Search

Two multiresidue methods based on different extraction procedures have been developed and compared for the liquid chromatography electrospray ionization tandem mass spectrometry analysis of 17 mycotoxins including ochratoxin A, aflatoxins (B(1), B(2), G(1), and G(2)), zearalenone, fumonisins (B(1) and B(2)), T-2 toxin, HT-2 toxin, nivalenol, deoxynivalenol, 3- and 15-acetyldeoxynivalenol, fusarenon-X, diacetoxyscirpenol, and neosolaniol in cereal-based commodities. The extraction procedures considered were a QuEChERS-like method and one using accelerated solvent extraction (ASE). Both extraction procedures gave similar performances in terms of linearity (r(2) > 0.98) and precision (both RSD(r) and RSD(iR) < 20%). Trueness was evaluated through participation in four proficiency tests and by the analysis of two certified reference materials and one quality control material. Satisfactory Z scores (|Z| < 2) and trueness values (73-130%) were obtained by the proposed procedures. Limits of quantification were similar by both methods and were within the 1.0-2.0 microg/kg range for aflatoxins, 0.5 microg/kg for ochratoxin A, and the 5-100 microg/kg range for all other mycotoxins tested. The QuEChERS-like method was found to be easier to handle and allowed a higher sample throughput as compared to the ASE method.

With the emergence of liquid chromatography coupled to tandem quadrupolar mass spectrometry (LC-QqQ) as a routine technique for quantitative analysis, analytical chemists claimed LC-QqQ to be the gold standard to reach the best compromise between versatility, high throughput, robustness, sensitivity and selectivity. In particular, a high selectivity is ensured when two or more transitions are monitored because not only the retention time and protonated molecule are controlled but also two or more product ions are. With the multiple-transition recording, the transition leading to the most intense signal is used for the quantification (quantifier), while the other one(s) is(are) aimed at confirming the detection of the analyte (qualifiers). The confirmation is based on the calculation of the relative intensity between the signal intensities of the quantifier and the qualifier(s). This useful approach raises the question of the validity of the limit of detection (LOD), initially employed for mono-channel detections such as HPLC combined with ultraviolet or fluorescence detection. Furthermore, it was shown that the multiple-transition recording leads to a confusing calculation of the decision limit (CCalpha) and detection capability (CCbeta). In the present article, the LOD is split in three concepts defined as the limit of suspicion (LOS), recognition (LOR), and confirmation (LOC). For these three limits, applications and drawbacks are shown, while determination methods are proposed.

Six ergot alkaloids belonging to the lysergic acid derivatives (ergonovine (EGN) and methysergide hydrogen maleinate (MHM)) and peptide-type derivatives (ergocristine (EGR), ergotamine (EGT), ergocornine (EGC) and α-ergokryptine (EGK)) were studied by positive electrospray tandem mass spectrometry. The fragmentation mechanisms of these compounds were studied by collision-induced dissociation (CID) using triple quadrupole and ion trap mass spectrometers, and the nature of the major product ions further confirmed by hydrogen/deuterium (H/D) exchange experiments. A common abundant product ion at m/z 223 was characteristic of the two classes of ergot alkaloids. Therefore, a precursor ion scan of m/z 223 that triggers information data acquisition (IDA) in combination with CID experiments was used to identify other potential ergot alkaloids. Using this approach, it was possible to confirm the presence of ergosine, another peptide-type ergot alkaloid, in a rye flour extract at trace levels. Copyright © 2006 John Wiley & Sons, Ltd.

A confirmatory method based on isotope dilution liquid chromatography-tandem mass spectrometry (LC-MS/MS) has been developed for the low-level determination of residues of four nitrofuran veterinary drugs in meat, e.g., furazolidone, furaltadone, nitrofurantoin, and nitrofurazone. The procedure entails an acid-catalysed release of protein-bound metabolites, followed by their in situ conversion into the 2-nitrobenzaldehyde (NBA) imine-type derivatives. Liquid-liquid extraction and clean-up on a polymeric solid phase extraction cartridge are then performed before LC-MS/MS analysis by positive electrospray ionisation (ESI) applying multiple reaction monitoring of three transition reactions for each compound. Reliable quantitation is obtained by using one deuterated analogue per analyte (d4-NBA derivative) as internal standard (IS). Validation of the method in chicken meat was conducted following the European Union (EU) criteria for the analysis of veterinary drug residues in foods. The decision limits (CCalpha) were 0.11-0.21 microg/kg, and the detection capabilities (CCbeta) 0.19-0.36 microg/kg, thus below the minimum required performance limit (MRPL) set at 1 microg/kg by the EU. The method is robust and suitable for routine quality control operations, and more than 200 sample injections were performed without excessive pollution of the mass spectrometer or loss of LC column performance.

