Analysis of Glyoxal-Induced DNA Cross-Links by Capillary Liquid

Jun 15, 2009 - linked products of glyoxal with 2′-deoxyribonucleosides have been characterized as dG-gx-dC, dG-gx-. dG, and dG-gx-dA. We herein deve...
0 downloads 0 Views 1MB Size
1334

Chem. Res. Toxicol. 2009, 22, 1334–1341

Analysis of Glyoxal-Induced DNA Cross-Links by Capillary Liquid Chromatography Nanospray Ionization Tandem Mass Spectrometry Hauh-Jyun Candy Chen* and Yu-Chin Chen Department of Chemistry and Biochemistry, National Chung Cheng UniVersity, 168 UniVersity Road, Ming-Hsiung, Chia-Yi 62142, Taiwan ReceiVed April 4, 2009

Glyoxal (gx) is an R-dicarbonyl species derived endogenously from the metabolism of carbohydrates or nitrosamines and from oxidation of lipids and nucleic acids. It is also widely distributed in foods and the environment. Glyoxal reacts with biomolecules, causing cross-links of proteins and DNA. The crosslinked products of glyoxal with 2′-deoxyribonucleosides have been characterized as dG-gx-dC, dG-gxdG, and dG-gx-dA. We herein develop a highly specific and sensitive capillary liquid chromatography nanospray ionization tandem mass spectrometry (capLC-NSI/MS/MS) assay for the simultaneous quantification of these three DNA cross-links using a triple-quadrupole mass spectrometer. The sample pretreatment procedures included enzyme hydrolysis of DNA and adduct enrichment by a reversed phase solid phase extraction column. We compared two enzyme hydrolysis conditions, and significantly different adduct levels were observed. This assay achieved attomole sensitivity with detection limits of 12–75 amol injecting each cross-link standard on-column. After calf thymus DNA was incubated with 1.0 mM of glyoxal at 37 °C for 30 days, the levels of dG-gx-dC, dG-gx-dG, and dG-gx-dA in this sample were determined as 6.52, 0.80, and 2.74 in 105 normal nucleotides, respectively, by capLC-NSI/MS/MS analysis after hydrolysis under optimized conditions. The identity of these cross-links in glyoxal-treated DNA was confirmed by MS2 and MS3 scan spectra using a linear ion trap mass spectrometer. In 20 µg of human placental DNA hydrolysate, the levels of dG-gx-dC, dG-gx-dG, and dG-gx-dA were quantified as 2.49, 1.26, and 3.50 in 108 normal nucleotides, respectively. These DNA cross-links, if not repaired, can be mutagenic, and they represent a type of damage to the integrity of DNA structure due to exposure of glyoxal. Introduction R-Dicarbonyl compounds, such as glyoxal (gx), methylglyoxal, and deoxyglucosones, are released from glycated proteins in the Maillard reaction (1). Among them, glyoxal is the most reactive species, and it is widely distributed in the environment and found in foods, beverages, and cigarette smoke (2-4). Glyoxal is produced from the metabolites of N-nitrosodiethanolamine, N-nitrosomorpholine, N-nitrosomethyethanolamine, and N-nitrosoethylethanolamine (5-7). It is also derived endogenously from the metabolism of carbohydrates and amino acids, and from oxidation of lipids and nucleic acids (8, 9). The concentrations of glyoxal are elevated in people with diabetes, peritoneal dialysis, and uremia (10-14). The reactions of glyoxal with the 2′-deoxyribonucleosides have been studied (15-21). The formation of a ring-closed adduct from 2′deoxyguanosine and glyoxal, 3-(2′-deoxy-β-D-erythro-pentofuranosyl)-5,6,7-trihydro-6,7-dihydroxyimidazo[1,2-a]purin-9one (dG-gx1), was reported as a major product of glyoxal with 2′-deoxyribonucleosides and DNA (15, 16). However, dG-gx reacts further with the 2′-deoxyribonucleosides in DNA forming cross-linked products (18). It was shown that the formation of * To whom correspondence should be addressed. Tel: (886) 5-242-8176. Fax: (886) 5-272-1040. E-mail: [email protected]. 1 Abbreviations: capLC-NSI/MS/MS, capillary liquid chromatography with nanospray ionization tandem mass spectrometry; CID, collision-induced dissociation; dG-gx, 3-(2′-deoxy-β-D-erythro-pentofuranosyl)-5,6,7-trihydro6,7-dihydroxyimidazo[1,2-a]purin-9-one; dR, 2-deoxyribose; ESI, electrospray ionization; gx, glyoxal; NSI, nanospray ionization; nt, normal nucleotides; % SE, percentage standard error; SRM, selective reaction monitoring; SPE, solid-phase extraction.

