Nucleic bases alkylation with acrylonitrile and cyanoethylene oxide: A

Article Cite This: Chem. Res. Toxicol. 2018, 31, 97−104

pubs.acs.org/crt

Nucleic Bases Alkylation with Acrylonitrile and Cyanoethylene Oxide: A Computational Study Martin Gladovic,†,‡ Eva Spaninger,† and Urban Bren*,†,§ †

Faculty of Chemistry and Chemical Technology, University of Maribor, Smetanova 17, SI-2000 Maribor, Slovenia Faculty of Chemistry and Chemical Technology, University of Ljubljana, Vecna pot 113, SI-1000 Ljubljana, Slovenia § National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Slovenia ‡

S Supporting Information *

ABSTRACT: Acrylonitrile (AN) is widely used in the manufacture of resins, plastics, and polymers, where workers are exposed to it during its production, transportation, and application. After intake a portion of AN is converted to cyanoethylene oxide (CEO) by cytochrome P450 2E1. Both AN and CEO represent possible chemical carcinogens leading to DNA damage mainly in the form of the major 7-(2oxoethyl)deoxyguanosine adduct. A kinetic model for its formation was devised and a corresponding second-order rate constant obtained from the experimental data on the reaction with CEO. A series of ab initio, density functional theory, and semiempirical calculations of activation free energies was then performed on the alkylation of nucleic bases with both CEO and AN. The combination of Hartree−Fock level of theory with the flexible 6-311++G(d,p) basis set and Langevin dipoles implicit solvation model gave the best agreement with the experimental activation barrier. It also predicted relative reactivities of all four nucleobases that are in agreement with the experimentally reported adduct yields. Moreover, this combination predicted higher reactivity of CEO than AN with all four nucleobases corroborating the experimental hypothesis that SN2 substitution of CEO rather than direct Michael addition of AN is responsible for the genotoxic properties of AN. In a broader context this paper points to the applicability of quantum chemical methods to the studies of carcinogenesis.

■

INTRODUCTION Acrylonitrile (AN) is a colorless volatile liquid used in the manufacture of resins, plastics, and polymers; humans can be exposed to it via air or water during its production, transportation, and application.1 An analysis performed in 2014 by Merchant Research & Consulting Ltd. reported a yearly production of 5.7 million tonnes of AN worldwide, with global production predicted to reach 7 million tonnes by 2017.2 AN is carcinogenic in rats3,4 and suspectedly carcinogenic in humans.5−8 The human carcinogenicity of AN is somewhat controversial, with some studies reporting its role in the development of lung, colon, and stomach cancers,9,10 while others report no significant association between AN and human carcinogenicity.11−15 After intake the major route of AN metabolism involves direct conjugation via Michael addition to glutathione.16 However, in the presence of oxygen and NADPH a portion of AN is converted to cyanoethylene oxide (CEO) via the epoxidation by cytochrome P450 2E1.17−19 CEO is a mutagenic and suspectedly a carcinogenic compound. In this article we present the calculation of the second-order rate constant and the corresponding activation free energy for alkylation of guanine with CEO from the available experimental data. Moreover, we also calculated the activation barriers for the reactions of CEO and AN with all four DNA bases leading to the most common covalent adducts observed in vitro and in vivo © 2017 American Chemical Society

at several ab initio, density functional theory (DFT), and semiempirical molecular orbital (MO) theory levels. Solvation effects were modeled using the self-consistent reaction field (SCRF) method of Miertuš et al.20 and with the Langevin dipoles (LD) model of Florian and Warshel.21 For semiempirical MO methods, the AM1-SM1 and PM3-SM3 models were applied.22−24 The experimental activation free energy was used as a measure to determine the most reliable combinations of QM method, basis set, and solvation model. Using these theory levels, we finally compared the calculated activation barriers for direct alkylation of nucleobases with AN via Michael addition versus the indirect alkylation with its metabolite CEO via SN2 substitution, which was determined to be much more reactive experimentally.25 Similar computational studies of DNA alkylation with small carcinogenic compounds have been carried out previously. Kranjc and Mavri performed a study of guanine alkylation with ethylene oxide.26 Bren and co-workers studied the alkylation of guanine with chloroethylene oxide27 and aflatoxin B1.28 Kržan and Mavri performed a QM study of the alkylation of guanine by styrene.29 Galeša et al. examined the alkylations with acrylamide.30 Mavri simulated DNA alkylation with propylene oxide.31 Lajovic et al. finally performed an extensive computaReceived: September 27, 2017 Published: December 22, 2017 97

DOI: 10.1021/acs.chemrestox.7b00268 Chem. Res. Toxicol. 2018, 31, 97−104

Article

Chemical Research in Toxicology Scheme 1. Proposed Mechanism of the Reaction between Guanine and Cyanoethylene Oxide

Scheme 2. Proposed Mechanism of the Reaction between Adenine and Cyanoethylene Oxide

Scheme 3. Proposed Mechanism of the Reaction between Cytosine and Cyanoethylene Oxide

Scheme 4. Proposed Mechanism of the Reaction between Thymine and Cyanoethylene Oxide

