JOURNAL OF MASS SPECTROMETRY, VOL. 31, 1271-1276 (1996) Quantitative Analysis of the DNA Adduct w,3Ethenoguanine Using Liquid Chromatography/ Electrospray Ionization Mass Spectrometry Ten-Yang Yen,? Nadia I. Christova-Gueoguieva, Nova Scheller, Sharon Holt, James A. Swenberg and M. Judith Charlest Department of Environmental Sciences and Engineering and Curriculum in Toxicology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7400, USA The need for specificity and sensitivity in the analysis of DNA adducts has led the development of GC/MS methods. Such methods require chemical derivatization (i.e. silylation, electrophore labelling), which can also bring its own sets of problems, including the production of artifacts, interferences and sample to sample variability in derivatization. To obviate such problems, a liquid chromatographic/electrospray ionization mass spectrometric (LC/ESI-MS) method was developed to quantify N2,3-ethenoguanine (eGua), a promutagenic DNA adduct of vinyl chloride exposure. The response of eGua to isotopically labelled internal standard [ "CC,]eGua was linear (us = 0.999) and reproducible from 0.027 to 0.538 pmol PI-'. We obtained an accuracy of 86 f 14% by analyzing chloroethylene oxide (CEO)-treated calf thymus DNA enriched with authentic eGua. The analysis of CEO-treated calf thymus DNA samples not enriched with authentic eGua provided a precision of 15%. The detection limits with a signal-to-noise ratio (S/N) 2.5:l were obtained in the determination of authentic eGua at 5 fmol per injection. The detection limit obtained in the routine analysis of the biological samples was 50 fmol eGua with S/N = 3: 1. The applicability of the method was established by determining ~ G u ain rats treated with CEO by portal vein injection and an unexposed human liver. It was observed that the concentration of EGua in the rat livers increased with increase in dose and was inversely related to the time after, CEO exposure. This trend suggests rapid repair of the adduct in rat livers. In the human liver DNA sample, sGua was quantitated at 0.06 f 0.01 pmol mg-' DNA. KEYWORDS: liquid chromatography/electrospray ionization mass spectrometry; DNA adduct; N2,3-ethenoguanine; vinyl chloride INTRODUCTION The formation of DNA adducts due to the electrophilic interactions between chemical carcinogens or their metabolites and DNA can be the initial step of chemical carcinogenesis. The molecular dose of such adducts represents a more accurate determinant of potential risk of exposure to carcinogens than those of external exposure such as air, water and diet. Several analytical techniques, such as high-performance liquid chromatography (HPLC) combined with fluorescence or electrochemical detection, 32P post-labeling and gas chromatography/mass spectrometry (GC/MS), have been employed to quantitate DNA adducts.' The greater need for specificity and sensitivity has led to the development of GC/MS method^.^-'^ Such methods require chemical derivatization (i.e. silylation, electrophore labelling), which introduces its own sets of problems. Different reagents have different derivatizing efficiencies and affinities for targeted analytes. For DNA adducts that are very polar, the derivatization is often difficult and less successful. Also, interferences and artifacts can arise from the derivatization process that can lead to difficulties in identifying and quantifying the anal~te.'