cross-links in DNA by glyoxal could induce deletions of large DNA fragments, frameshift mutations, and base substitution mutations in bacteria and mammalian cells (22, 23). Brock and co-workers extensively investigated the structures of glyoxal-induced DNA cross-links with individual stereoisomers at the nucleoside level, and the cross-links characterized include dG-gx-dC, dG-gx-dG, and dG-gx-dA (24) (Scheme 1). Only dG-gx-dC and dG-gx-dA cross-links were detected in single stranded DNA incubated with a high concentration of glyoxal using HPLC with UV detection (18). However, no quantification of these cross-links was provided. Over the past decade, advancement in miniaturized liquid chromatography and mass spectrometry has allowed sensitive measurement of low levels of DNA adducts (reviewed in ref 25). We herein develop a highly sensitive and specific assay based on capillary liquid chromatography with nanospray ionization tandem mass spectrometry (capLC-NSI/MS/MS) for simultaneous quantification of dG-gx-dC, dG-gx-dG, and dG-gx-dA in DNA hydrolysate. This assay was used to quantify these three DNA cross-links in double stranded calf thymus DNA incubated with a low concentration of glyoxal as well as in human tissue DNA from placenta.

Experimental Procedures Materials. Calf thymus DNA, human placental DNA, glyoxal (40% aqueous solution), micrococal nuclease, spleen phosphdiesterase, and nuclease P1 were obtained from Sigma Chemical Co. (St. Louis, MO). HPLC-UV System. The HPLC-UV system consisted of a Hitachi L-7000 pump system with a D-7000 interface (Hitachi, Tokyo,

10.1021/tx900129e CCC: $40.75  2009 American Chemical Society Published on Web 06/15/2009

Glyoxal-Induced Cross-Links by capLC-NSI/MS/MS

Chem. Res. Toxicol., Vol. 22, No. 7, 2009 1335

Scheme 1. Formation of Glyoxal-Induced Cross-Links in DNA

Japan), a Rheodyne injector, and a reversed phase C18 column [Luna C18 (2), 4.6 mm × 150 mm, 5 µm, Phenomenex, Torrance, CA]. It was eluted at a flow rate of 1.0 mL/min with a linear gradient of 0% to 25% aqueous methanol from 0 to 25 min, followed by a linear gradient of 25% to 55% aqueous methanol from 25 to 35 min. The methanol concentration was increased from 55% to 100% in 3 min and maintained in 100% for 5 min before conditioning with 100% water for 20 min. Synthesis of Standards. Synthesis of dG-gx, dG-gx-dC, dGgx-dG, and dG-gx-dA followed procedures modified from the report of Brock et al. (24), and the UV spectra of the standards were confirmed with those reported by Kasai et al. (18). dG-gx. A solution containing 2′-deoxyguanosine (0.5 mM) and glyoxal (5.0 mM) in potassium phosphate buffer (0.2 M, pH 7.4) was incubated at 37 °C for 2 h and collected by the HPLC-UV system from 15 to 16 min. The collected fraction was evaporated to dryness. ESI+ MS: m/z 308 ([M + H]+), 192 ([M + H - dR]+). dG-gx-dC. A solution containing 2′-deoxycytidine (78 µmol) and dG-gx (15 µmol) in dimethylsulfoxide (0.15 mL) was incubated at 60 °C for 6 days, collected by the HPLC-UV system from 22 to 24 min, and evaporated to dryness. ESI+ MS: m/z 535 ([M + H]+), 419 ([M + H - dR]+), 303 ([M + H - 2 dR]+), 228 ([dC + H]+). dG-gx-dG. A solution containing 2′-deoxyguanosine (30 µmol) and dG-gx (15 µmol) in 0.1 N sodium hydroxide (0.1 N, 0.5 mL) was incubated at 60 °C for 2 days. After neutralization, it was collected by the HPLC-UV system from 29 to 32 min and evaporated to dryness. ESI+ MS: m/z 557 ([M + H]+), 441 ([M + H - dR]+), 325 ([M + H - dR]+). dG-gx-dA. A solution containing 2′-deoxycytidine (78 µmol) and dG-gx (15 µmol) in dimethylsulfoxide (0.15 mL) was incubated at 60 °C for 6 days, collected by the HPLC-UV system from 29 to 31 min, and evaporated to dryness. ESI+ MS: m/z 559 ([M + H]+), 443 ([M + H - dR]+), 327 ([M + H - 2 dR]+), 252 ([dA + H]+). Incubation of Glyoxal with Calf Thymus DNA. A solution containing double stranded calf thymus DNA (1.0 mg/mL) and glyoxal (1.0 mM) in potassium phosphate buffer (0.2 M, pH 7.0) was incubated at 37 °C. One hundred microliters of the solution was removed at different time points, i.e., 0.6, 2.5, 5.5, and 30 days. The aliquot was precipitated with cold 95% ethanol (0.9 mL) and evaporated to dryness, and the precipitate was reconstituted in deionized water (100 µL). DNA Hydrolysis. Method A. DNA (100 µg) was digested with micrococal nuclease (2 units) and spleen phosphodiesterase (0.2 unit) in sodium acetate (2.0 mM, pH 6.0) and CaCl2 (10 mM) at 37 °C overnight, followed by the addition of nuclease P1 (1 unit) with incubation at 37 °C for 4 h, and the subsequent addition of adenosine deaminase (1 unit) and incubation at 37 °C for another 0.5 h.

