Food Sample Extracts Research Articles

Quantification of Cyclo-di-BADGE and Identification of Several BADGE Derivatives in Canned Food Samples

Most cans intended to come into contact with food require the application of an internal coating made from synthetic polymers. The most widely used type of coating are epoxy resins based on bisphenol A diglycidyl ether (BADGE). It has been found that some components like cyclo-di-BADGE (CdB) and other BADGE derivatives can migrate into food. In this study, the polymeric coating of forty-eight cans was assessed using an infrared spectrometer with attenuated total reflectance. The food samples were extracted, and a targeted analysis was carried out to quantify CdB using liquid chromatography with fluorescence detection (HPLC-FLD). The first estimation of the exposure of the adult population was estimated by combining the concentration of CdB in the samples and the Spanish consumption data. In addition, a non-targeted screening by LC-MS was performed in the food sample extracts for the identification of other BADGE derivatives. Twenty samples were positive for the presence of CdB reaching concentrations of 2623 µg/kg. However, the tolerable daily intake of 1.5 µg/kg bw/day recommended for chemical compounds with high toxicological risk was not exceeded. A total of 18 epoxy oligomers could be tentatively identified in the food extracts being BADGE.H2O.BuEtOH the derivative with the highest incidence.

A Smart Advanced Chemiluminescence-Sensing Platform for Determination and Imaging of the Tissue Distribution of Natural Antioxidants.

Antioxidants have gained marked attention owing to their ability to prevent the oxidation of biological components and to protect the body from reactive oxygen species, thereby maintaining human health. Thus, antioxidant-rich dietary supplements and natural foods can be effective against oxidative stress and can even act as chemopreventive agents. Therefore, a simple and rapid assay for evaluation of antioxidant capacity and assessment of their distribution profile in natural sources is vital. Herein, we report a rapid, innovative chemiluminescence (CL) platform for evaluation and visualization of antioxidant capacity. We found that intense and long-lasting CL was formed upon the redox reaction of quinones, e.g., menadione, with antioxidants, e.g., l-ascorbic acid, in the presence of luminol. The produced CL intensities were proportional to the antioxidants' concentrations with a detection limit of 0.18 μM for the model antioxidant, l-ascorbic acid. As the formed CL was long-lasting, it could be easily captured and detected with a charge-coupled device (CCD) camera. To evaluate the quantification ability of the CCD camera, we developed a smart and fast microplate-based assay based on photographing the generated CL with a cooled CCD camera. The photographed CL intensities were linearly proportional with the antioxidant concentrations, and then the method was applied for photographing multiple food sample extracts. Ultimately, we utilized our method for the distribution profiling of antioxidant capacity in food cut sections. Samples were dipped in luminol and then in quinone, followed by CCD camera photography, without the need for any pulverization/extraction procedure, giving precise antioxidant distribution information.

Stable isotope labeling-assisted GC/MS/MS method for determination of methyleugenol in food samples.

Methyleugenol is a common phenylpropanoid compound found in many plants and has been widely used as a flavoring agent in people's daily life. In this study, a stable isotope dilution assay-coupled gas chromatography/triple quadrupole mass spectrometry (SIDA-GC/MS/MS) method was developed for the quantitative determination of methyleugenol in food samples. Methyleugenol-D3 was synthesized and used as an isotope internal standard for the determination of methyleugenol. The QuEChERS (quick, easy, cheap, effective, rugged and safe) method was applied to the clean-up of food sample extracts. Confirmation and quantification were carried out by GC/MS/MS. The analytical performance of the method was validated. The determination range of methyleugenol was linear from 4 to 500 µg L-1 . Method detection limits for solid food samples, semi-solid food samples and liquid beverages were 50, 50 and 1 µg kg-1 respectively. Satisfactory recoveries in the range 94.29-100.27% were obtained. Intra- and inter-day precision was also validated and the values were all lower than 9%. The method was successfully applied to quantify methyleugenol in different kinds of food samples. This article describes a new method for the accurate quantification of methyleugenol in food samples based on a stable isotope labeling-assisted GC/MS/MS method. Methyleugenol-D3 was synthesized and used as an isotope internal standard for the determination of methyleugenol. Excellent results were generated with the method, and the detection sensitivity and accuracy of the method were good. © 2018 Society of Chemical Industry.

