A Stable Isotope Dilution Nanoflow Liquid Chromatography Tandem

Nov 25, 2017 - To our best knowledge, this is the first report of the identification and quantification of glyoxal-derived cross-linked DNA adducts in...
1 downloads 0 Views 1MB Size
Download PDF
Article Cite This: Anal. Chem. 2017, 89, 13082−13088

pubs.acs.org/ac

A Stable Isotope Dilution Nanoflow Liquid Chromatography Tandem Mass Spectrometry Assay for the Simultaneous Detection and Quantification of Glyoxal-Induced DNA Cross-Linked Adducts in Leukocytes from Diabetic Patients Hauh-Jyun Candy Chen,* Ya-Lang Chang, Yi-Chun Teng, Chiung-Fong Hsiao, and Tsai-Shiuan Lin Department of Chemistry and Biochemistry, National Chung Cheng University, 168 University Road, Ming-Hsiung, Chia-Yi 62102, Taiwan S Supporting Information *

ABSTRACT: Glyoxal (gx) is a bifunctional electrophile capable of cross-linking DNA. Although it is present in foods and from the environment, endogenous formation of glyoxal occurs through metabolism of carbohydrates and oxidation of lipids and nucleic acids. Plasma concentrations of glyoxal are elevated in in diabetes mellitus patients compared to nondiabetics. The most abundant 2′-deoxyribonucleoside adducts cross-linked by glyoxal are dG-gx-dC, dG-gx-dA, and dG-gx-dG. These DNA crosslinks can be mutagenic by damaging the integrity of the DNA structure. Herein, we developed a highly sensitive and specific assay for the simultaneous detection and quantification of the dG-gx-dC and dG-gx-dA cross-links based on stable isotope dilution (SID) nanoflow liquid chromatography nanospray ionization tandem mass spectrometry (nanoLC-NSI/MS/MS) under the highly selected reaction monitoring mode and using a triple quadrupole mass spectrometer. The entire assay procedure involved addition of the stable isotope standards [15N5]dG-gx-dC and [15N5]dG-gx-dA as internal standards, enzyme hydrolysis to release the cross-links as nucleosides, enrichment by a reversed-phase solid-phase extraction column, and nanoLC-NSI/MS/MS analysis. The detection limit is 0.19 amol for dG-gx-dC and 0.89 amol for dG-gx-dA, which is 400 and 80 times more sensitive, respectively, than capillary LC−NSI/MS/MS assay of these adducts. The lower limit of quantification was 94 and 90 amol for dG-gx-dC and dG-gx-dA, respectively, which is equivalent to 0.056 and 0.065 adducts in 108 normal nucleotides in 50 μg of DNA. In type 2 diabetes mellitus (T2DM) patients (n = 38), the levels of dG-gx-dC and dG-gx-dA in leukocyte DNA were 1.94 ± 1.20 and 2.10 ± 1.77 in 108 normal nucleotides, respectively, which were significantly higher than those in nondiabetics (n = 39: 0.83 ± 0.92 and 1.05 ± 0.99 in 108 normal nucleotides, respectively). Excluding the factor of smoking, an exogenous source of glyoxal, levels of these two cross-linked adducts were found to be significantly higher in nonsmoking T2DM patients than in nonsmoking control subjects. Furthermore, the levels of dG-gx-dC and dG-gx-dA correlated with HbA1c with statistical significance. To our best knowledge, this is the first report of the identification and quantification of glyoxal-derived cross-linked DNA adducts in human leukocyte DNA and their association with T2DM. This SID nanoLC-NSI/MS/MS assay is highly sensitive and specific and it requires only 50 μg of leukocyte DNA isolated from 2−3 mL of blood to accurately quantify these two cross-linked adducts simultaneously. Our assay thus provides a useful biomarker for the evaluation of glyoxal-derived DNA damage. lyoxal (gx) is the most reactive α-dicarbonyl compound generated from glycated proteins in the Maillard reaction.1 It can also be derived from the metabolism of Nnitrosamines, amino acids,2,3 and carbohydrates, and from the oxidation of lipids and nucleic acids.3−5 Being mutagenic in bacteria and mammalian cells, glyoxal is used in the industry, is widely distributed in the environment (including cigarette smoke),6 and is found in foods and beverages.7,8 Serum