CEO and guanine leading to the main adduct 7-(2-oxoethyl)guanine has not yet been directly measured. However, the concentration of this adduct as a function of time was determined and reported by Guengerich and co-workers.25 We devised a kinetic model for CEO alkylation of guanine and fitted it on these experimental data to obtain the corresponding reaction rate constant. The adduct concentrations for guanine alkylation with AN are also available,25 but could not be further analyzed due to insufficient data resolution. Computational Methods. All calculations were performed at the National Institute of Chemistry in Ljubljana on the CROW cluster.33,34 To obtain the Born−Oppenheimer hypersurfaces and the corresponding activation energies for the reactions between AN or CEO and nucleobases, we performed a series of ab initio, DFT, and semiempirical MO simulations using the Gaussian 09 program package.35 The ab initio calculations were executed at the Hartree−Fock (HF) and the Møller−Plesset perturbation theory of the second order (MP2) levels of theory in combination with flexible 6-31G(d), 631+G(d,p), and 6-311++G(d,p) basis sets. In addition, we considered the DFT method B3LYP consisting of Becke’s exchange functional36 combined with the correlation functional of Lee, Yang, and Parr.37 We also considered the M06-2X global hybrid functional of Zhao and Truhlar38 and MPW1K, a modified version of the Perdew−Wang gradient-corrected exchange functional developed by the Truhlar

tional and experimental investigation of urethane-induced carcinogenesis.32 The proposed mechanism for the formation of the main 7(2-oxoethyl)deoxyguanosine adduct is depicted in Scheme 1. The reaction proceeds via SN2 substitution. The achiral terminal epoxide carbon of CEO attacks the nucleophilic endocyclic nitrogen N7 of guanine. An unstable zwitterionic intermediate is formed, which quickly decomposes through the elimination of the good leaving cyano group. The SN2 substitution represents the rate-limiting step of the reaction. Similarly, the SN2 substitutions occur at nucleophilic nitrogens of other nucleic bases as well; they are depicted in Schemes 2−4. It was also experimentally determined that in the Michael additions of AN the cyano group does not leave during the reaction; instead it remains as a part of the final adduct. See Supporting Information, Schemes S1−S4, for the proposed mechanisms for the reactions with AN.

■

METHODS

Calculation of the Rate Constant for CEO Alkylation of Guanine from Experimental Data. Due to the intrinsic reactivity of CEO in aqueous solutions, the rate constant for the reaction between 98


Article

Chemical Research in Toxicology

Figure 1. Fitting of the solution of differential equations for the reaction kinetics between guanine and CEO to the experimental data. Best fit with the second-order rate constant ka = 0.05 M−1 s−1 is shown as a blue curve, while experimental data points are depicted in red.

■

research group.39 Again, the flexible 6-31G(d), 6-31+G(d,p), and 6311++G(d,p) basis sets were used also in combination with the DFT methods. Finally, two semiempirical methods, AM1 and PM3, were applied. These are favorable for mixed quantum mechanical/molecular mechanical (QM/MM) description and thermal averaging due to their low computational cost. To accurately calculate the activation energy, we first had to find a structure close to the transition state of the reaction using a relaxed potential-energy surface scan, as described previously by Lajovic et al.32 The structure with the highest energy was used as a starting point for the Berny algorithm40 which produced the optimized transition state (TS) structure. Vibrational analysis in the harmonic approximation was performed on the TS structure to ensure that there was only one imaginary frequency present and that its vibrational mode coincided with the breaking of the existing epoxidic C−O bond and the formation of the new C−N bond between CEO and the corresponding nucleobase. The reactant state was obtained by increasing the TS distance between the reactive centers by approximately 0.2 Å followed by the energy minimization procedure, during which the reacting fragments moved further apart. Again, vibrational analysis on the resulting structure was performed to ensure that only real frequencies were present, thus indicating that a true minimum was located. Finally, the activation energy of the SN2 reaction in vacuo was calculated from the energy difference between the TS and the reactant state. An analogous procedure was used for the Michael addition of AN on the corresponding nucleobases. The hydrogen atom at the N3 position in the most stable (keto) thymine tautomer acts as a steric barrier for the studied reactions and at the same time reduces the nucleophilic character of this nitrogen atom. Therefore, the second-most stable (enol) tautomeric form of thymine41,42 was chosen for obtaining the TS and reactant state geometries (Scheme 4). The calculated activation energies, solvation free energies, and activation free energies were corrected for this tautomerization with respect to the most stable thymine tautomer. Solvation free energies of reactants and TS were calculated with the SCRF method of Miertuš et al.20 and with the LD model of Florian and Warshel.21 The SCRF method included in the Gaussian 09 program package was applied to all ab initio and DFT levels using both the standard variational approach and the perturbational approach with external iteration. Merz−Kollman partial atomic charges served as an input for the LD model included in the ChemSol program.43 For the semiempirical MO methods, the AM1-SM1 and PM3-SM3 solvation models were used.24 Corresponding calculations were performed with the AMSOL 5.4.1 program of Truhlar and coworkers.24

RESULTS AND DISCUSSION Determination of the Experimental Reaction Rate Constant. The reaction between CEO and guanine is competitive with CEO hydrolysis, which leads to the formation of glycoaldehyde (GA). The main adduct 7-(2-oxoethyl)guanine is unstable and is further transformed through reactions like depurination leading to the final product (FP). We can thus summarize the reaction mechanism with the following scheme: kh CEO + H 2O → GA ka kd CEO + guanine → 7‐(2‐oxoethyl)guanine → FP

and write the corresponding system of differential equations: d[GA] = k h[CEO] dt d[guanine] = −ka[guanine][CEO] dt d[FP] = kd([guanine]0 − [guanine] − [FP]) dt