A method is described for the determination of residues of the illegal antibiotic chloramphenicol (CAP) in milk powders. The analyte is quantified by liquid chromatography coupled to electrospray ionisation tandem mass spectrometry (LC-ESI-MS-MS) operating in negative ion multiple reaction monitoring mode (MRM) after a liquid-liquid extraction followed by a clean-up step on solid phase extraction (SPE) cartridge. Because of the presence of two chlorine atoms in the CAP molecule, four specific transition reactions of CAP were monitored by MS-MS in selecting m/z 321 --257, 321 --152 (35Cl2) and m/z 323 --257, 323 --152 (37Cl35Cl). Two calibration curves were constructed by plotting the area ratio of m/z 321 --152 versus 326 --157 and m/z 321 --257 versus 326 --262 against their corresponding amount ratio. Indeed, even if m/z 321 --152 was found to give a higher MS-MS response (calibration curve used by default), an interfering chemical substance was sometimes observed for some milk extracts and not for the transition m/z 321 --257. The quantitation method was validated according to the European Union (EU) criteria for the analysis of veterinary drug residues at 0.1, 0.2 and 0.5 microg/kg concentration levels using d5-CAP as internal standard. The decision limit (CCalpha) and detection capability (CCbeta) of CAP in milk were calculated for m/z 321 --152 at 0.02 microg/kg and 0.03 microg/kg, respectively, and for m/z 321 --257 at 0.02 microg/kg and 0.04 microg/kg, respectively. At the lowest fortification level (i.e. 0.1 microg/kg), repeatability and within-laboratory reproducibility were calculated for m/z 321 --257 both at 0.02 microg/kg and for m/z 321 --152 at 0.03 and 0.05 microg/kg, respectively. Moreover, the measurement of uncertainty of the analytical method was calculated at the same spiking levels and falls within the precision values of the within-laboratory reproducibility. This method can be applied to several types of milk powders (e.g. full cream, skim) and can serve as a monitoring tool to avoid that unacceptable levels of residues of CAP enter the food chain.

A sensitive and selective isotope dilution liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESIMS/MS) method is presented for the simultaneous analysis of the metabolites of four nitrofuran veterinary drugs, that is, furazolidone, furaltadone, nitrofurantoin, and nitrofurazone, in honey samples. The method entails a combined hydrolysis of protein-bound drug metabolites and derivatization of the resulting metabolites with 2-nitrobenzaldehyde (NBA) during an overnight incubation, followed by a liquid-liquid extraction and a cleanup on a polymeric solid-phase extraction cartridge. Mass spectral acquisition is carried out in the positive ion mode by applying multiple reaction monitoring (MRM) of three diagnostic transition reactions for each analyte under survey. A reliable quantification is obtained by the use of one deuterated analogue per analyte (NBA-d(4) derivative). The method has been validated in honey according to the European Union criteria for the analysis of veterinary drug residues in food. Expressed in underivatized nitrofuran metabolite concentrations, the decision limits (CCalpha) ranged within 0.07-0.46 microg/kg, and the detection capabilities (CCbeta) were within 0.12-0.56 microg/kg. The method has been successfully applied in a survey of honeys of various geographical origins, showing that furazolidone is the main nitrofuran antibiotic administered to treat bacterial diseases of bees.