G

© 2017 American Chemical Society

concentrations of glyoxal are elevated in patients with diabetes, uremia, and peritoneal dialysis.9−12 Coincidentally, diabetic patients have been shown to have an increased risk of various types of cancers.13 Received: November 2, 2016 Accepted: November 25, 2017 Published: November 25, 2017 13082

DOI: 10.1021/acs.analchem.6b04296 Anal. Chem. 2017, 89, 13082−13088

Article

Analytical Chemistry

mL of 10% aqueous MeOH. For [15N5]dG-gx-dA, the SPE column was washed with an additional 3 mL of 15% aqueous MeOH, and [15N5]dG-gx-dA was eluted with 3 mL of 25% aqueous MeOH. The purity of [15N5]dG-gx-dC and [15N5]dGgx-dA obtained thereof was confirmed and quantified by comparing with dG-gx-dC and dG-gx-dA using nanoLC-NSI/ MS/MS described below (Figure S1, Supporting Information). The yield was 15% for both isotope-labeled standards. The daughter ion scan spectra were provided in Figure S2, Supporting Information. [15N5]dG-gx-dC: ESI+ MS: m/z 540 ([M + H]+), 424 ([M + H − dR]+), 308 ([M + H − 2 dR]+), 228 ([dC + H]+). [15N5]dG-gx-dA: m/z 564 ([M + H]+), 448 ([M + H − dR]+), 332 ([M + H − 2 dR]+), 252 ([dA + H]+). DNA Extraction from Blood. Human blood freshly drawn from the vein was added to a solution of of 10% (v/v) citratedextrose as an anticoagulant and stored at 4 °C. Within a week, the blood/anticoagulant solution was subjected to the Blood Genomic Midiprep System (Viogen, Sunnyvale, CA) or the Blood and Cell Culture DNA Maxi Kit (QIAGEN, Duesseldorf, Germany), following the protocols provided by the manufacturers. The amount of DNA was quantified by a NanoPhotometer (Implen, Inc. Westlake Village, CA). The purity of DNA was confirmed by the absorbance ratio at 260 and 280 nm being between 1.7 and 1.9. DNA Hydrolysis and Quantitation of dG. Typically, a solution containing DNA (50 μg) and 100 pg each of [15N5]dG-gx-dC and [15N5]dG-gx-dA was added nuclease P1 (0.5 unit) and calf spleen phosphodiesterase II (0.0025 unit) in 30 mM sodium acetate (pH 5.0) and incubated at 37 °C for 6 h. The solution was added snake venom phosphodiesterase I (0.025 unit) and alkaline phosphatase (10 units) in 0.15 M Tris-HCl (pH 8.9) and incubated at 37 °C for 4 h, followed by addition and incubation with adenosine deaminase (0.5 unit) at 37 °C for an additional 0.5 h.24 The amount of dG contained in the hydrolysate was analyzed by a portion (1/20) of the DNA hydrolysate by an HPLC-UV system consisting a Hitachi L7000 pump system with a D-7000 interface (Hitachi, Tokyo, Japan), a Rheodyne injector, and a reversed-phase C18 column [5 TC-C18(2), 4.6 mm × 250 mm, 5 μm, Agilent Technologies, Palo Alto, CA]. The column was eluted at a flow rate of 1.0 mL/min with water (mobile phase A) and methanol (mobile phase B) with a linear gradient from 0% to 100% methanol in 30 min; dG was eluted at 11.6 min and quantified using a calibration curve constructed from 0.1 to 1.0 μg dG. The amount of dG released from digestion was nearly quantitative. Removal of Hydrolytic Enzymes from the DNA Hydrolysate. Three different methods were used to remove hydrolytic enzymes in the DNA hydrolysate (131 μL, final volume). The DNA hydrolysate was filtered by (A) 0.22 μm Nylon syringe filter or (B) 0.22 μm PTFE spin column and centrifuged at 2400g for 5 min. In method C, DNA hydrolysate was added to cold acetone (1180 μL), allowed to stand at −20 °C for 20 min, and the precipitate was removed by centrifugation at 21 600g for 20 min. SPE Enrichment. The cross-linked adducts dG-gx-dC and dG-gx-dA in the DNA hydrolysate was enriched by a reversedphase C18 SPE column [Bond Elut Bonded phase C18, 100 mg, 1 mL, Varian (Harbor City, CA)]. The SPE column was conditioned with 5 mL of methanol, followed by 5 mL of water. The sample was loaded, and the SPE column was washed with 3 mL of water and 1 mL of 5% aqueous MeOH. Then, the cross-linked adducts were eluted with 1 mL of 30% aqueous

