Influence of Substrate Complexity on the Diastereoselective Formation of Spiroiminodihydantoin and Guanidinohydantoin from Chromate Oxidation

Synthesis of Spiroiminodihydantoin (Sp) and Guanidinohydantoin (Gh) Standards

8oxoG and sodium iridium(IV) chloride hexahydrate were purchased from Sigma-Aldrich (St. Louis, MO). The nucleoside was oxidized as described previously (27). Briefly, iridium(IV) chloride was added to 8oxoG at a 1:2 molar ratio in 10 mM potassium phosphate buffer, pH 7.4, to form Sp. The solution was incubated for 45 minutes in a mineral oil bath at 60 °C. To preferentially form Gh, the solution was unbuffered and chilled at 4 °C. Excess oxidant was removed via filtration through 20–24 mg of Sephadex A-25 anion exchange beads with 4–100 µL water washes. Gh was then purified by HPLC using an Agilent 1100 HPLC and a Zorbax SB-C18 column, dimensions 9.4 mm × 250 mm. A gradient of 0.5% acetonitrile/minute was used with a flow rate of 1.8 mL/min. Sp purification was accomplished using a gradient from 0% to 50% acetonitrile over 15 minutes. Collected fractions were lyophilized to dryness and reconstituted in Nanopure water. Concentrations were determined by UV-Vis spectroscopy at 230 nm using ε(Sp)= 4900 M⁻¹cm⁻¹ and ε(Gh)= 3000 M⁻¹cm⁻¹(14).

Oxidation of 8-oxo-7,8-dihydroguanine nucleoside

Potassium dichromate, ascorbic acid, and reduced glutathione were purchased from Sigma Aldrich (St. Louis, MO). Cr(V)-ehba was synthesized as previously published (28). Cr(V)-salen was formed by 1:2 reaction of Cr(III)-salen and iodosylbenzene in dry acetonitrile at room temperature for 15 minutes. Oxidations of 50 µL of 100 µM 8oxoG were carried out in a 10 mM tris buffer (pH 7.4) at 37 °C. Oxidants iridium(IV) chloride, Cr(V)-salen, and Cr(V)-ehba were used at a 1:8 8oxoG:oxidant ratio. Cr⁶⁺/ascorbate and Cr⁶⁺/glutathione were used at 1:4:40 and 1:10:1 ratios of 8oxoG:Cr⁶⁺: reductant, respectively. All oxidations were incubated for 1 hour with the exception of the Cr⁶⁺/glutathione system, which was incubated for 6 hours because of a slower reduction process. Nucleoside oxidations were purified with Sephadex columns as described for synthesis of standards, lyophilized, and reconstituted in water. Samples were analyzed by LC-ESIMS/MS without further processing.

Oxidation of the single-stranded 8oxoG containing oligonucleotide

A 22-mer oligonucleotide of sequence 5’-ACC AGC AGC GXC CGC ACC AGT G-3’, where X is 8oxoG, was purchased from TriLink BioTechnologies (San Diego, CA). The oligonucleotide was purified by HPLC with a Dionex column as previously published (26). Fractions containing oligonucleotide were collected in microcentrifuge tubes and lyophilized to approximately one-quarter volume. The DNA was ethanol precipitated and the oligonucleotide pellet was dried at room temperature before reconstituting in water. Concentrations were then determined UV-Vis spectroscopy and calculated from A₂₆₀using the extinction coefficient provided for the specific sequence. Oxidations were carried out in 50 µL volumes as detailed above, and products were desalted through the use of Micro Bio-Spin 6 columns purchased from Bio-Rad (Hercules, CA).

Oxidation of the double-stranded 8oxoG containing oligonucleotide

The complement oligonucleotide of sequence 5’-TCA CTG GTG CGG CCG CTG CTG G-3’ was purchased from TriLink BioTechnologies (San Diego, CA) and was HPLC purified, ethanol precipitated, and quantified as for the 8oxoG-containing single stranded oligonucleotide. The 8oxoG containing oligonucleotide was annealed with 20% excess complement using a Thermacycler in oxidation buffer. Double-stranded oligonucleotides were then oxidized and purified using identical conditions as described in the nucleoside section.

