Given the structural similarity of BMAA to other amino acids

Given the structural similarity of BMAA to other amino acids, our hypothesis was that N-nitrosation of BMAA might result in a toxic alkylating agent, providing a previously unrecognised mechanism for BMAA action.
Materials and methods
Results and discussion We first examined whether chemical nitrosation of BMAA with sodium nitrite resulted in the formation of an alkylating agent by using the NBP trapping assay (Shephard et al., 1987, Thomas et al., 1992, Gomez-Bomberelli et al., 2012). We used conditions where the amino Suramin hexasodium salt nitrosation and NBP alkylation were present simultaneously, in order to ensure the detection of labile intermediates (Shephard et al., 1987, Thomas et al., 1992). NBP-adduct formation was dose-dependent and N-BMAA proved to be a more potent alkylating agent than nitrosated alanine, based upon the response obtained with similar amounts of amino acid present in the nitrosation mix (Fig. 1). BMAA or alanine alone did not react with NBP (data not shown). The NBP assay has been widely used as a screening assay to detect the alkylating potency of many compounds (Garcia-Santos et al., 2001, Shephard et al., 1991, Shephard et al., 1987, Thomas et al., 1992, Gomez-Bomberelli et al., 2012), by the formation of adducts at the nucleophilic N4 position on NBP (Gomez-Bomberelli et al., 2012). Since the N4 has similar physicochemical properties to the N7 of guanine, a favourable site of alkylation (Boysen et al., 2009) this provided the initial indication that N-nitrosated BMAA was likely to be a DNA damaging agent. To further examine this, we used the plasmid nicking assay which is an extremely sensitive method for detecting DNA strand breakage (Dumax-Vorzet et al., 2015). Incubation of increasing concentrations of BMAA following nitrosation with supercoiled plasmid DNA induced a linear increase in the amount of relaxed plasmid, demonstrating the formation of DNA single strand breaks (Fig. 2). BMAA at the same concentrations also induced plasmid nicking but at lower levels (Fig. 2) a result consistent with the observation that BMAA induced cellular DNA damage detectable by the alkaline comet assay (Chiu et al., 2012). Incubation of N-alanine and alanine with plasmid DNA did not produce DNA nicks above those observed in controls under the conditions used. There was thus a clear difference between N-BMAA and N-alanine induced plasmid DNA strand break formation, in contrast to previous reports indicating that BMAA and alanine themselves were of similar toxicity in a number of cell lines: a result that was taken to indicate that BMAA did not contribute significantly to human neurodegenerative disease (Lee and McGeer, 2012). We finally examined whether N-BMAA is cytotoxic to SH-SY5Y cells. SH-SY5Y is an immortalised neuroblastoma cell line with many of the characteristics of neuronal cells. It is used commonly for investigations of BMAA induced neuronal toxicity and in other studies relating to AD, ALS, and ALS/PDC (Lopes et al., 2010, Kovalevich and Langford, 2013, Okle et al., 2013). Under conditions in which BMAA showed minimal toxicity, N-BMAA was toxic (Fig. 3) and the differences were statistically significant at all time points (0.05). N-BMAA toxicity was comparable to that of the positive control (5M NaCl) and was maintained over the 96hr of the experiment except for the lowest doses for which toxicity appeared to accumulate over this time period. We have not directly quantified the reactive alkylating species in this study and so do not know its level of formation from BMAA nitrosation. BMAA contains both a primary amine and a secondary amine that are potentially nitrosatable (Rostkowska et al., 1998). Nitrosation of the secondary amine will lead to a nitrosamine that typically requires metabolic activation to be reactive (Hecht, 1994) whereas nitrosation of the primary amine results in a direct acting alkylating species that is the corresponding lactone (Garcia-Santos et al., 2001). It is thus possible that the observed activity in vitro results from the nitrosation of the primary amine but we cannot rule out the involvement of a nitrosamine in the cellular toxicity. The biological potency of alkylating agents results in a large part from the formation of DNA adducts (Kaina et al., 2007, Drabløs et al., 2004) and hence it is likely that the cytotoxicity induced by nitrosated BMAA results from this mechanism although we cannot as yet rule out other mechanisms.