High-throughput screening for aroma production in food fermentations.

A microtiter plate (MTP) method was developed to screen 1064 unique microorganisms-substrate fermentations for production of 68 target aroma compounds. Based on the number of hits identified by GC-MS, 50 fermentations were repeated at 50-mL scale in flasks. Comparison of GC-MS data showed that scaling up from MTP to flask did not generally result in large differences between the volatile profiles, even with a wide variety of substrates (juice, food slurry and food side-streams) and microorganisms (yeast, bacteria and fungi) used. From the screening results, Lactobacillus plantarum fermentation of chilli pepper was further studied as a high amount of phenols, especially guaiacol and 4-ethylphenol, was produced after fermentation. From HPLC-MS and sensory analysis, capsaicin was shown to be a probable precursor for these phenols and a potential mechanism was proposed. The protocol described herein to screen aroma compounds from fermentation of agri-food products and side streams can support development of clean label flavourful food ingredients.

Furfuryl alcohol is a precursor for furfurylthiol in coffee.

This study investigated the role of furfuryl alcohol (FFA) in the formation of furfurylthiol (FFT), the most important odorant in roasted coffee, using in-bean and spiking experiments. Green beans were spiked with FFA, and after roasting FFT was quantified by stable isotope dilution analysis. The FFT level in the roasted beans increased dose-dependently with addition of FFA. Additionally, beans were spiked with isotopically labelled d2-FFA which generated isotopically labelled d2-FFT after roasting. However, no labelled furfural was observed. The results unambiguously show that FFA serves as a precursor of FFT in coffee. On the other hand, the data indicate that furfural stems not from oxidation of FFA and plays no major role as precursor for FFT formation during coffee roasting. The suggested formation pathway leads from FFA to the furfuryl cation, then protein-bound S-furfuryl-l-cysteine and by subsequent elimination to FFT.

Interaction and enrichment of protein on cationic polysaccharide surfaces.

In this study, the interaction of fluorescein isothiocyanate functionalized bovine serum albumin (FITC-BSA) with cellulose surfaces decorated with trimethyl chitosan (TMC) is investigated. Two types of TMC, one exhibiting a lower and one with a higher degree of cationization are used for protein adsorption. The adsorption is carried out at different pH values and concentrations of the protein solution. The amount, morphology and wettability of FITC-BSA coating on TMC/cellulose films are determined using quartz crystal microbalance with dissipation (QCM-D), atomic force microscopy, fluorescence microscopy and contact angle measurements. A lower pH and higher concentration of protein solution resulted in a greater amount of irreversibly adsorbed material owing to the reduced solubility and minimized electrostatic repulsion. A maximum adsorption of protein is observed on cellulose surfaces functionalized with TMC carrying a higher degree of cationization compared to TMC with a lower degree of cationization and pure cellulose surfaces at all applied concentrations and pH values. BSA is a commonly used model protein and is applied in this study to better understand its interaction with cationically rendered cellulose surfaces. Such knowledge is essential for creation of multifunctional polysaccharide-based biomaterials.

Formation of cysteine-S-conjugates in the Maillard reaction of cysteine and xylose.

Cysteine-S-conjugates (CS-conjugates) occur in foods derived from plant sources like grape, passion fruit, onion, garlic, bell pepper and hops. During eating CS-conjugates are degraded into aroma-active thiols by ß-lyases that originate from oral microflora. The present study provides evidence for the formation of the CS-conjugates S-furfuryl-l-cysteine (FFT-S-Cys) and S-(2-methyl-3-furyl)-l-cysteine (MFT-S-Cys) in the Maillard reaction of xylose with cysteine at 100°C for 2h. The CS-conjugates were isolated using cationic exchange and reversed-phase chromatography and identified by (1)H NMR, (13)C NMR and LC-MS(2). Spectra and LC retention times matched those of authentic standards. To the best of our knowledge, this is the first time that CS-conjugates are described as Maillard reaction products. Furfuryl alcohol (FFA) is proposed as an intermediate which undergoes a nucleophilic substitution with cysteine. Both FFT-S-Cys and MFT-S-Cys are odourless but produce strong aroma when tasted in aqueous solutions, supposedly induced by ß -lyases from the oral microflora. The perceived aromas resemble those of the corresponding aroma-active thiols 2-furfurylthiol (FFT) and 2-methyl-3-furanthiol (MFT) which smell coffee-like and meaty, respectively.

