Involvement of N-6 adenine-specific DNA methyltransferase 1 (N6AMT1) in arsenic biomethylation and its role in arsenic-induced toxicity.

BACKGROUND: In humans, inorganic arsenic (iAs) is metabolized to methylated arsenical species in a multistep process mainly mediated by arsenic (+3 oxidation state) methyltransferase (AS3MT). Among these metabolites is monomethylarsonous acid (MMAIII), the most toxic arsenic species. A recent study in As3mt-knockout mice suggests that unidentified methyltransferases could be involved in alternative iAs methylation pathways. We found that yeast deletion mutants lacking MTQ2 were highly resistant to iAs exposure. The human ortholog of the yeast MTQ2 is N-6 adenine-specific DNA methyltransferase 1 (N6AMT1), encoding a putative methyltransferase.
OBJECTIVE: We investigated the potential role of N6AMT1 in arsenic-induced toxicity.
METHODS: We measured and compared the cytotoxicity induced by arsenicals and their metabolic profiles using inductively coupled plasma-mass spectrometry in UROtsa human urothelial cells with enhanced N6AMT1 expression and UROtsa vector control cells treated with different concentrations of either iAsIII or MMAIII.
RESULTS: N6AMT1 was able to convert MMAIII to the less toxic dimethylarsonic acid (DMA) when overexpressed in UROtsa cells. The enhanced expression of N6AMT1 in UROtsa cells decreased cytotoxicity of both iAsIII and MMAIII. Moreover, N6AMT1 is expressed in many human tissues at variable levels, although at levels lower than those of AS3MT, supporting a potential participation in arsenic metabolism in vivo.
CONCLUSIONS: Considering that MMAIII is the most toxic arsenical, our data suggest that N6AMT1 has a significant role in determining susceptibility to arsenic toxicity and carcinogenicity because of its specific activity in methylating MMAIII to DMA and other unknown mechanisms.

Inorganic arsenic (iAs) compounds are considered known human carcinogens that target multiple sites, including the lung, skin, and urinary bladder [International Agency for Research on Cancer (IARC) 2004; Pershagen 1981; Smith et al. 1992; Smith and Steinmaus 2009; Straif et al. 2009]. In addition, chronic exposure to high levels of iAs has been associated with the development of multiple diseases and deleterious health effects in humans (Abernathy et al. 1999; Kapaj et al. 2006).

In humans, as in many animal species, iAs is metabolized to monomethylarsonous acid (MMA) and dimethylarsonic acid (DMA). The most cited conceptual model of arsenic methylation involves the reduction of pentavalent iAs (iAsV) to trivalent iAs (iAsIII), with subsequent methylation (Drobna et al. 2009). The general scheme is as follows:

Among these metabolites, MMAIII is the most toxic arsenic species (Drobná et al. 2005; Ferrario et al. 2008; Kligerman et al. 2003; Petrick et al. 2001). It is generally accepted that arsenic (+3 oxidation state) methyltransferase (AS3MT) is responsible for catalyzing methyl group transfer from S-adenosyl methionine (SAM) to iAs (Thomas et al. 2007). However, a recent study by Drobna et al. (2009) showed that knockout of As3mt in the mouse does not completely abolish the methylation of iAs, suggesting that there are alternative pathways for arsenic methylation in these animals. Although Bentley and Chasteen (2002) and Hall et al. (1997) suggested that arsenic methylation could be due to gastrointestinal tract microbiota, they also speculated that unidentified methyltransferses may be responsible for the methylated arsenicals found in As3mt-knockout mice.

We conducted a genomewide, parallel phenotypic screen of yeast deletion mutants to identify the genes required for the growth of yeast in the presence of MMAIII and iAsIII (Jo et al. 2009). We found a yeast strain with deletion of MTQ2, which was highly resistant to iAsIII. MTQ2 encodes a SAM-dependent methyltransferase and has been shown to be involved in the methylation of release factor eRF1 in yeast (Polevoda et al. 2006). The human ortholog of the yeast MTQ2 is N-6 adenine-specific DNA methyltransferase 1 (N6AMT1), a putative methyltransferase. Although the bacterial homologs of N6AMT1 have been shown to methylate DNA N6-adenine (Stephens et al. 1996), the current data do not indicate its function in the methylation of adenine in the DNA of mammalian cells (Ratel et al. 2006).