Simple and selective detection of quercetin in extracts of plants and food samples by dispersive-micro-solid phase extraction based on core–shell magnetic molecularly imprinted polymers

In the present study, a D-μ-SPE clean-up method was established for the analysis of quercetin in extracts of plants and food samples using a magnetic molecularly imprinted polymer as the sorbent by HPLC-UV detection.

Magnetic/non-magnetic argan press cake nanocellulose for the selective extraction of sudan dyes in food samples prior to the determination by capillary liquid chromatograpy

Two methods for the determination of Sudan dyes (Sudan I, Sudan II, Sudan III and Sudan IV) in food samples, by solid phase extraction - capillary liquid chromatography, are proposed. Both methods use nanocellulose (NC) extracted from bleached argan press cake (APC), as a nano-adsorbent recycled from an agricultural waste material. One of the methods involves the dispersion of NC in food sample extracts, along with the waste and eluents being separated by centrifugation. In the other method, NC was modified by magnetic iron nanoparticles before using it in the extraction of Sudan dyes. The use of a magnetic component in the extraction process allows magnetic separation to replace the centrifugation step in a convenient and economical way. The two proposed methods allows the determination of Sudan dye amounts at the 0.25–2.00µgL−1 concentration range. The limit of detections, limit of quantifications and standard deviations achieved were lower than 0.1µgL−1, 0.20µgL−1 and 3.46% respectively, when using NC as a nano-adsorbent, and lower than 0.07µgL−1, 0.23µgL−1 and 2.62%, respectively, with the magnetic nanocellulose (MNC) was used. Both methods were applied to the determination of Sudan dyes in barbeque and ketchup sauce samples, obtaining recoveries between 93.4% and 109.6%.

Celiac disease-specific prolamin peptide content of wheat relatives and wild species determined by ELISA assays and bioinformatics analyses

Enzyme-linked immunosorbent assays (ELISAs) are widely used to determine gluten contamination in gluten-free and low gluten food samples. ELISA assays developed using monoclonal antibodies against known toxic peptides have an advantage in the identification of toxic prolamin content in protein extracts of different food samples, as well as raw materials. R5 and G12 monoclonal antibodies specific for two known toxic peptides used in commercially available gluten ELISA assays were applied to test toxic peptide contents in wheat relatives and wild wheat species with different genome composition and complexity. Although the R5 peptide content showed some correlation with ploidy levels in Triticum species, there was a high variance among Aegilops species. Some of the analysed diploid Aegilops species showed extremely high R5 peptide contents. Based on the bioinformatics analyses, the R5 peptide was present in most of the sulphur rich prolamins in all the analysed species, whereas the G12 epitope was exclusively present in alpha gliadins. High variation was detected in the position and frequency of epitopes in sequences originating from the same species, thus highlighting the importance of genotypic variation within species. Identification of new prolamin alleles of wheat relatives and wild wheat species is of great importance in order to find germplasm for special end-use quality purposes as well as development of food with reduced toxicity.

Highly sensitive electrochemical sensor for Sudan I based on graphene decorated with mesoporous TiO2

A sensor for the highly sensitive determination of Sudan I based on the amplified electrochemical response of mesoporous TiO2-decorated graphene (GN–TiO2) was fabricated. The nanoparticles of TiO2 arrayed densely and uniformly on the GN sheets, as confirmed by field emission scanning electron microscopy and transmission electron microscopy images. The electrochemical behavior of Sudan I at this sensor was studied in detail, showing that this sensor exhibited electrocatalytic activity for the oxidation of Sudan I because of the significant peak current enhancement and the lowering of oxidation overpotential. Furthermore, the experimental parameters including supporting electrolyte, volume of GN–TiO2 suspension on electrode surface, accumulation potential, and time were optimized and the electrochemical reaction mechanism of Sudan I on this sensor was investigated. The linear range is from 3.3 nM to 0.66 μM, and the limit of detection is estimated to be 0.60 nM. At last, the sensor was used to determine Sudan I in food sample extracts, which are in good agreement with the results obtained by chromatographic method.

