Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Nov 24.
Published in final edited form as: Rapid Commun Mass Spectrom. 2021 Jul 15;35(13):e9095. doi: 10.1002/rcm.9095

DNA Adductomics by mass tag prelabeling

Poguang Wang 1, Elisabeth Roider 2, Michael E Coulter 3, Christopher A Walsh 4, Caitlin S Kramer 5, Penny J Beuning 5, Roger W Giese 1
PMCID: PMC10668917  NIHMSID: NIHMS1933272  PMID: 33821547

Abstract

Rationale:

As a new approach to DNA adductomics, we directly reacted intact, double-stranded (ds)-DNA under warm conditions with an alkylating mass tag followed by analysis by liquid chromatography/mass spectrometry. This method is based on the tendency of adducted nucleobases to locally disrupt the DNA structure (forming a “DNA bubble”) potentially increasing exposure of their nucleophilic (including active hydrogen) sites for preferential alkylation. Also encouraging this strategy is that the scope of nucleotide excision repair is very broad, and this system primarily recognizes DNA bubbles.

Methods:

A cationic xylyl (CAX) mass tag with limited nonpolarity was selected to increase the retention of polar adducts in reversed-phase high-performance liquid chromatography (HPLC) for more detectability while maintaining resolution. We thereby detected a diversity of DNA adducts (mostly polar) by the following sequence of steps: (1) react DNA at 45°C for 2 h under aqueous conditions with CAX-B (has a benzyl bromide functional group to label active hydrogen sites) in the presence of triethylamine; (2) remove residual reagents by precipitating and washing the DNA (a convenient step); (3) digest the DNA enzymatically to nucleotides and remove unlabeled nucleotides by nonpolar solid-phase extraction (also a convenient step); and (4) detect CAX-labeled, adducted nucleotides by LC/MS2 or a matrix-assisted laser desorption/ionization (MALDI)-MS technique.

Results:

Examples of the 42 DNA or RNA adducts detected, or tentatively so based on accurate mass and fragmentation data, are as follows: 8-oxo-dGMP, ethyl-dGMP, hydroxyethyl-dGMP (four isomers, all HPLC-resolved), uracil-glycol, apurinic/apyrimidinic sites, benzo[a]pyrene-dGMP, and, for the first time, benzoquinone-hydroxymethyl-dCMP. Importantly, these adducts are detected in a single procedure under a single set of conditions. Sensitivity, however, is only defined in a preliminary way, namely the latter adduct seems to be detected at a level of about 4 adducts in 109 nucleotides (S/N ~30).

Conclusions:

CAX-Prelabeling is an emerging new technique for DNA adductomics, providing polar DNA adductomics in a practical way for the first time. Further study of the method is encouraged to better characterize and extend its performance, especially in scope and sensitivity.

1 |. INTRODUCTION

DNA adducts are damaged nucleotides in DNA as a consequence of its exposure to genotoxic agents or conditions. Measurement of multiple (especially many) DNA adducts in a single procedure is referred to as “DNA adductomics”, a subject that has been reviewed.1–9 In the leading analytical technique for this purpose, typically the following sequence of steps takes place starting with a biosample: (1) isolate the DNA; (2) digest it to deoxynucleosides enzymatically; (3) remove the enzymes and inject the deoxynucleosides for analysis by ultra-performance liquid chromatography/tandem mass spectrometry (UPLC/MS/MS) using mild collision-induced dissociation (CID) conditions that release the sugar (116 u) as a neutral; and (4) observe peaks for adducts as protonated nucleobases. While this method is highly successful, including important extension in recent years to paraffin-fixed tissue samples,9,10 it has some shortcomings. First of all, response is adduct-dependent. While a limit of quantitation (LOQ) of about 3 adducts in 109 nucleotides can be reached using 2 μg of DNA for the most favorable (some bulky) adducts,9–14 limits of detection (LODs) can vary widely. For example, it was reported that LODs for different adducts ranged from 0.02 to 23.7 adducts in 108 nucleotides, a 1000-fold range.15 Variation in ionization efficiency in the ion source along with differences in ease of sugar loss probably explains most of this variation. Indeed, not all adducts give loss of sugar in the method,16 including phosphate adducts.17

Second, polar adducts in the above sugar-loss method elute early in the usual reversed-phase LC separation, where there is much noise, so they are not measured along with bulky adducts. Extra effort thereby may be necessary to measure polar adducts, such as two solid-phase extractions prior to the LC separation even for a single, targeted adduct.18 While 12 polar DNA adducts were measured in a single procedure,19,20 each adduct had to be collected separately from a first HPLC separation prior to subsequent injection again into a LC/MS system. An API 3000 triple quadrupole mass spectrometer was employed with analyte-dependent detection parameters. The LOD was about 1 adduct in 107 nucleotides. Third, different adducts tend to require different LC mobile phase conditions and/or different MS conditions for optimum sensitivity. Fourth, the neutral loss of 116 u for adduct detection can come from noise, especially at lower adduct levels. Fifth, delayed addition of a stable isotope nucleoside internal standard is usually employed, which can compromise absolute quantitation.

DNA adductomics also can be accomplished by mild acid depurination/LC/MS. This technique has been practiced, for example, by Hemeryck et al.21 Four targeted guanine adducts (methyl, carboxymethyl, malonaldehyde, and methylhydroxypropano) were detected at an LOQ in the range of 4 to 22 adducts in 108 nucleotides, based on spiking authentic, modified nucleobases at the ng level into 100 μg of DNA. Overall, in the samples tested (comprising chemically treated calf thymus DNA samples and several colon tumor tissues) there was tentative detection of 20 other small adducts.