Our goal in the present study was to explore the mechanism by which N6AMT1 confers resistance to arsenic toxicity. We enhanced N6AMT1 gene expression in UROtsa cells, given its relatively low expression in these cells. The UROtsa cell line, originally isolated from a primary culture of normal human uroepithelium, does not methylate arsenic because of the absence of AS3MT expression (Drobná et al. 2005; Styblo et al. 2000) and has been used as a model for bladder epithelium and arsenic-induced bladder cancer (Bredfeldt et al. 2006; Eblin et al. 2008; Sens et al. 2004). Here, we show that N6AMT1 is a human methyltransferase specifically involved in the biomethylation of MMAIII to DMA. Given that MMAIII is the most toxic arsenical and its implication in arsenic toxicity and carcinogenicity, N6AMT1 may have a significant role in modulating arsenic-induced toxicity and carcinogenicity.

Materials and Methods

Cultures of yeast strains and human UROtsa cells

The wild-type BY4743 yeast strain was purchased from Invitrogen (Carlsbad, CA), and the MTQ2 deletion strain has the same background as the wild-type strain. Growth was conducted in rich media [yeast extract-peptone-dextrose (YPD)] at 30°C with shaking at 200 rpm. UROtsa cells (generously provided by P. Simeonova, National Institute for Occupational Safety and Health, Morgantown, WV) were cultured at a starting cell density of 4–5 × 104 cells/mL in RPMI 1640 (Mediatech, Inc., Manassas, VA) with l-glutamine, 10% fetal bovine serum, 100 IU/mL penicillin, and 100 μg/mL streptomycin (Omega Scientific, San Diego, CA), under standard culturing conditions.

Arsenical exposures

We purchased sodium arsenite [NaAsO2 (iAsIII); purity > 99%] from Sigma-Aldrich (St. Louis, MO). Diiodomethylarsine [MMAIII iodide (MMAIII)] was a generous gift from J. Gandolfi (University of Arizona, Tucson, AZ). iAsIII and MMAIII solutions were freshly prepared using sterile water (Milli-Q; Millipore, Billerica, MA) and protected from light before use. Yeast cells were treated with either iAsIII or MMAIII at concentrations ranging from 0 to 300 μM.Once UROtsa cells reached 70–80% confluence in culture, they were treated with iAsIII at concentrations from 0 to 100 μM or MMAIII at 0–5 μM.

Yeast growth assay

Yeast strains were pregrown in YPD media to mid-log phase, diluted in fresh media to an optical density at 595 nm (OD595) of 0.0165, and inoculated into a 48-well microplate. Stock solutions of arsenicals were added to each culture with at least three replicate wells per dose.

Plates were incubated in a Tecan GENios spectrophotometer (Tecan Systems Inc., San Jose, CA) set to 30°C with intermittent shaking, and OD595 measurements were taken at 15-min intervals for 24 hr. Raw absorbance data were averaged for all replicates, corrected for background, and plotted as a function of time. The area under the curve (AUC) was calculated for the cultures in each well using Prism software (version 5.01; GraphPad Software, Inc., La Jolla, CA), and the treatments were averaged and expressed as a percentage of the control.

Human tissue array and real-time quantitative polymerase chain reaction (PCR) assay

We used TaqMan-based real-time quantitative polymerase chain reaction (rt-qPCR) to quantify N6AMT1 and AS3MT expression on a panel of 48 normal human tissues using the Human Rapid-Scan Plate (OriGene Technologies, Inc., Rockville, MD). The human tissues were selected from multiple individuals of different ethnicity and pooled together. We obtained the primers and probes used for amplification of N6AMT1, AS3MT, and ACTB (β-actin; control) from Applied Biosystems (Foster City, CA). Gene expression of N6AMT1 and AS3MT was calculated relative to ACTB using the ΔΔCT method.