Glyoxal can react with DNA and proteins to form single adducts as well as cross-links in DNA, in proteins, and between DNA and proteins.14−17 We recently characterized the sites and types of hemoglobin adducts of glyoxal and identified certain glyoxal-modified peptides as a biomarker for diabetes mellitus using the shot-gun proteomic quantitation approach.18 The reactions of glyoxal with the 2′-deoxyribonucleosides give rise to 3-(2′-deoxy-D-erythro-pentofuranosyl)-5,6,7-trihydro-6,7-dihydroxyimidazo[1,2-a]purine-9-one (dG-gx),15 which can be converted to the stable product N2-carboxymethyl-2′deoxyguanosine (N2-CMdG).14 In DNA, dG-gx reacts further with the nucleosides to form the cross-linked products dG-gxdC, dG-gx-dA, and dG-gx-dG.19,20 Among the various types of DNA damage, interstrand cross-links are the most harmful because they completely block DNA transcription and replication.21−23 In this laboratory, we previously reported quantification of these three cross-linked adducts in glyoxaltreated calf thymus DNA and in human placental DNA by capillary liquid chromatography tandem mass spectrometry (LC−MS/MS).24 However, stable isotope dilution (SID) method was not employed owing to the low yield in synthesis of these isotope standards. Alternatively, a calibration curve was constructed along with each set of samples. In this current study, the stable isotopomers of dG-gx-dC and dG-gx-dA were synthesized and used as internal standards. The isotope of dGgx-dG was omitted because its yield was the lowest, and the abundance of dG-gx-dG in DNA samples is also the lowest among the three cross-linked adducts. With the nanoflow LC system coupled to nanospray ionization tandem mass spectrometry (nanoLC−NSI/MS/MS), the sensitivity and specificity of the assay allowed accurate quantification of these two cross-linked adducts in leukocyte DNA from type 2 diabetes mellitus (T2DM) patients and nondiabetic subjects.

■

MATERIAL AND METHODS Materials. Glyoxal (40% aqueous solution), adenosine deaminase, micrococal nuclease, snake venom phosphodiesterase I, calf spleen phosphodiesterase II, nuclease P1, calf thymus DNA, and human placental DNA were obtained from Sigma Chemical Co. (St. Louis, MO). [15N5]2′-Deoxyguanosine ([15N5]dGuo) was from Cambridge Isotope Laboratories (Andover, MA). All reagents are of reagent grade or above. Synthesis of [15N5]dG-gx-dC and [15N5]dG-gx-dA. Standard isotopomers of dG-gx-dC and dG-gx-dA were individually synthesized from [15N5]gx-dG with dC or dA, respectively, and purified as described previously.24 Standard [15N5]gx-dG was synthesized from reaction of [15N5]dGuo (0.5 mM) with glyoxal (5.0 mM) in 0.4 M potassium phosphate buffer (pH 7.4) at room temperature with stirring for 15 h. The reaction mixture was purified by a disposable C18 solid-phase extraction (SPE) column [Bond Elut Bonded phase C18, 500 mg, 3 mL, Varian (Harbor City, CA)] conditioned with 15 mL of methanol, followed by 15 mL of water. After the sample was loaded, the SPE column was washed with 9 mL of water and eluted with 15% MeOH. [15N5]dG-gx-dC and [15N5]dG-gx-dA were synthesized by the same manner as reported for dG-gx-dC and dG-gx-dA24 except that they were purified separately by a disposable C18 SPE column to avoid contamination from the HPLC system. The C18 SPE column was conditioned with 15 mL of methanol and followed by 15 mL of water. After the sample was loaded, the SPE column was washed with 12 mL of water and 3 mL of 5% aqueous methanol, and [15N5]dG-gx-dC was eluted with 3 13083