Another technique that is useful for DNA adductomics is “32P-postlabeling”, which has been reviewed.3,22 In this method the following sequence of steps usually takes place once DNA has been isolated: (1) digest the DNA enzymatically to deoxynucleoside-3′-phosphates; (2) label the latter radio-enzymatically with [32P] adenosine triphosphate; (3) conduct a chromatographic separation, usually by multidimensional thin-layer chromatography (TLC) under conditions that first wash conventional deoxynucleotides out of a retention region of interest prior to migration of the adducts; and (4) measure DNA adducts as radioactive spots by storage phosphor imaging. This technique has been employed for many years and can provide high sensitivity. Its major disadvantages are that the yield of the labeling reaction is adduct-dependent; it is difficult to incorporate internal standards; and it is not easy to establish the identity of a radioactive adduct TLC spot (or radioactive HPLC peak when this technique is used instead of TLC).

Here we introduce a new method for DNA adductomics that we refer to as “CAX-Pre-labeling”. This name was selected largely because this method shares the feature of labeling with 32P-postlabeling. Also there are contrasts: labeling in our technique is done with a CAX mass tag rather than a radiolabel; our labeling is done prior to enzymatic digestion of the DNA to nucleotides; and detection in our method is done by mass spectrometry. CAX-Prelabeling is presented as a complementary technique to current practice for DNA adductomics, especially because it encompasses a diversity of polar DNA adducts for the first time in a single procedure under a single set of conditions throughout.

2 |. EXPERIMENTAL

2.1 |. Materials

N-(2-[Bromomethyl]benzyl)-N,N-diethylethanaminium bromide, that we designate as CAX-B, was synthesized as described.23 Benzyl bromide, methanesulfonic acid ethyl ester, chloroacetaldehyde, 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (Lomustine), styrene oxide, bromoacetic acid, calf thymus DNA, triethylamine, and α-cyano-4-hydroxycinnamic acid (CCA) were from Sigma (St Louis, MO, USA). Microcentrifuge tubes, pipette tips, HPLC grade methanol, and acetonitrile (ACN) were from Fisher Scientific (Pittsburgh, PA, USA). All materials were used as received. CCA matrix solution was 5 mg/mL in 50% ACN (v/v). Amicon ultra centrifuge filters, Ultracel–3K, were from Millipore (Billerica, MA, USA). DNA adducts in calf thymus DNA were formed in test tube experiments as described: benzyl, ethyl and hydroxyethyl,24 etheno,25 styrene oxide,26 benzo[a] pyrene,27 and benzoquinone.28 Oxidative DNA adducts were generated from treatment of DNA with bromoacetic acid: 100 μL of calf thymus DNA at 0.5 mg/mL was mixed with 2 μL of bromoacetic acid, and kept at 37°C for 2 h. Brain tissues were obtained from the NIH NeuroBioBank at the University of Maryland. Frozen postmortem tissues from three neurologically normal individuals, UMB4638, UMB1465, and UMB4643, were obtained as part of a previous study.29 Samples were processed according to a standardized protocol (https://www.medschool.umaryland.edu/btbankold/brain-protocol-methods/brain-sectioning—protocol-method-2/) under the supervision of the NIH NeuroBioBank ethics guidelines.

2.2 |. DNA extraction

DNA was extracted from human skin obtained under IRB approval using the Qiagen Blood & Cell Culture DNA Midi Kit (Cat No./ID: 13343).

2.3 |. Instrumentation and calculations

The model 5800 MALDI-TOF/TOF mass spectrometer was from AB-SCIEX (Framingham, MA, USA). The resolution is ~20,000 in the TOF positive reflectron mode with a delay time of 120 ns, and ~3000 in the TOF/TOF positive 1 kV mode with CID gas off and a 1 Da isolation window. The capillary LC (CapLC) system for detection of CAX-labeled DNA adducts by CapLC/MALDI-TOF-MS or CapLC/MALDI-TOF/TOF-MS was a Dionex Ultimate (Thermo Scientific). Column: Acclaim PepMed C18 (0.3 × 150 mm, 2 μm). Mobile phase solvent: A, 8% ACN (v/v); and B, ACN. Mobile phase conditions: 3% B for 0–4 min, then up to 90% B over 60 min at 4 μL/min. Post-LC steps: collect 4 droplets/min onto a MALDI plate (384 format), manually add 0.5 μL of CCA matrix solution to each spot, and conduct MALDI-TOF-MS or MALDI-TOF/TOF-MS. For detection of DNA adducts by LC/ESI-MS2, a Dionex Ultimate UPLC system fitted with an Aquasil C18 column (1 × 250 mm, 5 μm) was coupled to an LTQ Orbitrap XL: 0.1% formic acid for channel A and 0.1% formic acid, 99.9% acetonitrile for channel B, 8% B for 0–3 min, and up to 90% B over 40 min at 50 μL/min; ESI capillary temperature of 275°C, sheath gas flow of 10, source voltage of 4.9 kV, source current of 100 μA, capillary voltage of 49 V, and tube lens of 95 V; full-scan MS spectra (300–1500 m/z) at a resolution of 30,000, and two micro-scans of a maximum injection time of 80 ms, followed by top 2 data-dependent ion trap MS/MS with CID energy at 28, and a 2 Da isolation window. Energy minimization of gas-phase ions was performed using the MM2 method (Chem3D Pro 5.0: minimum RMS gradient of 0.1).