N6AMT1 gene expression vector constructs and stable cell lines

Human N6AMT1 cDNA (GenBank accession no. NM_013240; National Center for Biotechnology Information 2011) was PCR amplified with primers 5′-AACGCAGCGAAGGACTAT-3′ and 5′-CAGTAGTTCTGGGCACAC-3′. The PCR product was gel purified (Qiagen, Valencia, CA) and cloned into pcDNA 2.1 vector (Invitrogen) according to the manufacturer’s instructions, and the sequence was confirmed. The pcDNA 2.1 vector containing the N6AMT1 gene was excised using NotI/BamHI restriction enzymes (New England Biolabs, Ipswich, MA) and subjected to gel purification. The nucleotides of the N6AMT1 gene containing BamHI and NotI overhangs were annealed and ligated to a linearized pRetro X-IRES-ZsGreen vector (Clonetech, Mountain View, CA) digested with BamHI and NotI (New England Biolabs). The pRetro X-IRES-ZsGreen vector is a fluorescent retroviral expression vector that allows both a gene of interest and the ZsGreen gene to be expressed. The resultant constructs were amplified, purified, and sequenced. UROtsa cells were transfected with this constructed vector or a control vector using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s instructions. After incubation at 37°C for 8 hr, the supernatant fraction containing the retroviral vector was removed and replaced with normal growth medium. Cells grown for 48–72 hr were assessed by fluorescence microscopy. The ZsGreen fluorescent marker yields a bright green fluorescence, permitting direct monitoring of the delivery efficiency. Finally, the cell populations were sorted by the DAKO-Cytomation MoFlo High Speed Sorter (Dako North America, Carpinteria, CA), and the green fluorescent cells were purified and collected for continuing culture. The green fluorescent cells were used for additional experimentation.

Semiquantitative reverse-transcription (RT)-PCR

UROtsa cells with either N6AMT1 or plasmid vectors were collected, and total RNA was isolated from these cells using the Qiagen RNAEasy Mini kit. We performed a reverse transcription reaction using SuperScript II Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. The PCR conditions for DNA amplification in the linear range were established on the GeneAmp PCR System 7600 (PerkinElmer, Inc., Wellesley, MA). The primers for DNA amplification were as follows: N6AMT1, 5′-AACGCAGCGAAGGACTAT-3′ and 5′-CAGTAGTTCTGGGCACAC-3′; AS3MT, 5′-GTGTCTGGGTGGTGCTTTATACTG-3′ and 5′-TGGAGGGCAGAACCCAATT-3′; and the housekeeping gene ACTB, 5-TCACCCACACTGTGCCCATCTACGA-3 and 5-CAGCGGAACCGCTCATTGCCAATGG-3. RT-PCR products were analyzed on 1% agarose gels.

Cytotoxicity assay

We performed the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay to assess cell viability after arsenic treatment. Cells were cultured in 96-well plates in a volume of 100 μL medium/well at a density of 5 × 104 cells/mL. Twenty-four hours after incubation with iAsIII or MMAIII (six replicates/arsenical concentration), 10 μL sterile MTT dye (Sigma-Aldrich; 5 mg/mL) was added to each well and plates were incubated at 37°C for 4 hr. The culture medium was then removed, and 200 μL dimethyl sulfoxide was added and thoroughly mixed for 10 min. Spectrophotometric absorbance at 570 nm was measured in a microplate reader.

Arsenic species profile analysis by high-performance liquid chromatography/inductively coupled plasma–mass spectrometry (HPLC-ICP-MS) methods

