Article pubs.acs.org/JAFC
Impact of Mechanical and Thermal Energies on the Degradation of T‑2 and HT‑2 Toxins during Extrusion Cooking of Oat Flour Henning Sören Schmidt, Stefanie Becker, Benedikt Cramer, and Hans-Ulrich Humpf* Institute of Food Chemistry, Westfälische Wilhelms-Universität Münster, Corrensstrasse 45, 48149 Münster, Germany S Supporting Information *
ABSTRACT: The type A trichothecenes T-2 toxin (T-2) and HT-2 toxin (HT-2) are naturally occurring toxic food contaminants, with the highest concentrations found in contaminated oats. The influence of thermal food processing on these toxins is poorly understood, and only a few publications address the degradation rates. Therefore, we systematically investigated the degradation of T-2 and HT-2 during both laboratory and industrial-scale extrusion cooking of oats. Extrusion cooking under laboratory conditions was performed with oats fortified with T-2 or HT-2 as well as with naturally contaminated oat flour dust. The experiments were designed according to industrial conditions in terms of temperature, water content, pressure, residence time, and oat content. Flour mixtures containing naturally contaminated oats were used for industrial-scale processing. Degradation rates under laboratory conditions were up to 59.6 ± 1.51 and 47.2 ± 0.53% for T-2 and HT-2, respectively, in fortified extrudates but were decreased to 35.1 ± 1.55 and 22.0 ± 4.68% when naturally contaminated flour samples were used. The results show a higher degradation of T-2 during extrusion cooking than of HT-2. Moisture content, mechanical shear, and temperature showed an impact on the toxin degradation and can be optimized to counteract food contamination. KEYWORDS: mycotoxin, thermal food processing, extrusion, T-2 toxin, HT-2 toxin, thermal degradation, oats
■
INTRODUCTION T-2 toxin (3α-hydroxy-4β,15-diacetoxy-8α-(3-methylbutoxy)12,13-epoxytrichothec-9-ene, T-2), 1, and HT-2 toxin (3α,4βdihydroxy-15-acetoxy-8α-(3-methylbutoxy)-12,13-epoxytrichothec-9-ene, HT-2), 2 (Figure 1) are type A trichothecene
Therefore, the toxicity of T-2 is to some extent attributed to HT-2.2 Occurrence data from 22 European countries revealed that the highest concentrations for the sum of both toxins are found in oats,3 and the ratio of HT-2 to T-2 varies mainly in the range of 1.75−8.7.4−7 A survey on the harvest years 2005−2009 in European oat mills revealed average concentrations for the sum of T-2 and HT-2 in oats, oat flakes, oat meal, and oat byproducts of 94, 17, 11, and 293 μg/kg, respectively.8 European Union legislative authorities established maximum levels for some mycotoxins in 2006 to counter possible health risks to consumers caused by mycotoxins, but not for T-2 and HT-2.9 In European Commission (EC) recommendation 2013/165/EU indicative levels for the sum of T-2 and HT-2 in cereals and cereal products were set and cover inter alia unprocessed oats (1000 μg/kg); oat grains for human consumption, flaked oats, and oat bran (200 μg/kg); and bread, pastries, and cereal snacks (25 μg/kg), showing higher requirements regarding the levels of T-2 and HT-2 along with the level of processing.10 The EC is explicitly calling for more information on the effects of food processing on the presence of both toxins.10 The main source of exposure to T-2 and HT-2 for humans is cereals that are subjected to various food-processing steps including physical treatment, such as sorting, trimming, cleaning, or milling, and thermal treatment, such as kilning, cooking, baking, roasting, or extrusion cooking.11 Very little is known about the degradation of and mitigation strategies for
Figure 1. Chemical structures of the type A trichothecenes investigated.
