Determination of Acrylamide in Dried Fruits and Edible Seeds Using QuEChERS extraction and LC Separation with MS Detection

Determination of Acrylamide in Dried


Introduction
Acrylamide is a neo-formed contaminant (NFC), produced in food during manufacturing or homecooking (Capuano & Fogliano, 2011).It is absent in raw foods and in raw materials used to make food, and it is produced and accumulated during thermal processing (FSA, 2012).
Global levels of dietary exposure to acrylamide indicate a human health concern (FAO/ WHO, 2010).
Acrylamide was first classified as a potential carcinogen and neurotoxic to humans (Group 2A) based on its carcinogenicity in rodents in 1994 (IARC, 1994) and the suspicion was then endorsed in 2002 (WHO, 2002;SNFA, 2002;Tareke, Rydberg, Karlsson, Eriksson, & Tornqvist, 2002).For these reasons, acrylamide level has been strictly controlled by the authorities (EFSA, 2012) even if there is no existing legal limit for the concentration of this contaminant in foodstuffs.However, European Union fixed a maximum recommended level of 1000 μg kg -1 for potato chips (EU, 2013).
Acrylamide has been detected in some food products that are processed at temperatures in the 98-116 °C range, and in high moisture conditions [i.e., canned black olives (not cured oil) and prune juice] (Roach, Andrzejewski, Gay, Nortrup, & Musser, 2003).It is clear that other pathways of formation below 120 °C can yield acrylamide, and these are being further evaluated (JIFSAN, 2004).
Factors that are particularly important for the Maillard reaction are the starting reactants (Yaylayan & Stadler, 2005), i.e. kind of sugar and amino acid (protein), temperature, time and water activity.
The presence of metal salts (pro-oxidants), and inhibitors, such as antioxidants and sulfite, may have an impact.
During analysis, acrylamide, a small hydrophilic molecule, is usually extracted with water but a polar organic molecule, such as the more volatile acetonitrile, is a suitable alternative (Tateo & Bononi, 2003).To extract acrylamide, different authors used an aqueous solution with highconcentration of NaCl to inhibit the formation of emulsions (Young, Jenkins & Mallet, 2004), or water and 1-propanol on defatted samples (Biedermann, Biedermann-Brem, Noti, & Grob, 2002).
Other authors introduced a deproteinating step (Gertz & Klostermann, 2002).However, all these sample manipulations can be bypassed by a QuEChERS approach (Mastovska & Lehotay, 2006).In comparison with a traditional strategy based on solid phase technique (SPE), the proposed method allows "one-pot" sample preparation thus limiting the amount of solvent used for the extraction.
Remarkable time and money-per-sample savings are considerable, as well.LC-MS/MS is the most used and authoritative method for acrylamide determination.Because its high sensitivity, LC-MS/MS avoids the derivatization step, that is time consuming and potentially unhealthy.
The aim of this study is to set up and apply the more efficient QuEChERS approach in order to extract and determine acrylamide in packed dried fruits (dried prunes and raisins) and some edible seeds (almonds, hazelnuts, peanuts, pine nuts, pistachios, walnuts).

Sampling and grinding
Sixty-eight samples of packed dried fruits and edible seeds were purchased on the Italian market.In particular, dried prunes (13 samples) pitted (7 samples) and not pitted (6 samples), and raisins (7 samples) as dried fruits, and peeled almonds (2 samples), roasted and peeled hazelnuts (2 samples), roasted and salted pistachios (7 samples), pine nuts (7 samples) from Portugal (2 samples) and Italy (5 samples), walnuts (4 samples) with shell from USA (2 samples) and from Chile (2 samples), and roasted and salted peanuts (26 samples) with shell and from Israel (24 samples) and from Egypt (2 samples), as edible seeds.
A gross amount of 20 g of each sample was ground by Osterizer 12-speeds blender (Oster Manufacturing, Di Giovanni Srl, Bologna, Italy) for further elaboration.