2.4 |. Reaction of CAX-B with DNA

CAX-B in 50% ACN (20 mg/mL), with Et3N (30 μL/mL), is mixed 1:2 (v/v) with a solution of DNA (100 μL at ~1 mg/mL in water). After 2 h at 45°C, the reaction mixture is loaded into an Amicon ultra centrifuge filter (Millipore, UFC500396, 0.5 mL, 3000 NMWL), followed by centrifugation at 12,000 rpm for 15 min (13,800 g, Thermo-Fisher AccuSpin Micro 17) and washing similarly with 5 × 300 μL of 10% ACN in water (to remove low-mass chemical background), while retaining DNA. The retained DNA is recovered by rinsing the internal area of the filter with 50 μL of water, and centrifuging the reversed filter for 3 min. The recovered sample solution is subjected to a two-step enzymatic digestion (nuclease P1 and phosphodiesterase I) as described,30 loaded onto an OASIS HLB 1 cc cartridge (Waters, WAT094225), washed with 2 × 1 mL of water (to remove untagged normal nucleotides), and eluted with 0.8 mL methanol/water (8:2, v/v). After evaporation in a Speed Vac (Thermo, SPD111V), redissolving in 12 μL of ACN/water (6:94), and centrifuging at 12,000 rpm for 5 min, 5 μL of the clear solution is injected into the LC/MALDI-TOF/TOF-MS or LC/LTQ Orbitrap XL system described above. Alternatively, 2 μL were combined with 2 μL of matrix and 0.7 μL was subjected to analysis by MALDI-TOF/TOF-MS.

3 |. RESULTS AND DISCUSSION

CAX-Prelabeling takes place as illustrated in Figure 1, and relies on the tendency of DNA adducts to distort/destabilize the local structure of double-stranded (ds)-DNA into what is sometimes referred to as a “DNA bubble”. In principle, this can make this region of DNA, and its resident DNA adduct, more susceptible to further reaction, as with an alkylating mass tag, as practiced here. We selected CAX-B (cationic xylyl bromide) as the mass tag for this endeavor. As reported previously, CAX-B can enable highly sensitive, specific detection, especially in a tandem mass spectrometer, since it is an anchimeric-assisted, neutral loss mass tag.23 Thereby, it gives an analyte-characteristic [M + H – N(CH2CH3)3]+ ion under mild collision-induced dissociation (CID) conditions. An analyte-characteristic ion arises since a neutral loss group is built into the tag.

FIGURE 1.

FIGURE 1

Open in a new tab

Scheme for detecting a DNA adduct by CAX-Prelabeling

To conduct CAX-Prelabeling, we react DNA with CAX-B at 45°C (to promote DNA bubbling), and in the presence of triethylamine (to enhance labeling of active hydrogen sites). At the end of the 2-h reaction, the DNA is washed by filtration to remove residual CAX-B and its hydrolysis products, and digested enzymatically to nucleotides. This mixture is applied to an OASIS HLB cartridge, where non-tagged nucleotides, as highly polar anions, elute readily. CAX-labeled adducts and CAX-labeled canonical nucleotides tend to be retained since they are zwitterions, and CAX has some nonpolar structure. Subsequently, aqueous methanol is applied to the cartridge and the eluted zwitterions are analyzed by UPLC/MS/MS. DNA adducts tend to stand out in MS2 due to the loss of both the triethylamine moiety (a characteristic of CAX), and the phosphate-deoxyribose moiety, under CID conditions.

Five limitations of the CAX-Prelabeling assay for DNA adducts are immediately apparent. (1) Some DNA adducts stabilize the local structure of ds-DNA,31 and thereby are anticipated to resist labeling. (2) Normal nucleotides especially in the bubbled DNA regions are expected to be labeled as well, increasing background signals. This is indicated by the scheme shown in Figure 1. (3) A given DNA adduct may have multiple active hydrogens (whether on the base or adducted chemical) that can undergo labeling, decreasing the signal by dividing it over multiple chromatographic peaks. (4) CAX-labeling of a DNA adduct might interfere with subsequent enzymatic digestion of the DNA to monomers for detection by mass spectrometry. (5) The yield of the labeling step can be adduct-dependent, compromising sensitivity and absolute quantitation.

These disadvantages potentially can be mitigated by the following, corresponding considerations: (1) Adducts that stabilize the ds-DNA structure probably are rare.31 Indeed, nucleotide excision repair, which primarily recognizes DNA bubbles, recognizes a great diversity of DNA adducts.32–34 Further, nonphysiological conditions might increase exposure of such adducts. (2) Peaks from normal nucleotides provide calibration and carrier contributions. (3) Multiple peaks from an adduct can facilitate its characterization; further, each peak should come from a mono-labeled adduct due to charge repulsion (the positive charge deposited on the adduct by the first CAX moiety tends to repel other CAX-B molecules). (4) Potentially an adduct can be detected in a CAX-labeled dinucleotide or trinucleotide form; further, a diversity of nucleases is available to increase the opportunity to form a CAX-mononucleotide from a given adduct (or in planned future studies, from a CAX-mononucleoside). (5) The high sensitivity of the CAX tag can enable adequate sensitivity even when the labeling yield is low, and relative quantitation, which is valuable in DNA adductomics, is provided.