UROtsa cells with N6AMT1 and UROtsa cells with vector were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum and antibiotics. Culture medium was collected after exposure to iAsIII or MMAIII for 24 hr or 3 days and stored at −80°C until analysis. Cells treated with iAsIII or MMAIII for 3 days were collected, lysed in RIPA buffer, and extracted in methanol by incubating overnight at 4°C in a rotational shaker. After centrifugation at maximum speed for 5 min at 4°C, supernatants were transferred to microcentrifuge tubes and stored at −80°C until analysis. Before analysis, samples were diluted 1:5 with water–methanol to bring the methanol concentration to 2.5%, incubated at 5°C to precipitate poorly soluble material, and filtered (0.45 μm). Analysis was performed by HPLC-ICP-MS (Agilent 1090 HPLC and Agilent 7500CE ICP-MS run in normal mode; both from Agilent Technologies, Santa Clara, CA) under conditions that resolved neutral, trivalent, and pentavalent iAs species. The ion-pairing method (Le et al. 2000) was used with major modifications to improve the resolution of the species. Briefly, calibrants were prepared from neat materials [As2O3 (Aldrich) and As2O5 (Acros), Sigma-Aldrich; DMAV and MMAV · 6H2O, Chem Service, West Chester, PA] in deionized water (≥ 18 MΩ). For speciation, we used a Phenomenex Gemini-NX column (3 μm, C18, 110Å, 150 × 4.6 mm; Torrance, CA) with a corresponding guard column at 40°C. Concentrations of arsenic species in stock solutions were standardized against NIST traceable commercial ICP-MS standards (VWR BDH Aristar Plus; Ultra Scientific, Kingstown, RI). Serial dilutions were made into deionized water. iAsIII and iAsV species were quantified by separate calibrant series, and iAsIII concentration in the calibrants was corrected for any conversion to iAsV. The HPLC conditions were isocratic (5 mM tetrabutylammonium hydroxide, 10 mM ammonium carbonate, 2.5% methanol, pH 9.2, 1 mL/min) for 5 min; then a step gradient (5 mM tetrabutylammonium hydroxide, 30 mM ammonium carbonate, 2.5% methanol, pH 8.75, 1.2 mL/min) for 5 min to elute iAsV was followed by step gradients (5 mM tetrabutylammonium hydroxide, 30 mM ammonium carbonate, 2.5% methanol, pH 9.2, 1.2 mL/min) for a 5-min equilibration to the initial pH and finally to the initial mobile phase for 5 min (1.2 mL/min). The data were analyzed by LC ChemStation A.09.03 and ICP-MS ChemStation B.03.03 software (Agilent Technologies).

Data analysis

Statistical analyses were performed using one-way analysis of variance. Data represent mean ± SE of at least three independent experiments.

Results

Deletion of yeast MTQ2 leads to increased resistance to arsenic treatment

We evaluated the growth phenotype of the MTQ2-deletion mutants in the presence of either iAsIII or MMAIII (Figure 1). Deletion strains and their isogenic wild-type counterpart, BY4743, were treated with equioxic doses equivalent to the concentrations that resulted in 20% growth inhibition (IC20) and 2 × IC20, which were 300 and 600 μM for iAsIII and 150 and 300 μM for MMAIII, respectively. iAsIII treatments had no effect on the growth of MTQ2-deletion mutants but significantly decreased growth of the wild-type strain. In comparison, the growth of both MTQ2-deletion mutants and wild-type yeast treated with MMAIII decreased to the same degree despite the higher toxicity of MMAIII.

Figure 1

Deletion of the MTQ2 gene in yeast results in increased resistance to arsenite (iAsIII), shown as the growth phenotype of MTQ2 mutant yeast cells (B) and the wild-type BY4743 cells (A) treated with 300 or 600 μM iAsIII or 150 or 300 μM MMAIII. Growth curves show the OD595 for each treatment as a function of time for 24 hr. Bars represent the mean ± SE AUC for three technical replicates. At the doses tested, iAsIII treatment did not alter the growth pattern of MTQ2 mutants but led to a dose-dependent reduction in growth of the wild-type strain; the growth patterns of both yeast strains were similar after MMAIII exposure.

p< 0.001, compared with control.

