A reaction mechanism-based prediction of mutagenicity: α-halo carbonyl compounds adduct with DNA by S_N2 reaction

Abstract

Most of the α-halo carbonyl (AHC) compounds tend to be predicted as mutagenic by structure-activity relationship based on structural category only, because they have an alkyl halide structure as a structural alert of mutagenicity. However, some AHC compounds are not mutagenic. We hypothesized that AHC reacts with DNA by S_N2 reaction, and the reactivity relates to mutagenicity. As an index of S_N2 reactivity, we focused on molecular orbitals (MOs), as the direction and position of two molecules in collision are important in the S_N2 reaction. The MOs suitable for S_N2 reaction (SN2MOs) were selected by chemical-visual inspection based on the shape of the MO. We used the level gap and the energy gap between SN2MO and the lowest unoccupied molecular orbital as the descriptors of S_N2 reactivity. As the results, S_N2 reactivity related to mutagenicity and we were able to predict mutagenicity of 20 AHC compounds with 95.0% concordance. It was suggested that S_N2 reaction is a reaction mechanism of AHC compounds and DNA in the mutagenic process. The method allows for discrimination among structurally similar compounds by combination with quantitative structure-activity relationships. The combination approach is expected to be useful for the mutagenic assessment of pharmaceutical impurities.

INTRODUCTION

To predict mutagenicity, quantitative structure-activity relationship (QSAR) has been developed for pharmaceutical impurities (Valerio and Cross, 2012; Sutter et al., 2013) and cosmetic integrates (Aiba née Kaneko et al., 2015). The understanding for the mechanism is important for mutagenicity prediction and mechanism-based QSARs for mutagenicity were reviewed by Benigni and Bossa (2011). However, one of the problems of some QSARs or structural alerts is that it is not all related to chemical mechanisms in mutagenic processes directly. Because of this problem, it is difficult to discriminate between similar compounds based on QSAR based on structural alerts alone and such QSAR tends to be high sensitivity. Therefore, a new mutagenicity prediction method related with the mechanism of mutagenicity is desired for some chemical classes.

The α-halo carbonyl (AHC) is a structural alert for mutagenicity in some QSARs, e.g., CASE Ultra (Fig. 1, right). AHC belongs to the alkyl halide group (Fig. 1, left) which is known to react with DNA by S_N2 reaction. The alkyl halide tends to be positive in the Ames test. On the other hand, AHC has no clear tendency for mutagenicity. In addition, the opposite results were obtained among some similar compounds, e.g., methyl chloroacetate, ethyl chloroacetate, and isopropyl chloroacetate. Therefore, it is significant to predict mutagenicity of AHC compounds with high specificity and concordance.

Fig. 1

Examples of the MO levels and shapes of the reaction between alkyl halide and AHC. Gray: carbon, White: hydrogen, Green: chloride, Red: oxygen. Blue cloud: negative MO, Red cloud: Positive MO. The LUMO of alkyl halide was SN2MO. In AHC, the SN2MO was not LUMO, but LUMO+1.

AHC is able to react with nucleophiles by S_N1 (unimolecular nucleophilic substitution reaction) or S_N2 (bimolecular nucleophilic substitution reaction) mechanism; however, which mechanism is the major reaction for mutagenicity is unknown. On this background, we hypothesized that the mechanism of AHC reaction with DNA is S_N2 reaction, as with alkyl halides and epoxides, and the S_N2 reactivity of AHC compound relates to its mutagenicity in this research. Figure 2 indicates the mechanisms of S_N1 and S_N2 reaction. In S_N2 reaction, the direction and the position in collision of two molecules are important. The nucleophile agent attacks to the electrophile agent from back of leaving group, i.e. halogen of AHC (Fig. 2, green arrow). In the present study, we focused on molecular orbitals (MOs), as their graphic representations allow visualization of the position of the reaction.

