Electrospray Mass

Anal. Chem. 1997, 69, 1952-1955

High-Performance Liquid Chromatography/ Electrospray Mass Spectrometry for the Analysis of Modified Bases in DNA: 7-(2-Hydroxyethyl)Guanine, the Major Ethylene Oxide-DNA Adduct Laurent Leclercq,† Christian Laurent,† and Edwin De Pauw*,‡

Laboratoire d’Oncologie, Radiobiologie et Mutagene` se Expe´ rimentales (ORME), CHU B23, and Laboratoire de Spectrome´ trie de Masse, Institut de Chimie B6, Universite´ de Lie` ge, B4000 Liege, Belgium

A method was developed for the analysis of 7-(2-hydroxyethyl)guanine (7HEG), the major DNA adduct formed after exposure to ethylene oxide (EO). The method is based on DNA neutral thermal hydrolysis, adduct microconcentration, and final characterization and quantification by HPLC coupled to single-ion monitoring electrospray mass spectrometry (HPLC/SIR-ESMS). The method was found to be selective, sensitive, and easy to handle with no need for enzymatic digestion or previous sample derivatization. Detection limit was found to be close to 1 fmol of adduct injected (10-10 M), thus allowing the detection of approximately three modified bases on 108 intact nucleotides in blood sample analysis. Quantification results are shown for 7HEG after calf thymus DNA and blood exposure to various doses of EO, in both cases obtaining clear dose-response relationships. DNA adducts formed by the covalent attachment of a large number of xenobiotics to DNA have been more and more studied in the past 20 years because they are suspected to play an important role in the appearance of cancer.1 The identification of the carcinogenic agent (or its metabolite) and the nature of the lesion it induces is essential for the prediction of mutagenic and carcinogenic processes.2 There has thus been more and more interest in the development of analytical methods able to measure DNA modifications at the base or nucleoside level (most of the xenobiotics interact with nucleophilic centers of purine and pyrimidine DNA bases). The analytical challenge in the study of DNA adducts is the validation of a sensitive and selective method which could allow the measurement of at least one modification in 103-106 nucleotides within a very low quantity of DNA (a few micrograms). The method should also provide information on the molecular nature of the adduct because of its further potential mutagenic properties. Highly sensitive methods (32P-post labeling,3,4 immunoassays,5 etc.) allow detection of very small amounts †

Laboratoire ORME. Laboratoire de Spectrome´trie de Masse. (1) Hemminki, K.; Dipple, A.; Shuker, D. E. G.; et al. IARC Sci. Publ. 1994, 125. (2) Newbold, R. F.; Warren, W.; Medcalf, A. S. C.; et al. Nature 1980, 283, 596-599. (3) Kumar, R.; Staffas, J.; Fo¨rsti, A.; et al. Carcinogenesis 1995, 16, 483-489. (4) Szyfter, K.; Kru ¨ ger, J.; Ericsson, P.; et al. Mutat. Res. 1994, 313, 269-276. (5) Hsu, I. C.; Poirier, M.; Yuspa, S.; et al. Cancer Res. 1981, 41, 1091-1095. ‡