3.1 |. Assay advantages

There are also advantages or potential advantages of the CAX-Prelabeling for DNA adducts, in addition to those cited above. (1) It is a DNA adductomics method that simultaneously detects polar and nonpolar adducts, assuming that the detection of a benzo[a] pyrene adduct means that many nonpolar adducts will be detected. (2) The residual mass tag can be removed easily from the CAX-reacted DNA sample by membrane filtration or DNA precipitation (which helps to make the method practical). (3) When the labeled DNA is digested to nucleotides, CAX-labeled adducts will be zwitterions (unless the adducted chemical itself carries a charge), in contrast to the anionic canonical nucleotides, helping to achieve broad enrichment of adducts prior to detection, and further helping to make the assay practical. (4) Artifactual production of DNA adducts after the labeling step is much less of a concern, since it will tend to affect only a tiny fraction of the CAX-labeled adducts vs artifactual production on a larger scale from the overall DNA. This is important for detection of oxidative DNA adducts, which tend to form artifactually. (5) Triethylamine is an antioxidant, further reducing the likelihood of artifactual oxidative DNA adducts. (6) CAX-labeling up-shifts the masses of DNA adducts by about 200 Da, reducing interferences from low-mass background ions in the mass spectrometer. (7) When CAX-labeled nucleotides are subjected to CID in the mass spectrometer, the following two types of analyte-specific product ions tend to form: M – 196 from loss of deoxyribose phosphate, and M – 297 from combined loss of triethylamine and deoxyribose phosphate. These ions distinguish DNA adducts from ribonucleotide species (a feature similarly shared with the sugar-loss method). (8) Not only the simple steps in the method, but also the use of a single set of conditions throughout for at least many adducts additionally helps to make the method practical. (9) The method is anticipated to detect phosphate adducts, at least when digestion is extended to nucleosides,17 as long as one of the bases in the dinucleoside product provides an active hydrogen for CAX-labeling. (10) There is a good opportunity to shift the mass of a detected adduct to deal with noise as needed, since CAX-B-d8 is readily available (it is employed here in Figure S1, supporting information). (11) The method is anticipated to detect adducts that resist neutral loss of sugar, since a neutral loss moiety is built into the CAX mass tag. (12) CAX-B is easily synthesized in a single step from inexpensive, common reagents.23 (13) CAX has a moderate nonpolar structure, enough to increase the retention time of polar analytes into a less noisy region of a reversed-phase LC separation while retaining resolution. For example, as presented below, four isomers of CAX-labeled hydroxyethyl-dGMP are resolved in this way. (14) CAX-labeled adducts can be detected by both MALDI-MS and ESI-MS.

We now turn our attention to examples of DNA adducts that we have detected by the scheme shown in Figure 1. These examples illustrate some of the above disadvantages and advantages of our method.

3.2 |. 8-Oxoguanine

Detection of 8-oxo-dGMP in commercial calf thymus DNA by CAX-Prelabeling with LC/MALDI-TOF/TOF MS is shown in Figure 2B, along with detection of its isomer, GMP, in Figure 2A. (We have never encountered a DNA sample completely devoid of RNA.) For both spectra, the precursor ion at 567 Da is selected for CID. These precursor ion spectra are shown in the insets along with the CCA matrix blank spectra. In Figure 2A, CAX-GMP is detected as four product ions (m/z 254, 355, 466 and 487), while CAX-8-oxo-dGMP is detected similarly in Figure 2B by ions at m/z 270, 371, 466 and 487. As seen, the former pair of ions in each case (at lower masses) distinguishes these compounds. The fragmentation explaining these ions is shown in this figure. The placement of the CAX moiety on the O6 position of guanine in each case is arbitrary. Perhaps the 8-oxo atom of CAX-8-oxo-dGMP, by an inductive mechanism, enhances cleavage of the glycolytic bond relative to the corresponding bond in CAX-GMP when [M + H – neutral fragment]+ is generated (m/z 371 in B is much more intense than m/z 355 in A). The HPLC conditions separated CAX-labeled 8-oxo-dGMP and GMP (data not shown, retention times 17.7 and 20.3 min, respectively). While these compounds have similar polarities, their unique product ion spectra distinguish them. We made no effort to quantify the compounds.

FIGURE 2.

FIGURE 2

Open in a new tab

MALDI-TOF/TOF-MS spectra and peak assignments for CAX-labeled GMP and 8-oxo-dGMP, having the same precursor mass (C24H36N6O8P+, exact mass: 567.233), and derived from calf thymus DNA (which contains some RNA) according to the scheme shown in Figure 1. A. Inset: A precursor ion at m/z 567.232, along within M + 1 and M + 2 isotopes, by MALDI-TOF-MS; inset B, corresponding matrix blank. Peak assignments for the MALDI-TOF/TOF-MS spectrum of GMP are shown on the left. B, Inset: Corresponding data for 8-oxo-dGMP are shown in A and B, along with corresponding MS2 data and interpretation

Figure 3 shows the detection of 8-oxo-dGMP along with three other modified nucleotides that we detected when CAX-Prelabeling was applied to DNA from a sample of human skin. The selected ion monitoring (SIM) mass chromatograms are extracted from LC/LTQ Orbitrap full scan data in either the MS1 or MS2 mode. The four nucleotides detected, as CAX derivatives, are as follows: A, 8-oxo-dGMP; B, GMP; C, fapy-dGMP; and D, dUMP. An effective LC retention time (tR; around 19 min, escaping the early noise region while maintaining resolution) is seen for all of these compounds, largely due to the moderate nonpolar structure of the CAX moiety, and the fact that the CAX-labeled nucleotides are zwitterions. Note that CAX-GMP (tR = 18.9 min) once again (LC/MALDI above, LC/Orbitrap here) is readily separated from CAX-8-oxo-dGMP (tR = 19.4 min), even though they are isomers. Perhaps the peak for CAX-fapy-dGMP (C) is broader since, relative to the other compounds (A, B, D), this compound is less compact in a way that gives it some surfactant behavior.