~-'~ Direct analysis of DNA adducts by using liquid chromatography combined with electrospray ionization mass spectrometry (LC/ESI-MS) can obviate those problems arising from the employment of chemical deriv a t i z a t i ~ n . ~ ' -In ~ ~ this paper, we present data to demonstrate that DNA adducts in biological samples can be quantified precisely and accurately by using LC/ ESI-MS. N2,3-Ethenoguanine (EGua), a highly promutagenic adduct generated by vinyl chloride (a known human carcinogen) exposure,21was selected as a model for DNA adducts because the procedure for its isolation was well characterized in our l a b ~ r a t o r y Our . ~ findings are relevant to the analysis of other DNA adducts that are generated exogenously or endogenously. EXPERIMENTAL Materials t Author to whom correspondence should be addressed $ Present address: Department of Environmental Toxicology, University of California, Davis, CA 95616, USA. CCC 0030-493X/96/111271-06 0 1996 by John Wiley & Sons, Ltd. HPLC-grade water and methanol were obtained from Fisher Scientific (Fair Lawn, NJ, USA) and Mallinckrodt (Paris, KY, USA), respectively. Calf thymus DNA Received 30 June 1996 Accepted 23 July 1996 1272 T.-Y. YEN ET AL. was purchased from Sigma (St Louis, MO, USA). Other chemicals utilized were of analytical grade and were obtained from Fisher Scientific. The Amberlite IR 120 utilized in strong cation-exchange columns was acquired from Serva Feinbiochemica (Westbury, NJ, USA). Octadecyl (C,,, 40 pm particle size) used for solid-phase extraction columns was obtained from J. T. Baker (Phillipsburg, NJ, USA). All glassware used in sample preparation was silanized. N2,3-Ethenoguanine (EGua) and ['3C4]-N2,3ethenoguanine (['3C4]~Gua) were synthesized according to protocol described elsewhere.3722 We determined the residue of unlabeled EGua in the ['3C4]~Guato be < 2% by conducting an LC/ESI-MS experiment in which the [M -tH I + ion of EGua and [ ' 3 C 4 ] ~ G ~were a selectively ion monitored. Preparation of in vitro and in vivo samples Caution: Chloroethylene oxide (CEO) is hazardous and should be handled with protective clothing in a well ventilated hood. CEO-treated calf thymus DNA was prepared according to the protocol developed by G u e n g e r i ~ hBriefly, .~~ 25 mg of calf thymus DNA (2.5 mg ml-'), was incubated with 210 pmol of CEO in 0.1 M potassium phosphate buffer (pH 7.4) at 25°C for 20 min. The treated calf thymus DNA was then precipitated with 50 ml of cold ethanol, centrifuged at 3000 g for 10 min, dried under a stream of nitrogen, dissolved in HPLC-grade water and stored at 4°C. Aliquots (20-50 pg) of this CEO-treated calf-thymus DNA were taken for mild acid hydrolysis. Male Sprague-Dawley rats (0.4-0.8 kg) from Charles River Laboratories (Raleigh, NC, USA) were employed as control and exposed animals. The exposed animals were killed at 2, 4 or 6 h after exposure to 0.07 or 0.29 pmol of chloroethylene oxide per gram body mass by portal vein injection. Control rats were killed at the same time invervals during the experiment. Livers were removed, frozen and stored at - 80 "C. The human liver sample (H 116) was obtained from the Tennessee Donor Service and stored at - 80 "C. DNA was extracted from the livers by using an ABI 304A nucleic acid extractor (Applied Biosystems, Foster City, CA, USA). After two phenol-chloroform extractions and one chloroform extraction, DNA was precipitated by using sodium acetate and propan-2-01.~ The yields and purity of DNA were determined by measuring the UV absorption at 230, 260 and 280 nm. Aliquots of the rat liver DNA (1.88-4.42 mg) and the human liver DNA (7.4 mg) were prepared for mild acid hydrolysis. Extraction, isolation and enrichment of N2,3-et henoguanine Prior to the DNA mild acid hydrolysis, 4 pmol of ['3C4]eGua were added to each sample. The DNA was hydrolyzed in 0.2 M HC1 for 90 min at 80°C. EGua and ['3C,]~Gua, was isolated by applying each of the hydrolyzates to individual strong cation-exchange columns packed with Amberlite IR 120 AS 3545 resin.3 NaCl (0.25 M) was used as mobile phase. The fraction from 5 to 10 ml of the eluate was collected, applied to a C,, solid-phase extraction column, followed by 5 ml of water to remove NaCl (all materials for these lowpressure chromatographic separations were prepared in disposable glass Pasteur pipets (5.75 