1952 Analytical Chemistry, Vol. 69, No. 10, May 15, 1997

of adducts (one modification in 108-1010 nucleotides) but do not provide any structural information on the nature of unknown adducts. Mass spectrometry, despite its lower sensitivity compared to the previously mentioned methods, offers molecular weight identification and sometimes structural specificity.6 The use of MS for the characterization of DNA adducts has followed technical evolutions; early derivatization (permethylation or persilylation7,8) or field desorption methods have been replaced by techniques like fast atom bombardment,9-11 matrix-assisted laser desorption,12 and derivatization using electrophoric groups.13-16 Electrospray ionization and, more specifically, its interface with liquid chromatography or capillary electrophoresis provides the possibility of both ionization and liquid-to-gas transition for a wide variety of thermally labile compounds, including many biomolecules. Electrospray mass spectrometry (ESMS)17 has been used in the study of normal or chemically modified synthetic oligonucleotides18,19 and also for the analysis of unmodified nucleotides and nucleosides.20,21 Characterization of DNA adducts has been carried out recently at the nucleoside level by coupling electrospray to HPLC (after DNA exposure at a high single dose of (6) Chiarelli, M. P.; Lay, J. O. Mass Spectrom. Rev. 1992, 11, 447-493. (7) Von Minden, D. L.; McCloskey, J. A. J. Am. Chem. Soc. 1973, 95, 74807490. (8) Ashworth, D. J.; Baird, W. M.; Chang, C.; et al. Biomed. Mass Spectrom. 1985, 12, 309-318. (9) Bryant, M. S.; Lay, J. O.; Chiarelli, M. P. J. Am. Soc. Mass Spectrom. 1992, 3, 360-371. (10) Wolf, S. M.; Vouros, P. Anal. Chem. 1995, 67, 891-900. (11) De Pauw, E.; Marafante, E.; Riego, J.; et al. Nucleosides Nucleotides 1990, 9, 361-364. (12) Stemmler, E. A.; Buchanan, M. V.; Hurst, G. B.; et al. Anal. Chem. 1994, 66, 1274-1285. (13) Saha, M.; Kresbach, M.; Giese, R. W.; et al. Biomed. Environ. Mass Spectrom. 1989, 18, 958-972. (14) Allam, K.; Abdel-Baky, S.; Giese, R. W. Anal. Chem. 1993, 65, 1723-1727. (15) Chiu, C. S.; Saha, M.; Abushamaa, A.; et al. Anal. Chem. 1993, 65, 30713075. (16) Saha, M.; Abushamaa, A.; Giese, R. W. J. Chromatogr. A 1995, 712, 345354. (17) Smith, R. D.; Loo, J. A.; Edmonds, C. G.; et al. Anal. Chem. 1990, 62, 882899. (18) Pomerantz, S. C.; Kowalak, J. A.; McCloskey, J. A. J. Am. Soc. Mass Spectrom. 1993, 4, 204-209. (19) McLuckey, S. A.; Habibi-Goudarzi, S. J. Am. Soc. Mass Spectrom. 1994, 5, 740-747. (20) Banks, J. F.; Shen, S.; Whitehouse, C. M.; et al. Anal. Chem. 1994, 66, 406-414. (21) Wampler, F. M.; Blades, A. T.; Kebarle, P. J. Am. Soc. Mass Spectrom. 1993, 4, 289-295. S0003-2700(96)00767-6 CCC: $14.00