DOI: 10.1021/acs.analchem.6b04296 Anal. Chem. 2017, 89, 13082−13088

Article

Analytical Chemistry Table 1. H-SRM Transitions for the Glyoxal-Induced DNA Cross-Linked Adducts and Their Corresponding Isotopes

dG-gx-dC [15N5]dG-gx-dC dG-gx-dA [15N5]dG-gx-dA

SRM 1

SRM 2

([MH]+ → [MH − deoxyribose]+)

([MH]+ → [MH − 2 deoxyribose]+)

m/z m/z m/z m/z

535.2 540.2 559.2 564.2

→ → → →

m/z 419.0 m/z 424.0 m/z 443.0 m/z 448.0

m/z m/z m/z m/z

535.2 540.2 559.2 564.2

→ → → →

m/z m/z m/z m/z

303.0 308.0 327.0 332.0

3.2), and an aliquot (2 μL) was analyzed by the nanoLC−NSI/ MS/MS system described above. Study Subjects. Collection of blood from type 2 diabetic mellitus patients was approved by Institutional Review Board of Buddhist Dalin Tzu Chi General Hospital (IRB No. B10203014). The patients were 14 males and 24 females. The mean (±SD) age was 58.4 ± 12.1 (ranging from 23 to 77) years, and the mean body mass index (BMI) was 27.6 ± 4.0. The mean (±SD) HbA1C level was 9.6 ± 1.8%. The nondiabetic subjects included 27 males and 12 females, recruited from students and employees of the National Chung Cheng University under the approval of the Institutional Review Board of the National Chung Cheng University (IRB No. 100112902). The mean (±SD) age was 34.6 ± 15.3 (ranging from 21 to 60) years and the average BMI was 24.2 ± 6.2. The subjects were given a written warranty stating that the information they provided was for research purposes only and that their personal information would be kept confidential. Statistical Analysis. Statistical analyses were performed by GraphPad InStat version 3.00 for Windows 95, GraphPad Software (San Diego, CA, www.graphpad.com). The nonparametric Mann−Whitney test was used to analyze the difference in the adduct levels between diabetic patients and nondiabetic controls. The multiple regression analysis was used to correlate the level of each cross-linked adduct with HbA1c, age, number of cigarettes smoked per day, and BMI.