Differential level of N6AMT1 mRNA expression in human tissues

A direct search for sequence homology and conserved functional domains revealed that the human N6AMT1 gene is orthologous to the yeast MTQ2 gene. We used the web-based online tool Protein Function Prediction (PFP; Kihara Bioinformatics Laboratory 2010), to analyze and predict its potential functions (Hawkins et al. 2009). The suggested molecular functions of N6AMT1 include protein heterodimerization and methionine S-methyltransferase activity, with an almost certain (100%) predicted probability of being a methyltransferase. Given N6AMT1’s suggested function as a methyltransferase, we were interested in exploring its potential involvement in arsenic biomethylation. Considering that the primary methyltransferase responsible for arsenic metabolism in human cells is AS3MT, we did pairwise alignment analysis of N6AMT1 and AS3MT using EMBOSS Pairwise Alignment Algorithms, an online tool (European Bioinformatics Institute 2010). The two proteins shared about 25% similarity. Of the three sequence motifs found in most AS3MT homologs, only motif ILDLGSGSG is highly conserved in N6AMT1 [LEVGSGSG; see Supplemental Material, Figure 1(doi:10.1289/ehp.1002733)], whereas (D/N)PPY is present in N6AMT1 but not in AS3MT. These differences suggest that the mechanism by which N6AMT1 methylates arsenic may differ from that of AS3MT, if N6AMT1 is, in fact, involved in the methylation of arsenicals.

A search of the Expressed Sequence Tags (EST) Database (National Center for Biotechnology Information 2010) revealed sequences matching the cDNA of N6AMT1 in many human tissues with varied expression levels. To experimentally measure the mRNA expression of N6AMT1 across tissues, we performed rt-qPCR analysis using cDNA from a panel of 48 human tissues contained in a tissue array (Figure 2). We confirmed the amplification products to be N6AMT1 by DNA sequencing. Using the liver as a reference, the expression of N6AMT1 was normalized to the expression level of ACTB [with cycle threshold (Ct) values ranging from 18 to 20] and found to be relatively highly expressed in tissues such as the parathyroid, pituitary, adrenal gland, and kidney, and weakly expressed in tissues such as the skin, lung, and mammary gland. We also measured AS3MT mRNA levels using the same tissue panel in order to compare N6AMT1 expression in each tissue and AS3MT expression [see Supplemental Material, Figure 2 (doi:10.1289/ehp.1002733)]. The data showed that the level of N6AMT1 mRNA was relatively low in most measured tissues compared with AS3MT mRNA. The presence of detectable levels of N6AMT1 expression in all tissues analyzed suggests that this enzyme could be involved in methylating arsenicals. N6AMT1 may act in parallel to AS3MT and may be functional only under certain conditions or in certain tissues.

Figure 2

N6AMT1 mRNA expression in human tissues, quantified by rt-qPCR analysis a panel of 48 normal human tissues; transcript levels of N6AMT1 were normalized to ACTB expression and are shown as the fold change relative to liver.

Overexpression of N6AMT1 in UROtsa cells increases resistance to arsenic treatment

We also measured and compared the level of N6AMT1 mRNA in several cell lines, including 293 (human embryonic kidney cells), HeLa, UROtsa, and HL60 (human promyelocytic leukemia cells). N6AMT1 expression in UROtsa cells is relatively low, with a Ct value of about 33 (data not shown). This cell line also has almost no detectable level of AS3MT, making it an excellent model to study the role of N6AMT1 in arsenic toxicity and metabolism in mammals. We enhanced N6AMT1 expression in UROtsa cells using a retrovirus-based vector (Figure 3A) and found the level of N6AMT1 mRNA in UROtsa cells to be significantly increased by approximately 5-fold in clone 2, as measured by semiquantitative RT-PCR (Figure 3B). We also measured AS3MT gene expression in these two cell lines and found no detectable mRNA level in either cell line (Figure 3B). We further confirmed these PCR results by real-time PCR analysis (data not shown). Unfortunately, we could not detect N6AMT1 protein levels in these cells using two commercially available antibodies. Transfected UROtsa cells did not have an altered doubling time or morphology in culture. The UROtsa cells with N6AMT1 (N6AMT1-enhanced cells) and the UROtsa cells with vector cells (vector control) were treated with either iAsIII or MMAIII at a series of concentrations for 24 hr. Arsenical treatments induced a dose-dependent decrease in viability of both cell lines. However, increased expression of N6AMT1 in UROtsa cells resulted in higher viability after iAsIII and MMAIII treatment at almost all concentrations tested, compared with the UROtsa vector control cells (Figure 3C,D). This increased arsenic resistance was more apparent in cultures treated with MMAIII than with iAsIII, approximately 2- and 1.3-fold, respectively.