© 1997 American Chemical Society

xenobiotic)22 and also at the nucleotide level by coupling ESMS to capillary electrophoresis and analyzing standard adduct samples,23,24 with a detection limit of approximately 3 × 10-9 M using selected ion recording (SIR). In this work, we used electrospray ionization mass spectrometry coupled to reversedphase HPLC with high flow rate (100 µL/min) and classical mobile phase composition (water/methanol 50:50) for the characterization and the dose-dependent formation of a DNA adduct at the nucleic base level, 7-(2-hydroxyethyl)guanine (7HEG), the major adduct formed between DNA and ethylene oxide (EO). EO is a widely used sterilant and a chemical intermediate in the polymer industry. It is known to be genotoxic and carcinogenic in rodents and probably in human.25,26 EO interacts with DNA without previous metabolic activation by an SN2 reaction mechanism.27 Most of the hydroxyethylation (close to 95%) is focused on the N7 position of guanine due to its high electron density compared to other nucleophilic centers. The nucleoside and nucleotide forms of the N7-guanine adduct formed between DNA and EO are poorly stable,28 while the adduct at the base level is stable. EO-DNA adduct formation was followed in animal studies,29 and a background level of 7HEG was detected in smokers.30 We decided to study the rate of formation of 7HEG after in vitro exposures of calf thymus DNA solutions and human blood (followed by lymphocytes isolation). A large range of EO concentrations (from 5000 to 5 ppm EO) was investigated. EXPERIMENTAL SECTION Chemicals. Guanine and adenine bases were obtained from Sigma (St. Louis, MO). 7-(2-hydroxyethyl)guanine (7HEG) and O6-(2-hydroxyethyl)guanine (6HEG) were obtained from Campro Scientific Laboratories (The Netherlands). All solvents used were of analytical grade. DNA Exposure to EO. Calf thymus DNA dissolved in deionized water (1 mg/mL) was incubated with various doses of EO for 3 h at 37 °C. DNA was precipitated with 2 volumes of absolute ethanol, washed twice with 70% ethanol and absolute ethanol, dried under vacuum (Speed Vac), and dissolved overnight in 2 mL of deionized water. DNA concentration was determined by UV absorbance (assuming A260 ) 1 at 50 µg/mL). Human Blood Exposure to EO. Blood uptakes (10 mL) were incubated with various doses of EO for 3 h at 37 °C. Lymphocytes were isolated from blood (LeucoPREP) and washed twice with PBS (diluted 10 times) and 0.9% NaCl. DNA (approximately 200 µg) was extracted from mononuclear cells using guanidine isothiocyanate (Gibco, Gaithersburg, MD), precipitated by addition of 0.5 volume of absolute ethanol, washed twice with 70% ethanol and absolute ethanol, dried under vacuum (Speed Vac), and redissolved overnight in 200 µL of deionized water. DNA concentration was determined as described for DNA exposures. (22) Lemie`re, F.; Joos, P.; Vanhoutte, K.; et al. J. Am. Soc. Mass Spectrom. 1996, 7, 682-691. (23) Barry, J. P.; Norwood, C.; Vouros, P. Anal. Chem. 1996, 68, 1432-1438. (24) Deforce, D. L. D.; Ryniers, F. P. K.; Van den Eeckhout, E. G.; et al. Anal. Chem. 1996, 68, 3575-3584. (25) Ehrenberg, L.; Hussain, S. Mutat. Res. 1981, 86, 1-113. (26) Kolman, A.; Na¨slund, M.; Calleman, C. J. Carcinogenesis 1986, 7, 12451250. (27) Li, F.; Segal, A.; Solomon, J. J. Chem.-Biol. Interact. 1992, 83, 35-54. (28) Mu ¨ ller, N.; Eisenbrand, G. Chem. Biol. Interact. 1985, 53, 173-181. (29) Walker, V. E.; Fennell, T. R.; Boucheron, J. A.; et al. Mutat. Res. 1990, 233, 151-164. (30) Fo¨st, U.; Marczynski, B.; Kasemann, R.; et al. Arch. Toxicol. Suppl. 1989, 13, 250-253.

Figure 1. ESMS full-scan spectrum of 7HEG.

7HEG Isolation. DNA was submitted to neutral thermal hydrolysis (100 °C, 15 min) in order to selectively release N7guanine adducts. The solution was microconcentrated (3000 rpm, 20 min) using Centricon microconcentration filters (cutoff mass ) 30 000 Da) (Amicon, Beverly, MA). The resulting solution was injected onto HPLC for 7HEG quantification. Instrumentation. The mass spectrometer was a VG Platform (Fisons Instruments, Beverly, MA). Ion source temperature was set at 70 °C. Nitrogen was used as nebulizing gas. The 7HEG full-scan spectrum was taken at low cone voltage (20 V) in the positive ion mode (ES+) by scanning the ions in the mass range from m/z ) 50 to 500. High-performance liquid chromatography was performed with a Supelcosil C18 reversedphase column (stationary phase ODS deactivated base, internal diameter 2.1 mm (narrow-bore type), length 25 cm, particle size 5 µm). The HPLC pump was a Rheodyne type (Supelco, Bellefonte, PA) (two pumps acting out of phase to reduce pulse effect, each pump delivering 50% of the eluant). The mobile phase we used was a mixture of water and methanol (50:50 v/v). This phase was the best compromise we found between adduct solubility, surface tension, and conductivity problems in electrospray ionization. The flow rate was set at 100 µL/min. The injection volume for the analyzed samples was 10 µL. Amounts of 7HEG were quantified by comparison of SIR peak integration values with those of standard samples. RESULTS AND DISCUSSION The purpose of this study was to set up a selective and sensitive method for the detection of 7HEG after human exposure to ethylene oxide. As a preliminary step before the analysis of biological samples like DNA or blood, first experiments had to be carried out with standard samples to investigate the potential of the method. Purine bases (guanine and adenine) and standard EO-DNA adducts (7HEG and 6HEG) were injected in large amounts (1 nmol) directly in the electrospray ion source and their respective full-scan spectrum were recorded. Figure 1 shows the ESMS full-scan spectrum for 7HEG recorded in the positive ion mode at low cone voltage. The spectrum is characterized by a predominant protonated molecular ion with no intense fragmentation. The mass spectral behavior of unmodified purines (guanine and adenine) and 6HEG was comparable to that of 7HEG. Attempts were made to fragment the adducts at higher cone voltage values, but no significant fragment ions were formed while sensitivity was decreasing rapidly. Negative ion mode detection Analytical Chemistry, Vol. 69, No. 10, May 15, 1997

