Ochratoxin A (OTA, Figure 1) is a human toxin produced by Penicillium verrucosum and several Aspergillus species [1,2]. It is found in a variety of human foods, including cereal grains, e.g., wheat and rye, as well as coffee, beer, and wine [3,4]. OTA has been implicated as a cause of nephropathies and urothelial tract tumors in the Balkans [5,6,7] and in North African countries [8]. The International Agency for Research on Cancer (IARC) has classified OTA as a possible human carcinogen (group 2B) [9], and OTA exposure has been associated with cancer of the urothelial tract in rats [10, and reviewed in 5] and chicks [11]. OTA has also been proposed as a cause of testicular cancer in young men [12,13].

OTA is a substrate for photochemical [14,15,16], electrochemical [17], and transition metal ion-mediated [18,19] oxidation. OTA is also a substrate for enzymatic oxidation by microsomal mixed function oxidases [20,21,22] and enzymes with peroxidase activity [23,24]. Exposure to OTA is cytotoxic to cultured cells [25,26,27,28] as well as in vivo in rodent models [29,30,31]. The cytotoxicity of OTA shows a close correlation with the onset of oxidative DNA damage mediated by the toxin through production of reactive oxygen species (ROS) [25,27,28].

Studies of OTA-mediated mutagenicity have produced inconsistent results. The original assessment in S. typhimurium in the presence of rat liver post-mitochondrial supernatants was negative [32]. However, De Groene et al. observed OTA-mediated mutagenesis in NIH 3T3 cells stably expressing certain P450s [33,34]. Additional studies by the Obrecht-Plumio et al. in the presence of high dose OTA (403–1210 µg/plate), kidney microsomes, NADPH, and arachidonate caused the formation of mutagens detectable in S. typhimurium [35]. However, Zepnick and coworkers found negative evidence for OTA-mediated mutagenicity [36]. Palma et al. found that OTA-mediated mutagenicity is consistent with oxidative DNA damage and that bioactivation of OTA is not a requirement for the observed mutagenicity [37]. Several other studies have concluded that enzymatic oxidation of OTA does not cause covalent adduction of OTA to DNA [38,39,40] and that mutagenicity stems from by-products of cellular cytotoxicity [41]. However, other recent studies have produced conflicting results that indicate that oxidation of OTA in the presence of nucleosides or DNA yields covalent OTA-DNA nucleoside adducts [42,43,44]. Included in these studies is liquid chromatography-mass spectrometry (LC-MS) results that support the presence of the nonchlorinated OTB-2′-deoxyguanosine (dG) adduct (Figure 1) in male rat kidney [44]. Hibi and coworkers have also demonstrated the mutagenicity of OTA in the target tissue (outer medulla) of male rats [45]. In these studies the gpt delta transgenic rat model was employed and the reporter gene mutation assay showed significantly higher levels of deletion mutations compared to controls using DNA extracted from the outer medulla of male rat. The extracted DNA was also examined for 8-oxo-dG formation derived from ROS generation and no evidence for 8-oxo-dG formation was found and OTA exposure failed to increase the frequencies of GC: TA transversion mutations, which are characteristic of 8-oxo-dG-mediated mutagenicity. The in vivo mutagenicity assays reported by Hibi et al. suggest that oxidative DNA damage does not contribute to OTA-mediated mutagenicity and favor a direct genotoxic mechanism [45].

The positive results on OTA-mediated in vivo mutagenicity [45] combined with the finding that OTA generates the OTB-dG adduct in male rat kidney [44] demonstrates that DNA adduction and mutagenicity remains a viable mechanism of action for OTA-mediated renal carcinogenesis [46]. These results prompted us to report the current study, in which we address the mutagenicity of OTA in cell culture, using the human mutation reporter plasmid pSP189 developed by Seidman [47]. The data presented herein indicate that oxidation of OTA, either by microsomal enzymes or by transition metal ions, activates OTA to a directly genotoxic mutagen(s). Synthetic ochratoxin hydroquinone (OTHQ, Figure 1), an OTA metabolite that forms covalent DNA adducts [43], is also mutagenic.

There has been considerable debate whether OTA is directly genotoxic [38,39,44,46,57,58]. OTA can promote oxidative DNA damage through ROS generation that causes cytotoxicity [25,27,28] and oxidative DNA damage has been proposed to play an important role in carcinogenicity [41]. However, other lines of evidence favour DNA adduction by OTA [23,42,43,44,46,57]. The in vivo mutagenicity findings for OTA do not support oxidative base damage as the initiation event and favour direct genotoxicity in OTA-induced renal carcinogenicity [44]. However, the link between DNA adduction and OTA-mediated mutagenicity has not been firmly established [46].

The data presented here demonstrate that OTA is activated to a species that is a directly genotoxic mutagen. The free toxin lacked mutagenicity in the absence of external cofactors, such as RLM (Table 1) or redox-active transition metal ions (Table 3). These findings differ from the mutagenicity data presented by Palma et al., which suggested that bioactivation in not required for mutagenicity [37], and are more in line with the data from Tozlovanu et al., 2006 showing that DNA adduction by OTA is not observed directly with OTA in the absence of oxidative metabolism [43].