Figure 3

Enhancing expression of human N6AMT1 in UROtsa cells increases resistance to both iAsIII and MMAIII treatment. (A) Representative images of UROtsa cells with either plasmid alone or plasmid with N6AMT1; bars = 10 μm. (B) Semiquantitative RT-PCR analysis shows dramatically increased N6AMT1 expression in UROtsa cells with N6AMT1 relative to control cells or cells with plasmid. AS3MT mRNA was not detected in either cell line; ACTB was used as the loading control. UROtsa cells with vector and with N6AMT1 were treated with increasing concentrations of iAsIII (C) or MMAIII (D) for 24 hr, and cell viability was evaluated with MTT. Bars represent the mean ± SD of three independent experiments. Treatment with either iAsIII or MMAIII resulted in a dose-dependent decrease in viability. However, increased expression of N6AMT1 in UROtsa cells led to resistance to both iAsIII (C) and MMAIII (D); this effect was more significant when cells were treated with MMAIII than with iAsIII.

**p < 0.01, compared with control and vector control.

Enhanced N6AMT1 in UROtsa cells methylates MMAIII to DMA

We collected medium and cell extracts from cultures of UROtsa cells with N6AMT1 and UROtsa cells with vector treated with different concentrations of either iAsIII or MMAIII for up to 3 days and then measured and analyzed arsenic metabolic profiles using ICP-MS [for representative chromatograms, see Supplemental Material, Figure 3 (doi:10.1289/ehp.1002733)]. Methylated metabolites were undetectable either in the media (Table 1) or in the cell extract (Table 2) from cultures of UROtsa cells with N6AMT1 or of UROtsa cells with vector, after treatment of with iAsIII, suggesting that N6AMT1 does not methylate iAsIII. When UROtsa cells with vector control were treated with MMAIII, levels of MMAV but not dimethylarsinic acid (DMAV) increased in the media (Table 3) and cells (Table 2). Treatment with MMAIII of UROtsa cells with N6AMT1 resulted in similarly increased levels of MMAV andled to increased levels of DMAV in the media after 24 and 72 hr (Table 3) and, to a lesser degree, in the cell extracts after 72 hr (Table 2). At 24 hr, the level of DMAV was similar at each dose level, but after 3 days treatment levels in both media and cell extract increased in relation to the initial concentration of MMAIII, in a dose-dependent manner. Moreover, the amount of DMAV in the culture medium increased 5-fold after 3 days of treatment (1 μM MMAIII) compared with 1 day of treatment.

Table 1

Arsenic metabolic profile in cell culture medium after iAsIII treatment of UROtsa cells with vector and UROtsa cells with N6AMT1 (ng/mL; mean ± SE).

	UROtsa cells with vector				UROtsa cells with N6AMT1
iAsIII (μM)	iAs(III +V)	MMAIII	MMAV	DMAV	iAs(III +V)	MMAIII	MMAV	DMAV
1 day
0	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ
1	261.50 ± 126.22	BLOQ	BLOQ	BLOQ	417.00 ± 1.41	BLOQ	BLOQ	BLOQ
3	1109.00 ± 11.31	BLOQ	BLOQ	BLOQ	1159.00 ± 12.02	BLOQ	BLOQ	BLOQ
10	3695.50 ± 10.96	BLOQ	BLOQ	BLOQ	3751.00 ± 72.12	BLOQ	BLOQ	BLOQ
3 days
0	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ
1	425.50 ± 8.13	BLOQ	BLOQ	BLOQ	414.50 ± 5.30	BLOQ	BLOQ	BLOQ
3	1034.00 ± 2.12	BLOQ	BLOQ	BLOQ	1017.00 ± 7.07	BLOQ	BLOQ	BLOQ

BLOQ, below the limit of quantitation (the m/z 75 signal at the retention time of the species is not distinguishable from the baseline signal).

