Detection and identification of fungi and fungal metabolites in cereals, malt, and beer
Physical, chemical, and affinity- based methods
Visual and olfactory examination of ingredients for signs of fungal contamination has been the principal method of quality assessment since humans first made beer. Even today with the availability of advanced instrumentation, experienced brewers and maltsters can recognize a low-quality barley or malt by its colour, smell and hand feel. Musty smells are an indication of attack by typical storage fungi, which in turn can be related to improper storage of barley or malt. Also, changes in grain colour may be indicative for darker beer colours resulting from a malt lot or even for the potential to induce primary gushing. Many German brewers use a method in which the number of red-discoloured kernels is counted in a malt lot. The ‘relevant' grains show a discoloration typical for Fusarium contamination and malt lots with more than five such grains in
200 g will lead to rejection by the brewery (Niessen et al., 1991; Niessen et al., 1992; Engelmann et al., 2012). The method is rapid and requires little effort to be implemented. However, it requires some experience to differentiate ‘relevant' grains from ‘non-relevant' red discolorations. Kernels may be mistakenly stained in shades of red by growth of other moulds, e.g. Epicoccum nigrum, or red yeast belonging to genera Rhodotorula, Rhodosporidium, Sporidiobolus, Sporobolomyces, and Phaffia.
In order to assess fungal contamination and fungal secondary metabolites in brewing cereals and malt, physical and chemical sensors are increasingly applied in modern quality control because results are more objective and reliable as compared to visual and olfactory analysis (Logrieco et al., 2005). Visual and acoustic sensors make use of differences between contaminated and sound grains in regard to absorption or reflection of light or acoustic waves. Both types of analysis provide nondestructive measurement of fungal biomass and secondary metabolites in cereal samples. Acoustic wave sensors have been applied to the detection of trichothecene mycotoxins in wheat using either transmission of acoustic waves at frequencies of 5-36 kHz or reflection of an acoustic impulse in a frequency range 0 - 125 kHz (Juodeikiene et al., 2004). Sensors have been developed for the detection of DON (Juodeikiene et al., 2004, 2008) and DON in co-occurrence with T2-toxin/HT2-toxin (Juodeikiene et al., 2011) in wheat samples. Available visual techniques use different parts of the light spectrum from UV/visible to near- and far-infrared. Levasseur-Garcia (2012) provides an overview of the application of infrared spectroscopy for the identification and detection of fungi and fungal metabolites on cereal grain. The method makes use of the fact that infrared (IR, 2500 nm to 25 ^m) and near infrared (NIR, 760-2500 nm) light induces molecular vibration in organic molecules, which can be measured very sensitively. The analysis of resulting spectra both of acoustic wave-based analysis or IR- and NIR-based analysis uses multivariate, chemometric methods in which statistical correlations between the spectra and certain sample parameters such as mycotoxin concentrations or fungal biomass are established. In order to address such correlations specifically, the range of wavelengths has to be determined for each parameter at which a maximum correlation between deviation from a non-infected sample and the quantity of the assessed parameter, e.g. a mycotoxin or fungal biomass, can be observed. NIR spectroscopy has been applied to the assessment of DON concentrations concurrently with ergosterol and numbers of scabby grains as an indicator for fungal biomass in wheat (Dowell et al., 1999) and barley (Roberts et al., 1991; Borjessen et al., 2007). Also chitin is a compound typically produced by fungi and yeasts that has been applied to the detection of mould contamination in barley and other food sources (Roberts et al., 1991; Cousin, 1996). Since both parameters are prevalent in all fungal organisms, no species-specific detection of fungi in contaminated materials is possible with this method. Fourier-transformed infrared microscopy (FTIR microscopy) is a method using the same principles as described above but it is used to analyse pure cultures of microorganisms including moulds and yeasts and identify them at the species level by comparing sample spectra with reference spectra from a database (Santos et al., 2010; Wenning and Scherer,
2013). In hyperspectral imaging the reflectance of a sample is analysed at various wavelength bands ranging from UV to NIR. For analysis, colour and light intensity of each pixel of the image is analysed for each wavelength band and differences between colour and light distribution in images of contaminated and non-contaminated reference samples are compared. Systems have been commercialized for brewing applications such as detection of Fusarium contamination in wheat (Delwiche et al., 2011).
Chemical detection of fungal contamination and fungal secondary metabolites is a field that has been extensively studied for several decades and is widely used for the analysis of brewing raw materials and beer (Lattanzio et al., 2009). Analytical protocols for all known mycotoxins have been extensively reviewed by several authors ( Jarvis, 2003; Shephard, 2008; Rahmani et al., 2009; Turner et al., 2009). Protocols usually comprise sampling, sample preparation, extraction, cleanup, separation, detection and quantification of mycotoxins. Extraction from contaminated samples is performed using organic solvents or solvent mixtures of optimized polarity in order to separate the analytical target compound from the matrix and other compounds interfering with the analysis. Further cleanup by liquid-liquid extraction or solid phase extraction (reversed phase, ion exchange, immunoaffinity) can be applied to remove non-target compounds and to concentrate analytes. Thin-layer chromatography (TLC), high performance liquid chromatography (HPLC), gas chromatography (GC) and electrophoresis have been used as technical platforms to separate extracted analytes. Absorption of UV and visible light, fluorescence with or without derivatization as well as mass spectrometry are applied to detect the previously separated analytes. Identification of analytes has been accomplished by comparison of retention times with reference materials. HPLC and GC separation have been combined with mass spectrometry (LC-MS, GC-MS) for identification of individual compound peaks. Currently, the most sophisticated analytical systems use tandem mass spectrometry (LC-MS/MS) in which the first MS (MS1) is used to separate different compounds present in an LC-peak after electrospray ionization (ESI) and the second MS (MS2) provides further analysis of selected mass fragments from MS1 for identification. Analysis of each sample in positive and negative ionization mode enables detection and quantification of >130 different secondary metabolites including frequently occurring mycotoxins and their glycosylated derivatives (Vishwanath et al., 2009; Streit et al., 2013). Using an LC-MS/MS-based method that could detect 15 different mycotoxins in parallel analysis, Tamura et al. (2011) analysed samples of beer and beer- based drinks from the Japanese market and found NIV, DON and fumonisins in low concentrations. Romero-Gonzalez et al. (2009) used an LC-MS/ MS-based system to detect 12 different mycotox- ins in beer. Analysis of a small set of commercial samples revealed occurrence of T2- and HT-2 toxins, aflatoxin B1, and fumonisin B2 in low concentrations in some samples. Zachariasova et al. (2010) developed a multimycotoxin method for the screening of 32 different compounds in beer but did not show results of sample analyses. Quantification of analytes in all the chemical detection methods described above is possible by calibration with external standards. Calibration with internal standards has been demonstrated for several mycotoxins with isotopically labelled derivatives in stable isotope dilution assays (SIDA) (Rychlik and Asam, 2008).