That RLM could be used for the conversion of OTA to a genotoxic mutagen in our experimental system (Table 1), suggested that the genotoxic metabolite of OTA is an oxidation product. In this regard, we previously demonstrated that the oxidation of OTA (100 µM) by RLM (1 mg/mL) generated a GSH-conjugate that suggested the intermediacy of the quinone electrophile (OTQ) in the oxidation of OTA [59]. Pfohl-Leszkowicz and coworkers have also outlined a role for GSH in conjugation of OTA-derived electrophiles following bioactivation of the parent toxin [60]. The oxidation of OTA by activated RLM in the presence of a reducing agent (ascorbate) also generates the hydroquinone metabolite OTHQ [59]. OTHQ autoxidizes to OTQ that reacts covalently with DNA to generate DNA adducts [43]; although the structure(s) of such adducts are unknown. This process occurs spontaneously in aqueous media in the presence of molecular O₂, and no other cofactors are required for conversion of OTHQ into OTQ [22]. While this process has the potential to generate genotoxic ROS, we have recently demonstrated that OTHQ lacks cytotoxicity in mammalian kidney cells, suggesting that the metabolite is ineffective at stimulating ROS production [61]. Thus, DNA damage mediated by OTHQ is mainly due to covalent DNA adduction stimulated by OTQ formation [43]. This background information implies that OTQ formation is likely responsible for mutagenicity stimulated by the RLM/OTA system (Table 1).

Authentic OTHQ was also mutagen in our experimental system, albeit a weak one (Table 2). The weakness of OTHQ mediated mutagenicity may be a reflection of its slow rate of autoxidation [22]. Prolonged incubation of pSP189 with OTHQ may have enhanced its mutagenicity. However, interestingly, addition of cysteine enhanced the mutagenicity of OTHQ. A possible mechanism for enhancement of OTHQ-mediated mutagenicity by cysteine is suggested by our previous work [16] and outlined by Monks and Lau for the nephrotoxicity of polyphenolic-gluathione conjugates [62]. Thus, as outlined in Figure 2, autoxidation of OTHQ generates OTQ that reacts with cysteine to afford the conjugate OTHQ-Cys [16]. This conjugate is not stable [16] and can also undergo an autoxidative process to reform the quinone structure that reacts covalently with the amine group of the attached cysteine to afford a quinonimine that upon tautomerization will afford the imine electrophile in the new six-membered ring system. The imine functional group forms the basis for DNA adduction by the pyrrolobenzodiazepines, such as anthramycin, that have potential use in anticancer therapies [63]. This proposed pathway may play a role for the heightened mutagenicity of OTHQ in the presence of cysteine (Table 2).

In our experimental system, Fe(III)/OTA was the most potent mutagen showing a 32-fold MF above background levels (Table 3). This observation may correlate with earlier studies on the reactivity of OTA in the presence of dG to yield OTB-dG [53]. These studies were prompted by reports that OTA forms guanine-specific DNA adducts [64,65] and demonstrated that OTB-dG formation resulted from reaction of OTA/dG in the presence of Fe, Cu, or peroxidase enzymes [53]. Interestingly, Fe generated OTB-dG in the highest yield, ~5 orders of magnitude greater than Cu and peroxidase. The greater reactivity of OTA toward guanine in the presence of Fe, as opposed to Cu, may provide a rationale for the enhanced MF of the Fe/OTA system (Table 3). However, Fe(III) also oxidizes OTA to the quinone electrophile OTQ that reacts with DNA [43] and provides a rationale for the mutagenicity of OTHQ (Table 2). Thus, there are at least two possible mechanisms for Fe(III)/OTA mediated mutagenesis: (1) OTB-dG formation, and (2) OTQ formation with subsequent adduct formation. At present, the DNA adduct(s) formed by OTQ have not been structurally characterized and it is unknown whether the OTB-dG adduct is mutagenic. Our goal is to first characterize the DNA adducts formed by OTQ, and incorporate site-specifically the OTQ adducts and OTB-dG into oligonucleotide substrates so that their biological consequences can be assessed.

The observation of OTA-mediated mutations in the supF gene of plasmid DNA replicating in human Ad293 cells seemingly is not concordant with the work of Hibi et al. [45], who did not observe enhanced mutagenesis in the gpt gene in kidneys of OTA-exposed mice. The explanation for the disparate results is not known; however, methodologic differences may be contributory gpt mutations are detected phenotypically by resistance to 6-thioguanine due to loss of hypoxanthine-guanine phosphoribosyl transferase (HGPRT) catalytic activity. Phenotypically silent gpt mutations that do not affect HGPRT catalytic activity and are therefore not detected in a 6-thioguanine resistance assays may have contributed to underestimation of OTA-mediated mutations in mouse kidney DNA. In contrast, there are no phenotypically silent mutations in the supF gene in the reporter assay used in this study. Hibi et al. [45] did observe enhance OTA-mediated production of large (ca. 10 kb) deletions in mouse kidney; such mutations are not observable in the plasmid pSP189-based assay. The occurrence of large deletions in the mouse kidney vs. base substitutions and frameshift mutations in the in vitro plasmid based assay may reflect differences in the OTA activation pathways to a proximate mutagen. Base substitution and/or frameshift mutations mediated by OTA in vitro are enhanced by conversion to the hydroquinone in the presence of cysteine or to the phenoxyl radical in the presence of Fe(III), both of which produce DNA base-reactive OTA intermediates. It is possible that OTA activation pathways in the mouse kidney generate phosphodiester backbone reactive intermediates, producing double strand breaks that lead to large deletions, rather than DNA base reactive intermediates. In this regard, differences in OTA activation pathways in the kidney as compared to those utilized in the in vitro mutation reporting system may be attributable to the fact that liver, as opposed to kidney, microsomes were used for in vitro activation. Future experiments using the in vitro mutation reporting system described herein will focus on the impact of OTA bioactivation by kidney microsomes, enabling comparison with the mutagenicity of liver microsome-activated OTA in vitro and the in vivo OTA-mediated genetic alterations described by Hibi et al.