1953

Figure 3. (A) HPLC/SIR-ESMS chromatogram of 1 f/mol of 7HEG (SIR of m/z ) 196). (B) Logarithmic calibration curve for 7HEG.

Figure 2. (A) HPLC/SIR-ESMS chromatogram of a mixture of guanine, 7HEG, and adenine with simultaneous SIR of m/z ) 136, 152, and 196. (B) HPLC/SIR-ESMS chromatogram of a mixture of guanine, 7HEG, and adenine with SIR of m/z ) 196. (C) HPLC/SIRESMS chromatogram of a mixture of 7HEG and 6HEG with SIR of m/z ) 196.

was found to be less sensitive than positive ion mode detection. We decided to monitor each base or adduct by recording its protonated molecular ion signal (SIR). Figure 2A shows an HPLC/ESMS chromatogram of a mixture of guanine (SIR of m/z ) 152), adenine (SIR of m/z ) 136), and 7HEG (SIR of m/z ) 196), while Figure 2B shows an HPLC/SIR-ESMS chromatogram of the same mixture with SIR of 7HEG molecular ion only (m/z ) 196). Focusing on one particular ion (7HEG molecular ion in this case) provides a great enhancement in both selectivity and sensitivity and allows us to isolate a compound of interest from undesired compounds in a complex mixture (as is the case after DNA hydrolysis, see below). The possible limitation of monitoring 1954 Analytical Chemistry, Vol. 69, No. 10, May 15, 1997

only one ion (involving possible false positive or even negative results) was considerably reduced here by the purification step prior to HPLC/ESMS injection (see below) and by the invariance in the retention time of 7HEG on the LC column when analyzing 7HEG after DNA and blood in vitro exposures to EO. Figure 2C shows an HPLC/ESMS chromatogram for a mixture of 7HEG and 6HEG (SIR of m/z ) 196), showing a clear separation between these two isomers. Coupled to an unambiguous identification and isolation of the adduct, trace detection level is the major problem of interest in the study of DNA adducts (see the introductory paragraph); we recorded HPLC/ESMS chromatograms for various quantities of 7HEG using SIR of m/z ) 196 and proceeded to SIR signal integrations. Good linear correlation was observed from 1000 pmol of 7HEG to up to 1 fmol of 7HEG injected, this latter amount corresponding to a concentration of 10-10 mol/L (Figure 3A). Signal-to-noise ratio for SIR analysis of 1 fmol of 7HEG was approximately 2:1 (figure 3B), indicating this concentration to be the limit of detection of the method. HPLC/SIR-ESMS was thus found to be more or at least as sensitive as HPLC with EC detection, FAB, or GC/MS with derivatization using electrophoric groups. 7HEG SIR peak integration values sometimes varied 20% from day to day in the low picomole range, while integration values for higher 7HEG amounts were found to be constant. This reproducibility problem obliged us to proceed to complete calibration before every mass spectrometry session. We are investigating the use of internal standards to avoid this type of inconvenience. In our in vitro experiments conducted to quantitate 7HEG after

Figure 4. Logarithmic dose-response curve showing 7HEG formation vs EO dose after in vitro calf thymus DNA exposure to EO.

