The natural occurrence of the carcinogenic mycotoxin ochratoxin A (OTA) in wines sold in local retail outlets in South Africa and Italy was investigated by HPLC analysis with fluorescence detection following cleanup by immunoaffinity column. All 24 local South African wines tested (15 white and 9 red) were found to contain detectable levels (>0.01 microg/L) of OTA, with a mean of 0.16 microg/L in the white wines and a mean of 0.24 microg/L in the red wines. Results were subsequently confirmed by LC-MS analysis using positive ion electrospray ionization with collision-induced dissociation of the protonated molecular ion [M + H](+) at m/z 404 and selected reaction monitoring of the resultant product ions [M + H - H(2)O - CO](+) at m/z 358 and [M + H - H(2)O](+) at m/z 386. Comparison with the fluorescence method gave a significant correlation (r = 0.87; p < 0.01). Although OTA contamination was present in all of the South African samples analyzed, levels were well below the suggested European Union limit of 0.5 microg/kg. The highest level found in a locally purchased wine was 0.39 microg/L in a blend of local and imported Spanish red wine. Of the eight Italian wines analyzed, only two red wines were contaminated above the suggested maximum level.

This study shows for the first time the accumulation of fumonisin mycotoxins in human hair of population clusters exposed to contaminated maize, and thus the feasibility of human hair analysis for the assessment of past fumonisin exposure. Composite hair samples were obtained from the Bizana, Butterworth and Centane districts within the Transkei region of the Eastern Cape Province of South Africa. Following methanol extraction and strong anion exchange clean up, the fumonisins FB(1), FB(2) and FB(3) were detected using high performance liquid chromatography coupled to electrospray ionization-mass spectrometry (HPLC-ESI-MS). Hair from Centane and Butterworth showed mean levels of FB(1) of 26.7 and 23.5 microg kg(-1) hair, respectively. FB(2) was only detected in hair from Centane and in one sampling point in Butterworth, with mean levels of 6.5 and 5.7 microg kg(-1) hair, respectively. Hair samples from Bizana, on the other hand, were found to contain higher levels of FB(1) (mean 33.0 microg kg(-1) hair) and FB(2) (mean 11.1 microg kg(-1) hair). No samples contained more than trace levels of FB(3). Recoveries from spiked hair samples using this method ranged from 81% to 101%, demonstrating the applicability of hair analysis in assessing human exposure to fumonisin mycotoxins.

Six strains of Fusarium verticillioides, two of F. oxysporum, one strain of F. proliferatum, and a strain of an unidentified species were cultured on maize patties and rice and evaluated for their ability to simultaneously produce fumonisin B (FB) and C (FC) series analogues. Fumonisins were quantified by LC-MS-MS using positive ion electrospray ionization. FC1 provided characteristic fragment ions at m/z 690, 672, 654, 532, 514, and 338 corresponding to sequential loss of H2O and tricarboxylic acid moieties from the alkyl backbone, while FC3 and FC4 provided equivalent product ions 16 and 32 amu lower than the corresponding FC1 fragments, respectively. All isolates cultured on maize produced FC4. All isolates except for that of F. proliferatum also produced FC1, and three of the six strains of F. verticillioides produced FC3. All isolates except those of F. oxysporum produced detectable amounts of FB1, FB2, and FB3. Isolates that produced fumonisin B analogues produced at least 10 fold more of the B series analogues than they did of the C series analogues. The results confirm that at least some strains of F. oxysporum produce FC, but not FB, fumonisin analogues and also suggest that the genetics and physiological regulation of fumonisin production may be more complicated than previously envisaged since some strains of F. verticillioides and F. proliferatum as well as the strain of the unidentified species can simultaneously produce both FB and FC analogues.

No abstract