Hawk-Seq™ differentiates between various mutations in Salmonella typhimurium TA100 strain caused by exposure to Ames test-positive mutagens

Abstract

A precise understanding of differences in genomic mutations according to the mutagenic mechanisms detected in mutagenicity data is required to evaluate the carcinogenicity of environmental mutagens. Recently, we developed a highly accurate genome sequencing method, 'Hawk-Seq™', that enables the detection of mutagen-induced genome-wide mutations. However, its applicability to detect various mutagens and identify differences in mutational profiles is not well understood. Thus, we evaluated DNA samples from Salmonella typhimurium TA100 exposed to 11 mutagens, including alkylating agents, aldehydes, an aromatic nitro compound, epoxides, aromatic amines and polycyclic aromatic hydrocarbons (PAHs). We extensively analysed mutagen-induced mutational profiles and studied their association with the mechanisms of mutagens. Hawk-Seq™ sensitively detected mutations induced by all 11 mutagens, including one that increased the number of revertants by approximately 2-fold in the Ames test. Although the sensitivity for less water-soluble mutagens was relatively low, we increased the sensitivity to obtain high-resolution spectra by modifying the exposure protocol. Moreover, two epoxides indicated similar 6- or 96-dimensional mutational patterns; likewise, three SN1-type alkylating agents indicated similar mutational patterns, suggesting that the mutational patterns are compound category specific. Meanwhile, an SN2 type alkylating agent exhibited unique mutational patterns compared to those of the SN1 type alkylating agents. Although the mutational patterns induced by aldehydes, the aromatic nitro compound, aromatic amines and PAHs did not differ substantially from each other, the maximum total base substitution frequencies (MTSFs) were similar among mutagens in the same structural groups. Furthermore, the MTSF was found to be associated with the carcinogenic potency of some direct-acting mutagens. These results indicate that our method can generate high-resolution mutational profiles to identify characteristic features of each mutagen. The detailed mutational data obtained by Hawk-Seq™ can provide useful information regarding mutagenic mechanisms and help identify its association with the carcinogenicity of mutagens without requiring carcinogenicity data.

Figure 1.

Mutational spectra induced by (a) MNNG (30, 50 µg/tube), (b) MMS (30, 1500, 3500 µg/tube), (c) glyoxal (120, 200 µg/tube), (d) FA (60 µg/tube), (e) 4NQO (0.5, 1 µg/tube), (f) glycidol (10 000, 20 000 µg/tube) and (g) PO (20 000, 30 000 µg/tube) in TA100 cells. The BS frequencies per 10⁶ G:C or A:T base pairs are displayed (n = 3). Error bars represent standard deviation. Asterisks and daggers indicate P values (Dunnett’s multiple comparison test: *P < 0.05, ^†P < 0.01 and ^‡P < 0.001; Student’s t test: ^§P < 0.05).

Figure 2.

Analyses of aromatic amine mutagens. (a) OD660 values for TA100 cell suspensions cultured for 14 h in NB (white circles) or NB+S9 (black circles) after 2AA exposure are shown. The growth of TA100 cells cultured in NB+S9 was significantly inhibited. BS frequencies in TA100 cells cultured in NB+S9 after (b) 2AA (100, 200 µg/tube) or (c) 2AAF (600, 1000 µg/tube) exposure. The BS frequencies per 10⁶ G:C or A:T base pairs are displayed (n = 3). Error bars represent standard deviation. Asterisks and daggers indicate P values (Dunnett’s multiple comparison test; *P < 0.05, ^†P < 0.01 and ^‡P < 0.001).

Figure 3.

Enhancement of PAH-induced mutation sensitivity by modifying exposure to mutagens. BS frequencies in TA100 cells cultured in NB+S9 after (a) 3MC (1000, 2000 µg/tube) and (b) DMBA (1000, 2000 µg/tube) exposure. BS frequencies in TA100 cells cultured for 48 h on minimum glucose medium plate after (c) 3MC (1000, 2000 µg/plate) and (d) DMBA (1000, 2000 µg/plate) exposure. The BS frequencies per 10⁶ G:C or A:T base pairs are displayed (n = 3). Error bars represent standard deviation. Asterisks and daggers indicate P values (Dunnett’s multiple comparison test; *P < 0.05, ^†P < 0.01 and ^‡P < 0.001).

Figure 4.

The 96-dimensional mutation pattern in TA100 cells after mutagen exposure. The mean proportion of BS frequencies for each trinucleotide in cells exposed to MNNG (50 µg/tube), MMS (3500 µg/tube), glycidol (20 000 µg/tube), PO (30 000 µg/tube), glyoxal (200 µg/tube), FA (60 µg/tube), 4NQO (1.0 µg/tube), 2AA (100 µg/tube), 2AAF (1000 µg/tube), 3MC (1000 µg/plate) and DMBA (1000 µg/plate) are shown.

Figure 5.

PCA results using the 6-dimensional mutation spectra for each mutagen. (a) Cumulative proportion of the variance of each principal component (PC1–PC6). These values are added up to 1. (b) PCA loading of each base substitution type on PC1 and PC2. The PC1 score correlates positively with G:C > A:T and A:T > G:C and negatively with G:C > T:A and G:C > C:G. The PC2 score correlates positively with mutations on the A:T base pair. (c) The PCA score plot for each mutagen is based on their PC1 and PC2 values.

Figure 6.

The maximum total BS frequency induced by each mutagen is shown. The total BS frequencies per 10⁶ base pairs are displayed in log-scale (n = 3). Error bars represent standard deviation.