Optimized extraction protocol for acrylamide
An aliquot of ground sample (2.50 g) was transferred into a 50-mL Falcon tube together with a ceramic homogenizer for QuEChERS.Then 5 mL of Milli-Q water and 10 mL of acetonitrile were added, and the tube was vigorously hand shaken for 1 min after the addition of each solvent.A prepared mix of QuEChERS pouch composed by MgSO 4 4.0 g + NaCl 0.5 g was added and hand shaken for 1 min and for 3 min with a shaker (Unimax 2010, Heidolph Instruments Italia S.r.l., Milan, Italy) to separate acrylamide into the acetonitrile phase.The tubes were centrifuged for 3 min at 3000 rpm to separate the two layers.An aliquot of the upper layer (2 mL) was dried by a gentle nitrogen stream (20 min in a water bath at 40 °C).The sample was dissolved in 1 mL of Milli-Q water and filtered through a 0.22 μm PES membrane into a 2 mL vial that was loaded into autosampler chamber at controlled temperature, ready for analysis by LC-MS-MS.Each sample extraction was carried out in triple.

Acrylamide determination
The acrylamide determination was carried out by reverse phase liquid chromatography coupled with mass spectrometry system (Agilent Technologies, Waldbronn, Germany) consisting of a vacuum pump (Agilent 1200), a gas generator (API MM20 ZA; Peak Scientific Billerica, MA, USA), a degasser (Agilent 1200), a binary pump (Agilent 1200), an autosampler (Agilent 1200), a thermostated column compartment (Agilent 1200), and a mass spectrometer triple quadrupole (API 3200, AB Sciex Germany GmbH, Darmstadt, Germany).Samples (20 μL) were injected into a Gemini RP C 18 column (Phenomenex, Torrance, CA, USA) (25 cm × 2 mm i.d.× 5 μm particle size × 110 Å pore size).The solvent system was 0.1 % formic acid in water (99.5 %, solvent A) and 0.1 % formic acid in methanol (0.5 %, solvent B) and the elution was carried out in isocratic mode (total run 7 min), with a flow rate of 0.25 mL/min at ambient temperature.The analysis was performed in double for all the samples.
The analyses were carried out in positive electrospray ionization mode (ESI+) and using the following conditions: curtain gas, 20.0 psi; collision activated dissociation, 7.0 (arbitrary units); ion spray, 5500.0V; temperature, 700.0 °C; nebulizer gas, 70.0 psi; heater gas, 30.0 psi.The main MS parameters optimized for acrylamide determination were: declustering potential, 22.0 V; collision energy, 14.1 V; collision cell exit potential, 4.1 V; entrance potential, 6.0 V.The parent ion m/z was 72.1 and the qualifier ion m/z was 54.9.These conditions were optimized by injecting acrylamide solutions directly into the MS source.
Quantification was performed by external standard calibration and the chromatograms were acquired and processed by Analyst software version 1.5 (AB Sciex).

Effect of defatting by hexane and dispersive SPE clean-up steps
Mastovska et al. ( 2006) used hexane to defat samples and dSPE to clean up the extract.However, in our samples the addition of 5 mL of hexane or the use of dSPE reduced acrylamide recovery at about 20 %, and 36 %, respectively.For these reasons, these two steps were not included in the protocol of acrylamide extraction.

Effect of injection solvent on acrylamide determination
Direct injection of the acetonitrile extract after QuEChERS altered the retention time during chromatography.In addition, acetonitrile is a poor solvent for lipids that precipitated trapping part of the acrylamide, thus lowering its recovery.To overcome this problem, two ways were carried out: (i) a pre-dilution of the acetonitrile (250 μL) with water (750 μL), (ii) the complete acetonitrile evaporation and the resolubilization of the sample with water.
Best results were achieved with the latter solution that was used throughout this study.
QuEChERS with NaCl gave the best recoveries (Table 1), probably as a consequence of the reduced emulsions (Young et al., 2004).On the contrary, QuEChERS with CH 3 COONa had a negligible effect on recoveries despite the observations of Eriksson & Karlsson (2006).

Effect of water volume (MgSO 4 -NaCl pouches)
To verify the effect of water volume on acrylamide recoveries, some tests on an acrylamide solution in acetonitrile (250-μg/kg) were carried out (Table 2).Different volumes of water were used: 0, 2.5, and 5.0 mL.To meet a reasonable compromise between recovery and a good sample dispersion, 5.0 mL were used for further experiments.

Effect of sample weight
Comparison of the recoveries of different sample weights was carried out (Table 3).A 2.50-g sample amount was a reasonable compromise between the highest recoveries and the necessity to have a sample amount large enough to be representative.