Five mutagenic heterocyclic aromatic amines (HAAs) were quantified from meat extracts, and grilled and pan fried bacon samples using stable isotopically labeled internal standards. These compounds were isolated from the matrices by a tandem solid-phase extraction procedure, followed by separation on reversed-phase liquid chromatography (HPLC) and quantified by atmospheric pressure chemical ionization tandem mass spectrometry (APCIMS-MS). Tandem mass spectrometry (MS-MS) acquisition was done in selected reaction monitoring (SRM) mode to provide a high degree of sensitivity and selectivity for accurate quantification of HAAs. The detection and quantification limits of these HAAs approached 0.015 and 0.045 microg/kg (part-per-billion), respectively, with only 4 g of meat. The HAA levels ranged widely from 0.045 to 45.500 microg/kg, and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) was the predominant HAA found in these samples. The amount of HAAs formed was highly dependent upon the type of meat and method of preparation. An intralaboratory comparison of the extraction procedure showed that estimates of these HAAs obtained by three different individuals at HAA levels below 2 microg/kg were within 5% with coefficients of variation below 19%, indicating the robustness of the analytical method. Moreover, because all of these HAAs from this class of molecules undergo facile cleavage at the N-methylimidazole moiety under collision-induced dissociation (CID) conditions, MS-MS analysis in the constant neutral loss mode of [M+H]+-15 enabled the identification of two other HAAs, 2-amino-3-methylimidazo[4,5-f]quinoxaline (IQx) and 2-amino-1,7,9-trimethylimidazo[4,5-g]quinoxaline (7,9-DiMeIgQx), which have rarely been reported in cooked meats.

The quantitative aspect of the electrochemical detection method to detect 8-oxo-7,8-dihydroguanine (8-oxoGua) has been improved by using an internal standard. In addition, emphasis was placed on the reduction of artifactual oxidation of DNA during isolation and hydrolysis. Nuclear DNA was isolated from rat organs and purified on an anion-exchange column following treatment with proteinase K and RNase. DNA hydrolysis to nucleobases or nucleosides was performed using either formic acid treatment or enzymatic digestion, respectively. The levels of either 8-oxoGua or 8-hydroxy-7,8-dihydro-2'-deoxyguanosine were comparable. For accurate quantification, 2,6-diamino-8-oxopurine [(NH2)2-OH-Pur], added prior to hydrolysis, was used as an internal standard for the high-performance liquid chromatography with electrochemical detection assay. The baseline level of 8-oxoGua in DNA of Sprague-Dawley rats was estimated to be 2 to 5 8-oxoGua residues per 10(6) DNA bases, with slight differences depending on the tissue origin. In agreement with the results of previous observations, the level of the oxidized base in the kidney of animal treated with iron complexed to nitrilotriacetic acid (Fe-NTA) (15 mg/kg) was three- to fourfold higher than that of untreated rats or animals treated with a saline solution, while there was no change in 8-oxoGua levels in the liver and colon of these treated animals.

A rapid solid phase extraction (SPE) method was developed to analyze volatile N-nitrosamines (NAm) in foods. The currently used vacuum distillation technique is time and solvent consuming and is cumbersome for multiple analyses. Our objective was to develop a robust analytical method to measure the eight most common volatile NAm found in foods. The extraction procedure was based on two consecutive extraction/concentration steps using Extrelut and Florisil SPE. The amounts of food sample and solvent required were reduced compared to the conventional vacuum distillation method without affecting the sensitivity which was at a 0.3 ppb detection limit. A greater confidence in the analyte detection was obtained by the use of a capillary GC column, instead of the conventionally used open packed column, prior to thermal energy analyzer (TEA) detection. Another advantage of our method is the use of commercial cartridges, which ensures reproducibility and standardization of the method for multiple analyses. An appl...

A novel method utilizing 14C palmitic acid for image analysis and liquid scintillation for quantifying oil uptake of French fries undergoing deep-fat frying is described. Radiolabeled computerized image analysis of the fried product furnished comprehensive visual information on the distribution and localization of oil uptake which was limited to the crust. A significant high linear correlation (R2 = 0.977; P < 0.001) was found between oil uptake measured by liquid scintillation and differential scanning calorimetry, and between liquid scintillation and radiolabeled imaging (R2 = 0.861; P < 0.001). The high specificity and sensitivity of the method allows quantifying uptake of very low oil concentrations ranging from several milligrams. Oil quantification via liquid scintillation was more sensitive than image analysis and requires up to 3 order of magnitude lower concentration of the radiolabeled marker. The low concentration and inherent stability of the radiolabeled chemical make this technique uniquely ...