in) plugged with glass-wool). EGua and [13C4]~Guawere eluted from the C1, column with 4 ml of methanol. The solvent was evaporated and the dry residue was dissolved in 30 p1 of water for LC/ESI-MS analyses. LC/ESI-MS analyses LC,ESI-MS analyses were conducted using a Beckman Gold liquid chromatographic system (Beckman Instruments, Arlington Heights, IL, USA) coupled to a Finnigan 4000 quadrupole mass spectrometer. The mass spectrometer was retrofitted with a pneumatic electrospray source (Analytica of Branford, Branford, CT, USA). The solvent (water and/or methanol) was passed through a mixing tee. One end of the tee was connected to a pressure-balance column C,,, 150 x 4.6 mm id.), while the other end was connected to a Cl, capillary column (150 x 0.3 or 0.8 mm id.; Hypersil, 3 pm particles size) (LC Packings, San Francisco, CA, USA). The flow rate was 4 p1 min-' for the 0.3 mm i.d. column and 25 pl min-' for the 0.8 mm id. column. A piece of fused-silica capillary (30 cm x 50 pm i.d. x 375 pm 0.d.) directed the eluent of the capillary column to the electrospray needle. The sample injection volume was controlled by a Rheodyne Model 77253 injector (20 pl external loop) or a Valco CI4WS injector (0.5 pl internal loop). Chromatographic separations were accomplished by increasing the mobile phase from 100% water to 25% water-75% methanol in 18 min, altering the phase to 5% water-95% methanol in 4 min and passing 100% methanol through the column for 1 min. A voltage of 3.6 kV was applied to the electrospray needle and 70 psi of nebulizer gas (nitrogen) was employed to stabilize the spray. The voltage difference between the exit of the glass capillary and the first skimmer in the differential pumping region was optimized at 130 V for detection of the [M + H]+ ion of EGua (m/z = 176). Data were acquired and processed by a Technivent Vector data system (Teknivent Maryland Heights, MO, USA). Full-scan data were obtained by scanning from m/z 20 to 350 in 1 s. For selected-ion monitoring (SIM) analyses, two ions (m/z 176 and 180) were monitored with a dwell time of 0.7 s. Quantification of EGua was accomplished by measuring the peak areas of the [M + H I + ion of EGua (m/z 176) and the [M + H]+ ion of [13C4]eGua (m/z 180) in standard solutions containing 27, 54, 135, 269 or 538 fmol pl-1 of eGua and 500 fmol pl-1 of [13C4]~Gua. The calibration curve was constructed by plotting the response of the ratio of EGua to [13C4]sGua vs. the concentration of eGua. The equation of the calibration curve was determined by linear regressions (leastsquares method) of the data. During sample analyses, the reproducibility of the data was verified by analyzing a standard solution after every five samples. The relative difference throughout the day among these standards was less than 15%. For every eight samples, one 1273 QUANTITATION OF DNA ADDUCTS USING LC/ESI-MS rapid decrease in the ion signal with a small fraction of fragment ions [see Fig. l(c)]. In previous work, we investigated the factors that affect the electrospray response of nucleobases to determine how to optimize ESI-MS methods for the analysis of DNA ad duct^.^^ We found that for compounds with pK, d 4, the sample solution acidified with weak acid to promote the formation of protonated ions in solution can increase the ion signal less than threefold. However, the LC/ESI-MS ion signal for eGua in a watermethanol solution acidified with 0.1% formic acid was three times weaker than the response in watermethanol that was not acidified. Therefore, we did not utilize any modifiers in the mobile phase to separate eGua from other components in the sample extract. Because the electrospray response is dependent on the analyte concentration,2s*26it is desirable to employ small internal diameters of liquid chromatographic method blank containing 4 pmol of [13C,]~Gua was performed to assure the accuracy of