Method B. The procedures were modified from the previous reports (26, 27). Typically, to the DNA (100 µg) was added nuclease P1 (1.0 unit) and calf spleen phosphodiesterase (0.005 unit) in 30 mM sodium acetate (pH 5.0) and incubated at 37 °C for 6 h, followed by the addition of alkaline phosphatase (20 units) and snake venom phosphodiesterase (0.05 unit) in 0.15 M Tris-HCl (pH 8.9) at 37 °C for 4 h. Finally, adenosine deaminase (1 unit) was added and incubated at 37 °C for 0.5 h. Adduct Enrichment. Five micrograms each of 2′-deoxycytidine, 2′-deoxyinosine, 2′-deoxyguanosine, 2′-deoxythymidine, dG-gx-dC, dG-gx-dG, and dG-gx-dA were loaded on a preconditioned reversed phase C18 solid phase extraction (SPE) column [Bond Eut C18, 100 mg, 1 mL, Varian (Harbor City, CA)]. The SPE column was eluted and collected 1 mL each with 3 mL of water, followed by 1 mL of 5% aqueous methanol, 1 mL of 25% aqueous methanol, and 1 mL of 35% aqueous methanol. These six fractions were evaporated to dryness by a centrifuge concentrator, reconstituted in water, and analyzed by the HPLC-UV system described above for calculating the recovery and distribution. The enzyme digest was filtered through a 0.22 µm syringe filter, and the cross-links were enriched by the reversed phase C18 SPE column collecting the 1 mL of 25% aqueous methanol fraction after washing with 3 mL of water and 1 mL of 5% aqueous methanol. The collected fraction was evaporated under vacuum by a centrifuge concentrator, reconstituted in 10 µL of 0.1% acetic acid, and passed through a 0.22 µm Nylon syringe filter. Two microliters of the processed sample was subjected to the capLC-NSI/MS/MS analysis described below. capLC-NSI/MS/MS Analysis. A 2-µL injection loop was connected to a six-port switching valve injector into a LC system consisting of an UltiMate 3000 Nano and Capillary LC system (Dionex, Amsterdam, Netherlands) and a reversed phase column [(Bio-Basic C18, 180 µm × 150 mm, 5 µm (Thermo Electron Corp., Bremen, Germany)]. The pump output (100 µL/min) was split before the injection port to a flow rate of 1.0 µL/min. Mobile phase A was 0.1% acetic acid (pH 3.3), and mobile phase B was 0.1% acetic acid in acetonitrile. The elution started with a linear gradient of 0% mobile phase B to 30% mobile phase B in the 30 min. The effluent was subjected to analysis by a triple quadrupole mass spectrometer, TSQ Quantum Ultra EMR mass spectrometer (Thermo Electron Corp., San Jose, CA), equipped with a nanospray ionization (NSI) interface. The column effluent enters the spray chamber through a tapered emitter constructed from a 75-µm i.d. fused-silica capillary and is directly electrosprayed into the mass spectrometer under the positive ion mode for the NSI-MS/MS. The spray was monitored by a builtin CCD camera. The spray voltage was 1500 V, and the source temperature was at 220 °C. Argon was used as the collision gas in

