DNA damage induced in cultured human alveolar (L-132) cells by exposure to dimethylarsinic acid.

Gene damage in cultured human alveolar (L-132) cells induced by exposure to dimethylarsinic acid (DMAA), a major metabolite of inorganic arsenics in mammals, was studied. DNA single-strand breaks and DNA-protein cross-links were induced by the treatment of L-132 cells with 10 mM DMAA. These kinds of damage appeared at 8 hr after start of exposure to DMAA. As regards DNA-protein cross-links, the DNA was found to bind not only to core histone proteins but also linker histone (H1) and nonhistone proteins. Furthermore, the cross-links were formed by the binding to serine or threonine residues of H1 or nonhistone proteins through phosphate moieties of the DNA. The induction of the alkali-labile sites in DNA in DMAA-treated L-132 cells was observed prior to that of DNA single-strand breaks and DNA-protein cross-links. As one of the alkali-labile sites in DNA, we estimated apurinic/apyrimidinic (AP) sites in DNA. The present study suggests that the DNA single-strand breaks and DNA-protein cross-links induced by the treatment of L-132 cells with DMAA occurred via the formation of AP sites in the DNA and that the DNA-protein cross-links were produced by a Schiff-base reaction between amino groups of nuclear proteins and aldehyde groups of AP sites in the DNA and the DNA single-strand breaks, by a beta-elimination reaction on AP sites in the DNA.


Introduction
Although inorganic arsenics are epidemiologically proved to be carcinogenic for lung and skin (1,2), experimental studies so far have not succeeded in evidencing their carcinogenicity. On the other hand, among a number of investigations on the genotoxicity of inorganic arsenics, some data (3)(4)(5)(6) have suggested that inorganic arsenics are not direct-acting mutagens, but act as comutagens with other carcinogens, e.g., ultraviolet and alkylating agents, through modification of the DNA repair process. In fact, Li and Rossman reported that the activity of DNA ligase II was suppressed by arsenite (7). However, many more studies will be required to solve the genotoxic action of inorganic arsenics.
This article describes the gene damage induced by exposure of L-132 cells, an established human cell line of alveolar type-II cells, to DMAA. We found that DNA single-strand breaks (15) and DNA-protein crosslinks (16) in L-132 cells were induced by treatment with DMAA, as observed in in vivo experiments (10,11), and assumed that the damage might be induced via the formation of apurinic/apyrimidinic (AP) sites in DNA.

Materials and Methods
L-132 cells, an established human embryonic cell line of alveolar epithelial cells (17) The alkaline elution method established by Kohn et al. (18) was used to measure DNA single-strand breaks; and alkaline agarose gel electrophoresis (19), was used to measure alkali-labile sites in DNA. To detect AP sites in DNA, DNA was isolated by phenol extraction from L-132 cells. The DNA was treated with a restriction enzyme (Kpn I) and then subjected to alkaline agarose gel electrophoresis. The formation of DNA-protein cross-links was determined according to the filter binding assay described by Strniste and Rall (20). For separation of DNA-nuclear protein complexes, chromatin prepared from L-132 cells according to the method Environmental Health Perspectives of Haung and Haung (21) was chromatographed on Sepharose 4B (22). Induction of DNA Single-Strand Breaks and DNA-Protein Crosslinks To confirm if DNA single-strand breaks found in mouse lung in vivo can be observed also in an in vitro cell culture system, we determined the strand breaks in L-132 cells exposed to DMAA by the alkaline elution method (15). When L-132 cells were incubated in medium containing DMAA at concentrations of 5, 7.5, and 10 mM for 12 hr, concentration-dependent single-strand breaks occurred. Furthermore, as shown in Figure 1, in the case of exposure to 10 mM DMAA, the induction of the single-strand breaks was dependent on the duration of incubation with the DMAA. The breaks were not detected up to 6 hr incubation but were observed remarkably at 8 to 12 hr after start of DMVAA exposure. To evaluate the recovery from the breaks, we first exposed L-132 cells to 10 mM DMAA for 12 hr and then transferred them to DMAA-free medium; thereafter, the degree of the single-strand breaks was determined by the alkaline elution assay. The recovery proceeded timedependently; the breaks were partially recovered at 2 to 4 hr and fully at 6 hr after transfer into the DMAA-free medium.
On the other hand, for the measurement of DNA-protein cross-links, the filter binding assay was performed (16). As shown in Table 1, the amount of DNA retained on the filter was not changed up 10 20 30 40 Fraction number Figure  Sepharose 4B. 3H-radioactivity (0) of DNA and 14Cradioactivity (0) of nuclear proteins in each fraction were measured in a liquid scintillation counter. The chromatographic conditions were as follows: column size, 1.6 x 20 cm; eluent, 0.45 M NaCI/10 mM sodium bisulfite/10 mM EDTA/25 mM Tris-HCI (pH7.0); fraction size, 2 ml; flow rate, 0.2 ml/min. to 6 hr but increased by 12 hr after exposure to DMAA. These results indicate that DNA single-strand breaks and DNA-protein cross-links occur similarly in mouse lung in vivo (11,16) and in L-132 cells in vitro (15,16), suggesting that the present in vitro system is useful to analyze the DNA damage observed in the in vivo experiments.