DNA Hydrolysis

Phosphodiesterase I and phosphodiesterase II were purchased from Worthington Biochemicals (Lakewood, NJ); calf intestinal alkaline phosphatase was purchased from Promega (Madison, WI); butylated hydroxytoluene and deferoxamine were purchased from Sigma Aldrich (St. Louis, MO). Oxidized oligonucleotides were hydrolyzed prior to analysis as described previously (27), with a few minor adjustments. Antioxidants butylated hydroxytoluene and deferoxamine were added to the digestion mixture at final concentrations of 0.5 mM and 1.5 mM, respectively, to reduce artifactual oxidation due to the digestion conditions (29). Double-stranded oligonucleotides were heated at 95 °C for 5 minutes to denature DNA strands. Samples were then ice-chilled for 3 minutes, with the unbuffered addition of 0.10 units of phosphodiesterase II within the first 30 seconds. A 1-hour incubation at 37 °C followed with a repeat addition of enzyme and incubation. 10X sodium-free phosphodiesterase I buffer (100 mM tris-HCl, 100 mM NH₄Cl, and 100 mM MgCl₂) and 0.5 units of phosphodiesterase I were added and incubated for 45 minutes, with an additional repeat of enzyme and incubation time. Lastly, phosphates were removed with the addition of 5 units of calf intestinal alkaline phosphatase and 45 minute incubation, followed by one repeat of enzyme addition and incubation time. Samples were filtered with Microcon YM-30 centrifugal filters for 25 minutes at 2200 × g to remove enzymes and any undigested DNA. Samples were stored at −80 °C until analyzed.

LC-ESI-MS/MS analysis method

Samples were analyzed by selected reaction monitoring (SRM) using a 1200 series Agilent HPLC coupled to an Agilent 6300 XCT series LC/MSD Trap with electrospray ionization. Separations were achieved using a Thermo Scientific Hypercarb column with 5 µm particle size and dimensions of 100 mm × 2.1 mm. A mobile phase of 100% A and 0% B was ramped to 10% B at 6 minutes with a flow rate of 0.2 mL/minute, where mobile phase A is 0.1% aqueous acetic acid and mobile phase B is 100% MeOH. The mass spectrometer was run at a dry temperature of 350 °C, nebulizer pressure of 40 psi, and dry gas flow rate of 8 L/min. Other pertinent MS parameters are given in Table 1. The buffer front in the first 2.5 minutes of each run was diverted to waste, after which Gh nucleoside m/z of 274 was collected with an isolation window of 2.0 amu, fragmented, and monitored for the free base at m/z 158. After 6 minutes Sp was detected through a 2.0 amu isolation window of m/z 300 with fragmentation and monitoring of the free base at m/z 184. Quantification of Sp and Gh was conducted through the use of two standard addition points to create a curve ranging 1 to 2 orders of magnitude for each oxidant type and DNA substrate. Average baseline levels of lesion found in the unoxidized control samples were then subtracted from levels found in oxidized samples.

Rationale for oxidant systems

Oxidation by Ir(IV) was chosen as a positive control because of its effective formation of Sp and Gh from 8oxoG, and a well-defined mechanism of action (14, 17, 25). The high-valent chromium complexes, Cr(V)-salen and Cr(V)-ehba, were used in this study as they have been proposed to be models for the highly oxidizing reduction intermediates of chromate formed during intracellular reduction (26). More importantly, it has been shown that these model complexes do not produce oxygen radicals during their reduction to the final stable Cr(III) oxidation state (3). Endogenous reductants including ascorbate and glutathione have been identified as the primary reductants of intracellular chromate (30, 31) and act as pro-oxidants with regard to cellular toxicity (32, 33). Both ascorbate and glutathione are known to reduce Cr(VI) to the more highly oxidizing Cr(IV) and Cr(V) oxidation states, as well as producing a variety of radicals species under certain metal/reductant ratios (8, 34). These systems have been included in this work to approximate cellular oxidation systems, although no attempt to optimize lesion formation based on varying metal/reductant ratios was made.