(1)

where [ ] and [ ]0 indicate current and initial molar concentrations, respectively, kh is the pseudo-first-order rate constant for CEO hydrolysis, ka is the second-order rate constant for 7-(2-oxoethyl)guanine adduct formation, and kd is the first-order rate constant for its decomposition. The current CEO concentration can thus be calculated with the following mass-balance equation [CEO] = [CEO]0 − [GA] − [guanine]0 + [guanine] (2)

and the current concentration of 7-(2-oxoethyl)guanine adduct can thus be obtained using the relation [7‐(2‐oxoethyl)guanine] = [guanine]0 − [guanine] − [FP] 99

(3)


Article


Table 1. Activation Energies for the Reactions of Nucleic Bases with Cyanoethylene Oxide or Acrylonitrile: HF/6-311++G(d,p) Method guanine + CEO adenine + CEO cytosine + CEO thymine + CEO guanine + AN adenine + AN cytosine + AN thymine + AN

ΔE⧧ + ΔZPE [kcal/mol]a

ωTS [i cm−1]b

ωR [cm−1]c

dTS [Å]d

dR [Å]e

40.6 42.5 43.1 61.3 37.5 39.1 41.9 49.5

574 617 584 568 424 439 454 470

15.4 12.3 26.2 21.5 13.7 10.6 17.3 10.1

1.95 1.96 1.96 1.95 1.79 1.79 1.80 1.81

3.81 3.70 3.47 3.57 3.77 7.02 3.92 4.56

a

Gas-phase activation energy including zero-point vibrational energy correction. bImaginary vibrational frequency corresponding to the transition state. cLowest vibrational frequency corresponding to the reactant state. dDistance between the reacting N atom of nucleobase and the C atom of cyanoethylene oxide or acrylonitrile in the transition state. eDistance between the reacting N atom of nucleobase and the C atom of cyanoethylene oxide or acrylonitrile in the reactant state.

Table 2. Activation Free Energies for the Reactions of Nucleic Bases with Cyanoethylene Oxide or Acrylonitrile: HF/6-311+ +G(d,p)/LD Method ΔGLD hydr [kcal/mol] guanine + CEO adenine + CEO cytosine + CEO thymine + CEO guanine + AN adenine + AN cytosine + AN thymine + AN

TSa

Rb

c ΔΔGLD hydr [kcal/mol]

ΔG⧧LD [kcal/mol]d

−48.2 −34.4 −42.6 −41.5 −43.0 −26.9 −35.2 −32.9

−26.6 −18.3 −20.4 −19.9 −29.9 −20.0 −22.3 −28.1

−21.6 −16.2 −22.2 −21.6 −13.2 −6.9 −12.9 −4.8

19.0 (19.2 ± 0.2)e 26.4 20.9 38.0 24.4 32.2 29.0 42.6

a c

Hydration free energy of the transition state, obtained with the LD method. bHydration free energy of the reactants, obtained with the LD method. Hydration free energy of the transition state minus hydration free energy of the reactants. dActivation free energy. eExperimental value.

QM Calculations of Free Energy Barriers. Out of many ab initio, DFT, and semiempirical MO methods used (see Supporting Information, Tables S1−S32), the Hartree−Fock level of theory in combination with the flexible 6-311++G(d,p) basis set and the LD solvation model was found to reproduce the experimental activation free energy of the reaction between CEO and guanine in the most accurate way. Consequently, this combination was selected as a default for comparison of activation free energies for the reactions between the nucleic bases and CEO or AN. The corresponding activation energies in vacuo, imaginary vibrational frequencies of the transition state, lowest frequencies of the reactant states, and distances between the reacting atoms are collected in Table 1. We were able to obtain viable transition states for the reactions of all nucleic bases, because all vibrational modes with imaginary frequencies coincide with the desired creation of a new C−N bond between the reacting molecules as well as with the cleavage of the adjacent epoxidic C−O bond in the case of CEO. The corresponding distances between the reactive centers are very similar for all four nucleic bases, with a maximum spread of 0.01 and 0.02 Å for the reactions with CEO and AN, respectively. The reacting molecules in the transition state are on average 0.16 Å closer for the reactions with AN than with CEO. The energy barriers for the reactions in vacuo fail to predict the experimental order of reactivity. For all four bases the reaction with AN appears to be thermodynamically more favorable than with CEO. On the contrary, reaction rates calculated using our model based on the experimental data reported by Guengerich et al.25,44 for the reactions with CEO