1336

Chem. Res. Toxicol., Vol. 22, No. 7, 2009

Chen and Chen

Figure 2. Calibration curves for dG-gx-dC, dG-gx-dG, and dG-gx-dA under the selective reaction monitoring (SRM) mode. Untreated calf thymus DNA (20 µg) was added to various amounts of standard dGgx-dC, dG-gx-dG, and dG-gx-dA, hydrolyzed with (a) method A or (b) method B, enriched by the SPE column, and analyzed by the capLCNSI/MS/MS analysis as described in Experimental Procedures. The experiments were performed in duplicate. Figure 1. Collision-induced ionization spectra of (a) dG-gx-dC, (b) dG-gx-dG, and (c) dG-gx-dA.

MS/MS experiments. The pressure of the collision cell was 1.5 mTorr, and the collision energy was 15 V. The adduct enriched sample was analyzed by capLC-NSI/MS/MS with the transition from the parent ion [M + H]+ focused in quadrupole 1 (Q1) and dissociated in a collision cell (quadrupole 2, Q2) yielding the product ion [M + H - 232]+ ([M + H - 2 dR]+) or [M + H 116]+ ([M + H - dR]+), which was analyzed in quadrupole 3 (Q3) under the selective reaction monitoring (SRM) mode with the mass width of Q1 and Q3 being 0.7 m/z, and a dwell time of 0.1 s. The SRM method monitoring Q1 and Q3 were m/z 535 f m/z 303 and m/z 535 f m/z 419, respectively, for dG-gx-dC. For dG-gx-dG, the SRM method monitored Q1 and Q3 were m/z 557 f m/z 325 and m/z 557 f m/z 441, respectively. For dG-gx-dA, the SRM method monitored Q1 and Q3 were m/z 559 f m/z 327 and m/z 559 f m/z 443, respectively. Calibration Curves. The stock solutions of dG-gx-dC, dG-gxdG, and dG-gx-dA (1.0 mg/mL) in water were stored at -20 °C. The sample solutions for calibration were freshly prepared by diluting the stock solutions in water for each analysis. To untreated calf thymus DNA (20 µg or 100 µg) was added dG-gx-dC, dGgx-dG, and dG-gx-dA ranging from 0, 0.2, 0.5, 1.0, 2.0, 4.0, 10, 50, 100, and 200 to 400 pg each, followed by enzyme hydrolysis by method A or B. Each sample was enriched by the C18 SPE column, evaporated, and analyzed by capLC-NSI/MS/MS analysis

as described above. The experiments were performed in duplicate, and the equation for each analyte was obtained by linear regression. nanoLC-NSI/IT-MSn Analysis. An LTQ linear ion trap mass spectrometer (Thermo Electron Corp., San Jose, CA) equipped with a nanospray ionization source coupled online to an X’TremeSimple nanoflow LC system (Microtech, Orange, CA) was used. Two microliters of each sample was manually injected onto a C18 column (120 mm × 75 µm, 5 µm, 100 Å) packed in-house (magic C18, Michrom BioResource, Auburn, CA). Mobile phases A and B were composed of 5% and 80% acetonitrile in formic acid (0.1%, pH 2.6), respectively. A linear gradient was employed from 5% B to 80% B from 0 to 30 min and maintained at 80% B over the next 15 min at a flow rate of 300 nL/min. The MS2 and MS3 spectra were obtained at a heated capillary temperature of 200 °C with a capillary voltage of 2.0 kV, a source voltage of 1500 V, and a collision energy of 15 V. The collision gas was argon. Ions were isolated with a mass isolation width (m/ z) of 2.0. The selected reaction monitoring (SRM) experiments were performed by selecting the precursor ion and acquiring full-scan product ion spectra. The formation of a specific fragment ion from each precursor was used to construct the chromatogram.