measurement. The EGua signal was not detectable in the method blank. RESULTS Optimization of LC/ESI-MS of EGua The base peak of the LC/ESI mass spectrum for EGua is the [M + H I + ion at m/z 176 with a retention time of 10.5 min [see Fig. l(a) and l(b)]. We hoped to dissociate these ions collisionally to gain structural information about the adduct, but attempts to do so by increasing the voltage between the skimmer and the exit of the glass capillary regions from 130 to 200 V resulted in a 140oooO0 .nlr 12000000 - 176 A 1 ooO0ooo 8000000 a" 600oooO 400oooO 0 1 3 2 4 5 7 8 9 10 11 12 13 1 4 IS 10 0 SO 70 SO 110 130 lS0 170 190 210 230 250 SO 1 7 6 t3U+- I $j (M+H)'-a-1= 135 (M+H)'-a-b= 94 a 2o 10 135 0 , 70 90 110 130 110 170 190 110 230 210 d z Figure 1. Determination of EGua (3.8 pmol): (a) reconstructed ion chromatogram at m/z 176; (b) background-subtracted mass spectrum of EGua; (c) collision-induced dissociation mass spectrum of EGua. The relative intensity of the ion signal is normalized to the response of the analyte shown in (A). 1274 T.-Y. YEN ET AL. columns with low flow rates to increase the analyte peak concentration and the analyte electrospray r e s p o n ~ e . ~We ~ - ~compared ~ the response of EGua obtained by using C I Bcapillary columns (150 x 0.3 mm i.d. and 150 x 0.8 mm id.). With 25 fmol of EGua injected, S/N = 20: 1 was obtained by using the 0.3 mm id. column compared with S/N = 4 : 1 by using the 0.8 mm i.d. column. Although, as expected, we achieved greater sensitivity when using the 0.3 mm i d . column, we discovered two disadvantages. First, we observed 5-50 fmol of EGua in a solvent blank analyzed after the analysis of standards containing > 3 pmol of EGua. This carry-over of EGua was eliminated by washing the column with 95% MeOH-5% H,O for 1 h or more. Second, we observed a decrease in signal when we injected 3 4 pl of a sample solution extract from CEOtreated calf thymus DNA. This decrease, probably due to the presence of other electrolytes in biological samples, did not occur when we injected 6 pl of the sample solution on to the 0.8 mm i.d. column. Moreover, EGua was not detected in solvent blanks analyzed after the determination of < 5 pmol of EGua. Therefore, the 0.8 mm i.d. column was considered to be more appropriate for the analysis of biological samples. Evaluation of the method by quantification of EGua in CEO-treated calf-thymus DNA We investigated the accuracy, precision and detection limits of the LC/ESI-MS method by analyzing the response of EGua to [13C,]~G~afrom standard solutions and samples. We present the response factors for EGua (Table 1) from the analysis of 27-538 fmol pl-' standards with 0.5 p1 on-column injection to construct a five-point calibration curve on five different days. The reproducibility of the analyses varied less than 15% (relative standard deviation), and the response of EGua to [13C4]sGua was linear with a mean r2 = 0.999 from linear regression analysis. We further evaluated this response by analyzing blank samples spiked with 8 p1 of 27-538 fmol pl-1 EGua standards that were carried through the sample preparation procedures. The data are summarized in Table 1. These results are similar to those obtained by direct analysis. We evaluated the precision of the method by analyzing replicate samples of CEO-treated calf-thymus DNA (n = 14), and we determined the accuracy of the method by analyzing matrix spikes (sample contained 20 pg of CEO-treated calf-thymus DNA enriched with 3 pmol of authentic EGua). A mean f standard deviation of 55.3 f 8.3 pmol EGua mg-' CEO-treated calf-thymus DNA was obtained for the analysis of 14 samples of CEO-treated calf-thymus DNA that provided a precision of 15% among the samples. A mean f standard deviation of 2.57 k 0.35 pmol (n = 3) was obtained for the average increase in EGua concentration for CEOtreated calf-thymus DNA enriched