The experimental half-life for CEO hydrolysis was reported to be about 2 h at the temperature of 310.15 K,25 indicating a pseudo-first-order rate constant kh = 9.6 × 10−5 s−1. The rate constant for the adduct formation ka was obtained through a best fit to the experimental data from eqs 2 and 3 and a numerical solution of the system of differential eqs 1 using the Levenberg−Marquardt optimization algorithm in the Wolfram Mathematica 10 environment (see Supporting Information for the corresponding code). Experimental data and the best-fit solution are presented in Figure 1. The second-order rate constant for guanine alkylation with CEO was determined as ka = 0.05 ± 0.02 M−1 s−1. The rate constant for the decomposition of the 7-(2-oxoethyl)guanine adduct was also optimized as a fitting parameter and was determined as kd = 6 × 10−4 s−1. As a proof of model validity, when all three rate constants ka, kh, and kd were used as fitting parameters, we calculated kh = 3 × 10−5 s−1, which is close to the experimental value. The rate constant is related to the corresponding activation free energy via Eyring’s transition state theory, ka =

⎛ ΔG ⧧ ⎞ kBT a ⎟⎟ exp⎜⎜ − h k T ⎝ B ⎠

(4)

where kB is the Boltzmann constant, h the Planck constant, and T the thermodynamic temperature. Using eq 4, the experimental activation free energy was calculated as ΔG⧧a = 19.2 ± 0.2 kcal/mol. This value was used to determine the best combination of theory level, basis set, and solvation model for theoretical calculations of activation barriers. 100


Article

Chemical Research in Toxicology are on the order of hours, while the rates for the reactions with AN reported by Solomon and Segal45 are on the order of weeks. To obtain the correct prediction of the relative reactivities of nucleic bases, it is thus necessary to employ a biologically relevant environment using a water solvent. The hydration free energies of the transition state and the reactants, their differences, and the free energy barriers of the corresponding reactions calculated using the Langevin dipoles implicit solvation model are summarized in Table 2. Activation free energies using this level of theory give a much more accurate prediction of relative reactivities. The barriers are now lower for the reactions with CEO in all cases. The solvent stabilizes the transition state more than the reactants, resulting in a negative difference in the hydration free energies. This reduction in the activation free energy ranges from −4.8 to −22.2 kcal/mol. The observed solvent acceleration is in line with the more polar nature of the transition state close to the one of the unstable zwitterionic intermediates. The transition states with the same base are better hydrated in the reactions with CEO, while the reactants are better hydrated in the reactions with AN. In addition to the presented level of theory, alternative combinations of quantum mechanical methods, basis sets, and solvation models were also evaluated (see Supporting Information, Tables S1−S32). B3LYP, MP2, and MPW1K all gave significantly lower activation free energies using the same basis set and solvation model; true to its nature, the M06-26 method gave results in between those obtained with the pure HF method and the B3LYP functional. On the contrary, the semiempirical methods resulted in too high activation free energies, with PM3 being closer to the ab initio HF calculations. Formation of Guanine Adducts. The structures of the transition states for the reactions with CEO and AN are shown in Figure 2. The tables with the corresponding numerical

and it indeed also has the lowest calculated activation free energies for both cases. For comparison, the activation free energies obtained with other QM methods and solvent models range from 8.3 kcal/mol (MP2/LD2) to 35.5 kcal/mol (AM1/ SM1) for the reaction with CEO and from 5.4 kcal/mol (MPW1K/LD2) to 29.9 kcal/mol (HF/SCRF) for the reaction with AN. Out of all the other tested combinations B3LYP/ SCRF (19.5 kcal/mol) and M062X/LD1 (18.9 kcal/mol) also yield activation barriers close to the experimental value for the reaction with CEO. Formation of Adenine Adducts. The structures of the transition states for the reactions with CEO and AN are depicted in Figure 3. The tables with the corresponding

Figure 3. Structure of the transition state for the reaction of the N1 atom of adenine with cyanoethylene oxide (A) and acrylonitrile (B) as predicted at the HF/6-311++G(d,p) level of theory. Carbon is depicted in gray, oxygen in red, nitrogen in blue, and hydrogen in white. The newly forming and breaking bonds are represented with gray dashed lines and hydrogen bonds are represented with yellow dashed lines. The highlighted bond lengths are given in Å.

numerical results can be found in the Supporting Information (Tables S2, S6, S10, S14, S18, S22, S26, and S30). On the basis of the calculations at the HF/6-311++G(d,p) level of theory in conjunction with the implicit LD solvation model, the activation free energy amounts to 26.4 and 32.2 kcal/mol for the reaction with CEO and AN, respectively (Table 2). The calculated activation barriers are second highest, which is in agreement with the second lowest experimental reaction yields of the alkylation with CEO.44 On the other hand, the alkylation with AN experimentally yields the second largest amount of adenine adducts.45 For comparison, the activation free energies obtained with other QM methods and solvent models range from 10.4 kcal/mol (B3LYP/LD2) to 40.4 kcal/mol (AM1/ SM1) for the reaction with CEO and from 9.9 kcal/mol (M062X/LD2) to 35.0 kcal/mol (HF/SCRF) for the reaction with AN. Formation of Cytosine Adducts. The structures of the transition states for the reactions with CEO and AN are shown in Figure 4. The tables with the corresponding numerical results can be found in the Supporting Information (Tables S3, S7, S11, S15, S19, S23, S27, and S31). On the basis of the calculations at the HF/6-311++G(d,p) level of theory in conjunction with the implicit LD solvation model, the activation free energy is 20.9 and 32.2 kcal/mol for the reaction with CEO and AN, respectively (Table 2). Cytosine indeed gives the second-largest yield in the reaction with CEO, but gives the lowest yield in the reaction with AN.44,45 Its calculated activation free energy is, however, the second lowest in both cases. For comparison, the activation free energies obtained with other QM methods and solvent models range from 5.4 kcal/mol (MP2/LD2) to 32.9 kcal/mol (AM1/SM1) for the reaction with CEO and from 7.9 kcal/mol (M062X/ LD2) to 34.0 kcal/mol (HF/SCRF) for the reaction with AN.