Fig. 2

The mechanism of S_N1 and S_N2 reaction. S_N1 reaction: S_N1 reaction is stepwise reaction. The 1^st step is loss of leaving group (i.e. chloride ion). The 2^nd step is reaction with carbocation and nucleophile. S_N2 reaction: S_N2 reaction is a concertante reaction. The reaction proceeds through a backside attack of nucleophile (green arrow). Nu: nucleophile; Gray: carbon, White: hydrogen, Green: chloride, Red: oxygen. Blue cloud: negative MO, Red cloud: Positive MO, Green arrow: the direction of attack of nucleophile.

MATERIALS AND METHODS

Dataset

We used 20 AHC compounds, comprising 10 positive (including 2 marginal compounds which was defined by CASE Ultra) and 10 negative compounds included in the database of CASE Ultra (GT_EXPERT, version 1.6.0.2, MultiCASE Inc., Beachwood, OH, USA; the Ames data and references are listed in Table 1). All of these compounds were judged to be positive by GT_EXPERT (Gray zone was set on 40% ~ 60%). The dataset included some structural diversity, e.g., carboxylic acid, ketone, ester, amide, and aromatic ketone (Fig. 3). These compounds did not include any other structural alerts. We used the results of the Ames test in the database as the mutagenicity.

Table 1. The decisions of Ames test and their reference of compounds used in this research.

No.	Decision	Species	Strain	Metabolic activation	Method	Test Condition	Solvent	Reference
1	Positive	S	TA100	R	3	1-30 μg/mL	DMSO	Le Curieux et al., 1994
	Positive	S	-	N, R	1	-	-	Le Curieux et al., 1994
2	Positive	S	-	N, R	1	-	-	Gold and Lueschen, 1983
3	Positive	S	-	N, R	1	-	-	Kazius et al., 2005
	Positive	S	TA100	N	1	1.25-10 μg/plate	-	Meier et al., 1985
	Positive	S	-	N, R	2	-	-	Kier et al., 1986
	Positive	S	TA100	N	3	0.001-1 μg/mL	DMSO	Le Curieux et al., 1994
	Positive	S	TA100, TA1535	N, R	1	0-40 nmol/plate	-	Merrick et al., 1987
4	Positive	S	-	N, R	1	-	-	Kazius et al., 2005
	Positive	S	TA98, TA100	N	1	20-80 μg/plate	-	Meier et al., 1985
	Positive	S	-	N, R	1	-	-	Nestmann et al., 1985
5	Positive	E	WP2	N	1	-	-	Zijlstra, 1989
6	Positive	E	WP2	N	1	-	-	Zijlstra, 1989
7	Positive	S	TA100	N, M, R	1	10-4000 μg/plate	DMSO	Dolzani et al., 1986
8	Positive	S	TA100	N	1	100-4000 μg/plate	DMSO	Dolzani et al., 1986
9	Negative	S	-	N, R	1	-	-	Kazius et al., 2005
	Negative	S	TA102, TA2638	N	1	78-2500 μg/plate	DMSO	Watanabe et al., 1996
	Negative	E	WP2, WP2 uvrA/pKM101	N	1	-	-	Watanabe et al., 1996
10	Positive	E	WP2, WP2 uvrA/pKM101	N	1	-	-	Watanabe et al., 1996
11	Negative	S	-	N, R	1	-	-	Kazius et al., 2005
12	Positive	S	TA100	N, R	2	50-40 μmol/plate	-	Yaguchi et al., 1993
13	Negative	S	TA98, TA100	N, R	2	-	-	Saito et al., 1995
	Negative	S	-	N, R	2	-	-	Kier et al., 1986
	Negative	S	TA98, TA100, TA102, TA104, TA1537, TA1535	N, R	2	9.77-5000 μg/plate	water	JCIETI CENTER, 1996
	Negative	E	WP2 uvrA, WP2 uvrA/pKM101	N, R	2	-	-	JCIETI CENTER, 1996
	Negative	S	TA100	N, R	3	0.3-10000 μg/mL	water	Giller et al., 1997
14	Negative	S	-	N, R	1	-	-	Judson et al., 2005
15	Negative	S	-	R	1	-	-	Kazius et al., 2005
	Negative	E	WP2 uvrA	N	2	0-2.5 μg/plate	acetone	Azuma et al., 1997
	Negative	S	TA98, TA100	N, R	2	0-120 μg/plate	acetone	Azuma et al., 1997
16	Positive	S	-	N, R	1	-	-	Judson et al., 2005
	Positive	S	TA98, TA100	N	2	-	-	National Toxicology program
17	Negative	S	-	N, R	1	-	-	Kazius et al., 2005
	Negative	S	TA98, TA100	N, R	2	0-140 μg/plate	acetone	Azuma et al., 1997
	Negative	E	WP2 uvrA	N	2	0-3 μg/plate	acetone	Azuma et al., 1997
18	Negative	S	-	N, R	1	-	-	Undisclosed private communications
19	Negative	S	TA98	R	1	5-500 μg/plate	DMSO	Bertenyi and Lambert, 1996
	Negative	S	TA98, TA100, TA1535, TA1537	N, R	2	0.3-333 μg/plate	DMSO	Zeiger et al., 1987
	Negative	S	TA100, TA1538	N, R	1	0.5-500 μg/plate	DMSO	Ashby et al., 1996
	Negative	S	-	N, R	1	-	-	Judson et al., 2005
	Negative	S	TA98, TA100, TA1537, TA1535	N, R	2	-	-	Zeiger et al., 1987
20	Negative	S	-	N, R	1	-	-	Kazius et al., 2005
	Negative	S	TA98, TA100	N, R	2	0-75 μg/plate	acetone	Azuma et al., 1997
	Negative	E	WP2 uvrA	N	2	0-2 μg/plate	acetone	Azuma et al., 1997

