Chemical Contaminants in Plant Based Protein Survey

Fera has ISO17025 accreditation for several mycotoxin and plant toxins methods. These include a method for a suite of 17 Fusarium mycotoxins, by LC-MS/MS; a method for ten trichothecenes; as well as aflatoxins B1, B2, G1, G2 and OTA in a range of matrices, including cereals and cereal products using immunoaffinity column clean-up and HPLC with fluorescence detection. Ergot alkaloids and tropane alkaloids methods using LC-MS/MS are also accredited. Mycotoxin methods are required to meet, as a minimum, the performance characteristics given in Commission Regulation (EC) No 401/2006 (17). The general operation of LC-MS methods is accredited and Fera has Flexible Scope accreditation that allows accreditation to be claimed for certain analyte, matrix and instrument combinations, this was applied where the analyte matrix was not covered by the fixed scope. 

All samples in Phase 1 and 2 were analysed for mycotoxins using an in-house method for the determination of 11+ mycotoxins by IAC clean-up and LC-MS/MS analysis (11+Myco MS-PREP®, Art. No. RBRP128 / RBRP128B). This column was validated for the analysis of multiple mycotoxins in several matrices through the AOAC Research Institute, Certificate No. 112401. Fera was the independent laboratory that carried out some of the validation (18). The method uses immunoaffinity columns with antibodies specific to certain mycotoxins and has claimed performance for the following mycotoxins: 

• Aflatoxins B1, B2, G1 and G2
• Ochratoxin A (OTA)
• Deoxynivalenol (DON)
• T-2 and HT-2 toxin
• Fumonisins B1 and B2
• Zearalenone (ZON)

In addition, due to cross reactivity of the antibodies the columns also cross-react with: 

• Fumonisin B3
• Sterigmatocystin

The method had previously been validated in-house for oat products, dairy alternatives, cereals and animal feed for all of the above analytes. Method optimisation and then validation for plant protein products was completed before the survey samples were analysed. Application for flexible scope accreditation was made for plant protein products for this method using the validation data generated during this study and the method is now accredited.

7.1.1 Mycotoxins analysis by 11+ IAC method

For all samples, an aliquot of sample (10 g) was weighed into a centrifuge bottle. Water (40 mL) was added and the sample shaken by hand to mix. A mixture of acetonitrile: methanol, (1:2), was added and the tube shaken well by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 2 hours. Following this, samples were centrifuged, and an aliquot (10 mL) of supernatant was diluted with Phosphate Buffered Saline (PBS, 110 mL). An aliquot of the diluted extract was cleaned up by passing through the Immunoaffinity Column (IAC, R-Biopharm Rhone) under gravity. The IAC was washed with water, then dried by passing air through before the analytes were eluted by passing methanol (1 mL) through the IAC into a vial, followed by water (1 mL) which was collected in the same vial to ensure maximum recovery was obtained. This solution was mixed well using a vortex mixer, and if necessary, filtered through a syringe filter before being transferred to an autosampler vial for LC-MS/MS analysis.

Quality control samples including procedural blanks, and spiked samples, were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. In the absence of suitable matrix matched in-house reference materials, samples spiked at different concentrations were included in each batch. Limits of quantification (LOQ) for the 11+ IAC method are given in Table 6 (below). The target LOQ is defined as the lowest level at which validation using spiked samples was undertaken. The target LOQs were agreed with FSS at the start of the project. The LOQ (lowest level with acceptable signal: noise from the calibration curve) has been used when reporting results as some samples contained mycotoxins below the lowest validation concentration. 

Method performance requirements for mycotoxins are given in Assimilated Commission Regulation (EC) No 401/2006 (17), no criteria are given for sterigmatocystin, however average recovery of 70-120% is deemed to be acceptable for mycotoxins and plant toxins according to the EURL-MP Guidance doc_003 (v1.4) that sets out method performance criteria for methods of analysis (18). LC-MS analysis is covered by accreditation and quality parameters are set out in in-house document FSG 002. All criteria for method performance, i.e. calibration and confirmation of residues were met. 

Table 6. Limits of Quantification (LOQ, lowest validated level) for analytes in 11+ IAC method

Samples containing wheat protein were analysed for ergot alkaloids using In-house method FSG 601, the analytes included in the method are listed in Table 7. This method is UKAS accredited for cereals, cereal products and baby food. Additional validation for wheat containing plant protein products was carried out to ensure the method performed satisfactorily for these products. Products containing wheat in Phase 1 and all samples in Phase 2 were analysed by this method.

Table 7. Analytes included in In-house method FSG 601. 