Table 2

Arsenic metabolic profile in cell extract after 3-day iAsIII and MMAIII treatment of UROtsa cells with vector and UROtsa cells with N6AMT1 (ng/mL; mean ± SE).

	UROtsa cells with vector				UROtsa cells with N6AMT1
Treatment (μM)	iAs(III+V)	MMAIII	MMAV	DMAV	iAs(III+V)	MMAIII	MMAV	DMAV
iAsIII
0	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ
1	2.7 ± 0.3	BLOQ	BLOQ	BLOQ	5.7 ± 0.6	BLOQ	BLOQ	BLOQ
3	12.0 ± 2.0	BLOQ	BLOQ	BLOQ	14.5 ± 0.5	BLOQ	BLOQ	BLOQ
MMAIII
0	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ
0.5	BLOQ	BLOQ	3.5 ± 0.7	BLOQ	BLOQ	BLOQ	3.0 ± 0.1	2.8 ± 0.2
1	BLOQ	BLOQ	2.9 ± 0.8	BLOQ	BLOQ	BLOQ	4.7 ± 0.0	4.8 ± 0.6

BLOQ, below the limit of quantitation (the m/z 75 signal at the retention time of the species is not distinguishable from the baseline signal).

Table 3

Arsenic metabolic profile in cell culture medium after MMAIII treatment of UROtsa cells with vector and UROtsa cells with N6AMT1 (ng/mL; mean ± SE).

	UROtsa cells with vector				UROtsa cells with N6AMT1
MMAIII (μM)	iAs(III+V)	MMAIII	MMAV	DMAV	iAs(III+V)	MMAIII	MMAV	DMAV
Treated 1 day
0	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ
0.5	BLOQ	Present	112.00 ± 0.71	BLOQ	BLOQ	Present	104.00 ± 0.71	21.50 ± 0.35
1	BLOQ	Present	236.50 ± 15.20	BLOQ	BLOQ	Present	210.00 ± 2.12	22.50 ± 1.06
2	BLOQ	Present	468.00 ± 17.68	BLOQ	BLOQ	Present	424.50 ± 6.72	20.00 ± 0.71
Treated 3 days
0	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ	BLOQ
0.5	BLOQ	Present	115.00 ± 2.12	BLOQ	BLOQ	Present	106.00 ± 2.83	72.00 ± 4.24
1	BLOQ	Present	277.00 ± 5.66	BLOQ	BLOQ	Present	236.50 ± 9.55	105.50 ± 4.60

BLOQ, below the limit of quantitation (the m/z 75 signal at the retention time of the species is not distinguishable from the baseline signal). “Present” indicates that the peak was seen at MMAIII retention time; no concentration is given because the samples were not analyzed by the MMAIII assay.

Discussion

MMAIII is the most toxic arsenic metabolite in vivo and in vitro

In humans, iAs is metabolized to methylated arsenical species in a multistep process. Methylated arsenicals, especially MMAIII, may be more toxic than iAs both in vivo and in vitro (Drobná et al. 2005; Ferrario et al. 2008; Kligerman et al. 2003; Petrick et al. 2001). In cultured human cells, MMAIII is the most toxic arsenical (Drobná et al. 2005; Ferrario et al. 2008; Petrick et al. 2001) and inhibits several key cellular proteins, such as glutathione reductase (Styblo et al. 1997) and thioredoxin reductase (Lin et al. 1999, 2001). Several studies have shown that MMAIII is capable of inducing genetic damage and changes in signal transduction by either direct or indirect mechanisms (Ahmad et al. 2002; Kligerman et al. 2003; Nesnow et al. 2002). In addition, exposure to MMAIII for 52 weeks induced malignant transformation of UROtsa cells (Bredfeldt et al. 2006). Epidemiological studies have suggested that individuals who excrete a higher proportion of ingested arsenic as MMAIII are more susceptible to arsenic-related cancer (Steinmaus et al. 2006). MMAIII has been proposed as the ultimate genotoxic form of arsenic (Kligerman et al. 2003), and the existing evidence indicates that biomethylation of iAs to MMAIII is likely to alter the adverse effects of environmental arsenic exposure on human health.