FIGURE 3.

FIGURE 3

Open in a new tab

Detection of four nucleotides in human skin DNA using LC/LTQ Orbitrap MS in the SIM mode from full scan MS1 or MS2 data. A, SIM MS2 of CAX-8-oxo-dGMP (monitoring the product ion at m/z 270 from 567 Da); B, SIM MS2 of CAX-GMP (monitoring the product ion at m/z 254 from 567 Da); C, SIM MS1 of CAX-fapy-dGMP (m/z 569, exact mass 569.249); and D, SIM MS1 of CAX-dUMP (m/z 512, exact mass 512.216)

3.3 |. Uracil-glycol

Figure 4 shows the detection of CAX-dUMP-glycol by MALDI-TOF-MS at 546 Da in the inset, along with an [M + H – neutral fragment]+ ion by MALDI-TOF/TOF-MS in the main frame of the figure. In the latter spectrum assigned ions are observed at m/z 466, 445, 350, 249, 231 and 81. The minor ion at m/z 100 is from diethylethenylmethylamine (arising from loss of triethylamine via an elimination reaction23). Perhaps this adduct forms in DNA by hydrolysis of dCMP to dUMP35 followed by oxidation of the latter. The data shown are from analysis of human skin DNA.

FIGURE 4.

FIGURE 4

Open in a new tab

MALDI-TOF-MS spectrum (inset), and MALDI-TOF/TOF-MS spectrum (main frame), of CAX-glycol-dUMP

3.4 |. Hydroxymethylcytosine

It is known that the level of hmdCMP in some tissues is relatively high.36 The inset in Figure 5 shows the detection of CAX-hmdCMP from brain DNA by MALDI-TOF-MS. The corresponding MS2 spectrum in the main frame shows a product ion at 222 Da. This indicates that the tagging is on oxygen, but which oxygen is unknown. When CAX-dCMP and CAX-mdCMP are selected for MS2 analysis, no product ion at m/z 222 can be observed (data not shown).

FIGURE 5.

FIGURE 5

Open in a new tab

MALDI-TOF/TOF-MS spectrum of CAX-hydroxymethyl-dCMP from brain DNA. The inset shows the corresponding MALDI-TOF-MS spectrum

3.5 |. Benzoquinone (BQ) adducts of cytosine and hydroxymethylcytosine

Previously we reported the detection of BQ-dCMP in a BQ-treated DNA sample which was estimated to contain 1 residue of this adduct in 105 nucleotides.28 Testing this same DNA by CAX-Prelabeling with LC/MALDI-TOF/TOF MS also reveals this adduct (Figure S2, supporting information). Also detected now are corresponding BQ-methyl-dCMP (Figure 6A) and BQ-hydroxymethyl-dCMP (Figure 6B) as CAX derivatives. The latter adduct has not been reported previously. mdCMP is about 4% of dCMP in the DNA that we tested. If we arbitrarily assume that BQ-hydroxymethyl-dCMP is about 1% of BQ-methyl-dCMP, then we have detected about 4 BQ-hmdCMP in 109 nucleotides by CAX-Prelabeling at a signal-to-noise (S/N) ratio ~30 (Figure 6B). Peak assignments are shown in this figure.

FIGURE 6.

FIGURE 6

Open in a new tab

MALDI-TOF/TOF-MS spectra of A, CAX-BQ-mdCMP and B, CAX-BQ-hmdCMP

3.6 |. Hydroxyethyl-guanine

We reacted calf thymus DNA with 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (Lomustine24) to form hydroxyethyl adducts, and subjected it to CAX-Prelabeling with detection by LC/MALDI-TOF/TOF-MS. This resulted in four chromatographic peaks from four isomers of CAX-labeled hydroxyethyl guanine adducts at retention indices (spot numbers on the MALDI plate, where the central spot for a LC peak is reported) of 42 (C), 44 (A), 55 (D), and 58 (B), each having a precursor ion at 595 Da. The corresponding MS2 spectra for each of these spots are shown in Figure 7.

FIGURE 7.

FIGURE 7

Open in a new tab

MALDI-TOF/TOF-MS spectra of four CAX-hydroxyethyl-dGMP adducts

We will now present arguments for the tentative structures shown in Figure 7 based on the fragmentation patterns. The compounds are named according to the subparts of the figure: 7A, 7B, 7C, 7D. Compound 7A is assumed to be CAX-N7-hydroxyethyl-dGMP, having a net charge of +1 as illustrated. Cleavage of the glycolytic bond in this compound is favored in TOF/TOF, we assume, by the positive charge on N7, leading to the major product ion, M – 196 (m/z 399). Loss of phosphate is minimal, forming m/z 515, probably because this loss is hindered by the nearby positive charge on the nucleobase. An MM2 energy minimizing calculation predicts a negatively charged oxygen of the phosphate group close to the positively charged N7 position (shown in Figure 8). In compound 7B by steric effect mainly, the hydroxyethyl moiety is proposed to be on N1, the CAX moiety on N2, and the phosphate is assumed to be neutral, in agreement with the observed fragments and the proposed fragmentation pathway. The observation of m/z at 355 (CAX-G) and 254 (CAX-G – 101) in Figures 7C and 7D, respectively, instead of the ions at m/z 399 and 298 shown for compounds 7A and 7B, indicates that the hydroxyethyl group is on the phosphate in compounds 7C and 7D. This is further supported by the proposed fragmentation patterns. Charge repulsion may explain the relatively higher intensity of m/z 494 in 7C than in 7D (MM2 calculation, as shown in Figure 8 and Figure S3, supporting information, predicts 7C and 7A to have a similar configuration, and likewise 7D and 7B). The relative chromatographic retention times also are consistent with these assignments: 7A elutes 3.5 min earlier than 7B, and 7C elutes 3.25 min earlier than 7D. Cations, being more polar, should elute earlier than the corresponding zwitterions from a C18 column.