Figure 2. Structure of the transition state for the reaction of the N7 atom of guanine with cyanoethylene oxide (A) and acrylonitrile (B) as predicted at the HF/6-311++G(d,p) level of theory. Carbon is depicted in gray, oxygen in red, nitrogen in blue, and hydrogen in white. The newly forming and breaking bonds are represented with gray dashed lines. The highlighted bond lengths are given in Å.

results can be found in the Supporting Information (Tables S1, S5, S9, S13, S17, S21, S25, and S29). On the basis of the calculations at the HF/6-311++G(d,p) level of theory in conjunction with the implicit LD solvation model, the activation free energy for the reaction with CEO is 19.0 kcal/ mol, which is very close to its experimental value (19.2 ± 0.2 kcal/mol), while the activation barrier for the reaction with AN amounts to 24.4 kcal/mol (Table 2). This translates to pseudofirst order reaction half-times of 3.4 min and 44 days, respectively. Guanine is the main adduct experimentally determined for the reactions with both CEO and AN,44,45 101


Article


for vinyl carbamate epoxide, chloroethylene oxide, glycidamide, ethylene oxide, propylene oxide, and styrene oxide range from 19.1 to 27.7 kcal/mol,26,27,29−32 and the corresponding experimentally determined values lie between 19.5 and 26.5 kcal/mol.26,27,29−32 On the contrary, the well-known aflatoxin B1 possesses an even lower free energy barrier, calculated at 14.3 kcal/mol and experimentally determined at 15.1 kcal/mol, due to its ability to form favorable stacking interactions that lead to the intercalation of the toxin in between the Watson− Crick base pairs.28

■

CONCLUSIONS We performed a quantum mechanical evaluation of biologically relevant chemical interactions between nucleic bases and possible human carcinogens CEO and AN. We considered the alkylations from a kinetic standpoint by assessing the ab initio calculated activation free energies and comparing them to the experimentally determined reaction rate constants.25 With the selection of the right combination of the quantum theory level (Hartree−Fock), basis set (6-311++G(d,p)), and appropriate solvation model (LD), we could obtain accurate predictions of absolute guanine reactivity and of relative reactivities of remaining nucleic bases toward the alkylation with CEO.44 This holds also in the case of the Michael addition of AN, with cytosine being the only exception. The relative reactivity for cytosine was overestimated, placing it right after guanine in the reactivity series, while experimentally the yield of cytosine adducts was the lowest among all nucleic bases.45 An alternative combination of B3LYP/6-311++G(d,p) level of theory with the standard solvent reaction field (SCRF) method giving accurate predictions of experimental activation barriers for guanine alkylation with CEO made a somewhat better placement of cytosine reactivity with AN behind both guanine and adenine. The quantum mechanical calculations also correctly predicted a significantly higher reactivity of all nucleobases toward CEO than toward AN.25 Finally, it has to be emphasized that in real biological system, the nucleic bases are confined in a relatively rigid and sterically much more crowded environment, whereas our model systems consisted of the free methylated nucleobase floating in implicit solvent. Steric hindrance of the nascent Watson−Crick base pairs could indeed explain some discrepancies between the experimental and theoretical values. This was checked by superposing the transition state of the reaction between cytosine and AN onto a DNA model from X-ray crystallography. Visual inspection of the superposed geometry reveals a clear overlap of AN and the complementary guanine base (Supporting Information, Figure S1). For future studies it would therefore be prudent to consider these reactions using a QM/MM model system with the alkylation taking place in a biologically more accurate environment. In addition, the epoxidation of AN by cytochrome P450 2E1 and the interaction of the obtained DNA adducts with DNA polymerases should be computationally and experimentally further investigated to obtain a full understanding of the early AN-related carcinogenesis.46−51 Nevertheless, our quantum mechanical study of the alkylation of nucleobases provided us with valuable insights in the reaction mechanisms and the geometries of the transition states. Alkylated bases represent genetic mutations, which are removed by the base excision repair mechanism, where endonucleases play a vital role.52,53 Consequently, computational studies of the interactions between damaged nucleobases and endonucleases are envi-

Figure 4. Structure of the transition state for the reaction of the N3 atom of cytosine with cyanoethylene oxide (A) and acrylonitrile (B) as predicted at the HF/6-311++G(d,p) level of theory. Carbon is depicted in gray, oxygen in red, nitrogen in blue, and hydrogen in white. The newly forming and breaking bonds are represented with gray dashed lines, and hydrogen bonds are represented with yellow dashed lines. The highlighted bond lengths are given in Å.