Electrochemical sensor based on graphene and mesoporous TiO2 for the simultaneous determination of trace colourants in food

Currently, synthetic colourants draw much attention as food additives. This paper investigated the simultaneous electrocatalytic oxidation of sunset yellow and tartrazine, two yellow colourants commonly present in food together, with a novel voltammetric sensor based on graphene and mesoporous TiO2 modified carbon paste electrode. Due to the high accumulation effect and great catalytic capability of graphene and mesoporous TiO2, the developed sensor exhibited well-defined and separate square wave voltammetric peaks (i.e., 272mV) for sunset yellow tartrazine. The peak currents of sunset yellow and tartrazine increased linearly with their concentration in the ranges of 0.02–2.05μM and 0.02–1.18μM, respectively. And the detection limit was 6.0 and 8.0nM for sunset yellow and tartrazine, respectively. This new sensor was applied to determine sunset yellow and tartrazine in several food sample extracts. Results suggested that the proposed sensor was sensitive, rapid and reliable.

Graphene Decorated with Nickel Nanoparticles as a Sensitive Substrate for Simultaneous Determination of Sunset Yellow and Tartrazine in Food Samples

AbstractIn this paper, ultrathin graphene (GN) nanosheets were uniformly decorated with nickel (Ni) nanoparticles, as confirmed by scanning electron microscopy and transmission electron microscopy. This GNNi hybrid exhibited excellent accumulation and catalytic capacity for the simultaneous detection of sunset yellow and tartrazine, two colorants commonly found mixed in foodstuffs. The electrochemical reaction mechanism of sunset yellow and tartrazine were studied in detail on the GNNi nanocomposite modified glassy carbon electrode. In square wave voltammetry measurements, the oxidation peak potentials of sunset yellow and tartrazine were separated by about 250 mV. This novel proposed voltammetric method has good selectivity which was successfully applied for the simultaneous determination of sunset yellow and tartrazine in some food sample extracts.

Use of a new functional dual coating (FDC) assay to measure low toxin levels in serum and food samples following an outbreak of human botulism

Clostridium botulinum type A toxin is the most prevalent cause of naturally occurring outbreaks of human botulism in the world. The active dichain neurotoxin molecule is composed of a heavy chain (H-chain) of ~100 kDa with the carboxy-terminal end consisting of a receptor-binding (HC) domain, while the amino-terminal (HN) domain is linked by a critical disulfide bond to a light chain (L-chain) of ~50 kDa. Although the mouse bioassay (MBA) is traditionally used to confirm the presence of toxin in serum or food, its sensitivity is insufficient to detect low toxin levels in approximately 30 to 60 % of botulism patients. A novel FDC (functional dual coating) microtitre plate immuno-biochemical assay, which quantifies botulinum toxicity by measuring the HC domain linked with L-chain endopeptidase activity, was modified to allow human serum (lysed or unlysed) to be tested without interference from the matrix, with toxin detection down to 0.03 mouse LD50 per ml serum or 0.13 pg ml(-1) using just 100 µl of clinical samples. The assay was specific for type A toxin and could additionally be applied to whole blood and food samples. Low levels of 1 to 2 mouse LD50 per ml serum of type A toxin were quantified for the first time using the modified FDC assay in two severely intoxicated UK patients who required mechanical ventilation and antitoxin. Toxin levels in recovered food sample extracts were also detected and one MBA-negative sample was found to contain 0.32 LD50 per ml extract. The FDC assay provides a real alternative for public health laboratories to unambiguously confirm all cases of type A botulism and, due to its sensitivity, a promising new tool in toxin pharmacokinetic studies.

Analysis of seven folates in food by LC–MS/MS to improve accuracy of total folate data

An LC–MS/MS method for analyzing seven folates in food was developed and validated. 5-Methyltetrahydrofolate, 5-formyltetrahydrofolate, 10-formylfolic acid, tetrahydrofolate and folic acid were quantified using a stable isotope dilution assay (SIDA) with deuterated analogues as internal standards. Additionally, 10-formyldihydrofolate and 5,10-methenyltetrahydrofolate were quantified using deuterated internal standards different in structure. Due to interconversion of 5,10-methenyltetrahydrofolate to 5-formyltetrahydrofolate and 10-formyldihydrofolate to 10-formylfolic acid during sample preparation, a SIDA was not considered because of a resulting double calculation of the amounts interconverting. [2H4]-5-methyltetrahydrofolate was used as internal standard for 5,10-methenyltetrahydrofolate, due to a similar retention time, and [2H4]-10-formylfolic acid as well as [2H4]-5-methyltetrahydrofolate was used for 10-formyldihydrofolate, because no internal standards co-elute. To confirm that no matrix effects affect the quantitation of 5,10-methenyltetrahydrofolate and 10-formyldihydrofolate, postcolumn infusion experiments were performed. Validation of the assay was accomplished by determining linearity, precision, recovery, limit of detection and limit of quantitation. The latter parameters were partly obtained by application of a dual-isotope label design including [13C5]-labeled folates. The amounts of 5,10-methenyltetrahydrofolate in the purified extracts of different food samples ranged between 0.3 and 1.3 % and for 10-HCO-H2folate between 0.05 and 8 % of the total folate amount. Correction for incomplete recovery of the latter folate during cleanup indicates even higher contents. Therefore, especially 10-formyldihydrofolate should not be neglected to obtain accurate results for folates.

