Single laboratory method validation comparing MS3 with FI/MS fingerprinting, and quantitation strategies for the accurate determination of ochratoxins in beer.

摘要

The current mycotoxin safety concerns demand the availability of reliable and selective control systems to measure the content of mycotoxins present in daily food. To address this, we report a prospective analytical tool involving the application of a flow injection mass spectrometry (FI/MS) tandem artificial neural network (ANN) strategy to predict the amount of ochratoxin A and B (OTA and OTB) in beer. Triple stage mass spectrometry (MS3) aided by chromatographic separation was applied as a reference method for comparison. 0.1% formic acid and methanol were used to convert ochratoxins into their respective ions under negative MS polarity. For experiments involving MS3, ochratoxins were separated by reversed-phase liquid chromatography in a 6 min run, ionised using electrospray ionisation, and detected by tandem mass spectrometry. Analyte-specific mass-to-charge ratios were used to perform quantitation in MS3 mode. For experiments involving FI/MS, no chromatographic separation was performed. Approximately 2% of the mass spectra was used for model construction. ANN models representing each ochratoxin were individually trained and validated using three sets of matrix-matched and matrix-free calibration curves at nine concentration levels of 2.5, 10, 25, 50, 100, 200, 300, 400 and 500 micro g/l. Quintuplicate analyses were made in FI/ MS mode providing a total of 270 spectra for both OTA and OTB. Single measurement was made for each sample in MS3 mode. A root-mean-square error value of <1% was reported for both ochratoxin models in beer. Limits of quantitation were determined to be 0.2 micro g/kg for both MS3 and FI/MS mode. Recovery assessment was performed over two days using beer blanks (n=6) spiked at three concentration levels of 5, 100 and 200 micro g/kg. Extraction using acetonitrile provided excellent recovery ranges of 88 to 102% for both MS techniques. Relative standard deviations of 10% or better were achieved for interday spike recovery experiments. The successful utilisation of FI/MS without performing chromatographic separation implies the availability of a new analytical tool relevant to the field of mycotoxin analysis, possibly offering analyte specificity exceeding the capability of MS3 through chemometry.