Cross-finks
To characterize the type of the cross-links induced by the DMAA treatment, the chromatin, with its DNA and nuclear proteins prelabeled with [3H]thymidine and [14C]lysine, respectively, prepared from L-132 cells, was sheared in a buffer containing 0.45 M NaCl using a Polytron homogenizer. Under this condition, linker histone (HI) and nonhistone proteins are dissociated from DNA but core histone proteins are not (22). The soluble chromatin sample was separated on Sepharose 4B column chromatography, and each fraction was measured for 3H-radioactivity of DNA and 14C-radioactivity of protein (16). The control sample gave two peaks ( Figure 2A); one was DNA-core histone complex in the void volume and the other, free nuclear proteins (HI and nonhistone proteins). The sample from the cells exposed to 10 mM DMAA for 12 hr, on the other hand, showed a marked change; some 3H-radioactivity appeared in the fraction of free nuclear proteins, indicating the presence of DNA fragments associated with the proteins ( Figure 2B). This result indicates that the formation of DNA cross-links occurred not only with core histone, but also with HI or nonhistone proteins. The cross-link between DNA and nuclear proteins was dissociated by the treatment with 6 M guanidine hydrochloride (16). As it is known that the bonds of phosphoserine and phosphothreonine are labile to 6 M guanidine hydrochloride (23), the crosslink formation is thus very likely due to the binding between phosphate moieties of DNA and serine or threonine residues of H1 histone or nonhistone proteins (16).

Formation ofAP Sites in DNA in DMAA-Treated L-132 Cells
To clarify DNA damage induced at the early stage after exposure to DMAA, we used alkaline agarose gel electrophoresis to examine whether the exposure to DMAA of L-132 cells formed alkali-labile sites in DNA. As shown in Figure 4, DNA in the cells exposed to DMAA for 6 hr underwent a greater degree of alkaline hydrolysis than control DNA. This result indicates that, by exposure of L-132 cells to 10 mM DMAA, alkali-labile sites were formed in the DNA prior to the induction of DNA singlestrand breaks and DNA-protein cross-links, suggesting the formation of AP sites in DNA at an early stage of DMAA exposure. On the other hand, Liuzzi and Talpaert-Borle (24) reported that methoxyamine bound to aldehyde groups of AP sites in DNA and that the methoxyamine-bound AP sites were more resistant to the hydrolysis than nonbound AP sites under alkaline conditions. In fact, when L-132 cells were exposed to DMAA for 6 hr, the DNA preteated with methoxyamine became alkali-resistant (Figure 3)  formed AP sites in the DNA prior to the induction of DNA strand scissions (15) and DNA-protein cross-links (16).

Discussion
Based on the present study, we propose a possible pathway of DNA damage induced by dimethylarsenics, as schematically shown in Figure 4. We had assumed that lung-specific DNA damage induced by DMAA administration in mice might be attributed to the dimethylarsenic peroxyl radical [(CH3)2AsOO*] and active oxygens produced in the metabolic processing of DMAA (12). We even speculated that DNA single-strand breaks and DNA-protein cross-links were directly caused by these radicals, as in the case of ionizing radiation (11). However, the results obtained from the present study may not fully support our previous speculation. Here we assume that DNA single-strand breaks and DNAprotein cross-links in L-132 cells exposed to DMAA may be indirectly caused by these radicals via the formation of AP sites in DNA. As regards the induction mechanism of DNA damage, we estimated that the cross-link formation occurred by a Schiff-base reaction between amino groups of nuclear proteins, particularly of H 1 and nonhistone proteins, and aldehyde groups of AP sites in DNA and that the induction of DNA single-strand breaks occurred after a :-elimination reaction at AP sites in the DNA. On the other hand, some of the base damage induced by exposure to alkylating agents such as N-methyl-N-nitrosourea and dimethylsulfate is known to be repaired through the excision of damaged bases by the action of DNA glycosylase, which hydrolyses N-glycosyl bonds to form AP sites in DNA (25). The formation of AP sites is considered one of the important processes of excision repair of not only alkylated bases, but also a variety of DNA adducts. We therefore speculate that the exposure of L-132 cells to DMAA might cause enzymatic formation of AP sites in DNA via the production of base damage such as DNA-adduct formation (15).