-: no information in CASE Ultra Database; S: S.typhimurium, E: E. Coli; R: rat liver S-9, M: mouse liver S-9, N: None; 1: Standard plate, 2: Preincubation, 3: Functional test.

Fig. 3

The chemical structures of AHC compounds used in this research. These compounds did not include any other structural alerts.

Calculation of MOs

We used Chem3D^®Pro (version 16.0, PerkinElmer Informatics, Tokyo, Japan) to calculate the MOs. Each structure was drawn on Chem3D, and then it was optimized by MM2 method to obtain a local minimum structure (parameter of Chem3D: Display Every Iteration, Minimum RMS Gradient: 0.100, Properties: pi Bond Orders and Steric Energy Summary). The MM2 method was a modified method of Allinger NL (Burkert and Allinger, 1982). The MOs were obtained by the extended Hückel method. In the Chem3D, following extension of “Hückel surface” of compounds, “Molecular Orbital Surface” was shown, and then orbital energies were read. The MOs adaptable to S_N2 reaction were selected by chemical-visual inspection.

Receiver operating characteristic (ROC) analysis

ROC analysis was performed using JMP^® (version 11.0.1, SAS Institute Japan, Tokyo, Japan). We used the results of the Ames tests as the objective variables and each descriptor as an explanatory variable for nominal scale logistic regression. In brief, the predictive potential of a descriptor (explanatory variable) for the objective variable was evaluated by ROC analysis, and the obtained area under of the curve (AUC) was used as the result.

RESULTS

MO selection

Each compound has multiple MOs, such as the highest occupied molecular orbital (HOMO; unit: eV) and the lowest unoccupied molecular orbital (LUMO; unit: eV). It is well known that LUMO relates to electrophilicity, because the orbital has the lowest energy among MOs unoccupied with electrons; HOMO relates to nucleophilicity, because the orbital has the highest energy among MOs occupied with electrons (Clayden et al., 2003). There are some reports that HOMO or LUMO relates to mutagenicity for aromatic amine (Debnath et al., 1992b; Hatch and Colvin, 1997; Lopez de Compadre et al., 1990), furanone and halopropenal (Tuppurainen et al., 1991), quinolone (Debnath et al., 1992a; Hu et al., 2007), pyrrol (Freeman et al., 2001), and phenalenone (Misaki et al., 2008), and some under review (Benigni, 2005; Tuppurainen, 1999). In the case of alkyl halide, the shape of LUMO is suitable for S_N2 reaction (Fig. 1, left); however, in some of the other compounds, i.e., some AHC compounds, LUMOs are not suitable for S_N2 reaction (Fig. 1, center). In such cases, LUMOs cannot be used for prediction of mutagenicity. Therefore, selection of the MO which was suitable for S_N2 reaction (SN2MO) was required. In this research, SN2MO was selected by chemical-visual inspection based on the shape and position of MOs.