Linearity and sensitivity determination
Increasing concentrations of acrylamide, from 1 μg/kg to 500 μg/kg, were used to study the range of linearity.As nor EFSA neither FSA suggested limits for dried fruits, we used those ones recommended for biscuits, wafers, and crisp bread (EFSA 2012;FSA 2012).
The coefficient of determination R 2 was 0.999, showing a very good linearity.
Sensitivity of the method was determined with the limit of detection (LOD, 2.0 μg/kg ) and the limit

Accuracy evaluation (recoveries) and inter-laboratory essay
One sample for each matrix was spiked with 250 μg/kg of acrylamide.The recoveries were calculated for each sample (Table 4).For calculations of real samples, correction factors based on the recoveries for each matrix were applied.Calibration was verified each day with three spiked samples at different concentrations.
To test the adaptability of the method to extract acrylamide from other matrices, it was used a European reference matrix (ERM), crisp bread (ERM-BD272) used in Food Analysis Performance Assessment Scheme (FAPAS) for proficiency test 3024 (FAPAS, 2009;Koch, Bremser, Koeppen, Siegel, Toepfer, & Nehls, 2009) to evaluate inter-laboratory variability.Even applying the poorest recovery (61 %), the concentration of acrylamide was in the range of acceptability (Table 5).The same test was then applied to rusks, as this matrix was the most similar to ERM (Koch et al., 2009).
Rusks were added with 250 μg/kg acrylamide and, even in this case, results were in the range of acceptability (Table 5).

Evaluation of precision: repeatability
For the determination of the precision, four levels of contamination were considered: (i) a sample not contaminated (pine nuts, acrylamide < LOD), (ii) a sample with a not quantifiable concentration of acrylamide (pistachios, acrylamide < LOQ), (iii) a sample with a medium amount of acrylamide (peanuts, 42.86 μg/kg), and (iv) a sample with a high concentration of acrylamide (dried prunes, 124.26 μg/kg).Each sample was extracted and injected three times.With acrylamide lower concentrations (5 -25 μg/kg) relative standard deviation (RSD) was 20 %, while with higher acrylamide concentrations (26 -124 μg/kg) RSD was 10 %.

Results obtained on real samples
Raisins, almonds, hazelnuts, pine nuts, and walnuts showed values lower than LOD, as well as pistachios, except for one sample that was lower than LOQ.For this reason the results of these samples are not shown.
These very low concentrations were expected because dried fruits and edible seeds are generally not subjected to high temperature treatments.In literature, it was reported that acrylamide levels in almonds were below 200 μg/kg if the roasting temperature was below 146 °C (Zhang, Huang, Xiao, Seiber, & Mitchell, 2011).These data were recently confirmed by Schlörmann and coll. (2015).
Dried prunes and peanuts (Table 6) showed acrylamide content ranged from 14.74 μg/kg to 124.26 μg/kg, and from 6.16 μg/kg to 42.86 μg/kg (with two samples < LOQ and one < LOD), respectively.The RSD were within 20 %, very good results if compared with 47 % found for roasted kernel (Amrein, Lukac, Andres, Perren, Escher, & Amadò, 2005) and 60 % for dried prunes (Amrein, Andres, Escher, & Amadò, 2007).The lower manipulation of the sample was likely the reason of these good results.Amrein and coll. (2007) detected from 730 to 1680 μg/kg of acrylamide in dried prunes.They also heated again dried prunes at 120 °C for 40 min and they found an increase in acrylamide content (up to 2240 μg/kg).In this case, time is the key factor for acrylamide formation.In fact, even if the drying process of the prunes is carried out at quite low temperature (70-80 °C), the time of exposure is long enough (24-36 h) to allow acrylamide formation.Moreover, the drying process (up to about 18 % moisture) enhances sugar and asparagine concentration (Amrein et al., 2007), thus fostering acrylamide formation.
In the case of peanuts, the roasting temperature is the key factor for acrylamide accumulation.In fact, temperatures are considerably higher (160-180 °C) for 25-30 min of roasting time.

Conclusions
The study describes a quick and easy method for acrylamide determination by QuEChERS approach coupled to LC-ESI-MS-Triple Quadrupole technique with a "one-pot" sample preparation.The method was applied for the first time to food matrices such as dried fruit (prunes and raisins) and some edible seeds.Besides solvent saving, the proposed approach is more sensitive and repeatable, with lower LOD and LOQ.
Results confirm that acrylamide was found with significant concentrations in dried prunes and peanuts as a consequence of the thermal process they were submitted.In general, data agree with those of literature.Other products have shown minimal acrylamide amounts, thus posing reduced harm to humans.
For these reasons and for its flexibility to different products, this method is very promising for acrylamide determination in other matrices, as well.