MeOH. The eluent was evaporated to dryness and reconstituted in 10 μL of 0.1% aqueous acetic acid (v/v) before analysis by nanoLC−NSI/MS/MS. nanoLC System. The LC system consists of a 2 μL injection loop connected to a six-port switching valve injector into an UltiMate 3000 Nano Nano and Capillary LC system (Dionex, Amsterdam, The Netherlands) and a reversed phase tip-column packed in-house (Magic C18AQ, 5 μm, 200 Å, Michrom BioResource, Album, CA). System A employed a 75 μm × 11 cm column with mobile phase A of 0.1% acetic acid (v/v, pH 3.2) and mobile phase B of 0.1% acetic acid (v/v) in acetonitrile eluting at the flow rate of 300 nL/min. System B used a 180 μm × 11 cm column with mobile phase A being 0.1% acetic acid (v/v, pH 3.2) and mobile phase B being 0.1% acetic acid (v/v) in methanol. The elution started with a linear gradient of 1% mobile phase B (v/v/) to 100% mobile phase B from 0 to 30 min, and it was maintained at 100% B in the next 20 min at the flow rate of 500 nL/min. NSI/MS/MS Analysis. The column 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 and performed under the positive-ion mode. The source temperature was at 220 °C, and the spray voltage was 1.5 kV. For the MS/MS experiments, argon was the collision gas with the collision energy of 10 eV, and the pressure of the collision cell was 1.5 mTorr. The adductenriched sample was analyzed by either nanoLC-NSI/MS/MS with the transition from the parent ion [M + H]+ focused in quadrupole 1 (Q1) and dissociated in a collision cell (Q2), yielding the product ion analyzed in quadrupole 3 (Q3) under the highly selective reaction monitoring (H-SRM) mode. The mass width at Q1 was 0.2 m/z and that at Q3 was 0.7 m/z, and the dwell time was 0.1 s. For dG-gx-dC and [15N5]dG-gx-dC, the SRM 1monitoring Q1 and Q3 were m/z 535.2 → m/z 419.0 and m/z 540.2 → m/ z 424.0, respectively, with the collision energy of 10 eV. For dG-gx-dA and [15N5]dG-gx-dA, the SRM 1 monitored Q1 and Q3 were m/z 559.2 → m/z 443.0 and m/z 564.2 → m/z 448.0, respectively, with the collision energy of 15 eV. The SRM 2 monitored Q1 and Q3 for dG-gx-dC and [15N5]dG-gx-dC at m/z 535.2 → m/z 303.0 and m/z 540.2 → m/z 308.0, respectively, with the collision energy of 15 eV. For dG-gx-dA and [15N5]dG-gx-dA, the SRM 2 monitored Q1 and Q3 were m/z 559.2 → m/z 327.0 and m/z 564.2 → m/z 332.0, respectively, with the collision energy of 15 eV (Table 1). Calibration Curve. The solutions containing 100 pg each of [15N5]dG-gx-dC and [15N5]dG-gx-dA with various amounts (0, 0.05, 0.1, 0.2, 0.5, 1.0, 10, and 50 pg) of the dG-gx-dC and dG-gx-dA were prepared in triplicates. Each sample went through a C18 SPE column as described above. The fraction containing these cross-linked adducts was evaporated to dryness and reconstituted in 10 μL of 0.1% acetic acid (pH

■

RESULTS AND DISCUSSION Synthesis and Characterization of Stable IsotopeLabeled Standards. Standard [15N5]dG-gx-dC and [15N5]dGgx-dA were individually synthesized from [15N5]gx-dG with dCyd and dAdo, respectively, using the reported procedures for dG-gx-dC and dG-gx-dA.24 To avoid cross-contamination, the isotope standards were purified through a disposable SPE column. [15N5]dG-gx-dC and [15N5]dG-gx-dA were characterized by their distinctive UV spectra and retention time on HPLC and by comparing their collision-induced dissociation mass spectra with those of standard dG-gx-dC and dG-gx-dA (Figure S1, Supporting Information).24 The purity of the two isotopomers [15N5]dG-gx-dC and [15N5]dG-gx-dA was confirmed by nanoLC−NSI/MS/MS analysis (described below), showing no signals for dG-gx-dC and dG-gx-dA. The quantities of [15N5]dG-gx-dC and [15N5]dG-gx-dA were estimated by the respective UV molar responses of standard dG-gx-dC and dGgx-dA and calibrated by their peak areas in the nanoLC−NSI/ MS/MS.24 [15N5]dG-gx-dC and [15N5]dG-gx-dA were chemically pure with an isotopic purity of 100% and 99.8%, respectively (Figures S2, Supporting Information). Method Development for Detecting DNA CrossLinked Adducts by Stable Isotope Dilution Approach. The procedures for analyzing dG-gx-dC and dG-gx-dA include the (1) isolation and quantification of DNA, (2) addition of isotope-labeled [15N5]dG-gx-dC and [15N5]dG-gx-dA to DNA as internal standards, (3) enzyme hydrolysis of DNA to its 13084