Figure 4 shows one example of LUMO (lower) and SN2MO (upper) of ethyl chloroacetate. In SN2MO, the MO exists at the opposite position of the bond (C-Cl bond) that is cleaved by S_N2 reaction. Among the MOs in a molecule, those occupied with electrons work as nucleophiles, and unoccupied MOs work as electrophiles. In addition, AHC acts as an electrophile in mutagenic process and MOs with lower energy or level react more easily. Therefore, we focused on LUMOs and SN2MOs to predict mutagenicity. In this research, we defined the level gap between SN2MO and LUMO (LG; unit: no dimension) and the energy gap between SN2MO and LUMO (EG; unit: eV). For example, in Fig. 4, LG of ethyl chloroacetate was 2, which meant SN2MO existed at LUMO + 2.

Fig. 4

An image of MO and an example of an MO at LUMO and SN2MO of Ethyl chloroacetate. Arrows indicate electrons. In this case, the SN2MO existed at LUMO+2. HOMO: highest occupied molecular orbital.

Descriptor selection

Individual LG and EG data for 21 AHC compounds are shown in Table 2. To confirm our hypothesis, we assessed ROC analysis results to evaluate the mutagenicity predictive ability of each property. As shown in Table 3, LG was an ideal descriptor for mutagenicity prediction, because it displayed a better AUC value than did LUMO or SN2MO energies (AUC value, LG = 0.89, EG = 0.70, LUMO = 0.76, SN2MO = 0.57). This result indicated that the relationship between SN2MO and LUMO, but not the value of SN2MO, was a significant factor to predict mutagenicity of AHC compounds. Therefore, we used LG and EG as descriptors for prediction of mutagenicity.

Table 2. Individual data of the AHC compounds investigated.

No.	Structure Class	Name	LUMO (eV)	SN2MO (eV)	LG	EG (eV)	Prediction	Ames
1	Ketone	chloro acetone	1.721	33.356	1	31.635	Positive	Positive
2	Ketone	2-chloro cyclohexanone	−1.535	29.192	1	30.727	Positive	Positive
3	Ketone	1,3- dichloro acetone	−1.812	33.761	1	35.573	Positive	Positive
4	Ketone	1,1,3,3-tetrachloroacetone	−1.823	33.916	1	35.739	Positive	Positive
5	Amide	chloro acetamide	3.251	34.135	1	30.884	Positive	Marginal*
6	Amide	2-chloro-N-(hydroxymethyl) acetamide	3.354	34.235	2	30.881	Positive	Marginal*
7	Amide	ethyl (2-bromopropanoyl)glycinate	0.040	27.080	2	27.04	Positive	Positive
8	Ester	ethyl 2-bromo propionate	0.066	26.980	1	26.914	Positive	Positive
9	Ester	ethyl chloroacetate	0.025	34.878	2	34.853	Negative	Negative
10	Ester	isopropyl chloroacetate	0.058	27.259	1	27.201	Positive	Positive
11	Ester	methyl chloroacetate	0.022	35.258	2	35.236	Negative	Negative
12	Ester	methyl 2-chloro propionate	0.103	27.787	1	27.684	Positive	Positive
13	Acid	chloro acetic acid	0.280	36.913	2	36.633	Negative	Negative
14	Acid	2-chloro propionic acid	0.298	34.516	2	34.218	Negative	Negative
15	Aryl ketone	4-bromo phenyl bromomethyl ketone	−4.117	25.481	4	29.598	Negative	Negative
16	Aryl ketone	2,4-dichlorophenyl chloromethyl ketone	−3.571	32.165	4	35.736	Negative	Positive
17	Aryl ketone	3-methoxy phenyl bromomethyl ketone	−4.663	25.478	4	30.141	Negative	Negative
18	Aryl ketone	4-methylphenyl chloromethyl ketone	−4.524	33.284	4	37.808	Negative	Negative
19	Aryl ketone	phenyl chloromethyl ketone	−4.744	33.280	4	38.024	Negative	Negative
20	Aryl ketone	phenyl bromomethyl ketone	−4.745	25.489	4	30.234	Negative	Negative

Italic: false prediction.