Formation of thymine adducts. The structures of the transition states for the reactions with CEO and AN are depicted in Figure 5. The tables with the corresponding

Figure 5. Structure of the transition state for the reaction of the N3 atom of thymine with cyanoethylene oxide (A) and acrylonitrile (B) as predicted at the HF/6-311++G(d,p) level of theory. Carbon is depicted in gray, oxygen in red, nitrogen in blue, and hydrogen in white. The newly forming and breaking bonds are represented with gray dashed lines, and hydrogen bonds are represented with yellow dashed lines. The highlighted bond lengths are given in Å.

numerical results can be found in the Supporting Information (Tables S4, S8, S12, S16, S20, S24, S28, and S32). On the basis of the calculations at the HF/6-311++G(d,p) level of theory in conjunction with the implicit LD solvation model, the activation free energy barrier is 38.0 and 42.6 kcal/mol for the reaction with CEO and AN, respectively (Table 2). This is in agreement with experimental results of CEO alkylations, where the yield of thymine adducts yield is the lowest among all nucleic bases.44 For the reaction with AN, however, the computationally predicted relative reactivity seems to be too low.45 For comparison, the activation free energies obtained with other QM methods and solvent models range from 18.9 kcal/mol (MP2/LD2) to 49.0 kcal/mol (AM1/SM1) for the reaction with CEO and from 21.3 kcal/mol (MPW1K/LD2) to 46.7 kcal/mol (HF/SCRF) for the reaction with AN. Reactivity of CEO. In addition to the analysis of the regioselectivity, we can also establish the relative reactivity of CEO in comparison with the other previously studied ultimate carcinogens of the epoxy type. Focusing on the main reaction with guanine, it appears that the free energy barrier of CEO of around 19 kcal/mol is among the lowest of all studied carcinogens, which indicates that CEO represents one of the most genotoxic chemicals. The calculated free energy barriers 102


Article


(7) O’Berg, M. T., Chen, J. L., Burke, C. A., Walrath, J., and Pell, S. (1985) Epidemiologic study of workers exposed to acrylonitrile: an update. J. Occup. Environ. Med. 27, 835−840. (8) Felter, S. P., and Dollarhide, J. S. (1997) Acrylonitrile: A Reevaluation of the Database to Support an Inhalation Cancer Risk Assessment. Regul. Toxicol. Pharmacol. 26, 281−287. (9) Thiess, A. M., and Fleig, I. (1978) Analysis of chromosomes of workers exposed to acrylonitrile. Arch. Toxicol. 41, 149−152. (10) Blair, A., Stewart, P. A., Zaebst, D. D., Pottern, L., Zey, J. N., Bloom, T. F., Miller, B., Ward, E., and Lubin, J. (1998) Mortality of industrial workers exposed to acrylonitrile. Scand. J. Work, Environ. Health, 25−41. (11) Benn, T., and Osborne, K. (1998) Mortality of United Kingdom acrylonitrile workers - an extended and updated study. Scand. J. Work, Environ. Health 24 (Suppl. 2), 17−24. (12) Swaen, G. M., Bloemen, L. J. N., Twisk, J., Scheffers, T., Slangen, J. J. M., Collins, J. J., ten Berge, W. F. J. P., and Sturmans, F. (1998) Mortality update of workers exposed to acrylonitrile in The Netherlands. Scand. J. Work, Environ. Health 24 (Suppl. 2), 10−16. (13) Wood, S. M., Buffler, P. A., Burau, K., and Krivanek, N. (1998) Mortality and morbidity of workers exposed to acrylonitrile in fiber production. Scand. J. Work, Environ. Health 24 (Suppl. 2), 54−62. (14) Marsh, G. M., Youk, A. O., and Collins, J. J. (2001) Reevaluation of lung cancer risk in the acrylonitrile cohort study of the National Cancer Institute and the National Institute for Occupational Safety and Health. Scand. J. Work, Environ. Health 27, 5−13. (15) Schulz, M. R., Hertz-Picciotto, I., Todd, L., and Ball, L. M. (2001) Reconciling animal and human data in a cancer risk assessment of acrylonitrile. Scand. J. Work, Environ. Health 27, 14−20. (16) Lambotte-Vandepaer, M., and Duverger-van Bogaert, M. (1984) Genotoxic properties of acrylonitrile. Mutat. Res., Rev. Genet. Toxicol. 134, 49−59. (17) Kedderis, G. L., Batra, R., and Koop, D. R. (1993) Epoxidation of acrylonitrile by rat and human cytochromes P450. Chem. Res. Toxicol. 6, 866−871. (18) Sumner, S. C. J., Fennell, T. R., Moore, T. A., Chanas, B., Gonzalez, F., and Ghanayem, B. I. (1999) Role of Cytochrome P450 2E1 in the Metabolism of Acrylamide and Acrylonitrile in Mice. Chem. Res. Toxicol. 12, 1110−1116. (19) Wang, H., Chanas, B., and Ghanayem, B. I. (2002) Cytochrome P450 2E1 (CYP2E1) is Essential for Acrylonitrile Metabolism to Cyanide: Comparative Studies Using CYP2E1-Null and Wild-Type Mice. Drug Metab. Dispos. 30, 911−917. (20) Miertuš, S., Scrocco, E., and Tomasi, J. (1981) Electrostatic interaction of a solute with a continuum. A direct utilizaion of AB initio molecular potentials for the prevision of solvent effects. Chem. Phys. 55, 117−129. (21) Florián, J., and Warshel, A. (1997) Langevin Dipoles Model for ab Initio Calculations of Chemical Processes in Solution: Parametrization and Application to Hydration Free Energies of Neutral and Ionic Solutes and Conformational Analysis in Aqueous Solution. J. Phys. Chem. B 101, 5583−5595. (22) Dewar, M. J. S., Zoebisch, E. G., Healy, E. F., and Stewart, J. J. P. (1985) Development and use of quantum mechanical molecular models. 76. AM1: a new general purpose quantum mechanical molecular model. J. Am. Chem. Soc. 107, 3902−3909. (23) Stewart, J. J. P. (1989) Optimization of parameters for semiempirical methods I. Method. J. Comput. Chem. 10, 209−220. (24) Hawkins, G. D., Lynch, G. C., Giesen, D. J., Rossi, I., Storer, J. W., Liotard, D. A., Cramer, C. J., and Truhlar, D. G. AMSOL, version 5.4.1; 1996. (25) Guengerich, F. P., Geiger, L. E., Hogy, L. L., and Wright, P. L. (1981) In vitro metabolism of acrylonitrile to 2-cyanoethylene oxide, reaction with glutathione, and irreversible binding to proteins and nucleic acids. Cancer Res. 41, 4925−4933. (26) Kranjc, A., and Mavri, J. (2006) Guanine alkylation by ethylene oxide: Calculation of chemical reactivity. J. Phys. Chem. A 110, 5740− 5744.