DOI: 10.1021/acs.analchem.6b04296 Anal. Chem. 2017, 89, 13082−13088

Article

Analytical Chemistry

column. Two SPE columns were examined for their performance, namely, a reversed phase C18 column and a polymerbased Strata-X SPE column. The former gave cleaner chromatograms for dG-gx-dC and dG-gx-dA in the DNA hydrolysate and was therefore used for enrichment of these cross-linked adducts. The advantage of using SPE enrichment before LC-MS/MS analysis is that the sample-to-sample carryover interference is avoided. Method Performance and Validation. To achieve high sensitivity and specificity of the assay, a nanoflow LC system was connected to a triple quadrupole mass spectrometer monitored under the H-SRM mode, in which a narrow window (0.2 Da) was used for the precursor ion at the first quadrupole. This setting allows fewer ions to pass through the first quadrupole and thus lowers the background signals, which is particularly helpful in analyzing complex mixture. The adducts were analyzed by nanoLC−NSI/MS/MS under the H-SRM mode using the transition from the parent ion [M + H]+ focused in the first quadrupole (Q1) and dissociated in a collision cell (Q2) to yield the product ion [M + H − deoxyribose]+ (SRM 1) or [M + H − 2 deoxyribose]+ (SRM 2), which was then analyzed in the third quadrupole (Q3). The SRM transitions with loss of dual or single deoxyribose for dGgx-dC, dG-gx-dA, [15N5]dG-gx-dC, and [15N5]dG-gx-dA are listed in Table 1. The sensitivity of dG-gx-dC and dG-gx-dA using SRM 1 was approximately twice of that using SRM 2.24 The transitions with less interference were chosen as quantifier and the other transitions were used for qualitative assurance. As a result, dG-gx-dC was quantified by monitoring the dual loss of deoxyribose and dG-gx-dA was quantified by monitoring the loss of single deoxyribose. Assay Sensitivity. Using nanoLC−NSI/MS/MS system A with a 75 μm-id column, the limit of detection (defined as a signal-to-noise ratio ≥3) of the standard dG-gx-dC and dG-gxdA in aqueous solution was 0.1 fg (0.19 amol) and 0.5 fg (0.89 amol), respectively, which was 400 and 80 times lower, respectively, than that using capillary LC−NSI/MS/MS.24 The on-column detection limit of dG-gx-dC and dG-gx-dA was 1.0 fg (1.9 amol) and 3.0 fg (5.4 amol), respectively, using a 180 μm-id column (nanoLC system B). The calibration curves were constructed by adding a fixed amount (100 pg each) of the internal standards [15N5]dG-gx-dC and [15N5]dG-gx-dA and various amounts of dG-gx-dC and dG-gx-dA (0−50 pg) to calf thymus DNA (50 μg), in which both adducts are not detectable. The mixture was subjected to enzyme hydrolysis, SPE enrichment, and nanoLC−NSI/MS/MS analysis. Only an equivalent of 10 μg of DNA hydrolysate was subjected to analysis by nanoLC-NSI/MS/MS. The lower limit of quantification was 50 fg for both adducts in either nanoLC system, corresponding to 0.056 dG-gx-dC and 0.065 dG-gx-dA in 108 normal nucleotides in 50 μg DNA (Figure S3, Supporting Information), which was 26 and 8 times, respectively, more sensitive than the published report.24 The limit of quantification was not improved by using nanoLC system A, mainly due to the ratio of analyte vs isotope standard in the positive control experiments. Assay Validation. The precision of the assay was evaluated by analyzing dG-gx-dC and dG-gx-dA in human leukocyte DNA (50 μg) in triplicates per day for experiment performed on three different days. The relative standard deviations of intraday analyses for both adducts were within 10% and those for interday analyses were within 5%, indicating the high reproducibility of the assay (Table 2). The accuracy of the assay

nucleosides, (4) enrichment of the cross-linked adducts through a disposable reversed phase SPE column, and (5) nanoLC−NSI/MS/MS analysis under the highly selected reaction monitoring (H-SRM) mode (Scheme 1). Scheme 1