Evaluation of matrix effect in determination of some bioflavonoids in food samples by LC–MS/MS method

In the present work the LC–MS/MS method with solid phase extraction for simultaneous determination of bioflavonoids rutin, quercetin, hesperidin, hesperetin and kaempferol in some food samples (red onion, orange peel and honey) was developed and the matrix effect accompanying this determination was quantified.The matrix effect evaluated using a postextraction addition method was found to be negative in the range −44 to −0.5%, indicating ionization suppression and strongly depended on bioflavonoid concentration. The observed matrix effect was explained taking into account the co-elution of phenolic acids, in terms of their acid–base and hydrophilic properties. The efficacy of extraction expressed as the absolute recoveries of flavonoids were 88–96%, indicating very good efficiency of extraction.The extracts of food samples obtained either by Soxhlet or ultrasonic extraction were analyzed for bioflavonoid content by the LC–MS/MS method in selected reaction monitoring mode using a triple quadrupole detector and standard addition method, which was found to be the most suitable calibration approach for these samples. The optimized separation was achieved on a Phenomenex Gemini C18 column with gradient elution and mobile phase composition A: 2% acetic acid in water and B: acetonitrile. Rs values were in the range from 1.3 to 3.1, indicating good selectivity of the method. The obtained results (mg/100g fresh weight) for different bioflavonids were for rutin 0.16, for quercetin in the range 0.65–56, for hesperidin 0.016–24, for hesperetin 0.0068–36.4 and for kaempferol 0.14–1.63 and generally show good agreement with published data. Low detection limits (0.014–0.063μg/mL) were obtained with acceptable recoveries (86–114%). Total time of analysis was less than 40min, therefore the proposed method represents significant improvement over existing methods.

QUANTITATIVE ANALYSIS OF THE BACILLUS CEREUS EMETIC TOXIN, CEREULIDE, USING MICELLAR ELECTROKINETIC CHROMATOGRAPHY‐CAPILLARY ELECTROPHORESIS

ABSTRACT A micellar electrokinetic chromatography capillary electrophoresis (MEKC‐CE) method was developed for quantitative analysis of cereulide, the emetic toxin produced by Bacillus cereus. Valinomycin, a highly hydrophobic cyclic dodecadepsipeptide with a chemical structure similar to that of cereulide, was used as a surrogate for method development. A series of buffer systems and electrophoretic conditions were investigated and optimized. The final method, yielding good peak resolution and reproducibility of analysis, utilized 20 mM borate buffer, pH 8.5, containing 75 mM SDS. The capillary had a 50 µm internal diameter and 36.4 cm effective length, and was maintained at 25C throughout column preparation and analysis. The sample compartment was maintained at 22C. A constant voltage of 15 kV was applied, after hydrodynamic injection of the sample at 34.5 KPa for 2 s. Detection used ultraviolet light, at 195 nm. A calibration curve, using peak area rather than peak height, was obtained with purified cereulide, with linearity between 2 and 30 ng/µL.PRACTICAL APPLICATIONSDetection of the emetic toxin cereulide, as well as enumeration of B. cereus, is important for determining the risk of emetic food poisoning. The novel capillary electrophoresis method developed in this study, employing the principles of MEKC‐CE, can be used for the quantitative analysis of cereulide. MEKC‐CE gives results quickly and simply, and offers advantages over reversed phase high performance liquid chromatography (RP‐HPLC), through simple sample handling and preparation, less sample required and lower cost of operation. After calibration with purified cereulide was obtained, the applicability of the method to model and real food samples was evaluated with rice samples inoculated with an emetic B. cereus strain and samples from an outbreak of emetic food poisoning. The results showed that the method can be successfully applied to both semi‐purified toxin extracts of cultures and food samples.