sioned as future perspectives. The results of the study can also be used for the design of novel molecular scavengers that could prevent DNA alkylation damage by covalently binding to the ultimate carcinogens via a lower activation barrier.54

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemrestox.7b00268. Schemes for proposed mechanisms of the reactions between nucleic bases and AN; tables with activation energies, activation free energies using all combinations of implicit solvent models, imaginary frequencies of the transition state, lowest frequencies of the reactant state, and distances between the reacting atoms for all combinations of nucleic bases, reactive species, quantum mechanical methods, and basis sets; transition state of the reaction between cytosine and AN superposed onto a DNA model; the code numerically solving the set of differential equations (1) (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +386 2 2294 421. ORCID

Martin Gladovic: 0000-0002-4862-3659 Funding

Financial support through Slovenian Research Agency project grant J1-6736 and Slovenian Ministry of Education, Science and Sports program grant F4F is gratefully acknowledged. Notes

The authors declare no competing financial interest.

■

ABBREVIATIONS AN, acrylonitrile; CEO, cyanoethylene oxide; DFT, density functional theory; DNA, deoxyribonucleic acid; FP, final product; GA, glycoaldehyde; HF, Hartree−Fock; LD, Langevin dipoles; MO, molecular orbital; MP2, Møller−Plesset perturbation theory of the second order; QM/MM, quantum mechanical/molecular mechanical; SCRF, self-consistent reaction field; TS, transition state

■

REFERENCES

(1) Baus, D. K., Hsu, R. S., Neal, M. W., Santodonato, J., Sugatt, R. H., Bayard, S., Bayliss, D. L., Hiremath, C. B., and Vaughn- Dellarco, V. Health Assessment Document for Acrylonitrile; Technical Report EPA600/8-82-007F, 1983. (2) Merchant Research & Consulting Ltd. Acrylonitrile (ACN): 2014 World Market Outlook and Forecast up to 2018; 2014. (3) Maltoni, C., Ciliberti, A., and Di Maio, V. (1971) Carcinogenicity bioassays on rats of acrylonitrile administered by inhalation and by ingestion. Med. Lavoro 68, 401−411. (4) Quast, J. F. (2002) Two-year toxicity and oncogenicity study with acrylonitrile incorporated in the drinking water of rats. Toxicol. Lett. 132, 153−196. (5) Chen, J. L., Walrath, J., O’Berg, M. T., Burke, C. A., and Pell, S. (1987) Cancer incidence and mortality among workers exposed to acrylonitrile. Am. J. Ind. Med. 11, 157−163. (6) O’Berg, M. T. (1980) Epidemiologic study of workers exposed to acrylonitrile. J. Occup. Med. 22, 245−252. 103