To achieve accurate quantification of trace analyte in an assay, an appropriate internal standard is needed to compensate for the analyte loss in the multistep assay procedures as well as for the matrix effect. The ideal internal standard in MS-based analysis is the stable isotope analogue of the analyte, because it has identical physical and chemical properties as the analyte except for its mass.25−27 The isotope-labeled internal standard functions to identify the analyte peak in the complex chromatogram. It also serves as a carrier for the small amount of analyte through the multistep procedures because the amount of analyte in samples can vary widely, and a trace amount of analyte would suffer from poor recovery without the presence of the isotope-labeled internal standard.28 A fixed amount of [15N5]dG-gx-dC and [15N5]dG-gx-dA (100 pg each) was added to a solution containing DNA, which was then enzymatically hydrolyzed to its nucleosides using the appropriate enzymes under the optimized conditions.24 The hydrolytic enzymes were removed by three different methods: (A) 0.22 μm nylon syringe filter, (B) 0.22 μm PTFE spin column, or (C) precipitation with cold acetone. Method C was chosen for the assay because the recovery of [15N5]dG-gx-dC and [15N5]dG-gx-dA (65% and 68%, respectively) was the highest and the chromatograms showed the least interference. Two DNA extraction kits from different venders (Qiagen and Viogen) were used in this study. The efficacies of these two kits on the same blood sample were evaluated, and similar adduct levels were obtained (data not shown). The average amount of DNA isolated from these kits was 23.1 ± 7.9 (±SD) μg/mL in 30 blood samples. The isolated leukocyte DNA was enzymatically hydrolyzed to its nucleosides. Because of structural alteration of the DNA by the cross-linking, hydrolysis of the cross-linked adducts by the hydrolytic enzymes can be much less efficient than that of unmodified normal nucleosides. The hydrolysis efficiency varies greatly according to our previous finding.24 The optimized conditions employed in this study were modified from those used for other DNA crosslinks.29,30 The isotope-labeled standards can adjust for loss of the target analytes during the hydrolysis and enrichment steps by a SPE 13085

DOI: 10.1021/acs.analchem.6b04296 Anal. Chem. 2017, 89, 13082−13088

Article

Analytical Chemistry Table 2. Precision in Quantification of dG-gx-dC and dG-gxdA by nanoLC-NSI/MS/MS Analysisa cross-link levels (adducts/108 nucleotides) mean ± SD (RSD)b

dG-gxdC dG-gxdA

day 1

day 2

day 3

interday variation

0.29 ± 0.03 (9.2%) 1.07 ± 0.08 (7.5%)

0.27 ± 0.01 (5.2%) 1.03 ± 0.09 (8.8%)

0.28 ± 0.02 (6.0%) 1.00 ± 0.09 (9.1%)

0.28 ± 0.01 (2.0%) 1.03 ± 0.04 (3.4%)

Each experiment started with 50 μg of human leukocyte DNA, and an equivalent of 10 μg of DNA hydrolysate was subjected to the nanoLC−NSI/MS/MS analysis. bLevels of these cross-links are presented as mean ± standard deviation (SD) from triplicate experiments. The relative standard deviation (RSD) is expressed in parentheses. a