with 3 pmol of authentic EGua samples which provided an accuracy of 86 & 14%. We also conducted a more stringent test of the accuracy of the method by determining EGua from unexposed rat liver DNA (4 mg) spiked with 8 p1 of 27-538 fmol p1-l EGua. In Fig. 2, we present the data as the response factor of EGua to [13C4]eGua for these samples and for the data presented in Table 1. The data acquired from the analysis of rat liver DNA (open squares) are close to the calibration curve, demonstrating high accuracy for the analysis of spiked control rat liver DNA samples. The limit of detection for an authentic standard of EGua was 5 fmol using the 0.3 mm capillary column with S/N = 2.5: 1, and 15 fmol using the 0.8 mm capillary column with S/N = 3 : 1. Application of method to quantify EGua in CEO-treated rat livers and in an unexposed human liver Rats were treated with CEO via portal vein injection (0.07 or 0.29 pmol CEO per gram rat body mass). Two, four or six hours after exposure, the concentration of EGua was determined in the exposed rat liver DNA. The results are summarized in Table 2. The lowest concentration of EGua, 0.15 f 0.05 pmol mg-' DNA, was found in rat 4 treated with 0.07 pmol of CEO 6 h after exposure. The highest concentration of EGua 1 1.0 Table 1. Response factors of EGua to I''CC,J&Guaby LC/ESI-MS r2 = 0.999 PI 0.6 Response factorb Direct analysis of standard solutions EGua (pmol PI-')' 0.027 0.054 0.135 0.269 0.538 - (n 5 ) 0.07f 0.01 0.1 3 *0.01 0.29 0.01 0.55f 0.02 1.06f 0.03 * Standard solutions processed through sample preparation procedures (n 3) - 0.08f 0.01 0.15 f 0.01 0.32f 0.02 0.57f 0.03 1.08f 0.03 Each standard solution contains the same concentration of ''C.+Gua (0.5pmol PI-'). bThe response factor is the ratio of the m/z 176 area to the m/z 180 area, and the value given is the mean f the standard deviation of the number of replicates indicated. a 0.01 0.0 " 0.1 " 0.2 " 03 ' 0.4 [email protected]) ' I 0.5 . ' 0.6 Figure 2. Plot of the response factor of cGua in control rat liver DNA (4 mg) spiked with the standard solutions (open squares) vs. the concentration of cGua; the solid line is the calibration curve obtained from the direct analysis of standard solutions (open circles; data presented in Table 1 ), and the open triangles are for the analysis of standard solutions processed through the sample preparation procedures. 1275 QUANTITATION OF DNA ADDUCTS USING LC/ESI-MS Table 2. Concentrations of N2,3-ethenoguanine in biological samples Sample Oose of CEO (pmol g- ')' Time after exposure (h) 0.29 0.29 0.07 0.07 2 4 4 6 Rat liver DNA Human liver DNA Mean f SD Concentration (pmol mg-' DNA) 1.49*0.19( n = 8 ) 0.35k0.04 (n "4) 0.30f0.04 (n = 3) 0.15 f 0.05(n = 5) 0.06f 0.01 Detection limit (S/N- 3 : 1 ) by extrapolation (fmol) 77 62 36 36 42 51 *18 'Values are the amount for chloroethylene oxide per gram rat body mass administered by portal vein injection. bValue is the mean *standard deviation for three measurements of the sample extract. (1.49 f 0.19 pmol mg-' DNA) was found in rat 1 treated with 0.29 pmol of CEO 2 h after exposure. We also applied LC/ESI-MS to measure endogenous EGua present in human liver. The chromatograms obtained with SIM at m/z 176 (EGua) and 180 (['3C,]eGuaj are shown in Fig. 3. The concentration of ~ G u ain the human liver DNA sample was 0.06 f 0.01 pmol mg-' DNA (we employed 7.4 mg of DNA and obtained S/ N = 5: 1 on 70 fmol). If we extrapolate the data to obtain S/N = 3 : 1 on these measurements, we obtain a detection limit of about 50 fmol in biological samples. This limit of detection is about 50 times greater than the detection limit that can be routinely obtained by using " 5 6 7 8 9 10 11 12 13 14 1s 16 Tima (mill) ml7A 80 0 , , , , , 5 6 7 8 9 ( 10 , 11 , 12 I 13 14 15 16 (mill) Figure 3. Selected-ion monitoring chromatograms of human liver DNA analyzed for cGua and [lSC,]cGua. derivatization with trifluoromethyltetrafluorobenzyl bromide and gas chromatography/electron-capture negative chemical ionization high-resolution mass spectrometry (GC/ECNI-HRMS), the technique routinely employed in our laboratory to determine ~ G u a . ~ ' DISCUSSION The isolation and detection of DNA adducts (modified bases) from unmodified bases in biological samples is a great challenge. The concentration of EGua is about 0.1-2.5 per lo6 of guanine for in viuo samples; therefore, it is impossible to analyze DNA hydrolyzates directly by using capillary LC/ESI-MS without employing analyte isolation and enrichment procedures. The employment of a cation-exchange resin followed by C, solid-phase extraction method is able to separate the analyte and remove most of guanine and adenine from the DNA hydrolyzates. However, the sample solutions are still complex mixtures. This is evident by the suppression of the ion signal when > 4 p1 of solutions of extracts were injected on to a 0.3 mm capillary column. Because of this sample loading limitation and carryover problems associated with the usage of the 0.3 mm capillary column, we employed a 0.8 mm capillary column for routine sample analyses, even though the determination of cGua using the 0.3 mm capillary column provided greater sensitivity. Using eGua as a model compound, we demonstrated that LC/ESI-MS provides for accurate and precise quantification of the DNA adduct. The response of EGua to [13C,]~Gua is linear (r2 = 0.999) and reproducible from 27 to 538 fmol PI-'. We obtained an accuracy of 86 f 14% by analyzing CEO-treated calf thymus DNA enriched with EGua. The analysis of CEO-treated calf thymus DNA samples not enriched with EGua provided a precision of 15%. We also demonstrated the applicability of the LC/ ESI-MS method by determining EGua in livers of rats treated with CEO by portal vein injection. Using a small number of samples, we observed that the concentration of EGua in DNA from CEO-treated SpragueDawley rat livers increases with increasing dose of CEO, and decreases with time after exposure. Such a 1276 T.-Y.YEN ET AL. reduction in the concentration of EGua suggests rapid repair of EGua in rat livers. Additional support for this conclusion comes from evidence for urinary excretion of the related adduct 1,N6-ethenoadenine and the demonstration of 1,N6-ethenoadenine and EGua repair by glyc ~ s y l a s e . We ~ ' ~also ~ ~ succeeded in the quantification of endogenous EGua from a human liver DNA (7.4 mg) sample. The detection limit of this method for EGua in biological samples is about 50 fmol with S/N = 3 : 1. This is about 50 times less sensitive than the GC/ECNI-HRMS method employed in our laboratory. Further work is needed to increase the sensitivity. This may be accomplished by using immunoaffinity chromatography for sample enrichment to increase analyte recovery and reduce other complex mixtures and by sharpening the electrospray needle to enhance ionization efficiency.33934 Also, the amount of DNA required for the analysis could be reduced by using the column switching (LC/LC) method with injecting a greater proportion of the sample.35 The method as presented can be employed to determine EGua in vitro and to investigate dose-response relationships in animals at levels above the detection limit of our method. In our laboratory, we have used the method to investigate artifactual formation of EGua during chemical derivatization. While at this time the GC/ECNI-HRMS method is capable of detecting attomolar quantities of EGua, in cases where the sample size is not a limitation or the analyte is very polar, the LC/ESI-MS method may be the method of choice. Whichever method is chosen, good-quality assurance measures should be employed (e.g. the analysis of procedural blanks and matrix spikes, and confirmation of the identity of the compound by another method when one ion is monitored) to ensure accurate quantification of low levels of DNA adducts. Acknowledgements This research was supported by a grant from NIEHS, P42 ES05948. We thank Dr R.Sangaiah for the synthesis of chemical standards. We also thank Dr Craig Whitehouse for providing the electrospray ionization source. REFERENCES 1. H. Bartsch. K. Hemminki and I. K. O'Neill (Eds), Methods for Detecting DNA Damaging Agents in Humans: Application in Cancer Epidemiology and Prevention, IARC Scientific Publication No. 89. IARC, Lyon (1988). 