Gregor Kos and Rudolf Krska
Chemical structures of fumonisins B1-B4 (FB1 – FB4) are given in figure 1. Fumonisin B1 is the diester of propane-1,2,3-tricarboxylicacid and 2S-amino-12S, 16R-dimethyl-3S,5R,10R,14S,15R-pentahydroxyeicosane in which the C-14 and C-15 hydroxygroups are esterified with the terminal carboxy group of propane-1,2,3-tricarboxylic acid. FB2 – FB4 show different hydroxylation patterns. CAS numbers and the molar weights of all four fumonisins are given in table 1.
CAS numbers and molar weights of (FB1 – FB4)
Figure 1: Structural formula of fumonisin B1-B4:
Fumonisin B1: R1= OH; R2= OH; R3= OH;
Fumonisin B2: R1= H; R2= OH; R3= OH;
Fumonisin B3: R1= OH; R2= OH; R3= H;
Fumonisin B4: R1= H; R2= OH; R3= H;
Fumonisins mostly occur in maize and maize based products and the development of analytical methods has concentrated on this commodity. FB1 and FB2 are the main target molecules although the methods have need to be found valid for FB3 as well, although no reliable standards exist for this molecule. Little is known about FB4 and its natural occurrence. Fumonisins are produced by Fusarium spp. from the Liseola section including F. moniliforme and F. proliferatum.
Levels of fumonisins in food and feed range from 30 µg/kg to a level of several mg/kg, so methods of analysis have to focus on the µg/kg range.
Analytical procedures differ in extraction, clean up and determination steps, and the method of choice depends on available equipment and analytical requirements such as sensitivity and time of required for analysis. Most analytical methods fall into the category techniques of thin layer chromatography (TLC), liquid chromatography (LC), gas chromatography (GC) and immunochemical methods. Of these, TLC is the simplest, but like all other methods, extraction and clean-up make a major contribution to accuracy and precision of obtained data. Derivatisation is necessary Before before fluorescent detection can be performed, derivatisation is necessary as fumonisins do not contain a chromophore to exhibit radiation.
Detailed reviews of available analytical methods have been given by Shephard , Dutton , Scott  and Norred .
TLC and immunochemical methods such as ELISA are among the most frequent screening methods. There are a number of ELISA kits, which are commercially available. Additionally dipstick tests have been developed, which claim a detection limit of FB1 as low as 0.04-0.06 µg/g in corn based foods (Schneider et al, 1995). In connection with clean-up procedures prior to analysis, the sensitivity of TLC methods can be improved considerably. Reversed phase TLC (on C18 modified silica plates) has also been employed with acidic vanillin or fluorescamine/sodium borate buffer as a spray reagent.
Methods suitable for quantification usually include extraction and clean-up steps prior to determination of the analyte. Various A range of methods can be employed, but all require varying degrees of experience in the field of analytical chemistry and use considerable amounts of reagents.
Extraction and Clean-up
While a clean up is usually not necessary for immunoassays, extensive clean-up procedures are required before physicochemical methods can be employed. Among the commonly used solvent mixtures for extraction are methanol/water (3:1) and acetonitirile/water (1:1). Mixtures can be used either with reversed phase cartridges (C18) or strong anion exchange columns (SAX). More recently immunoaffinity columns (IAC) have become available featuring high selectivity. The analyte molecules are bound to antibodies and the toxin can be eluted subsequent tousing a washing step the toxin can be eluted. Reversed phase cartridges (C18) are useful for the determination of hydrolysis products, although yielding less pure extracts. They can also be used, when SAX columns give poor recoveries. In order to enable quantitative extraction and clean-up from other matrices than maize, a number of different solvent mixtures have been proposed: methanol/borate buffer (3:1, pH 9.2), acetonitrile/methanol/water (1:1:2) and methanol/0.1 M HCl (3:1).
Separation and detection techniques
Reversed Phase (RP-) HPLC separation und and detection is the most widely used technique for the detection of fumonisins, because of their polar character. It was also adopted as a Final Action AOAC Official Method (995.15) for the determination of FB1, FB2 and FB3 in maize. The majority of researchers reported using pre-column derivatisiation with o-phthaldialdehyde / mercaptoethanol (OPA), despite its limited stability. Naphthalene-2,3-dicarboxaldehyde/potassium cyanide (NDA) has been proposed as a viable alternative, although handling is more tedious requiring additional safety precautions.
In an international collaborative study Reported detection limits reported with the usage of OPA in an international collaborative study were 0.05 µg FB1/g and 0.1 µg FB2/g.
GC methods have also been developed by determining aminopolyol after clean-up on a XAD-2 column. Aminopentol (from FB1) and aminotetrol (FB2) react with trimethylsilyl (TMS) or trifluoroacetate (TFA) and are detected by flame ionisation or mass spectroscopy. The advances being made in LC methods caused a shift away from GC methods, as the latter require multiple sample handling steps and more sophisticated equipment.
With the availability of liquid chromatography systems with mass spectrometric detection (LC-MS), a new powerful new method has been established for identification and quantification. High sensitivity can be achieved, although disadvantages include high running and maintenance costs. Limits of quantification have been reported to be as low as 0.001 µg/g.
A project funded by the EU conducted between 1993 and 1995 (MAT1920029) focused on the improvement in the determination of fumonisins (FB1 and FB2) in maize and maize based feeds and was coordinated by CNR/Italy. The intercomparison study generated data (after correction for recoveries) with very high precision for FB1 and FB2. Samples that were spiked with fumonisin standards lead to higher recoveries than naturally contaminated samples, possibly due to an association with other sample matrix components.
Table 2: Results from an intercomparison study conducted as part of the MAT1920029 project (SD: Standard Deviation)
|Precision (mean ± 1SD)
||2.29 ± 0.27
||1.25 ± 0.17
|within-laboratory SD (RSDr)
|Between-laboratory SD (RSDR)
However, most participants obtained low recoveries (average, 70% for FB1 and 69% for FB2), which were considerably affected by the extraction mode. Average recoveries for laboratories using blending were 62% and 60%, whereas for laboratories using shaking were 85% and 86% for FB1 and FB2, respectively.
In addition, the following factors improving recoveries were identified: consecutive extractions, extraction with higher solvent to maize ratio, and SAX clean-up instead C18 clean-up.
Besides, aAnother method based on the principle of immunoaffinity cleanup followed by HPLC detection was also developed and collaboratively validated by CNR/Italy (Dr Visconti) in the frame of an EU-SMT project (EUR report 19451, 2001). The validated scope of the method extends to maize flour and cornflakes. The interlaboratory performance criteria were considered as satisfactory by the CEN/TC 275 WG5 Biotoxins and in its last meeting (November 2001), it was decided to endorse this protocol as a EN Standard which will be published at the beginning of next year. The reproducibility relative standard deviations extended from 26.1 to 34.8 when recoveries were found to be 97-110 % for cornflakes and the reproducibility relative standard deviations ranged from 21.9 to 29.7 with recoveries from 72-75 % for maize flour.
 G.S. Shephard, Chromatographic Determination of the Fumonisin Mycotoxins, Journal of Chromatography A, 815, 31-39 (1998)
 M.F. Dutton, Fumonisins, Mycotoxins of Increasing Importance: Their Nature and Their Effects, Pharmacological and Therapeutics, 70, 2, 137-161 (1996)
 P.M. Scott, Fumonisins, International Journal of Food Microbiology, 18, 257-270 (1993)
 W.P. Norred, Fumonisins – Mycotoxins produced by Fusarium moniliforme, Journal of Toxicology and Environmental Health, 38, 309-328 (1993)