Results and Discussion The goal of this study was to develop a highly sensitive MSbased assay for quantification of DNA cross-links induced by glyoxal. Three cross-linked products, dG-gx-dC, dG-gx-dG, and dG-gx-dA, have been characterized previously at the nucleoside

Glyoxal-Induced Cross-Links by capLC-NSI/MS/MS

Chem. Res. Toxicol., Vol. 22, No. 7, 2009 1337

Table 1. Levels of Cross-Linked Products in Glyoxal-Treated Calf Thymus DNA and Human Placental DNA by Hydrolysis Methods A and B level (mean ( SE)a cross-linked product dG-gx-dC dG-gx-dG dG-gx-dA

gx-treated calf thymus DNAb,c (method A)

gx-treated calf thymus DNAb,c (method B),

human placental DNAd (method A)

human placental DNAc (method B)

4.35 ( 0.03 in 108 nt (% SE: 0.7%) 2.96 ( 0.14 in 108 nt (% SE: 4.7%) 3.50 ( 0.03 in 108 nt (% SE: 0.9%)

6.52 ( 0.05 in 105 nt (% SE: 0.8%) 8.00 ( 0.07 in 106 nt (% SE: 0.9%) 2.74 ( 0.06 in 105 nt (% SE: 2.3%)

7.46 ( 0.73 in 109 nt (% SE: 9.8%) 4.49 ( 0.01 in 109 nt (% SE: 0.1%) 8.56 ( 0.24 in 109 nt (% SE: 2.8%)

2.48 ( 0.08 in 108 nt (% SE: 3.3%) 1.26 ( 0.04 in 108 nt (% SE: 3.4%) 3.50 ( 0.04 in 108 nt (% SE: 1.3%)

a The cross-links levels were expressed as mean ( standard error (SE) in normal nucleotides (nt) calculated from duplicated experiments. The percentage standard error (% SE) was expressed in parentheses. b Calf thymus DNA (1.0 mg/mL) was incubated with glyoxal (1.0 mM) at 37 °C for 30 days. c Hydrolysate of 20 µg of DNA was analyzed after processes described in the Experimental Procedures. d Hydrolysate of 100 µg of DNA was analyzed after processes described in the Experimental Procedures.

Figure 3. Accuracy of the capLC-NSI/MS/MS assay was confirmed by addition of various amounts of standard dG-gx-dC, dG-gx-dG, and dG-gx-dA to the hydrolysate of human placental DNA (20 µg). The cross-link levels were obtained from the y-intercept by linear regression and calculated via the calibration curve.

Figure 4. capLC-NSI/MS/MS analysis of dG-gx-dC (upper panel), dGgx-dG (middle panel), and dG-gx-dA (lower panel) in calf thymus DNA (20 µg) incubated with glyoxal (1.0 Mm) at 37 °C for 30 days after enzyme hydrolysis and SPE enrichment.

level (17). They were individually synthesized from a reaction of purified dG-gx with dCyd, dGuo, or dAdo and purified by collecting the elution from HPLC-UV. Each cross-linked product is present as a pair of diastereomers with identical UV spectra (Figure S1 in Supporting Information). The collision-induced dissociation (CID) spectrum of dG-gx-dC showed the [M + H]+ ion at m/z 553 and the [M + H - dR]+, i.e., the protonated molecular ion with loss of a 2-deoxyribose (- 116), at m/z 419. Also present were the ions of m/z 303 and 228; the former being the protonated molecular ion with a loss of two 2-deoxyribose (- 232) ([M + H - 2 dR]+) and the latter being the protonated dCyd, i.e., [M + H - dG-gx]+ (Figure 1a). In the CID spectrum

