other (e.g., gene expression) analyses. The amount of extracted DNA is another critical parameter that affects artifactual DNA oxidation23,24,25. Therefore, working with rather large (~80 mg/sample) that we suggest here is consistent with previous recommendations of using > 30 mg to minimize artifactual DNA oxidation25. For smaller tissues (e.g., the hippocampus), one could scale down the protocol to start with 20-30 μg DNA, since this should be sufficient for 2-3 runs. Larger quantities of DNA would presumably facilitate the resolution of peaks above background noise, but this remains to be experimentally verified The presence of potentially interfering substances can be addressed on a case-by-case basis, and it is essential to always include DNA sample from simultaneous in vitro or in vivo treatments with a positive control that is known to induce oxidative DNA damage (see Figure 7).
Although the phosphate in phosphate buffers has been reported to reduce the activity of alkaline phosphatase, method optimization experiments (Figure 5) confirmed that phosphate buffer could readily be used. Since phosphate buffer is contained in the mobile phase, its use in sample preparation reduced detector noise due to solvent mixing. Thus voltage, buffer composition, DNA extraction and digestion were optimized and positive controls were included. Moreover, a summary outlining how these variables can be further optimized by individual laboratories is provided. Our goal was to produce a method that could be useful for specialist laboratories aimed at developing an assay for 8-oxo-dGuo detection and quantification.
However, we note that formal validation of the method is required, as outlined under the validations steps recommended by the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) for validation of analytical procedures (i.e., ICH Q2(R1)26). Briefly, the elements of quantitative validation should examine the assay’s: 1) accuracy (which can be accomplished by including samples with different amounts of 8-oxo-dGuo spiked-in; 2) range and linearity (linear response across extended analyte concentration; 3) precision (repeatability and reproducibility); 4) limit of detection (generally, the point at which the signal to noise ratio is greater or equal 2-3; this may be matrix-dependent); 5) limit of quantitation (concentration at which the response meets some predetermined levels such as 2 x standard deviation of control; may be matrix-dependent); 6) selectivity/specificity (ability to detect the analyte in complex samples versus neat solutions); 7) reproducibility (the ability to get the same result in different labs with different operators); and 8) robustness and system suitability. As the purpose of the validation recommended by the ICH is to demonstrate that it is “suitable for its intended purposes” and many laboratories will likely have different purposes, the prospective users of this assay would have to validate this to the extent deemed necessary. Accurate quantification of 8-oxo-dGuo is desirable in toxicology and molecular medicine since it can enhance the understanding of how oxidative stress, and more specifically DNA oxidation, is mechanistically and empirically linked to adverse health effects.
The authors have nothing to disclose.
This research was funded by the Health Canada Genomics Research and Development Initiative (GRDI) and the Canadian Regulatory Strategy for Biotechnology (CRSB). The authors have no conflict of interests.
8-oxo-dGuo standard | Cayman Chemical Company | 89320 | Inappropriately referred to as "8-hydroxy-2'-deoxy Guanosine" – see Fig. 1 and text for details |
Alkaline phosphatase | Sigma-Aldrich | P5931 | From E.coli |
Chelex 100 | Sigma-Aldrich | C7901 | Chelates heavy metals |
Desferoxamine mesylate | Sigma-Aldrich | D9533 | |
dGuo standard | Sigma-Aldrich | D7145 | |
Dibasic sodium phosphate | Sigma-Aldrich | S9390 | |
DNA from salmon sperm | Sigma-Aldrich | D1626 | Sodium salt |
DNase I | Sigma-Aldrich | D4527 | TypeII, from bovine pancreas |
DNAzol | Invitrogen | 10503-27 | |
Ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA) | Sigma-Aldrich | E4884 | The compound would not completely dissolve until solution pH is adjusted to 8.0 with e.g. NaOH |
F12-K media | ATCC | 30-2004 | |
Foetal bovine serum | ATCC | 30-2020 | |
Guard column | Chromatographic Specialties | YBA 99S03 0204GC | Protects colum from contamination; may also lead to pressure build-up |
Magnesium chloride | Sigma-Aldrich | M8266 | |
Monobasic sodium phosphate | Sigma-Aldrich | S9638 | |
Penicillin-Streptomycin | Invitrogen | 15140-122 | |
Phosphate buffered saline | Invitrogen | 15190-250 | |
Phosphodiesterase I enzyme | Sigma-Aldrich | P3243 | Type II from Crotalus adamaneus venom |
Teflon homogenizer | Thomas Scientific | 7724T-1 or 7724T-5 for 1 or 5 mL, respectively | Volume (holding capacity) depends on the amount of sample to be processed. |
Trypsin | Invitrogen | 15050-065 | |
YMC-BASIC column with bonded spherical silica | Chromatographic Specialties | YBA 99S03 1546WT |