DNA and blood exposure to EO, the first step after DNA isolation was the release of 7HEG from the DNA backbone. Classical methods used in the isolation of DNA units are enzymatic digestion (going to the nucleoside or nucleotide level) and DNA neutral or acid hydrolysis (going to the base level). DNA neutral thermal hydrolysis16,28,30 was found to be the most appropriate isolation method in our case because it permits selective release of N7-guanine adducts by simple heating of the DNA solution, with only a low rate of release for unmodified guanine and adenine (1 or 2%). Further adduct isolation was made by microconcentration using Centricon cutoff filters (cutoff mass ) 30 000 Da). NTH followed by microconcentration was seen to reduce the rate of contamination and the usual purification steps by TLC or HPLC. Although NTH is only useful in N7 adduct isolation, it is appropriate for a large number of alkylating agents whose electrophilic properties often focus their action on this highly nucleophilic position in DNA. Figure 4 shows results of 7HEG quantification after in vitro DNA exposure to various doses of EO. A clear dose-response relationship was observed in the whole dose range, with a decrease in linearity at low EO doses (from 25 to 5 ppm EO). A level of hydroxyethylation of approximately six modified guanines on 107 intact nucleotides was obtained for an EO concentration of 10-4 mol/L. Release of free guanine bases during NTH was calculated by integration of guanine molecular ion SIR signal (see Figure 3). The rate of free purine contamination (approximately 2%) was in good agreement with that determined by Van Delft et al. by HPLC with EC detection.31 Variations in time of hydrolysis showed no further increase in 7HEG after 15 min. Blood incubation results are shown in Figure 5 for different EO doses; a clear dose-response relationship was (31) Van Delft, J. H. M.; Van Weert, E. J. M.; Van Winden, M. J. M.; et al. Chem.Biol. Interact. 1991, 80, 281-289. (32) Tates, A. D.; Grummt, T.; To ¨rnqvist, M.; et al. Mutat. Res. 1991, 250, 483497.

Figure 5. Dose-response curve showing 7HEG formation vs EO dose after in vitro blood exposure to EO.

established as for DNA incubation. Lower amounts of 7HEG were observed for equivalent EO doses due to the lower DNA amounts in blood compared to DNA solutions and to the presence of “biological barriers” and detoxification mechanisms in biological medium. A rate of hydroxyethylation of approximately four 7HEGs on 107 nucleotides was found for blood exposure to 10-3 M EO. CONCLUSIONS AND PERSPECTIVES HPLC/ESMS on a benchtop spectrometer was found to be an interesting alternative for the routine quantitative analysis of 7HEG after DNA and blood exposure to EO. Combination of DNA neutral thermal hydrolysis, adduct microconcentration, and HPLC/ ESMS using SIR of the molecular ion provided a rapid, clean, selective, and sensitive method. Electrospray ability to identify DNA adducts at the base level opens the way for the analysis of a large number of alkyl DNA N7-guanine adducts that are often poorly stable in the nucleoside or nucleotide form. Furthermore, it was found appropriate for both standard and biological samples exposed to very low doses of xenobiotics. Attempts are now under study to lower the detection limit of the method, including sample concentration techniques and tandem mass spectrometry using constant neutral loss. 7HEG molecular ion fragmentation in MS/ MS could also avoid the inconvenience of monitoring only one ion in SIR. The gain in sensitivity obtained by the inherent reduction of the “chemical noise” may be compensated by the loss of signal due to collisional activation energy. With the abovementioned increase in sensitivity, the method could become competitive for further biomonitoring of occupationally exposed populations.32 Received for review July 30, 1996. Accepted January 16, 1997.X AC9607673 X

Abstract published in Advance ACS Abstracts, March 15, 1997.

Analytical Chemistry, Vol. 69, No. 10, May 15, 1997

1955