of dG-gx-dG, two fragment ions at m/z 441 and 325 were characterized as [M + H - dR]+ and [M + H - 2 dR]+, respectively (Figure 1b). For dG-gx-dA, the fragment ions were assigned as [M + H - dR]+, [M + H - 2 dR]+, and [M + H - dG-gx]+ (or [dAdo + H]+) at m/z 443, 327, and 252, respectively (Figure 1c). Sample Pretreatment: Enzyme Hydrolysis of DNA Samples to 2′-Deoxyribonucleosides. In the presence of adenosine deaminase, 2′-deoxyadenosine was converted to 2′-deoxyinosine, which eluted earlier than 2′-deoxyadenosine in the reversed phase column. The three cross-links were enriched from the DNA hydrolysate by a reversed phase solid phase extraction (SPE) column before the MS analysis. The majority of 2′-deoxycytidine (86%) and a small portion of 2′deoxyinosine (5%) were eluted with water, while the rest of the nucleosides were eluted by 5% of aqueous methanol and were discarded. The more hydrophobic analytes were eluted with 25% aqueous methanol, together with 3% of 2′deoxyguanosine and 2′-deoxythymidine. Analyzed by HPLCUV, the recoveries of the SPE column were nearly quantitative with microgram quantities of the 2′-deoxyribonucleosides and the cross-link standards. The recovery of each analyte might decrease with reducing quantity (for example, in picogram quantity), but the distribution does not change provided that the column is not overloaded. Thus, this SPE enrichment procedure was proven to be appropriate and efficient in enrichment of the analyte. Alternatively, one can collect the analytes from the biological sample by HPLC. However, the process is labor-intensive, and special attention should be paid to avoid cross-contamination between samples (28-30). Although the resolution of the SPE column is lower than that of HPLC, the disposable SPE column is free of cross-contamination, and many samples (up to more than 20 samples) can be processed at the same time. capLC-NSI/MS/MS Assay. To increase assay sensitivity for low-abundant analytes, it is a trend to use liquid chromatography with a small diameter column at a low flow rate, although the drawbacks include low optimum injection volume, low column capacity, and short column lifetime because the column tends to give high back pressure after extended use. Extra care on sample handling should be taken. The use of nanoscale liquid chromatography is accompanied by a nanospray (or nanoelectrospray) ionization (NSI) source, which has been widely used in proteomic research with an ion trap mass spectrometer. A limited number of reports adopted nano liquid chromatography with nanospray ionization tandem mass spectrometry analysis for quantification of DNA adducts (31-35). In this study, we use a column with 180 µm inner diameter coupled with a nanospray ionization source in a triple quadrupole mass spectrometer. The flow rate (1.0 µL/min), which falls in the

1338

Chem. Res. Toxicol., Vol. 22, No. 7, 2009

Chen and Chen

Figure 5. Structrual confirmation of dG-gx-dC, dG-gx-dG, and dG-gx-dA in glyoxal-treated calf thymus DNA by MS2 and MS3 spectra. The MS2 spectra were obtained by acquiring the scan of peaks corresponding to the [M + H]+ ions at (a) m/z 535, (b) m/z 557, or (c) m/z 559. The MS3 spectra of dG-gx-dC, dG-gx-dG, and dG-gx-dA were obtained by acquiring the scan of the corresponding [M + H - dR]+ ions at (d) m/z 419, (e) m/z 441 or (f) m/z 443 in (a), (b), or (c), respectively.

category of capillary LC, is the upper limit for the NSI source. In this setup, a tapered spray emitter constructed from a 75-µm inner diameter fused-silica capillary provides a stable spray,

which is not achievable using the conventional electrospray ionization (ESI) source at this flow rate. Overall, this assay optimizes the sensitivity offered by the NSI source with a low-

Glyoxal-Induced Cross-Links by capLC-NSI/MS/MS

Chem. Res. Toxicol., Vol. 22, No. 7, 2009 1339

Figure 6. Time-dependent formation of dG-gx-dC, dG-gx-dG, and dGgx-dA in calf thymus DNA incubated with glyoxal (1.0 mM) at 37 °C.