*: Marginal was defined in CASE Ultra and is used as positive in this research.

LG: the level gap between SN2MO and LUMO; EG (eV): the energy gap between SN2MO and LUMO

All compounds were predicted as positive by GT_EXPERT.

Table 3. The result of ROC analysis of LG, EG, LUMO, and SN2MO.

Physicochemical Property	ROC AUC
LG	0.89
EG	0.70
LUMO energy	0.76
SN2MO energy	0.57

Mutagenicity prediction with MO

Figure 5a shows the relationship between EG and mutagenicity. Mutagenic compounds tended to distribute to the areas with lower EG value. The borderline probably existed around EG = 31 eV, i.e., the values of marginal compounds (No.6: 30.884 eV, No.7: 30.881 eV; Table 2). Therefore, we set EG of 31 ± 2 eV as a “gray zone.” The gray zone was defined as co-existence of positive and negative compounds at similar rates (Haranosono et al., 2014, 2016). In a similar way, Fig. 5b shows the relationship between LG and mutagenicity. Mutagenic compounds tended to distribute to the areas with lower LG value. In other words, compounds with low LG tended to be positive. The borderline probably existed around LG = 1, and we therefore set the gray zone at LG = 1. A 2D plot of EG and LG is shown in Fig. 6. The higher EG and higher LG compounds tended to be negative. This result supported our hypothesis. The positive compounds existed mainly at LG = 1 or EG ≤ 29 eV in this dataset.

Fig. 5

2D-plot of EG and mutagenicity (5a) and bubble plot of LG and mutagenicity (5b) of AHC compounds. The compounds with lower EG or lower LG tended to be mutagenic. In Fig. 5a, filled diamond: positive, empty diamond: marginal, empty circle: negative. In Fig. 5b, the numbers in circles indicate the bubble size (= the number of compounds). Back lines indicate the cut-off lines.

Fig. 6

The relationship between the S_N2 reactivity and mutagenicity of AHC compounds. The compounds with lower LG and lower EG tended to be mutagenic. This figure indicates that the compounds with high S_N2 reactivity tended to be mutagenic. The chemical structure indicates the false negative compound, which had two electron-withdrawing groups on a benzene ring. Filled diamond: positive, empty diamond: marginal, empty circle: negative. Black line: cut-off line.

To predict mutagenicity, we used a “scoring” method (Haranosono et al., 2016). Each compound was given a “score” for each descriptor as follows: positive zone = +1, negative zone = −1, and gray zone = 0. The scores of EG and LG were summed to combine these descriptors in this prediction method, and the summed score was defined as “total score” for each compound. We defined “total score of 0 or above” as the positive area, and “total score lower than zero” as the negative area. The prediction results with our method based on above rule were showed in Table 2, we obtained 95.0% concordance (1 false negative) for 20 AHC compounds (Table 4).

Table 4. The prediction power of our method.

		Our method
		Positive	Negative	Sum
Ames test	Positive	9	1	10
	Negative	0	10	10
	Sum	9	11	20
Sensitivity		90.0%
Specificity		100.0%
Positive rate		100.0%
Negative rate		90.9%
Concordance		95.0%

DISCUSSION

Our MO method was able to predict the mutagenicity with 95.0% concordance for AHC compounds. The dataset included very similar compounds, e.g., ethyl chloroacetate and isopropyl chloroacetate, and it is likely to be difficult to discriminate such compounds using QSARs (GT_EXPERT, concordance = 50.0%). Therefore, we consider that knowledge and discussion of the reaction mechanism with DNA are required to predict the mutagenicity, and our method is a powerful tool for AHC compounds. In drug development, our method in combination with QSARs (both of knowledge-based and statistical-based) will be a useful tool for an expert judge (ICH M7 Guideline, 2014) in the assessment of mutagenicity of impurities in pharmaceuticals to avoid misjudging of AHC compounds because AHC compounds are basically predicted as positive with knowledge-based QSAR.