7.2.1 Ergot alkaloids method

For samples tested for ergot alkaloids, an aliquot of sample (10 g) was weighed into a sample bottle. Extraction solvent, a mixture of acetonitrile and ammonium carbonate solution (84:16, 100 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 30 minutes. Following this, samples were shaken by hand then poured through Whatman 12.5 cm hardened No. 54 ‘fast’ filter paper. The filtrate was collected in a conical flask. An aliquot (1 mL) of the filtrate was transferred into a 4 mL amber glass vial containing ca. 50 mg of Bondesil dispersive solid phase extraction (SPE) material. The mixture was filtered through a 13 mm PTFE 0.22 µm syringe filter before being transferred to an autosampler vial for LC-MS/MS analysis.

The test samples were analysed in the same analytical batch as the spiked validation samples. Samples were spiked at three different concentrations (1x target LOQ, 2 x target LOQ and 10 x target LOQ). Six replicates of each concentration and three blank samples were analysed with the test samples. Performance characteristics for mycotoxins are given in Commission Regulation (EC) No 401/2006 (17), no criteria for ergot alkaloids are given, however average recovery of 70-120% is deemed to be acceptable for all mycotoxins and plant toxins (11). 

All Phase 1 samples were analysed for tropane alkaloids atropine and scopolamine using In-house method FSG 827 LC-MS/MS analysis of tropane alkaloids (TAs) in cereals, processed cereal based foods, herbal dry tea, herbal tea infusions, and vegetable products. This method is UKAS accredited for cereal and vegetable products however a set of verification analyses were carried out for the products included in the survey to ensure the method performance was acceptable. 

7.3.1 Tropane alkaloids method

For all samples, an aliquot of sample (4 g) was weighed into a polypropylene tube. Internal standard solution containing deuterium labelled atropine and scopolamine was added to all samples. Extraction solvent (methanol/water/formic acid solution (75/25/0.4, 40 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 30 minutes. Following this, samples were centrifuged, and an aliquot (10 mL) of supernatant was taken for SPE clean-up using cation exchange. After conditioning the cartridge, the sample extract was loaded. The SPE cartridge was washed with a solution of formic acid in methanol and water, and the cartridge dried by vacuum. The cartridges were eluted with a solution of 0.5% ammonia in methanol. The collected eluate was evaporated to dryness under a nitrogen atmosphere in a water bath at 50°C using a Turbovap. The dry residue was redissolved in 500 µL of sample resuspension solvent (10% methanol in water) before being transferred to a filter vial and closed. 

Quality control samples including procedural blanks, and spiked samples were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. In the absence of suitable matrix matched in-house reference materials, samples spiked at three different concentrations (1 µg/kg, 2 µg/kg and 10 µg/kg) were included in each batch. The limit of quantification (LOQ) for both analytes was 0.1 µg/kg. Performance characteristics for plant toxins are given in EURL-MP Guidance doc_003 (v1.4) (19), average recovery of 70-120% is deemed to be acceptable, however average recoveries can be acceptable within the range 50-130% when precision criteria for RSDr are met. 

All Phase 1 samples were analysed for acrylamide using In-house method FSG 262 GC-MS analysis of acrylamide in foods. This method is UKAS accredited for foods and beverages. It is based on the PD CEN/TS 17083:2017 method for determination in of acrylamide (20) and has been used extensively for many years in range of foods, and therefore no further validation or verification was required. 

7.4.1 Acrylamide method

For all samples, an aliquot of sample (5 g) was weighed into a bottle. Internal standard solution was added to all samples. Freshly boiled hot water (100 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for at least one hour. Following this, samples were removed from the shaker and allowed to cool. An aliquot (approximately 40 mL) of extract was transferred to a centrifuge tube. Carrez solution I (1 mL) and Carrez II (1 mL) were added with shaking after each addition. The tubes were centrifuged and a portion of supernatant (25 mL) transferred to a clean centrifuge tube. This was shaken with n-hexane (approximately 15 mL) and centrifuged again. An aliquot of the aqueous extract (15 mL) was derivatised by bromination by addition of 10% sulphuric acid solution (600 µL), 0.1M potassium bromate solution (1 mL) and 0.4mg/ml potassium bromide solution (6.25 mL). The tube was swirled to mix the sample after each addition, capped securely and shaken well to mix. 

Bromination took place in the dark in a refrigerator (<10°C) for at least 1 hour. After incubation, 1M sodium thiosulfate solution (200 µL) was added until the yellow colour disappeared after shaking. This extract was shaken with ethyl acetate and salt, after which the ethyl acetate was removed by pipette and then evaporated to a volume of ca. 0.5 mL.  Triethylamine (50 µL) was added, the vial capped securely and shaken to dissolve the residue. If required the extract was filtered through a 0.45µm PTFE syringe filter. All extracts were analysed by GC-MS. 