Pentachlorophenol (PCP) was measured in wood following extraction with methanol:0.1% acetic acid. The extract was derivatized with acetic anhydride in a biphasic solvent containing toluene and 0.1 M Na2CO3 (pH 11.4), and pentachlorophenyl acetate was quantitated by gas chromatography−mass spectrometry (GC−MS) using [13C6]PCP as an internal standard. A second detection method employed thin layer chromatography (TLC) after nitric acid-mediated oxidation of PCP to tetrachloro-1,4-benzoquinone. The third method was by colorimetry and required a liquid−liquid partitioning of PCP under acidic and alkaline pH against toluene. Then, PCP was coupled to 4-aminoantipyrine in the presence of sodium persulfate to form a dye which was measured at 580 nm. The limit of detection approached 50 ppb by GC−MS and 1 ppm by TLC when 100 mg of wood was used for analysis and 1 ppm for the colorimetric method when 1 g of wood was used for detection. The estimates of PCP were comparable for all three methods. Keywords: Chloropheno...

A method making use of turbulent flow chromatography automated online extraction with tandem mass spectrometry (MS/MS) was developed for the analysis of 4 quinolones and 12 fluoroquinolones in honey. The manual sample preparation was limited to a simple dilution of the honey test portion in water followed by a filtration. The extract was online purified on a large particle size extraction column where the sample matrix was washed away while the analytes were retained. Subsequently, the analytes were eluted from the extraction column onto an analytical column by means of an organic solvent prior to chromatographic separation and MS detection. Validation was performed at three fortification levels (i.e., 5, 20, and 50 microg/kg) in three different honeys (acacia, multiflower, and forest) using the single-point calibration procedure by means of either a 10 or 25 microg/kg calibrant. Good recovery (85-127%, median 101%) as well as within-day (2-18%, median 6%) and between-day (2-42%, median 9%) precision values was obtained whatever the level of fortification and the analyte surveyed. Due to the complexity of the honey matrix and the large variation of the MS/MS transition reaction signals, which were honey-dependent, the limit of quantification for all compounds was arbitrarily set at the lowest fortification level considered during the validation, e.g., 5 microg/kg. This method has been successfully applied in a minisurvey of 34 honeys, showing ciprofloxacin and norfloxacin as the main (fluoro)quinolone antibiotics administered to treat bacterial diseases of bees. Turbulent flow chromatography coupled to LC-MS/MS showed a strong potential as an alternative method compared to those making use of offline sample preparation, in terms of both increasing the analysis throughput and obtaining higher reproducibility linked to automation to ensure the absence of contaminants in honey samples.

A simple and fast selective extraction of the antibiotic chloramphenicol (CAP) from milk (raw milk, skimmed milk, and milk powder) using a molecularly imprinted polymer (MIP) sorbent is described. The method entails a single centrifugation step prior to loading the supernatant onto the MIP cartridge and subsequent elution with a mixture of solvents. CAP was further analyzed by isotope dilution liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) operating in negative ionization acquisition mode. The advantages of the MIP approach were assessed by comparing the data generated from a classical solid-phase and liquid-liquid extractions procedure, previously developed in our laboratory. A better recovery of CAP due to an enhanced selectivity and a faster turnaround time (18 samples processed within 3 h compared to 8 h with the classical approach) were evidenced when using the MIP cleanup. The analysis of CAP in raw milk was further validated according to the 2002/657/EC European Union criteria for the analysis of veterinary drug residues at the minimum required performance limit (MRPL) of 0.3 microg/kg, using CAP-d(5) as internal standard. Non-internal-standard corrected recovery values ranged between 50% and 87% over the range of concentrations considered. The decision limit (CCalpha) and detection capability (CCbeta) were calculated to be 0.06 and 0.10 microg/kg, respectively.