There was one false negative compound, an aromatic AHC, 2,4-dichlorophenyl chloromethyl ketone (Fig. 6). Aromatic ring increases electron density of carbonyl of aromatic AHC by the resonance effect, which decreases the reactivity at the α-position of aromatic AHC. We consider that this is the reason the aromatic AHC compounds were generally negative. In the case of one false negative compound, the benzene ring is substituted with two electron-withdrawing groups, which decreases the electron density of carbonyl of AHC and increases the reactivity at the α-position. Therefore, the electron density or acidity at α-position probably is an index of the S_N2 reactivity in aromatic AHC compounds. The S_N2 reactivity contributes on the mutagenicity; however, the kind of index of S_N2 reactivity might be different in the class of chemical structure and its sterical/electrical environment. In addition, the permeability of a compound should relate to mutagenicity. Because the compounds in this research were small (molecular weight < 230) and do not have any free amino groups, we thought that the permeability of them may be critical in Ames test condition without two compounds with carboxylic acid.

In this research, we hypothesized that AHC reacts with DNA by S_N2 reaction as with alkyl halides and epoxides. The compounds with low EG and/or LG tended to be mutagenic (Fig. 6), which means that they have high potential for S_N2 reaction. This tendency suggests that S_N2 reaction contributes on the reaction mechanism of AHC and DNA in the mutagenic process. Although more investigations (such as the S_N2 reactivity, the electron density, the sterical effect, or cross-validation) are needed to define the main reaction of AHC with DNA, our result suggests an approach for avoiding mutagenicity by decreasing the S_N2 reactivity of AHC compounds.

In conclusion, we presented a new mutagenicity prediction method for AHC compounds, which was based on the S_N2 reactivity by MOs. Our research suggests that S_N2 reaction contributes on the reaction mechanism of AHC and DNA in the mutagenic process. By developing the research of reaction mechanisms of structural alerts and DNA, the mutagenicity prediction should come to have high concordance when combined with QSARs.

ACKNOWLEDGMENT

The authors are grateful to Dr. Tetsuya Tajika of Senju Pharmaceutical Co., Ltd., for reading the manuscript and providing constructive comments.

Conflict of interest

The authors declare that there is no conflict of interest.