Application of an Ultracentrifugation-based Method for Detection of Feline Calicivirus (a Norovirus Surrogate) in Experimentally Contaminated Delicatessen Meat Samples

The aim of this study was the application of an ultracentrifugation-based method for detection of feline calicivirus (FCV) in experimentally contaminated meat samples of roast pork chop, salami and gammon. Virus particles were liberated from food items by washing in 0.5-M glycine containing 1% bovine albumin. Food debris were removed by slow-speed centrifugation, and viruses were subsequently sedimented by ultracentrifugation. The recovery of infectious virus particles was determined by cell culture. Furthermore, RNA was extracted from virus particles and reverse-transcription polymerase chain reaction ((RT)PCR) was performed on the nucleic acid extracts. To demonstrate the efficiency of this method in removal of inhibitors from food sample extracts, an internal amplification control was constructed and incorporated into FCV (RT)PCR. Virus recovery for smoked meat, salami and cooked ham was 12.5%, 3.4% and 5.9%, respectively. The ultracentrifugation-based method is efficient in eliminating or reducing the level of inhibitory substances and can be applied for calicivirus extraction and concentration from delicatessen meat samples.

In Situ Sample Cleanup during Immunoassay: A Simple Method for Rapid Detection of Aflatoxin B1 in Food Samples

A strategy for rapid in situ elimination of interfering substances that are present in extracts of food samples during assay is described in this article. The novel feature of this method is that the sample purification is carried out as a part of the assay, and a separate sample cleanup step is not required. The assay procedure involves the sequential addition of standard or sample, cleaning solutions, and aflatoxin B1–horseradish peroxidase conjugate (AFB1-HRP) over antibody-spotted zones of a membrane, and 3,3′-diaminobenzidine was used as the substrate for visualization. We have determined that trifluoroacetic acid and propionic acids at concentrations of 100mM are highly effective for cleaning groundnut, wheat, corn, and poultry feed samples and that NaHCO3 (100 mM) is successful in cleaning processed soybean. In all cases, subsequent washing was performed with phosphate-buffered saline solution to facilitate the removal of traces of adhering interfering substances. A batch of 12 samples can be analyzed within 8 min either by visual comparison of the color intensity (inversely related to the analyte concentration) of a sample spot with those of reference standards or, more precisely, by densitometry. The method was tested for the analysis of AFB1 in groundnut, wheat, corn, processed soybean, chili, and poultry feed. The detection limit obtained was 5 μg/kg, except for chili, where it was 10 μg/kg. The average recoveries from different noninfected food samples spiked with AFB1 at concentrations of 5 to 100 μg/kg were between 99 and 105%. The values obtained for infected corn and groundnut samples correlated well with the estimates obtained by high-pressure liquid chromatography. The absence of a sample extraction step reduces the cost and labor involved in the assay. The method may be potentially applicable to the assay of other mycotoxins and environmental pollutants.

Australian survey of acrylamide in carbohydrate-based foods

A method was developed and validated for the determination of acrylamide in carbohydrate-based foods. Solid-phase extraction employing a mixed-bed anion and cation exchange cartridge in series with a C18 extraction disk was used to clean-up water extracts of food samples before analysis by liquid chromatography coupled with tandem mass spectrometry detection. The limit of detection was calculated as approximately 25 μg kg−1 and the limit of reporting was 50 μg kg−1. The average method recovery for 84 samples from a range of matrices reporting was 99% with a relative standard deviation of 11.2%. A survey was conducted of 112 samples of carbohydrate-based foods composited from 547 products available in the Australian market. The analytical results were used in conjunction with Australian food consumption data derived from the 1995 National Nutrition Survey (NNS) to prepare preliminary dietary exposure estimates of Australians to acrylamide through only the food groups examined. Mean dietary exposure to acrylamide resulting from consumption of the foods tested, for Australians aged 2 years and above, was estimated as 22–29 µg day−1 (equivalent to 0.4–0.5 µg kg−1 bodyweight day−1) and between 73 and 80 µg day−1 (1.4 and 1.5 µg kg−1 bodyweight day−1) for 95th percentile consumers. Young children (2–6 years) consuming acrylamide-containing foods had a higher acrylamide exposure on a per kilogram bodyweight basis (mean 1.0–1.3 µg kg−1 bodyweight day−1). The estimated exposure of Australians to acrylamide is similar to that estimated for other countries.