A multi-screening approach for monitoring potential chemical contaminants in honey by liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) has been developed. A total of 42 veterinary drugs (5 tetracyclines, 7 macrolides, 3 aminoglycosides, 8 beta-lactams, 2 amphenicols and 17 sulfonamides) were surveyed with the ultimate goal of unambiguously confirmed and quantified these analytes at a concentration level of 20 microg/kg. A basic sample preparation including four subsequent liquid/liquid extraction steps was necessary to adequately extract the compounds of interest from the honey. The four extracts were injected into the LC-ESI-MS/MS using a stacking injection procedure. Validation of the entire procedure was carried out according to the European Union directive 2002/657/EC at three concentration levels, i.e., 10, 20 and 30 microg/kg. Good performance data were obtained for 37 analytes, out of the 42 studied. Limit of compliance and detection limit were calculated based on an internal limit set at 20 microg/kg for all the compounds and ranged between 24-30 and 27-80 microg/kg, respectively. A limited survey on honeys of different geographical origins has demonstrated that positive honey samples were often contaminated by more than one class of drugs, thus highlighting the usefulness of such multi-screening approach to ensure and warrants the quality of honey.

An isotope dilution liquid chromatography-electrospray ionization-tandem mass spectrometry method is presented for the simultaneous analysis of several 5-nitroimidazole-based veterinary drugs, which are dimetridazole (DMZ), ronidazole (RNZ), metronidazole (MNZ), ipronidazole (IPZ), and their hydroxylated metabolites (DMZOH, MNZOH, and IPZOH), in egg (fresh egg, whole egg powder, and egg yolk powder) and chicken meat. Data acquisition was achieved by applying multiple reaction monitoring, and quantitation was performed by means of five deuterated internal standards (ISs), namely, DMZ-d3, RNZ-d3, IPZ-d3, DMZOH-d3, and IPZOH-d3, whereas MNZ and MNZOH were quantitated using DMZOH-d3. At the lowest fortification levels (i.e., 0.5 microg/kg for fresh egg and chicken meat and 1.0 microg/kg for other egg-based matrices) and for compounds having their own corresponding deuterated analogue used as an IS, acceptable performance data were obtained (corrected recoveries, 88-111%; decision limits, 0.07-0.36 microg/kg; detection capabilities, 0.11-0.60 microg/kg; and within-lab precision,or = 15%). The method failed to give acceptable quantitative results for MNZ and MNZOH due to the unavailability of the corresponding deuterated ISs. Nevertheless, a reliable identification of these two analytes at levelsor = 1 microg/kg was still feasible.

A confirmatory method for detecting five ergot alkaloids, ergocristine, ergotamine, ergonovine, ergocornine and α-ergokryptine, in rye flour is described using high performance liquid chromatography coupled to tandem mass spectrometry detection by monitoring two transition reactions per analyte. The procedure entails a liquid–liquid extraction followed by a clean-up step using a C18 solid-phase extraction (SPE) cartridge. An analogue compound, methysergide hydrogen maleinate, was used to assess both repeatability sample preparation and potential MS response fluctuations. The method was fully validated according to the European Union (EU) criteria. Detection and quantification limits of all analytes were calculated ranging from 7 to 11 μg/kg and from 23 to 37 μg/kg, respectively. Fifteen rye flour samples were investigated with the newly developed method, and none of them were above the current Swiss limits of 200 mg/kg for total ergot alkaloids.

Evidence is presented for the first time showing that semicarbazide (SEM) is a minor thermal decomposition product of the blowing agent azodicarbonamide (ADC). A novel direct analytical method based on liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESIMS/MS) has been developed to determine SEM in foamed polyvinyl chloride (PVC) seals of metal lids, as well as in commercially available ADC. The direct LC-MS/MS method for gaskets entails extraction of the gaskets in hot water, addition of ((15)N(2)(13)C)-SEM as internal standard, and injection of an aliquot directly into the LC-MS system, achieving good sensitivity (S/N = 348 for 2 ng injected on-column) and monitoring three characteristic mass transitions (m/z 76--31; 76 --44; 76--59). Semicarbazide can be detected in thermally treated ADC, reaching up to 0.93 mmol mol(-1) at 220 degrees C, as determined by the direct LC-MS/MS method. This new method is also compared to the classical derivatization method using 2-nitrobenzaldehyde (2-NBA) that is routinely employed to determine SEM as an indicator of the usage of the antimicrobial drug nitrofurazone, the use of which is not authorized in the European Union (EU). Both methods revealed proportional results, with approx. 3-fold higher levels recorded by the direct SEM approach, probably due to differences in the extraction procedures used. A limited survey of plastic seals from used press twist and twist-off metal lids on food jars (non-foamed and foamed) revealed levels of SEM ranging from 2 to 8689 microg kg(-1)(average = 1593 microg kg(-1), n= 57 determinations).