flow-rate LC system to concentrate the analyte on-column as well as the triple quadrupole mass spectrometer under the selective reaction monitoring (SRM) mode. Under the SRM mode, the parent ion [M + H]+ was focused on quadrupole 1 (Q1) and dissociated in a collision cell (quadrupole 2, Q2) yielding the product ion [M + H 232]+ ([M + H - 2 dR]+) (method 1) or [M + H - 116]+ ([M + H - dR]+) (method 2), which was analyzed in quadrupole 3 (Q3). The slopes of the standard curves constructed using various amounts of standard dG-gx-dC, dGgx-dG, and dG-gx-dA indicate that the assay sensitivity was lower for dG-gx-dC and dG-gx-dA using SRM method 1 (Figure S2 in Supporting Information) than SRM method 2 (Figure S3 in Supporting Information). For dG-gx-dG, the slope in Figure S2b (Supporting Information) was approximately twice that in Figure S3b (Supporting Information). It was interesting to note that the base peaks in the CID spectra were the [M + H - dR]+ ions for dG-gx-dC and dG-gx-dA (Figure 1a and c), while that for dG-gx-dG was the [M + H - 2 dR]+ ion (Figure 1b). The high abundance of the [M + H - 2 dR]+ ion explains why it is more sensible to use SRM transition method 1 than method 2 for analysis of dG-gx-dG. Thus, SRM method 2 was used for the analysis of dG-gx-dC and dG-gx-dA, while SRM method 1 was used for dG-gx-dG in DNA hydrolysate. Using this capLC-ESI/MS/MS, the signal-to-noise ratios (S/N) of 40 fg of the three forms of DNA cross-link standards dGgx-dC, dG-gx-dG, and dG-gx-dA were 5, 19, and 3, respectively. Thus, the detection limits were approximately 75, 12, and 72 amol for dG-gx-dC, dG-gx-dG, and dG-gxdA, respectively. Assay Calibration. The most accurate and precise quantification of an analyte by mass spectrometry is the incorporation of the stable isotope of the analyte as the internal standard to monitor analyte recovery during the assay procedures and to adjust the matrix effect on signal reduction. In this study, the stable isotopes of dG-gx-dC, dG-gx-dG, and dG-gx-dA are not used mainly due to the low yield in their synthesis (especially for dG-gx-dG) and the high cost of the isotopes. Instead, calibration curves were constructed to account for sample loss and matrix effect with various amounts of the standards in the presence of untreated calf thymus DNA (20 µg), followed by hydrolysis, C18 SPE enrichment, and capLC-NSI/MS/MS analysis. These three cross-links were not detectable in untreated calf thymus DNA. Two hydrolysis methods (A and B) were employed in this study, and they might have different matrix effects. Therefore, two calibration curves were constructed for each method. The results showed that the slopes of each calibration curve using hydrolysis methods A and B were

Figure 7. (a) capLC-NSI/MS/MS analysis of dG-gx-dC (upper panel), dG-gx-dG (middle panel), and dG-gx-dA (lower panel) in human placental DNA. (a) DNA (100 µg) was enzyme hydrolyzed by method A, while (b) DNA (20 µg) was enzyme hydrolyzed by method B, followed by SPE enrichment. (c) Chromatograms showing coinjection of the sample in b, equivalent to 10 µg of DNA hydrolysate, with 0.5 pg each of standard dG-gx-dC, dG-gx-dG, and dG-gx-dA.

similar. Compared to the standard curves (Figures S2 and S3, Supporting Information), 38%, 32%, and 35% reduction in the

1340

Chem. Res. Toxicol., Vol. 22, No. 7, 2009

slope was observed in adduct levels ranging from 0.5 to 4.0 pg for dG-gx-dC, dG-gx-dG, and dG-gx-dA, respectively, using method A (Figure 2a), while those using method B were 50%, 47%, and 47% in adduct levels ranging from 0.2 to 400 pg (Figure 2b). The assay using hydrolysis method B allowed quantification of 0.5 pg (0.94 fmol), 0.2 pg (0.36 fmol), and 0.2 pg (0.36 fmol) of dG-gx-dC, dG-gx-dG, and dG-gx-dA, respectively, corresponding to 14.3, 5.5, and 5.5 adducts in 109 normal nucleotides (nt) starting from 20 µg DNA. The relatively high quantification limit for dG-gx-dC was due to the high background of its SRM channel. Without the use of isotopes as internal standards, the percentage standard error (% SE) in the calibration curve (Figure 2) and for DNA analysis (Table 1) ranged from 0.7% to 4.7% (except for one with 9.8%), suggesting the high precision of the assay. The variability of the MS instrumentation response day to day was dealt with by constructing a standard curve with the synthetic standards each day before analyzing the samples to justify the calibration curve. Furthermore, adduct levels of a sample were analyzed on different days, and the interday variations were acceptable (