Identification of 5-hydroxy-3-mercapto-2-pentanone in the maillard reaction of thiamine, cysteine, and xylose.

5-Hydroxy-3-mercapto-2-pentanone is claimed in the scientific literature as a key intermediate in the degradation of thiamine and the related generation of aroma compounds; however, there are no analytical NMR and MS data available. We have identified the compound in a thermally treated mixture of thiamine, cysteine, and xylose and characterized it by MS and NMR.

The aroma side of the Maillard reaction.

The Maillard reaction in food produces, among others, a diversity of sensory-active compounds (aroma, taste, color). The resulting key aroma compounds are often present only in trace concentrations of 1 microg/kg to 1 mg/kg. Nevertheless, they contribute to the respective flavor because of their low odor-perception thresholds. While Maillard intermediates, such as Amadori compounds and deoxyosones, are formed at percentage levels during model reactions, the yield of aroma compounds, in particular nitrogen and sulfur-containing ones, is often as low as 0.001-0.01 mol%, thus indicating their formation through chemical side reactions. The elucidation of the relevant precursors in food and the identification of previously unknown intermediates can throw light on these minor pathways. Also, model reactions with isotopically labeled precursors are of great value in gaining insight into the relevant formation mechanisms. Several examples of these studies are illustrated including work to elucidate the role of the solvent glycerol in the formation of pyrazines, trials to reveal the relative significance of 4-hydroxy-5-methyl-3(2H)-furanone as intermediate in the reaction between ribose and cysteine, and experiments to assess the proportional contribution of the precursors cysteine, xylose, and thiamine to the formation of the resulting aroma compounds in the thermal reaction.

Effect of pH on the Maillard reaction of [13C5]xylose, cysteine, and thiamin.

The influence of different pH values, ranging from 4.0 to 7.0, on the formation of sulfur volatiles in the Maillard reaction was studied using a model system with [13C5]xylose, cysteine, and thiamin. The use of 13C-labeled xylose allowed, by analysis of the mass spectra, volatiles that incorporated xylose carbons in the molecule from other carbon sources to be discerned. For 2-furaldehyde and 2-furfurylthiol, which were favored at low pH, the labeling experiments clearly indicated that xylose was the exclusive carbon source. On the other hand, xylose was virtually not involved in the formation of 3-mercapto-2-butanone, 4,5-dihydro-2-methyl-3-furanthiol, and 5-(2-hydroxyethyl)-4-methylthiazole, which apparently stemmed from thiamin degradation. Both xylose and thiamin seemed to significantly contribute to the formation of 2-methyl-3-furanthiol, 3-mercapto-2-pentanone, and 2-mercapto-3-pentanone, and therefore different formation pathways must exist for each of these molecules. In general, the pH determined strongly which volatiles were formed, and to what extent. However, the relative contribution of xylose to the C-skeleton of a particular compound changed only slightly within the investigated pH range, when both xylose and thiamin were involved in the formation.

Role of the solvent glycerol in the Maillard reaction of d-fructose and l-alanine.

The volatiles in the headspace above a solution of [(13)C(6)]fructose and alanine in glycerol/water, heated in a closed vial at 130 degrees C for 2 h, were analyzed by solid-phase microextraction in tandem with GC-MS. Carbonyl compounds and pyrazines were among the detected components. The examination of their mass spectra showed that most of the 1-hydroxy-2-propanone and 2,3-pentanedione were (13)C(3)-labeled, the majority of the 2-methylpyrazine and 2-ethyl-3-methylpyrazine were (13)C(5)-labeled, and 2,5-dimethylpyrazine and 3-ethyl-2,5-dimethylpyrazine were mainly (13)C(6)-labeled. This is in agreement with the literature, and corresponds to the incorporation of fructose carbons, and in the case of 2,3-pentanedione, 2-ethyl-3-methylpyrazine, and 3-ethyl-2,5-dimethylpyrazine alanine carbons, into the molecules. However, minority fractions of 1-hydroxy-2-propanone (10%) and 2,3-pentanedione (14%) were found unlabeled, 2-methylpyrazine (10%) and 2-ethyl-3-methylpyrazine (11%) only doubly labeled, and 2,5-dimethylpyrazine (20%) and 3-ethyl-2,5-dimethylpyrazine (27%) only triply labeled, suggesting they contain carbons originating from the solvent glycerol. This could be confirmed by reaction of fructose and alanine in [(13)C(3)]glycerol/water, which produced the same volatiles, with 11-27% existent in their (13)C(3)-labeled form. Hence, glycerol participated not only as a solvent but also as a precursor in the reaction.