FIGURE 8.

FIGURE 8

Open in a new tab

Energy minimization by the MM2 method (Chem3D pro 5.0: Minimum RMS gradient of 0.1) of the N1 and N7 isomers of the CAX-hydroxyethyl-dGMP adducts

3.7 |. Apurinic/apyrimidinic sites in DNA

Figure 9 shows the detection of apurinic/apyrimidinic (AP) sites in human skin DNA by CAX-Prelabeling with detection by LC/MALDI-TOF-MS and LC/MALDI-TOF/TOF-MS. The inset shows the mass spectrum of an LC peak recorded from one of the three MALDI spots observed to contain this compound at about 14.7 min (data not shown) when m/z 418 is monitored. This relatively low retention time corresponds to a very polar adduct; indeed, the peak elutes earlier than that of CAX-dCMP (about 16.9 min). As seen, the detected mass (418.199) fully agrees with that of the assignment (418.199). In the main frame of the figure, the fragmentation data by LC/MALDI-TOF/TOF-MS confirms the structure shown (aside from the stereochemistry of the glycosidic bond), based on peaks at m/z 121, 222 and 317. The abundance of this peak at m/z 418 for AP is about two-fold higher than that from fC or hmC, but about two-fold lower than from fapy A. AP sites also have been detected in other studies by another form of mass tag prelabeling. For example, Chen et al37 used a pyridinyl-hydroxylamine mass tag reagent to label such sites in their ring-opened aldehyde form in DNA, yielding an oxime for subsequent detection by LC/MS2.

FIGURE 9.

FIGURE 9

Open in a new tab

Detection of apurinic/apyrimidinic (AP) sites in human skin DNA by CAX-Prelabeling with detection by LC/MALDI-TOF-MS (inset) and LC/MALDI-TOF/TOF-MS (main frame)

3.8 |. Benzo[a]pyrene-guanine

A sample of DNA containing benzo[a]pyrene-guanine adducts, prepared as described,27 was subjected to CAX-Prelabeling. The CAX-labeled nucleotides derived from 0.5 μg of DNA were injected into an LC/Orbitrap LTQ setup. Monitoring the product ion (from the precursor ion at 853.3 ± 0.5 Da) at m/z 657.319 ± 0.05 Da, extracted from the Orbitrap MS2 full scan spectra, gave the SIM mass chromatogram shown in Figure 10A. Two peaks are seen, in agreement with the prior study.27 The full MS2 mass spectrum for the peak at 21.78 min, obtained at its maximum intensity, is shown in Figure 10B. A product ion, from loss of deoxyribose phosphate, is seen at m/z 657.318 (accurate mass 657.319). The peak at 23.29 min in Figure 10A gives the same result.

FIGURE 10.

FIGURE 10

Open in a new tab

A, SIM LC/LTQ Orbitrap chromatogram from CAX-Prelabeling applied to DNA treated with B[a]P-diolepoxide, monitoring the product ion at m/z 657.319 ± 0.05 Da derived from the precursor ion at m/z 853.3 ± 0.5 Da. B, MS2 spectrum obtained at the maximum of the peak at 21.78 min in A

3.9 |. Other adducts

Aside from data that we have already discussed above (Figures S1, S2, S3) other figures in the supporting information show the detection of other DNA adducts that we formed by conducting reactions of chemicals with calf thymus DNA in the test tube, along with some RNA species arising from the fact that RNA always (in our experience) is a contaminant of such DNA. The tentative detection of four oxidative adducts is shown in Figure S4: 6-oxo-TMP, 5-OH-dCMP, 5-OH-dUMP, and 5-OH-CMP. The detection of ethyl-dGMP, styrene oxide-dAMP, ethyl-fapy-dGMP, and benzyl-methyl-dCMP adducts is shown in Figure S5 and Figure S6 presents the detection of benzyl-hydoxymethyl-dCMP. Figure S7 shows the detection of benzoquinone-CMP and benzoquinone-5-hydroxy-dCMP from DNA treated with benzoquinone.