Quality control samples including procedural blanks, and spiked samples were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. The limit of quantification (LOQ) was 30 µg/kg.

This method is similar to the method for tropane alkaloids described in section 7.3, it uses acidified extraction followed by cation exchange SPE clean-up. It allows the analysis of up to 41 pyrrolizidine alkaloids and atropine and scopolamine in a single analysis. The method has undergone in-house validation for a range of foodstuffs. Flexible scope application is in progress for these matrices. All Phase 2 samples were analysed with this method. The analytes requested by FSS are listed in Table 5. It should be noted that lycopsamine does not separate chromatographically from another pyrrolizidine alkaloid, indicine. Results for lycopsamine are therefore reported as a combined value for lycopsamine and indicine. 

7.5.1 PA and TAs method

For all samples, an aliquot of sample (4 g) was weighed into a polypropylene tube. Internal standard solution was added to all samples. Extraction solvent, (methanol/water/formic acid solution (60/40/0.4, 40 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 60 minutes. Following this, samples were centrifuged, and an aliquot (5 mL) of supernatant was taken for SPE clean-up using cation exchange. After conditioning the cartridge, the sample extract was loaded. The SPE cartridge was washed with a solution of formic acid in methanol and water, and the cartridge dried by vacuum. The cartridges were eluted with a solution of 5% ammonia in methanol. The collected eluate was evaporated to dryness under a nitrogen atmosphere in a water bath at 50°C using a Turbovap. The dry residue was redissolved in 100 µL of methanol followed by 900 µL water before being transferred to a filter vial and closed. 

Quality control samples including procedural blanks, and spiked samples were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. In the absence of suitable matrix matched in-house reference materials, samples spiked 10 µg/kg (before extraction and immediately before analysis) were included in each batch. The limit of quantification (LOQ) for pyrrolizidine alkaloid analytes was 1 µg/kg and was 0.25 µg/kg for atropine and scopolamine. As lycopsamine and indicine are reported together higher reporting limit (LOQ) of 2 µg/kg has been used for these analytes. 

The method is accredited for the analysis of 17 Fusarium mycotoxins in cereals and animal feed. All Phase 2 samples were analysed using this method. 

7.6.1 Fusarium mycotoxins analysis FSG/818

For all samples, an aliquot of sample (5 g) was weighed into a polypropylene tube. Internal standard solution was added to all samples. Extraction solvent, acetonitrile/water (84/16, 20 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 120 minutes. Following this, samples were centrifuged, and an aliquot (10 mL) of supernatant was taken for SPE clean-up using Oasis Prime HLB cartridges. The cartridge is conditioned with a small volume of sample, then 1 mL of the sample extract was loaded. The sample extract is allowed to pass through the cartridge and collected in a glass tube. An aliquot (0.5 mL) of cleaned extract was evaporated to dryness under a nitrogen atmosphere in a water bath at 50°C using a Turbovap. The dry residue was redissolved in 500 µl of sample resuspension solvent (100 µl of internal standard solution and 400 µl water) before being filtered into an autosampler vial. 

Quality control samples including procedural blanks, and spiked samples were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. In the absence of suitable matrix matched in-house reference materials, samples spiked at three different concentrations (1 x LOQ, 2 x LOQ and 10 x LOQ) were included in each batch. The limits of quantification (LOQ) for the 17 analytes in the method ranged from 2.5 – 50 µg/kg.

The method used is based on CEN method EN 17521:2020 – Determination of Alternaria toxins in tomato, wheat and sunflower seeds by SPE clean-up and HPLC-MS/MS. This method is not UKAS accredited. Fera participated in the interlaboratory method validation study for this method. A set of verification analyses were carried out to ensure the method performance was acceptable. The method includes altenuene (ALT), alternariol (AOH), alternariol monomethyl ether (AME), tentoxin (TEN) and tenuazonic acid (TEA) and isotopically labelled standards are used to internally standardise the method. All Phase 2 samples were analysed by this method. 