AS3MT is primarily responsible for methylating iAs to MMAIII and DMA in humans

The arsenic methyltransferase AS3MT is recognized as the primary enzyme responsible for conversion of iAs to its methylated metabolites MMAIII and DMA (Lin et al. 2002; Wood et al. 2006). Studies have shown that single-nucleotide polymorphisms in AS3MT lead to different urinary arsenical profiles (Agusa et al. 2009; Schläwicke Engström et al. 2009; Wood et al. 2006), some of which are associated with increased risk of premalignant skin lesions (Valenzuela et al. 2009). Although these data suggest that AS3MT plays a critical role in arsenic methylation and toxicity, a recent study showed that As3mt-knockout mice retain some ability to methylate arsenicals, suggesting the existence of other methyltransferases that could be involved in alternative arsenic metabolism pathways (Drobna et al. 2009).

N6AMT1 is capable of methylating MMAIII to DMA

In this study, we found that N6AMT1 has the capacity to methylate MMAIII to DMA. The expression of N6AMT1 is generally low compared with the expression level of AS3MT in most human tissues. In addition, the low sequence homology shared between these two proteins, about 25%, supports differences in substrate specificity and, possibly, in mechanisms of arsenic methylation. In contrast to AS3MT, which methylates iAs to the more toxic MMAIII, N6AMT1 methylates MMAIII to the less toxic DMA. This is consistent with the increased resistance to MMAIII of UROtsa cells overexpressing N6AMT1 compared with vector control cells, but other mechanisms likely also contribute to this increased resistance. Thus, our results suggest that N6AMT1 may play a role in modulating arsenical-induced toxicity and that decreased N6AMT1 expression or activity could have a significant impact on arsenic-induced toxicity and perhaps carcinogenicity under certain conditions or in certain tissues.

N6AMT1 in human cells responded to arsenicals differently from MTQ2 in yeast

We noted differences in response to arsenical treatments between yeast and human cells. Specifically, the MTQ2-deletion yeast strain is resistant only to iAsIII. We did not find evidence of iAsIII methylation in yeast wild-type; that is, DMA and MMA levels in cells and culture media were below the limit of quantitation (data not shown). Therefore, the function of MTQ2 in arsenic toxicity may not be related to iAsIII methylation. In contrast, overexpression of N6AMT1 in UROtsa cells leads to resistance to both iAsIII and MMAIII, which provides evidence of its involvement in protection from these arsenicals. These results indicate that the orthologous genes—MTQ2 in yeast and N6AMT1 in humans—have different roles in the cellular response to arsenic toxicity.

Conversion of MMAIII to DMA by N6AMT1 needs to be further confirmed biochemically

Our analyses showed that enhanced expression of N6AMT1 in UROtsa cells converts MMA to DMA. However, purified recombinant N6AMT1, in the presence of SAM and other cofactors, was unable to methylate iAsIII or MMAIII (data not shown). N6AMT1 dimerizes with tRNA methyltransferase 11-2 homolog (TRMT112), which appears to be necessary for proper N6AMT1 function (Figaro et al. 2008) and is consistent with heterodimerization activity as predicted with PFP. Thus, the lack of N6AMT1-dependent MMAIII methylation in this test tube experiment may be due to the absence of TRMT112 or another unknown protein. In addition, N6AMT1 overexpression increased resistance to iAsIII, although UROtsa cells were not able to methylate iAsIII to MMAIII or to other methylated species, suggesting that mechanisms other than participation in arsenic methylation might be involved. It is not clear whether interaction between N6AMT1 and TRMT112 has a role in arsenic toxicity, but it is certainly worthy of further investigation.

Conclusions

Our data suggest an important potential role of N6AMT1 in modulating arsenical-induced toxicity by methylating MMAIII to the less toxic DMA. However, further investigation is warranted to determine whether N6AMT1 can methylate MMAIII in vivo and also to identify the genetic and environmental factors that can alter N6AMT1 expression and/or activity. Our ongoing experiments are focused on the biochemical characterization of N6AMT1, specifically its capacity to methylate MMAIII, as well as its ability to modulate arsenic toxicity and carcinogenicity.