Alpha-mercaptoketone formation during the maillard reaction of cysteine and [1-(13)C]ribose.

The volatiles formed from [1-(13)C]-ribose and cysteine during 4 h at 95 degrees C in aqueous phosphate buffer (pH 5) were analyzed by headspace SPME in combination with GC-MS. The extent and position of the labeling were determined using MS data. The identified volatiles comprised sulfur compounds such as 2-[(13)C]methyl-3-furanthiol, 2-[(13)CH(2)]furfurylthiol, [1-(13)C]-3-mercaptopentan-2-one, [1-(13)C]-3-mercaptobutan-2-one, [4-(13)C]-3-mercaptobutan-2-one, and 3-mercaptobutan-2-one. The results confirm furan-2-carbaldehyde as an intermediate of 2-furfurylthiol, as well as 1,4-dideoxypento-2,3-diulose as an intermediate of 2-methyl-3-furanthiol and 3-mercaptopentan-2-one. Loss of the C-1 and C-5 carbon moieties during the formation of 3-mercaptobutan-2-one suggests two different mechanisms leading to the key intermediate butane-2,3-dione.

Hydrophilic interaction liquid chromatography coupled to electrospray mass spectrometry of small polar compounds in food analysis.

Reversed-phase liquid chromatography (RPLC) is commonly used to analyze nonvolatile components in food. However, polar low-molecular-weight compounds such as hydrophilic amino acids, di- and tripeptides, and organic acids are often not sufficiently retained and represent a challenge for RPLC. Hydrophilic interaction liquid chromatography in combination with electrospray mass spectrometry (HILIC-ESI-MS) on a carbamoyl-derivatized stationary phase was successfully employed to separate free amino acids and small polar peptides in complex food matrixes such as wheat gluten hydrolysate and Parmesan cheese. Glutamyl dipeptides were separated in a sequence-specific order with peptides with N-terminal glutamic acid residues eluting prior to their reverse sequence analogues. ESI-MSn detection in the positive ionization mode provided the necessary information to unambiguously identify isobaric peptides due to their characteristic fragmentation patterns. The technique also proved useful to separate and identify glycoconjugates between amino acids and reducing sugars (Amadori compounds). The investigation of organic acids present in food used a mobile phase comprising ammonium acetate buffer at pH 7 and mass spectrometric detection in the negative ionization mode.

Formation of aroma compounds from ribose and cysteine during the Maillard reaction.

The headspace volatiles produced from a phosphate-buffered solution (pH 5) of cysteine and a 1 + 1 mixture of ribose and [(13)C(5)]ribose, heated at 95 degrees C for 4 h, were examined by headspace SPME in combination with GC-MS. MS data indicated that fragmentation of ribose did not play a significant role in the formation of the sulfur aroma compounds 2-methyl-3-furanthiol, 2-furfurylthiol, and 3-mercapto-2-pentanone in which the carbon skeleton of ribose remained intact. The methylfuran moiety of 2-methyl-3-(methylthio)furan originated from ribose, whereas the methylthio carbon atoms came partly from ribose and partly from cysteine. In 3-mercapto-2-butanone one carbon unit was split from the ribose chain. On the other hand, all carbon atoms in 3-thiophenethiol stemmed from cysteine. In another trial cysteine, 4-hydroxy-5-methyl-3(2H)-furanone and [(13)C(5)]ribose were reacted under the same conditions. The resulting 2-methyl-3-furanthiol was mainly (13)C(5)-labeled, suggesting that it stems from ribose and that 4-hydroxy-5-methyl-3(2H)-furanone is unimportant as an intermediate. Whereas 2-mercapto-3-pentanone was found unlabeled and hence originated from 4-hydroxy-5-methyl-3(2H)-furanone, its isomer 3-mercapto-2-pentanone was formed from both 4-hydroxy-5-methyl-3(2H)-furanone and ribose. A new reaction pathway from ribose via its 1,4-dideoxyosone is proposed, which explains both the formation of 2-methyl-3-furanthiol without 4-hydroxy-5-methyl-3(2H)-furanone as an intermediate and a new way to form 3-mercapto-2-pentanone.