3.10 |. CAX-Prelabeling vs other adductome methods

In section 1 we compared CAX-Prelabeling with other methods for DNA adductomics in general terms. A more detailed comparison will require more experience with our emerging method, especially in regard to scope, sensitivity, ability to provide quantitation, and reproducibility. Additional DNA adducts need to be tested to better define the scope of our method, and its ability to cope with the problem of noise that holds back all DNA adductome methods at lower adduct levels. The scope of our method is particularly uncertain for bulky adducts since we have only tested one such species to date (benzo[a]pyrene). For sensitivity, it is encouraging that some background adducts have been detected in DNA from native biosamples (human brain, human skin, and calf thymus). Also, a benzoquinone-hydroxymethyl-dCMP adduct has been detected for the first time, and we provide the first adductome method that encompasses AP sites. A related limitation of our studies to date with CAX-Prelabeling is the limited information about its sensitivity. No doubt sensitivity will be adduct-dependent, although this is true for all DNA adductomic methods. A second aspect of sensitivity is the amount of DNA that can be processed. In this study we tested amounts of DNA ranging from about 1 to 100 μg. While a larger amount of DNA is not always available, it is helpful that our method can be readily applied to a larger amount of DNA. In regard to quantitation, it is difficult to obtain an authentic DNA sample with a known level of a given adduct, and this problem faces all adductomic methods including ours. (This explains the compromise of relying on delayed addition of monomeric adducts as internal standards in the sugar-loss method.) We therefore expect relative rather than absolute quantitation will characterize most DNA adductomics methods for some time. (Potentially this is quite useful, nevertheless, since some people with higher adduct levels might lower their cancer risk by lowering their DNA adducts.) This places a high demand on reproducibility. Our method might be advantageous in this respect since it detects all adducts within its scope in a single method under a single set of conditions.

4 |. CONCLUSIONS

CAX-Prelabeling is a new technique for DNA adductomics, providing polar DNA adductomics in a practical way for the first time. Based on its simultaneous detection of a benzo[a]pyrene adduct under the same conditions, the method might provide broader DNA adductomics than this, encompassing many nonpolar as well as polar adducts (to be studied). CAX-Prelabeling, especially with updated equipment, promises to become valuable as a complementary procedure in the challenging field of DNA adductomics, where various methods, each having certain advantages and disadvantages, will probably always be needed.

Supplementary Material

Prelabeling SI

ACKNOWLEDGMENTS

This work was funded by NIEHS Grant P42ES017198 to RWG; American Cancer Society Grant RSG-12-161-01-DMC to PJB; NU Office of the Provost Grant to PJB, RWG, and CSK; and an NU Office of the Provost Grant to Konstantin Khrapko, Dori Woods, and RWG.

Funding information

American Cancer Society, Grant/Award Number: RSG-12-161-01-DMC; National Institute of Environmental Health Sciences, Grant/Award Number: P42ES017198; NU Office of the Provost Grant

Footnotes

SUPPORTING INFORMATION

Additional supporting information may be found online in the Supporting Information section at the end of this article.

DATA AVAILABILITY STATEMENT

Data available on request from the authors.