7.7.1 Alternaria toxins method

A test portion (2 g) of the sample was spiked with the isotopically-labelled internal standards. The sample was extracted with a methanol, water and acetic acid solution. The sample/extraction solvent mixture was centrifuged and an aliquot of the supernatant collected. The extract was diluted with an equal volume of 1 % aqueous acetic acid solution, and concentrated on a polymeric solid-phase extraction (SPE) cartridge (Phenomenex Strata X2, 6mL, 200 mg sorbent). The extract was eluted from the SPE column with methanol and ethyl acetate solution (75/25, v/v, 7 mL). The eluate was then evaporated under a gentle stream of nitrogen and reconstituted, filtered through a polytetrafluoroethylene (PTFE) syringe filter and analysed by HPLC-MS/MS.

The LOQs were 1 µg/kg each for altenuene (ALT), alternariol (AOH), alternariol monomethyl ether (AME), and 5 µg/kg for tentoxin (TEN) and tenuazonic acid (TEA). 

A method verification exercise was carried out for this method before it was used to analyse the survey samples. Replicate spiked samples (n=6) were prepared at three spiking levels, equivalent to 1 x, 2 x and 5 x LOQ for all 5 analytes. These were analysed by LC-MS/MS and the results are presented in Table C31. 

During the analysis of the survey samples, spiked samples at the LOQs were carried out both before extraction and immediately before analysis to determine recovery and assess any effects from matrix suppression or enhancement. 

This method is based on a multi-mycotoxin ‘dilute and shoot’ method that uses LC-MS/MS analysis. The analytes included were Enniatin A, Enniatin A1, Enniatin B, Enniatin B1, Beauvericin and Fusaric acid. The analytes were extracted in the one extraction, however the LC-MS/MS analysis was carried out using 2 sets of LC-MS/MS conditions as fusaric acid required a different LC column to obtain satisfactory chromatography and peak shape. 

7.8.1 Fusarium toxins analysis

For all samples, an aliquot of sample (5 g) was weighed into a polypropylene tube. Extraction solvent, acetonitrile/water/acetic acid (79/20/1, 20 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 120 minutes. An aliquot (0.5 mL) of extract was diluted with 0.5 mL of dilution solution (acetonitrile/water/acetic acid (20/79/1). This was well mixed then filtered through a syringe filter into an autosampler vial. 

Quality control samples including procedural blanks, and spiked samples were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. In the absence of suitable matrix matched in-house reference materials, samples spiked at 2.5 µg/kg for enniatins and beauvericin and 10 µg/kg for fusaric acid (before extraction and immediately before analysis) were included in each batch. The limits of quantification (LOQ) for enniatins and beauvericin were 2.5 µg/kg each and was 10 µg/kg for fusaric acid.

All Phase 2 samples were analysed for citrinin using an in-house method using IAC clean-up and LC-MS/MS analysis. The method also uses an isotopically labelled citrinin standard to internally standardise the method. 

7.9.1 Citrinin analysis

For all samples, an aliquot of sample (5 g) was weighed into a polypropylene tube. Isotopically labelled citrinin was added to all samples as internal standard. Extraction solvent, methanol/water (75/25, 250 mL) was added and the sample shaken by hand to mix. Samples were placed on a rotary shaker and extracted by shaking for 120 minutes. The sample was sonicated then filtered through Whatman 113V filter paper. An aliquot (2 mL) of extract was diluted with PBS, and 10 ml of the diluted extract was applied to a citrinin IAC. The IAC was washed with 10mM phosphoric acid (10 mL), and the cartridge dried by pushing through 10 mL of air. 

The IAC was eluted with 2 x 1 mL portions of methanol, water (2 mL) was added and the solution mixed before being filtered through a syringe filter into an autosampler vial.  

Quality control samples including procedural blanks, and spiked samples were included in the analytical batches to check accuracy (recovery), and blank control, i.e. no contribution from reagents or laboratory environment. In the absence of suitable matrix matched in-house reference materials, samples spiked at 25 µg/kg citrinin (before extraction and immediately before analysis) were included in each batch. The limit of quantification (LOQ) for citrinin was 2.5 µg/kg

Firstly, the samples are freeze dried. Fat was extracted from the samples using hexane. The extracts were dissolved in toluene before a methylation reagent was added. The vial was sealed and heated overnight to convert into fatty acid methyl esters (FAME’s). These fatty acids were then extracted using hexane and analysed by Gas Chromatography-Flame Ionisation Detection (GC-FID). Fatty acid peaks were identified by comparison to the retention time of the 37-component FAME Standard peaks. The peak areas observed in the samples were used to create a profile of the fatty acid composition of the extracted fat. 

The results are reported showing the amount of fat extracted from 2 g lyophilised sample, and the % fatty acid composition of the extracted fat expressed as erucic acid methyl ester. These results are converted to values in g/kg, erucic acid in fat to allow comparison with the MLs in Commission Regulation EC (No) 1881/2006 (2).