REFERENCES

• Aiba née Kaneko, M., Hirota, M., Kouzuki, H. and Mori, M. (2015): Prediction of genotoxic potential of cosmetic ingredients by an in silico battery system consisting of a combination of an expert rule-based system and a statistics-based system. J. Toxicol. Sci., 40, 77-98.
• Ashby, J., Kier, L., Wilson, A.G., Green, T., Lefevre, P.A., Tinwell, H., Willis, G.A., Heydens, W.F. and Clapp, M.J. (1996): Evaluation of the potential carcinogenicity and genetic toxicity to humans of the herbicide acetochlor. Hum. Exp. Toxicol., 15, 702-735.
• Azuma, S., Kishino, S., Katayama, S., Akahori, Y. and Matsushita, H. (1997): Mutagenicity of 12 HPLC labeling reagents for analysing carboxyl compounds. KANKYO KAGAKU, 7, 249-255.
• Benigni, R. (2005): Structure-Activity Relationship studies of chemical mutagens and carcinogens: mechanistic investigations and prediction approaches. Chem. Rev., 105, 1767-1800.
• Benigni, R. and Bossa, C. (2011): Mechanisms of chemical carcinogenicity and mutagenicity: A review with implications for predictive toxicology. Chem. Rev., 111, 2507-2536.
• Bertenyi, K.K. and Lambert, I.B. (1996): The mutational specificity of furazolidone in the lacI gene of Escherichia Coli. Mutat. Res., 357, 199-208.
• Burket, U. and Allinger, N.L. (1982): Molecular mechanics. ACS monograph Series (Book 117). pp.1-340. American Chemical Society. Washington D.C.
• Clayden, J., Greeves, N. and Warren, S. (2003): Organic Chemistry. p.116. Tokyo Kagaku Dojin. Tokyo.
• Debnath, A.K., de Compadre, R.L. and Hansch, C. (1992a): Mutagenicity of quinolines in Salmonella typhimurium TA100. A QSAR study based on hydrophobicity and molecular orbital determinants. Mutat. Res., 280, 55-65.
• Debnath, A.K., Debnath, G., Shusterman, A.J. and Hansch, C. (1992b): A QSAR investigation of the role of hydrophobicity in regulating mutagenicity in the Ames test: 1. Mutagenicity of aromatic and heteroaromatic amines in Salmonella typhimurium TA98 and TA100. Environ. Mol. Mutagen., 19, 37-52.
• Dolzani, L., Tamaro, M., Monti-Bragadin, C., Cavicchioni, G., Vecchiati, G. and D’Angeli, F. (1986): Selective mutagenic activity of 2-bromo-propanamides on Salmonella typhimurium. Mutat. Res., 172, 29-36.
• Freeman, B.A., Wilson, R.E., Binder, R.G. and Haddon, W.F. (2001): Halogenated 2,5-pyrrolidinediones: synthesis, bacterial mutagenicity in Ames tester strain TA-100 and semi-empirical molecular orbital calculations. Mutat. Res., 490, 89-98.
• Giller, S., Le Curieux, F., Erb, F. and Marzin, D. (1997): Cmparative genotoxicity of halogenated acetic acids found in drinking water. Mutagenesis, 12, 321-328.
• Gold, B. and Leuschen, T. (1983): α-chloroepoxides. 2. mutagenicity of 1-chlorocyclohexene oxide and its thermal isomerization product, 2-chlorocyclohexanone. Chem. Biol. Interact., 45, 305-314.
• Haranosono, Y., Kurata, M. and Sakaki, H. (2014): Establishment of an in silico phototoxicity prediction method by combining descriptors related to photo-absorption and photo-reaction. J. Toxicol. Sci., 39, 655-664.
• Haranosono, Y., Nemoto, S., Kurata, M. and Sakaki, H. (2016): Establishment of an in silico phospholipidosis prediction method using descriptors related to molecular interactions causing phospholipid-compound complex formation. J. Toxicol. Sci., 41, 321-328.
• Hatch, F.T. and Colvin, M.E. (1997): Quantitative structure-activity (QSAR) relationships of mutagenic aromatic and heterocyclic amines. Mutat. Res., 376, 87-96.
• Hu, J., Wang, W., Zhu, Z., Chang, H., Pan, F. and Lin, B. (2007): Quantitative structure-activity relationship model for prediction of genotoxic potential for quinolone antibacterials. Environ. Sci. Technol., 41, 4806-4812.
• JCIETI CENTER. JAPAN. (1996): Mutagenicity test data of existing chemical substances based on the toxicity investigation system of the industrial safety and health law. JETOC. Tokyo.
• Judson, P.N., Cooke, P.A., Doerrer, N.G., Greene, N., Hanzlik, R.P., Hardy, C., Hartmann, A., Hinchliffe, D., Holder, J., Müller, L., Steger-Hartmann, T., Rothfuss, A., Smith, M., Thomas, K., Vessey, J.D. and Zeiger, E. (2005): Towards the creation of an international toxicology information centre. Toxicology, 213, 117-128.