Mass spectrometry detection method

We have developed a LC-ESI-MS/MS method as the primary approach to lesion analysis throughout this study. The Sp diastereomers observed with this approach were assigned based on optical methods from previously published protocols (35,36) as validated in our own laboratory (see supporting information). Because optical assignments for the Gh diastereomers have not yet been made, the diastereomers called Gh1 and Gh2 in this study were assigned based solely on their elution profile from the HPLC. Selected reaction monitoring (SRM) was utilized, that allowed for more selective and more sensitive measurements of the lesions of interest. Sodium and potassium salts were substituted in enzyme buffer solutions whenever possible throughout the experimental procedures, allowing only the molecular ion, (M+H of 274 and 300 for Gh and Sp) to be trapped and the ensuing fragmented free base of each to be selectively monitored at m/z of 158 and 184 for Gh and Sp respectively. These parameters gave detection limits for Gh and Sp of respectively 200 fmol and 400 fmol from a 250–1000 pmol 8oxoG nucleoside sample. The presence of higher salt concentrations from the oligonucleotide digests decreased the detection limits for single-stranded and double-stranded DNA to approximately 300 fmol for Gh and 700 fmol for Sp in a sample with an original concentration of 500–1300 pmol 8oxoG. A typical chromatogram for this system showing the formation of all four diastereomers is given in Figure 2.

Oxidation of the 8oxoG nucleoside and hydantoin lesion formation

The 8oxoG nucleoside was reacted with each of the oxidants and analyzed for Sp and Gh formation by SRM using LC-ESIMS/MS. Oxidations were carried out in triplicate, and standard addition curves over an order of magnitude of lesion formation were used for quantification. Because of the unique matrix of each oxidation system, a standard addition curve was created for each set of conditions. Levels of Sp and Gh found in the unoxidized standard (background) were subtracted from levels found in oxidized samples, allowing for comparison of lesion formed only by the oxidative conditions listed above.

Gh was the predominant lesion under all oxidizing conditions except Ir(IV) (Figure 3). Despite little expected steric influence from the nucleoside substrate, Gh2 was formed preferentially over Gh1. The formation of Gh1 was approximately 75% of Gh2 formation in all but the Cr⁶⁺/ascorbate system, where Gh1 formation was half that of Gh2. In contrast, background levels of oxidation showed opposite diastereomeric preference, with the amount of Gh1 exceeding the amount of Gh2. While the pattern of Gh formation was fairly consistent among all oxidative systems, the efficiency of formation varied between oxidants. Cr⁶⁺ with ascorbate was the least effective at forming Gh, followed by Cr⁶⁺/glutathione. Ir(IV) and Cr(V)-ehba were slightly more effective, forming Gh at very similar levels, and Cr(V)-salen formed more than triple the amount of Gh formed under any other condition.

Sp formation at the nucleoside level was less consistent. Nearly equal formation of diastereomers was seen in the background oxidation and with oxidants Ir(IV) and Cr⁶⁺/glutathione. Oxidation by Cr(V)-salen showed a clear preference for the formation of R-Sp, while Cr(V)-ehba and Cr⁶⁺/ascorbate favored formation of S-Sp. The Cr⁶⁺/ascorbate system was clearly the least-effective at forming Sp, followed by similar levels of formation by Cr(V)-ehba, Cr⁶⁺/ascorbate and Cr⁶⁺/glutathione, with Ir(IV) forming twice as much Sp as the previous systems and Cr(V)-salen forming the most Sp.