Rapid detection of neurotoxic insecticides in food using disposable acetyicholinesterase-biosensors and simple solvent extraction.

The extensive use of pesticides to protect agricultural crops necessitates reliable tools for the detection of residues in food and water, thus ensuring environmental protection and consumer safety. Neuroinhibitors such as organophosphates and carbamates in particular, represent a potential hazard to human health. These compounds are frequently found in food, but conventional methods of analysis are limited as they are either time consuming or not sufficiently sensitive. As a result, a rapid and sensitive biosensor test based on AChE-inhibition was developed. The disposable AChE-biosensor was directly applied in solvent extracts of food samples using isooctane as extraction solvent. A complete assay could be performed in less than 2 h. Recovery rates of 84% were obtained in tests with spiked orange juice samples. Tests in food samples with a lower water content resulted in reduced recovery rates (44% for peach pap baby food). Phosphorothionate insecticides in food could be detected after direct oxidation with N-bromosuccinimide and solvent extraction. The assay displayed a detection limit of 2 microg/kg paraoxon, which was sufficient for the monitoring of maximum residue limits in food according to EU regulations.

Progress in herbicide determination with the thylakoid bioassay

Chloroplast thylakoids are used as biological units to determine herbicides in different kinds of water samples as well as in aqueous extracts of compost, soil or food samples. The thylakoid bioassay shows clearly inhibition of fluorescence yield in the presence of photosystem II specific herbicides. Due to this method the ecotoxicological effect of samples with unknown pollutants can be tested fast and cost effective. It has been proven that all photosynthetic active compounds are recorded at the same time because only additive interactions occur. Therefore, the contamination level can be expressed as cumulative parameter for photosystem II active substances. Application was improved clearly by the addition of the radical scavenger sodium ascorbate to the isolation media and by a higher concentration of the measuring medium. A new data evaluation method is described yielding in a lower detection limit of 0.4 microg diuron/1. The guidelines for the quality of water for human consumption with an allowable concentration of pesticides in groups is 0,5 microg/1 and can be controlled with the thylakoid bioassay without performing any preconcentration steps.

Investigation of photochemical behavior of pesticides in a photolysis reactor coupled on-line with a liquid chromatography–electrospray ionization tandem mass spectrometry system: Application to trace and confirmatory analyses in food samples

The photochemical behavior of pesticides in a photolysis reactor coupled on-line with a liquid chromatography–electrospray ionization mass spectrometer (LC–hν-MS) was investigated. This paper describes the application of LC–hν-MS, in combination with tandem mass spectrometry (MS–MS), to identification of phototransformation products and to the establishment of possible photolytic pathways of pesticides. In addition, the applicability of LC–hν-MS as an alternative to LC–MS–MS, for trace and confirmatory multiresidue analysis in food samples was investigated. To demonstrate the potential of this technique, a series of N-heterocyclic compounds, phenylureas and carbamates, was studied. Several parameters, such as irradiation time and nature of photosensitizers, were investigated, and their impact on the photolytic transformation is presented here. The technique's versatility is also exhibited by using it for identification of triazine isomers, and for detection of pesticide residues in food sample extracts. Illustrative applications for analysis in lettuce and blueberry extracts are described.

Sample Preparation Optimization for the Analysis of Gliadins in Food by Matrixassisted Laser DesorptionIonization Timeofflight Mass Spectrometry

Analysis by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI/TOF-MS) of wheat gliadin extracts from distinct types of wheat varieties displays a similar characteristic gliadin mass pattern within the 30–40 kDa range. Direct observation of this pattern in food samples was used to identify gluten gliadins in food rapidly and easily. A modified sample preparation for gliadin analysis by MALDI/TOF-MS has improved sensitivity of gliadin detection by a factor of more than 50. This procedure allows the direct detection of complex mixtures of gluten gliadins in food sample extracts down to a concentration of 0.01 mg ml-1, i.e. down to levels which are toxic for coeliac patients. This highly sensitive improved method has been also exemplified by the facile detection of individual proteins at low concentrations (100–400 pmol ml-1). © 1997 by John Wiley & Sons, Ltd.