Article

Chemical Research in Toxicology (27) Bren, U., Zupan, M., Guengerich, F. P., and Mavri, J. (2006) Chemical Reactivity as a Tool to Study Carcinogenicity: Reaction between Chloroethylene Oxide and Guanine. J. Org. Chem. 71, 4078− 4084. (28) Bren, U., Guengerich, F. P., and Mavri, J. (2007) Guanine alkylation by the potent carcinogen aflatoxin B1: Quantum chemical calculations. Chem. Res. Toxicol. 20, 1134−1140. (29) Kržan, M., and Mavri, J. (2009) Carcinogenicity of Styrene Oxide: Calculation of Chemical Reactivity. Croat. Chem. Acta 82, 317− 322. (30) Galeša, K., Bren, U., Kranjc, A., and Mavri, J. (2008) Carcinogenicity of acrylamide: A computational study. J. Agric. Food Chem. 56, 8720−8727. (31) Mavri, J. (2013) Can the chemical reactivity of an ultimate carcinogen be related to its carcinogenicity? An application to propylene oxide. Toxicol. In Vitro 27, 479−485. (32) Lajovic, A., Nagy, L. D., Guengerich, F. P., and Bren, U. (2015) Carcinogenesis of Urethane: Simulation versus Experiment. Chem. Res. Toxicol. 28, 691−701. (33) Borštnik, U., Hodošcě k, M., and Janežič, D. (2004) Improving the Performance of Molecular Dynamics Simulations on Parallel Clusters. J. Chem. Inf. Comput. Sci. 44, 359−364. (34) Borštnik, U., and Janežič, D. (2005) Symplectic molecular dynamics simulations on specially designed parallel computers. J. Chem. Inf. Model. 45, 1600−1604. (35) Frisch, M. J.; et al.. Gaussian 09, Revision E.01; Gaussian Inc.: Wallingford, CT, 2009. (36) Becke, A. D. (1993) Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648−5652. (37) Lee, C., Yang, W., and Parr, R. G. (1988) Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B: Condens. Matter Mater. Phys. 37, 785− 789. (38) Zhao, Y., and Truhlar, D. G. (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other function. Theor. Chem. Acc. 120, 215−241. (39) Lynch, B. J., Fast, P. L., Harris, M., and Truhlar, D. G. (2000) Adiabatic Connection for Kinetics. J. Phys. Chem. A 104, 4811−4815. (40) Peng, C., and Bernhard Schlegel, H. (1993) Combining Synchronous Transit and Quasi-Newton Methods to Find Transition States. Isr. J. Chem. 33, 449−454. (41) Fan, J. C., Shang, Z. C., Liang, J., Liu, X. H., and Jin, H. (2010) Systematic theoretical investigations on the tautomers of thymine in gas phase and solution. J. Mol. Struct.: THEOCHEM 939, 106−111. (42) Maximoff, S. N., Kamerlin, S. C. L., and Florián, J. (2017) DNA Polymerase λ Active Site Favors a Mutagenic Mispair between the Enol Form of Deoxyguanosine Triphosphate Substrate and the Keto Form of Thymidine Template: A Free Energy Perturbation Study. J. Phys. Chem. B 121, 7813−7822. (43) Florián, J., and Warshel, A. (1999) Calculations of Hydration Entropies of Hydrophobic, Polar, and Ionic Solutes in the Framework of the Langevin Dipoles Solvation Model. J. Phys. Chem. B 103, 10282−10288. (44) Solomon, J. J., Singh, U. S., and Segal, A. (1993) In vitro reactions of 2-cyanoethylene oxide with calf thymus DNA. Chem.-Biol. Interact. 88, 115−135. (45) Solomon, J. J., and Segal, A. (1985) Direct alkylation of calf thymus DNA by acrylonitrile. Isolation of cyanoethyl adducts of guanine and thymine and carboxyethyl adducts of adenine and cytosine. Environ. Health Perspect. 62, 227−230. (46) Martínek, V., Bren, U., Goodman, M. F., Warshel, A., and Florián, J. (2007) DNA polymerase β catalytic efficiency mirrors the Asn279-dCTP H-bonding strength. FEBS Lett. 581, 775−780. (47) Brown, K. L., Bren, U., Stone, M. P., and Guengerich, F. P. (2009) Inherent stereospecificity in the reaction of aflatoxin B1 8,9epoxide with deoxyguanosine and efficiency of DNA catalysis. Chem. Res. Toxicol. 22, 913−917.

(48) Zhang, H., Bren, U., Kozekov, I. D., Rizzo, C. J., Stec, D. F., and Guengerich, F. P. (2009) Steric and Electrostatic Effects at the C2 Atom Substituent Influence Replication and Miscoding of the DNA Deamination Product Deoxyxanthosine and Analogs by DNA. J. Mol. Biol. 392, 251−269. (49) Bren, U., and Oostenbrink, C. (2012) Cytochrome P450 3A4 Inhibition by Ketoconazole: Tackling the Problem of Ligand Cooperativity Using Molecular Dynamics Simulations and FreeEnergy Calculations. J. Chem. Inf. Model. 52, 1573−1582. (50) Bren, U., Fuchs, J. E., and Oostenbrink, C. (2014) Cooperative Binding of Aflatoxin B 1 by Cytochrome P450 3A4: A Computational Study. Chem. Res. Toxicol. 27, 2136−2147. (51) Klvaňa, M., Bren, U., and Florián, J. (2016) Uniform FreeEnergy Profiles of the P−O Bond Formation and Cleavage Reactions Catalyzed by DNA Polymerases β and λ. J. Phys. Chem. B 120, 13017− 13030. (52) Konc, J., Hodošcě k, M., Ogrizek, M., Trykowska Konc, J., and Janežič, D. (2013) Structure-Based Function Prediction of Uncharacterized Protein Using Binding Sites Comparison. PLoS Comput. Biol. 9, e1003341. (53) Ogrizek, M., Konc, J., Bren, U., Hodošcě k, M., and Janežič, D. (2016) Role of magnesium ions in the reaction mechanism at the interface between Tm1631 protein and its DNA ligand. Chem. Cent. J. 10, 41. (54) Brglez Mojzer, E., Knez Hrnčič, M., Škerget, M., Knez, Ž ., and Bren, U. (2016) Polyphenols: Extraction Methods, Antioxidative Action, Bioavailability and Anticarcinogenic Effects. Molecules 21, 901.

104


Nucleic bases alkylation with acrylonitrile and cyanoethylene oxide: A

Recommend Documents