2. P. F. Crain and J. A. McCloskey, Anal. Biochem. 132, 124 (1983). 3. N. Fedtke, J. A. Boucheron, V. E. Walker and J. A. Swenberg, Carcinogenesis 11, 1287 (1990). 4. N. Fedtke and J. A. Swenberg, in Molecular Dosimetry and Human Cancer: Analytical. Epidemiological and Social Con siderafion, edited by J. D. Groopman and P. L. Skipper, p. 171. CRC Press, Boca Raton, FL (1991). 5. M. P. Chiarelli and J. 0. Lay Jr, Mass Specfrom. Rev. 11, 447 (1 992). 6. J. A. McCloskey and P. F. Crain, lnf.J. Mass Specfrom. /on Processes 118/119, 593 ( 1992). 7. V. Prevost. D. E. G. Shuker. M. D. Friesen, G. Eberle, M. F. Rajewsky and H. Bartsch, Carcinogenesis 14, 199 (1993). 8. D. C. Malins. E. H. Holmes, N. L. Polissar and S. J., Cancer 71,3036 (1993). 9. M. Dizdaroglu, F€BS Left. 315, 1 (1993). 10. M. D. Friesen, K. Kaderlik, D. Lin, L. Garren, H. Bartsch, N. Lang and F. F. Kadlubar, Chem. Res. Toxicol. 7,733 (1994). 11. A. K. Chaudhary, M. Nokubo, L. J. Marnett and I. A. Blair, Biol. Mass Specfrom. 23, 457 (1994). 12. P. B. Farmer and G. M. A. Sweetman, J. Mass Specfrom. 30, 1369 (1995). 13. A. J. R. Teixeira, M. R. Ferreira, W. J. van Dijk, G. van de Werken and J. M. de Jong,Anal. Biochem. 226, 307 (1995). 14. J.-L. Ravanat, R. J. Turesky, E. Gremaud, L. J. Trudel and R. H. Stadler, Chem. Res. Toxicol. 8, 1039 (1995). 15. T. Douki, T. Delatour, F. Bianchini and J. Cadet, Carcinogenesis 17,347 (1996). 16. K. Allam, S. Abdel-Baky and R. W. Giese, Anal. Chem. 64,238 and 2882 (1992). 17. J. F. Banks Jr., S. Shen, C. M. Whitehouse and J. B. Fenn, Anal. Chem. 66,406 (1 994). 18. T.-Y. Yen, N. I. Christova-Gueorguieva, N. Scheller, J. A. Swenberg and M. J. Charles, in Proceedings of the 43rd ASMS Conference on Mass Specfromefry and Allied Topics. Atlanta, GA, 1995, p. 497. 19. D. Rindgen, R. J. Turesky and P. Vouros, Chem. Res. Toxicol. 8,1005 (1995). 20. A. K. Chaudhary, M. Nokubo, T. D. Oglesby, L. J. Marnett and 1. A. Blair, Biol. Mass Specfrom. 30, 1157 (1995). 21. B. Singer, J. T. Kusmierek, W. Folkman, F. Chave and M. K. Dosanjh, Carcinogenesis 12, 745 (1991 ). 22. N. Scheller, R. Sangaiah, A. Ranasinghe, A. Venkataraman, A. Gold and J. A. Swenberg, Chem. Res. Toxicol. 8,333 (1995). 23. F. P.Guengerich, Chem. Res. Toxicol. 5, 2 (1992). 24. T.-Y. Yen, R. D. Voyksner and M. J. Charles, in Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics,Atlanta, GA. 1995, p. 247. 25. J. 6. Fenn, J.Am. SOC.Mass Spectrom. 4, 524 (1993). 26. P. Kebarle and L. Tang.Ana1. Chem. 65, 972A (1993). 27. M. R. Emmett and R. M. Caprioli, J. Am. SOC. Mass. Specfrom. 5,605 (1994). 28. K. B. Tomer, M. A. Moseley, L. J. Deterding and C. E. Parker, Mass Specrrom. Rev. 13,431 (1994). 29. D. B. Kassel, B. Shushan, T. Sakuma and J.-P. Salzmann, Anal. Chem. 66,236 (1994). 30. A. Ranasinghe, N. Scheller, K. Y. Wu and J. A. Swenberg, in Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, 1995, p. 496. 31. S. Holt and J. A. Swenberg, Toxicologisf 30, 236 (1996). 32. M. K. Dosanjh, E. Chenna, E. Kim, H. Fraenkel-Conrat, L. Samson and B. Singer, Proc. Nafl. Acad. Sci. USA 91, 1024 (1994). 33. M. S. Wilm and M. Mann, Inr. J. Mass Specrrom. Ion Processes 136,167 (1994). 34. G.A. Valaskovic, N. L. Kelleher, D. P. Little, D. J. Aaserud and F. W. McLafferty,Anal. Chem. 67, 3802 (1995). 35. P. 0. Edlund, L. Bowers and J. D. Henion, J. Chromatogr. 529. 1 (1 989).