Oxidation of single-stranded oligonucleotides containing an 8oxoG moiety

A 22-mer oligonucleotide containing an 8oxoG was oxidized and digested to its respective nucleosides prior to LC-ESI-MS/MS analysis. Samples were run in triplicate with standard additions in each oxidation matrix for quantification. Baseline lesion levels were quantified and subtracted from that observed in oxidized samples prior to comparison. As seen in Figure 4, the single-stranded oxidations differ substantially from the nucleoside oxidations. Gh diastereomeric preference was consistent across the set of oxidants, but now showed preference of Gh1 as opposed to Gh2 preference in the nucleoside oxidations. Gh was no longer the predominant lesion under all conditions. Instead, Sp showed preferential formation with Cr(V)-salen and Cr⁶⁺/ascorbate, as well as in the unoxidized/background sample. Gh was still formed over Sp by Ir(IV), Cr(V)-ehba, and Cr⁶⁺/glutathione, and showed a dramatic increase in the cases of Ir(IV) and Cr(V)-ehba. Efficiency of Gh formation increased in the order Cr⁶⁺/ascorbate < Cr⁶⁺/glutathione ~ Cr(V)-salen < Cr(V)-ehba < Ir(IV).

Sp formation in single-stranded oligonucleotides showed a significant change in the magnitude of diastereomeric preference. S-Sp was on average formed twice as often as R-Sp by Cr(V)-ehba, Cr⁶⁺/glutathione, and Cr⁶⁺/ascorbate, as well as in the background sample. R-Sp was formed preferentially in Ir(IV) and Cr(V)-salen, amounting to 3 and 2.5 times S-Sp respectively. Overall formation of Sp increased in the order Cr⁶⁺/glutathione ~ Cr⁶⁺/ascorbate < Ir(IV) < Cr(V)-salen < Cr(V)-ehba.

Oxidation of double-stranded oligonucleotides containing an 8oxoG moiety

A 22-mer oligonucleotide containing an 8oxoG was annealed to its complement containing cytosine opposite the 8-oxoG moiety, oxidized, and digested to its respective nucleosides prior to LC-ESI-MS/MS analysis. Samples were run in triplicate with standard additions in each oxidation matrix for quantification. Background lesion levels were quantified and subtracted from that in oxidized samples prior to comparison.

There were notable differences between lesion formation in the double-stranded oligonucleotide and the previous, less sterically hindered, substrates (Figure 5). Only Gh was observed to form from Ir(IV) oxidation and Cr(V)-ehba oxidation, while Sp was formed at greater levels than Gh under all other condition including in the background samples. The diastereomers of Gh did not separate well in all samples (presumably from adverse buffer effects in the electrospray source), and values given for diastereomeric preference by Ir(IV), Cr(V)-salen, and Cr(V)-ehba are from a single sample in which we were able to get reasonable separations. Under these conditions, Gh1 was formed less than Gh2, particularly from Ir(IV) oxidation. Gh was not detected in the Cr⁶⁺/ascorbate system and was only minimally formed by Cr⁶⁺/glutathione, with the Gh1 diastereomer being preferred.

Sp formation was seen in some but not all double-stranded samples; Sp was not detected as a result of Ir(IV) nor Cr(V)-ehba oxidation. In Cr(V)-salen, Cr⁶⁺/ascorbate, Cr⁶⁺/glutathione, and background samples, R-Sp was formed preferentially over S-Sp. Cr(V)-salen exhibited the largest amount of Sp formation as well as the greatest difference of formation between diastereomers, with almost 10 times as much R-Sp as S-Sp being formed.

Overall Gh and Sp formation with increasing substrate complexity are compared in Figure 6 and Figure 7, respectively. Background levels of Sp and Gh were fairly consistent, regardless of substrate complexity. Gh formation was also approximately the same across all substrates for Cr(V)-ehba. However, there was a clear decrease in Gh formation with substrate complexity for oxidation with Cr(V)-salen, Cr⁶⁺/ascorbate, and Cr⁶⁺/glutathione systems. Ir(IV) was the only oxidant to show an increase in Gh formation from oxidation of nucleosides to single-stranded oligomers to double-stranded oligomers. Sp formation was also observed to decrease with complexity for oxidation with Ir(IV), Cr(V)-salen, and Cr⁶⁺/glutathione. Cr(V)-ehba showed an increase in Sp formation from the nucleoside to single-stranded DNA, but no Sp was detected for the double-stranded oligonucleotide. Sp formation increased with substrate complexity only for oxidation with Cr⁶⁺ and ascorbate.