REFERENCES

  • 1.Tretyakova N, Villalta PW, Kotapati S. Mass spectrometry of structurally modified DNA. Chem Rev. 2013;113(4):2395–2436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Geacintov NE, Broyde S (Eds). The Chemical Biology of DNA Damage. Weinheim: Wiley-VCH; 2010. [Google Scholar]
  • 3.Balbo S, Turesky RJ, Villalta PW. DNA Adductomics. Chem Res Toxicol. 2014;27(3):356–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Pottenger LH, Andrews LS, Bachman AN, et al. An organizational approach for the assessment of DNA adduct data in risk assessment: Case studies for aflatoxin B1, tamoxifen and vinyl chloride. Crit Rev Toxicol. 2014;44(4):348–391. [DOI] [PubMed] [Google Scholar]
  • 5.Liu S, Wang Y. Mass spectrometry for the assessment of the occurrence and biological consequences of the DNA adducts. Chem Soc Rev. 2015;44(21):7829–7854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Stornetta A, Zimmermann M, Cimino GD, Henderson PT, Sturla SJ. DNA adducts from anticancer drugs as candidate predictive markers for precision medicine. Chem Res Toxicol. 2017;30(1):388–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Sturla SJ. DNA adduct profiles: Chemical approaches to addressing the biological impact of DNA damage from small molecules. Curr Opin Chem Biol. 2007;11(3):293–299. [DOI] [PubMed] [Google Scholar]
  • 8.Villalta PW, Balbo S. The future of DNA Adductomic analysis. Int J Mol Sci. 2017;18(9):1870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Yun BH, Guo J, Bellarmi M, Turesky RJ. DNA adducts: Formation, biological effects, and new specimens for mass spectrometric measurements in humans. Mass Spectrom Rev. 2020;39(1–2):55–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Yun BH, Rosenquist TA, Nikolíc J, et al. Human formalin-fixed paraffin-embedded tissues: An untapped specimen for biomonitoring of carcinogen DNA adducts by mass spectrometry. Anal Chem. 2013; 85(9):4251–4258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Yun BH, Rosenquist TA, Sidorenko V, et al. Biomonitoring of aristolactam-DNA adducts in human tissues using ultra-performance liquid chromatography/ion-trap mass spectrometry. Chem Res Toxicol. 2012;25(5):1119–1131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Gu D, Turesky RJ, Tao Y, et al. DNA adducts of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine and 4-aminobiphenyl are infrequently detected in human mammary tissue by liquid chromatography/tandem mass spectrometry. Carcinogenesis. 2012;33(1):124–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Bessette EE, Spivack SD, Goodenough AK, et al. Identification of carcinogen DNA adducts in human saliva by linear quadrupole ion trap/multistage tandem mass spectrometry. Chem Res Toxicol. 2010; 23(7):1234–1244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Goodenough AK, Schut HAJ, Turesky RJ. Novel LC-ESI-MS/MSn method for the characterization and quantification of 2′-deoxyguanosine adducts of the dietary carcinogen 2-amino-1-methyl-6phenylimidazo[4,5-b]pyridine by 2-D linear quadrupole ion trap mass spectrometry. Chem Res Toxicol. 2007;20(2):263–276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Monien BH, Schumacher F, Herrmann K, Glatt H, Turesky RJ, Chesne C. Simultaneous detection of multiple DNA adducts in human lung samples by isotope-dilution UPLC-MS/MS. Anal Chem. 2014;87: 641–648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Singh R, Teichert F, Seidel A, et al. Development of a targeted adductomics method for the determination of polycyclic aromatic hydrocarbon DNA adducts using online column-switching liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom. 2010;24(16):2329–2340. [DOI] [PubMed] [Google Scholar]
  • 17.Ma B, Zarth AT, Carlson ES, et al. Identification of more than 100 structurally unique DNA-phosphate adducts formed during rat lung carcinogenesis by the tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone. Carcinogenesis. 2018;39(2):232–241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Balbo S, Meng L, Bliss RL, Jensen JA, Hatsukami DK, Hecht SS. Time course of DNA adduct formation in peripheral blood granulocytes and lymphocytes after drinking alcohol. Mutagenesis. 2012;27(4):485–490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Pang B, Zhou X, Yu H, et al. Lipid peroxidation dominates the chemistry of DNA adduct formation in a mouse model of inflammation. Carcinogenesis. 2007;28(8):1807–1813. [DOI] [PubMed] [Google Scholar]
  • 20.Taghizadeh K, McFaline JL, Pang B, et al. Quantification of DNA damage products resulting from deamination, oxidation and reaction with products of lipid peroxidation by liquid chromatography isotope dilution tandem mass spectrometry. Nat Protocols. 2008;3(8):1287–1298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Hemeryck LY, Decloedt AI, Bussche JV, Geboes KP, Vanhaecke L. High resolution mass spectrometry based profiling of diet-related deoxyribonucleic acid adducts. Anal Chim Acta. 2015;892:123–131. [DOI] [PubMed] [Google Scholar]
  • 22.Randerath K, Randerath E. 32P-Postlabeling methods for DNA adduct detection: Overview and critical evaluation. Drug Metab Rev. 1984; 26:67–85. [DOI] [PubMed] [Google Scholar]
  • 23.Wang PG, Zhang Q, Yao YY, Giese RW. Cationic xylene tag for increasing sensitivity in mass spectrometry. J Am Soc Mass Spectrom. 2015;26(10):1713–1721. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Moschel RC, Hudgins WR, Dipple A. Selectivity in nucleoside alkylation and aralkylation in relation to chemical carcinogenesis. J Org Chem. 1979;44(19):3324–3328. [Google Scholar]
  • 25.Sattsangi PD, Leonard NJ, Frihart CR. 1,N2-Ethenoguanine and N2,3-ethenoguanine. Synthesis and comparison of the electronic spectral properties of these linear and angular triheterocycles related to the Y bases. J Org Chem. 1977;42(20):3292–3296. [DOI] [PubMed] [Google Scholar]
  • 26.Schrader W, Linscheid M. Styrene oxide DNA adducts: In vitro reaction and sensitive detection of modified oligonucleotides using capillary zone electrophoresis interfaced to electrospray mass spectrometry. Toxicology. 1997;71:588–595. [DOI] [PubMed] [Google Scholar]
  • 27.Straub KM, Meehan T, Burlingame AL, Calvin M. Identification of the major adducts formed by reaction of benzo(a)pyrene diol epoxide with DNA in vitro. Proc Natl Acad Sci U S A. 1977;74(12):5285–5289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Wang PG, Gao J, Li G, Shimelis O, Giese RW. Nontargeted analysis of DNA adducts by mass-tag MS: Reaction of p-benzoquinone with DNA. Chem Res Toxicol. 2012;25(12):2737–2743. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Evrony GD, Cai XY, Lee E, et al. Single-neuron sequencing analysis of L1 retrotransposition and somatic mutation in the human brain. Cell. 2012;151(3):483–496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Wang PG, Fisher D, Rao A, Giese RW. Nontargeted nucleotide analysis based on benzoylhistamine labeling-MALDI-TOF/TOF-MS: Discovery of putative 6-oxo-thymine in DNA. Anal Chem. 2012;84(8): 3811–3819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Geacintov NE, Broyde S. Repair-resistant DNA lesions. Chem Res Toxicol. 2017;30(8):1517–1548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Reardon JT, Sancar A. Nucleotide excision repair. Prog Nucleic Acid Res Mol Biol. 2005;79:183–235. [DOI] [PubMed] [Google Scholar]
  • 33.Kisker C, Kuper J, Van Houten B. Prokaryotic nucleotide excision repair. Cold Spring Harb Perspect Biol. 2013;5:a012591. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Choi JH, Kim SY, Kim SK, Kemp MG, Sancar A. An integrated approach for analysis of the DNA damage response in mammalian cells. J Biol Chem. 2015;290(48):28812–28821. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Kreutzer DA, Essigmann JM. Oxidized, deaminated cytosines are a source of C to T transitions in vivo. Proc Natl Acad Sci U S A. 1998;95 (7):3578–3582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Kriaucionis S, Heintz N. The nuclear DNA base 5-hydroxymethylcytosine is present in purkinje neurons and the brain. Science. 2009;324(5929):929–930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Chen H, Yao L, Brown C, Rizzo CJ, Turesky RJ. Quantitation of apurinic/apyrimidinic sites in isolated DNA and in mammalian tissue with a reduced level of artifacts. Anal Chem. 2019;91(11):7403–7410. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Prelabeling SI
NIHMS1933272-supplement-Prelabeling_SI.pdf (503.9KB, pdf)

Data Availability Statement

Data available on request from the authors.

RESOURCES