Agnieszka Owczarek-Fendor

Mechanistic insights into furan formation in Maillard model systems.

Furan has recently received considerable attention as a possibly carcinogenic compound occurring in thermally processed foods. Although several food constituents have been identified as furan precursors, multiple formation pathways remain unclear. Therefore, the mechanisms of furan formation in Maillard model systems were studied by means of the carbon module labeling (CAMOLA) technique. Under both roasting and pressure-cooking conditions, furan was formed from glucose via the intact skeleton, and its formation pathways from glucose alone were not amino acid-dependent. However, some amino acids, especially alanine and serine, did influence the furan production by providing an additional formation pathway. Furthermore, most amino acids enhanced the furan production from glucose. Roasting conditions produced 25-100 times higher amounts of furan as compared to pressure-cooking conditions. Surprisingly, in the alanine/glucose model systems, the relative importance of furan production from glucose alone and from the combination of a glucose-derived and an alanine-derived fragment changed completely over a limited time course of 60 min.

Importance of Fat Oxidation in Starch-Based Emulsions in the Generation of the Process Contaminant Furan

The formation of the possibly carcinogenic process contaminant furan was studied in starch-based emulsions during heat treatments as applied for sterilization. Fresh and oxidized soybean, sunflower, high-oleic sunflower, olive, linseed, and rapeseed oils were compared. Results indicated that both the oil type, in particular, the fatty acid composition, and the oxidation degree of the oil determined the susceptibility of the oils to generate furan upon heating. Thus, oils containing the nutritionally relevant omega-3 unsaturated alpha-linolenic acid proved to be able to generate significant amounts of furan if the oils were oxidized. No clear relationship between p-anisidine values of various oils and the amount of generated furan could be observed. However, in the case of soybean oil, significantly more furan was produced upon an increase in oxidation degree. Surprisingly, furan formation in food-relevant systems containing fresh lipids proved to be a minor route (up to 1.5 ppb furan) compared to a previously studied vitamin C containing model system (up to 13 ppb furan).

Furan formation from lipids in starch-based model systems, as influenced by interactions with antioxidants and proteins.

The formation of furan upon sterilization of a lipid-containing starch gel was investigated in the presence of various antioxidants, namely, α-tocopherol, β-carotene, and ascorbic acid, with and without proteins. Results indicated that α-tocopherol did not significantly influence furan formation from oxidized lipids. β-Carotene, suggested previously to be a furan precursor itself, did influence the generation of furan in a concentration-dependent manner, although to a limited extent. Surprisingly, the presence of lipids seemed to limit the furan generation from β-carotene. Interestingly, the addition of ascorbic acid to the emulsions containing soybean or sunflower oils considerably enhanced the formation of furan from these oils. This was also the case when fresh oils were applied, shown previously to be nearly unable to generate furan. This observation can be explained by an intensified ascorbic acid degradation stimulated by the presence of lipids.

On-fiber furan formation from volatile precursors: a critical example of artefact formation during Solid-Phase Microextraction.

For the analysis of furan, a possible carcinogen formed during thermal treatment of food, Solid-Phase Microextraction (SPME) is a preferred and validated sampling method. However, when volatile furan precursors are adsorbed on the carboxen/PDMS fiber, additional amounts of furan can be formed on the fiber during thermal desorption, as shown here for 2-butenal and furfural. No significant increase in furan amounts was found upon heating the furan precursor 2-butenal, indicating that the furan amounts formed during precursor heating experiments are negligible as compared to the additional amounts of furan formed during fiber desorption. This artefactual furan formation increased with increasing desorption time, but especially with increasing desorption temperature. Although this effect was most pronounced on the Carboxen/PDMS SPME-fiber, it was also noted on two other SPME-fibers tested (PDMS and DVB/Carboxen/PDMS). The general impact on furan data from food and model systems in literature will depend on the amounts of volatile precursors present, but will probably remain limited. However, considering the importance of this worldwide food contaminant, special care has to be taken during SPME-analysis of furan. Especially when performing precursor studies, static headspace sampling should preferably be applied for furan analysis.