A confirmatory method based on isotope dilution liquid chromatography-electrospray ionization tandem mass spec- trometry is described for the determination of the antibiotic chloramphenicol (CAP) in foods. The method is quantitative and entails liquid-liquid extraction followed by a clean-up step on a silica gel solid-phase extraction cartridge. Mass spectral acquisition is done in the negative ion mode applying multiple reaction monitoring of two diagnostic transition reactions for CAP (m/z 321→ 257 and m/z 321→ 152). In addition, the presence of two chlorine atoms in the CAP molecule provides 37 35 further analyte certainty by assessing the Cl / Cl ratio using the transition reactions m/z 323→ 257 and m/z 323→ 152. Validation of the method in chicken meat is conducted according to the latest European Union criteria for the analysis of 2 veterinary drug residues at levels of 0.05, 0.10, and 0.20 mg / kg, employing ( H )-chloramphenicol as internal standard. The 5 decision limit and the detection capability were calculated at 0.01 mg / kg and 0.02 mg / kg, respectively. At the lowest fortification level (i.e. 0.05 mg / kg), precision values below 14 and 17% were achieved under repeatability and within- laboratory reproducibility conditions, respectively. The accuracy of the method was within 20, 15, and 5% of the target values at the 0.05, 0.10, and 0.20 mg / kg fortification levels, respectively. The applicability of this procedure was demonstrated by the analysis of other meat (turkey, pork, beef) and seafood (fish, shrimps) products. The method is robust and suitable for routine quality control operations, and more than 200 sample injections were performed without excessive pollution of the mass spectrometer or loss of LC column performance.  2003 Elsevier Science B.V. All rights reserved.

Two mass spectrometric methods were established for the quantitative analyses of alpha-tocopherol (TH) and its oxidation product alpha-tocopherolquinone (TQ) in human plasma. Both methods make use of isotopically labeled internal standards of different levels of deuteration (d3-TH and d6-TQ). Plasma (100 microl) was saponified in the presence of a mixture of antioxidants, and then TH and TQ were extracted with hexane. With the GC-MS method, the analytes were first converted into O-trimethylsilyl derivatives before analysis in the selective ion monitoring mode. The derivatization procedure led to the quantitative conversion of TQ into the O-trimethylsilyl derivative of tocopherolhydroquinone, giving rise to a more stable molecule with less fragmentation than for TQ. The increased stability of the molecule resulted in an enhanced contribution of the base peak to the total observed ions and therefore an increased sensitivity of the base peak for quantification. With the liquid chromatography-tandem mass spectrometry (LC-MS/MS) method, TH and TQ were detected by multiple reaction monitoring after positive electrospray ionization. The GC-MS and LC-MS/MS methods showed nearly the same accuracy (>95%) and the same within-day precisions, with less than 5 and 10% for TH and TQ, respectively. The between-day precision and the limit of quantification for TQ in plasma were better by LC-MS/MS (4%; 3 nM) than by GC-MS (21%; 10 nM). Analysis and method validation were carried out with plasma samples obtained from a male volunteer pre- and postexercise. Both techniques showed that the ratio of TQ/TH was elevated by 35% immediately after exercise and had returned to basal levels when measured 24 h later.