• Harmonised Tripartite Guidline, I.C. Assessment and Control of DNA Reactive (Mutagenic) Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk. (2014): International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH M7 Guideline.
• Kazius, J., McGuire, R. and Bursi, R. (2005): Derivation and validation of toxicophores for mutagenicity prediction. J. Med. Chem., 48, 312-320.
• Kier, L.E., Brusick, D.J., Auletta, A.E., Von Halle, E.S., Brown, M.M., Simmon, V.F., Dunkel, V., McCann, J., Mortelmans, K., Prival, M., Rao, T.K. and Ray, V. (1986): The Salmonella typhimurium/mammalian microsomal assay: A report of the U.S. Environmental Protection Agency Gene-Tox Program. Mutat. Res., 168, 69-240.
• Le Curieux, F., Marzin, D. and Erb, F. (1994): Study of the genotoxic activity of five chlorinated propanones using the SOS chromotest, the Ames-fluctuation test and the newt micronucleus test. Mutat. Res., 341, 1-15.
• Lopez de Compadre, R.L., Debnath, A.K., Shusterman, A.J. and Hansch, C. (1990): LUMO energies and hydrophobicity as determinants of mutagenicity by nitroaromatic compounds in Salmonella typhimurium. Environ. Mol. Mutagen., 15, 44-55.
• Meier, J.R., Ringhand, H.P., Coleman, W.E., Munch, J.W., Streicher, R.P., Kaylor, W.H. and Schenck, K.M. (1985): Identification of mutagenic compounds formed during chlorination of humic acid. Mutat. Res., 157, 111-122.
• Merrick, B.A., Smallwood, C.L., Meier, J.R., McKean, D.L., Kaylor, W.H. and Condie, L.W. (1987): Chemical reactivity, cytotoxicity, and mutagenicity of chloropropanones. Toxicol. Appl. Pharmacol., 91, 46-54.
• Misaki, K., Hisamatsu, Y., Suzuki, H. and Takamura-Enya, T. (2008): Evaluation of the mutagenicity of nitration products derived from phenalenone (1H-phenalen-1-one). Mutagenesis, 23, 359-366.
• National Toxicity Program. Testing Status of 2,2’,4’-Trichloroacetophenone M882477. URL: https://ntp.niehs.nih.gov/testing/status/agents/ts-m882477.html
• Nestmann, E.R., Douglas, G.R., Kowbel, D.J. and Harrington, T.R. (1985): Solvent interactions with test compounds and recommendations for testing to avoid artifacts. Environ. Mutagen., 7, 163-170.
• Saito, H., Isoda, S., Kato, M. and Nagaoka, N. (1995): Mutagenic activity of indoor swimming pool water. KANKYO HENIGEN KENKYU, 17, 169.
• Sutter, A., Amberg, A., Boyer, S., Brigo, A., Contrera, J.F., Custer, L.L., Dobo, K.L., Gervais, V., Glowienke, S., van Gompel, J., Greene, N., Muster, W., Nicolette, J., Reddy, M.V., Thybaud, V., Vock, E., White, A.T. and Müller, L. (2013): Use of in silico systems and expert knowledge for structure-based assessment of potentially mutagenic impurities. Regul. Toxicol. Pharmacol., 67, 39-52.
• Tuppurainen, K., Lötjönen, S., Laatikainen, R., Vartiainen, T., Maran, U., Strandberg, M. and Tamm, T. (1991): About the mutagenicity of chlorine-substituted furanones and halopropenals. A QSAR study using molecular orbital indices. Mutat. Res., 247, 97-102.
• Tuppurainen, K. (1999): Frontier orbital energies, hydrophobicity and steric factors as physical QSAR descriptors of molecular mutagenicity. A review with a case study: MX compounds. Chemosphere, 38, 3015-3030.
• Valerio, L.G. Jr. and Cross, K.P. (2012): Characterization and validation of an in silico toxicology model to predict the mutagenic potential of drug impurities. Toxicol. Appl. Pharmacol., 260, 209-221.
• Watanabe, K., Sakamoto, K. and Sasaki, T. (1996): Comparisons on chemically-induced mutagenicity among four bacterial strains, Salmonella typhimurium TA102 and TA2638, and Escherichia Coli WP2/pkm101 and WP2 uvrA/pkm101: collaborative study I. Mutat. Res., 361, 143-155.
• Yaguchi, Y., Shiozawa, T., Matsushita, H. and Terao, Y. (1993): Mutagenic activities of chiral epoxides, halides, and N-nitrosamines. Biol. Pharm. Bull., 16, 942-944.
• Zeiger, E., Anderson, B., Haworth, S., Lawlor, T., Mortelmans, K. and Speck, W. (1987): Salmonella mutagenicity tests: III. Results from the testing of 255 chemicals. Environ. Mol. Mutagen., 9 Suppl 9, 1-109.
• Zijlstra, J.A. (1989): Liquid holding increases mutation induction by formaldehyde and some other cross-linking agents in Escherichia coli k12. Mutat. Res., 210, 255-261.