was validated by adding known amounts of dG-gx-dC (2.0, 4.0, and 6.0 pg) and dG-gx-dA (6.0, 12, and 18 pg) to human placental DNA (50 μg) and analyzing for their levels in triplicates as described above. Linear regression gave a yintercept of 0.46 pg and 2.35 pg for dG-gx-dC and dG-gx-dA, respectively, with high correlation coefficients (Figure S4, Supporting Information). These amounts are very close to 0.44 pg and 2.33 pg of dG-gx-dC and dG-gx-dA, respectively, in 50 μg of human placental DNA without the addition of standards, which correspond to 0.52 and 2.65 adducts in 108 total normal nucleotides. Quantification of dG-gx-dC and dG-gx-dA in Human Leukocyte DNA. Measuring DNA adducts from circulating blood leukocyte DNA can offer a minimally invasive biomarker to reflect the extent of DNA damage of the whole body. The use of leukocyte DNA for biomonitoring requires a highly specific, sensitive, and accurate assay because of the low adduct levels. Only 50 μg of leukocyte DNA was used in this SID nanoLC−NSI/MS/MS system. As shown in Figure 1A, the stereoisomers of dG-gx-dC separated into two peaks eluting at 20.32 and 20.62 min, whereas those of dG-gx-dA eluted as one peak at 21.72 min. A selected reaction monitoring transitions from m/z 535.2 → 303.0 ([MH]+ → [MH − 2 deoxyribose]+) was chosen for quantification of dG-gx-dC, whereas m/z 559.2 → 443.0 ([MH]+ → [MH − deoxyribose]+) was used for dG-gx-dA because they showed the lowest amount of interference in the chromatograms of human leukocyte samples. The corresponding transitions were used for their isotopomers. Thus, dG-gxdC and [15N5]dG-gx-dC were quantified by the loss of two deoxyribose moieties, whereas dG-gx-dA and [15N5]dG-gx-dA were quantified by loss of a single deoxyribose unit. The adduct levels of this sample were 4.05 dG-gx-dC and 1.68 dG-gx-dA in 108 normal nucleotides. Another set of SRM transitions was used for qualitative assurance (qualifier) for the presence of these adducts (Scheme 1 and Figure 1B). As we have reported previously, the types and amounts of hydrolytic enzymes used can affect the amount of DNA adducts released.31−33 Thus, fixed amounts of DNA and hydrolytic enzymes were used for each sample throughout this study. Because of the low levels of the cross-linked adducts, 50 μg of leukocyte DNA was used, which can be obtained from 2−3 mL of blood, making the assay clinically feasible. Leukocyte DNA was isolated from venous blood of 38 T2DM patients and 39 nondiabetic control subjects, using a blood DNA extraction kit, and quantified. A positive control experiment was performed,

Figure 1. nanoLC−NSI/MS/MS chromatograms for dG-gx-dC and dG-gx-dA in a human leukocyte DNA (sample No. N2) analyzed using the H-SRM transitions for (A) quantification and (B) qualification. (RT, retention time; MA, manual integration; NL, normalized intensity level).

which contained the buffer solutions and the enzymes, but not DNA, and was incubated under the same conditions as the samples. After the DNA had been analyzed using the assay procedures described above, the adduct levels were quantified by subtracting the ratio of the analyte versus the internal standard of the sample from that of the positive control experiment, and interpolated into the calibration curve as listed in Table S1, Supporting Information. The presence of a small but detectable amount of adducts in the positive control experiment indicates that the adducts may have been formed artificially during sample workup. Samples with a ratio of analyte versus the internal standard equal to or less than that of the positive control experiment were marked as being not detectable (ND) and calculated as zero for the mean and for statistical analysis. All samples with a quantifiable adduct level also showed detectable adducts in the SRM transitions for the qualifier. 13086

DOI: 10.1021/acs.analchem.6b04296 Anal. Chem. 2017, 89, 13082−13088

Article

Analytical Chemistry Statistical analysis showed that the levels of dG-gx-dC and dG-gx-dA are significantly higher in the T2DM patients than in the healthy subjects (p < 0.0001 and p = 0.0067, respectively, using the Mann−Whitney U-test) (Figure 2A). Because glyoxal

Table 3. Levels of dG-gx-dC and dG-gx-dA in Human Leukocyte DNA adducts level (mean ± SD) (adduct in 108 nucleotide)

dG-gx-dC dG-gx-dA

Mann− Whitney U-test

multiple regression HbA1cb

T2DM (n = 38)

healthy (n = 39)

p valuea

p valuea

1.94 ± 1.20 2.10 ± 1.77

0.83 ± 0.92 1.05 ± 0.99