The aim of ESCODD, a European Commission funded Concerted Action, is to improve the precision and accuracy of methods for measuring 8-oxo-7,8-dihydroguanine (8-oxoGua) or the nucleoside (8-oxodG). On two occasions, participating laboratories received samples of different concentrations of 8-oxodG for analysis. About half the results returned (for 8-oxodG) were within 20% of the median values. Coefficients of variation (for three identical samples) were commonly around 10%. A sample of calf thymus DNA was sent, dry, to all laboratories. Analysis of 8-oxoGua/8-oxodG in this sample was a test of hydrolysis methods. Almost half the reported results were within 20% of the median value, and half obtained a CVof less than 10%. In order to test sensitivity, as well as precision, DNA was treated with photosensitiser and light to introduce increasing amounts of 8-oxoGua and samples were sent to members. Median values calculated from all returned results were 45.6 (untreated), 53.9, 60.4 and 65.6 8-oxoGua/10(6) Gua; only seven laboratories detected the increase over the whole range, while all but one detected a dose response over two concentration intervals. Results in this trial reflect a continuing improvement in precision and accuracy. The next challenge will be the analysis of 8-oxodG in DNA isolated from cells or tissue, where the concentration is much lower than in calf thymus DNA.

The measurement of oxidative damage to cellular DNA is a challenging analytical problem requiring highly sensitive and specific methods. In addition, artefactual DNA oxidation during its extraction and subsequent work-up may give rise to overestimated levels of oxidized DNA bases. In the present study, we have used (18)O-labelled 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxodGuo) as an internal standard to evaluate the extent of artefactual DNA oxidation during the critical steps preceding the measurement. The labelled oxidized purine nucleoside was specifically generated in cellular DNA using the recently available generator of (18)O-labelled singlet oxygen. Artefactual DNA oxidation that could take place during the work-up increases the level of 8-oxodGuo but not of the (18)O-oxidized nucleoside. Therefore, the ratio between the two compounds, as measured by high performance liquid chromatography coupled to tandem mass spectrometry, allows an unambiguous comparison of different methodologies. The comparison of different DNA extraction protocols led to the conclusion that artefactual DNA oxidation during the extraction step could be minimized if: (i) nuclei are isolated after cell lysis; (ii) desferrioxamine, a transition metal chelator is added to the different extraction buffers; and (iii) sodium iodide (or alternatively guanidine thiocyanate) is used for DNA precipitation. It was also demonstrated that sodium iodide does not decompose the targeted oxidized purine nucleoside. In addition, three different DNA digestion protocols were evaluated and they were found to give rise to similar results. Using the best-studied protocol, the steady-state cellular background level of 8-oxodGuo, in a lymphocyte cell line, was determined to be approximately 0.5 lesions/10(6) DNA nucleosides.

Two analytical methods, one involving the combined use of reverse-phase HPLC and electrochemical detection (HPLC-EC) and one involving a mass spectrometric detection after gas chromatography separation (GC/MS), were developed for the detection of 8-oxoguanine in DNA. In order to obtain quantitative results, 2,6-diamino-8-oxopurine, whose chemical structure and electrochemical response are very similar to 8-oxoguanine, has been employed as an internal standard in the HPLC-EC assay. In the case of the GC/MS method, an isotopically stable (M + 4) 8-oxoguanine has been employed as an internal standard. Both methods are able to detect approximately 1 modification per 10(6) DNA bases. The background level of 8-oxoguanine in DNA as determined by GC/MS is approximately 50-fold higher than that determined by the HPLC-EC assay. The discrepancy between the two methods is due to an artifactual oxidation of guanine during the derivatization reaction as demonstrated by using pure guanine. The amount of 8-oxoguanine in guanine, determined by GC/MS, increases linearly with the time of derivatization, indicating that an oxidation occurs during the silylation reaction. Derivatization under nitrogen atmosphere reduces but does not suppress the artifactual oxidation. The amount of 8-oxoguanine in DNA, quantified by GC/MS, is comparable to that obtained by HPLC-EC when 8-oxoguanine is prepurified by HPLC or by immunoaffinity chromatography, prior to the silylation reaction. The artifactual formation of 8-oxoguanine during the derivatization reaction may explain, at least in part, why the values reported for 8-oxoguanine determination by GC/MS are generally about 1 order of magnitude higher than that determined by HPLC-EC. Prepurification of 8-oxoguanine from guanine is recommended in order to obtain reliable results by GC/MS which may be compared to HPLC-EC.