mycotoxins. They are prominent food contaminants produced by fungi of the genus Fusarium, especially Fusarium sporotrichioides, Fusarium langsethiae, and Fusarium poae. Structure−activity relationship studies revealed that the 12,13epoxide group and the C9−C10 double bond are essential for their toxicity.1 Reported toxicological properties include emesis, compromised resistance to infection, diarrhea, lethargy, necrosis, apoptosis, and death. T-2 undergoes rapid deacetylation to its major metabolite HT-2 in vitro and in vivo. © 2017 American Chemical Society
Received: Revised: Accepted: Published: 4177
March 31, 2017 April 26, 2017 April 30, 2017 April 30, 2017 DOI: 10.1021/acs.jafc.7b01484 J. Agric. Food Chem. 2017, 65, 4177−4183
Article
Journal of Agricultural and Food Chemistry
thermal degradation products of T-220 were synthesized in our laboratories. All reference compounds were verified by comparing data acquired via high-resolution mass spectrometry (HRMS) and NMR spectroscopy to the literature. A purity ≥98% was determined via HPLC-ELSD. Standard solutions of T-2 (1000 μg/mL) and HT-2 (600 μg/mL) and a combined standard solution of d9-T-2 and d3-HT-2 (20 ng/mL) were prepared in acetonitrile and stored at −25 °C in the dark. Raw Material Preparation. The moisture content of the flour samples was determined in duplicate via a standard dry-oven method and resulted in values of 10.2% for oat flour and 12.7% for rice flour. Oat flour, rice flour, and oat dust were analyzed for their T-2 and HT-2 contents using the method described below. Levels of 1.8 ± 0.1 μg/kg for T-2 and 3.2 ± 0.1 μg/kg for HT-2 were found in oat flour, rice flour showed no detectable levels, and 1860 ± 67 μg/kg T-2 as well as 7270 ± 315 μg/kg HT-2 were found in oat dust. Furthermore, the oat dust was proved to contain T-2 triol, neosolaniol, T-2 glucoside, HT-2 glucosides, and HT-2 diglucosides. In the case of the glucosides this was a tentative assumption, done via exact mass measurements and HRMS fragmentation patterns. Oat and rice flours were blended in a ratio of 3:2, and moisture was adjusted to 20, 25, 30, and 35%, respectively, by the addition of tap water. This mixture was spiked with T-2 and HT-2 individually to levels of 500 μg/kg, calculated on a dry weight basis (dwb) by adding the toxins to the water used for moistening. The final concentrations of both mycotoxins were consequently 450.4, 422.3, 394.1, and 366.0 μg/kg. For each water content non-toxin-spiked flours were prepared, acting as a blank reference. All flours were sieved after moistening to achieve a homogeneous toxin distribution and to avoid clumping. Homogeneity was assured by analysis of three subsamples per batch with the method described below. Naturally contaminated samples were prepared by replacing one-sixth of the oat flour with contaminated oat dust, yielding a mixture with a toxin content of 187 μg/kg for T-2 and 729 μg/kg for HT-2, which was preconditioned to a water content of 20 and 30%. These doughs were sealed airtight and equilibrated overnight at 4 °C in the dark. To monitor the degradation rates on an industrial extruder, 17 samples were collected before and after extrusion cooking in 30 min intervals, yielding in total 34 samples of 500 g. The mixtures for extrusion were prepared by blending oat flour (75.1%), fine sugar, liquid malt, dextrose, wheat flour, corn starch, and salt in lots of 830 kg. Inside the extruder 12% of water was added to this mixture and subsequently removed during the drying process. The water content was 8.21 ± 0.25 and 4.69 ± 0.59% for the flour mixture and the extrudates, respectively. Extruder Operation. Extrusion experiments on a laboratory scale were carried out on a single-screw compact extruder KE 19/25 (Brabender, Duisburg, Germany) with a cylindrical, grooved barrel of 2.0 cm in diameter and an engine power of 1.5 kW. Compression screws (3:1 and 4:1) were driven at maximum rotation speed (150 rpm) and with a round-section profile die of 3 mm and 2 mm diameter, respectively. In the following, the former instrument configuration is referred to as low-pressure configuration and the latter as high-pressure configuration. The feed hopper was equipped with a feeding screw to assist constant feeding rates at 30 rpm of 3.3 kg/h. Screw speeds were kept constant within all experiments. The die temperature was set to 120, 130, 140, 160, 180, or 200 °C for each screw and die setup. The die pressure was recorded using a pressure transducer located at the end of the barrel. The temperature at the nozzle, the barrel temperature, and the pressure inside the barrel were recorded every 1 s. For each experimental set of conditions 300 g of preconditioned flour mixture was fed to the extruder, and the system was allowed to equilibrate for twice the maximum residence time before sampling was started. Forty grams of sample was collected at different time points, ground for 20 s using a Büchi mixer B400 (Büchi Labortechnik, Essen, Germany), sealed in polyethylene bags, and stored at −20 °C until analysis. The water content of each extrudate was determined using the dry-oven method. Naturally contaminated samples were extruded at 120, 160, and 200 °C with the highcompression screw and 2 mm die (high-pressure configuration). Screw
T-2 and HT-2 during the processing of food. Most researchers have focused on the redistribution of T-2 and HT-2 that accompanies cleaning, screening, dehulling, and milling.5,7,12−16 The thermal degradation of T-2 and HT-2 has in comparison not been the subject of many studies. So far, data on the degradation during cooking, baking, and kilning operations have been reported in more detail, revealing a greater thermal degradation of T-2 compared to HT-2.5,7,17−19 Aimed at the fate of thermally degraded T-2, model heating experiments with T-2 in the presence of glucose were carried out. They revealed three degradation products that all lack the 12,13-epoxide group, which is required for trichothecene toxicity,1 and therefore possess negligible cytotoxicity in cell culture experiments.20 During the production of cereals, extrusion cooking represents an efficient technology that combines pumping, kneading, mixing, shearing, cooking, and forming in one machine. It has become an important tool in the grainprocessing industry with a number of food and feed applications. As several operations are carried out simultaneously, it is plausible that these interact with each other.21 The water content plays an important role in extrusion cooking because it is essential for starch gelatinization and it strongly affects the fluid viscosity and expansion ratio. Extrusion cooking was shown to decrease the mycotoxin content at rates depending on moisture level, screw configuration, extruder geometry, die temperature, die size, screw speed, and additives, as summarized by Castells et al.22 The temperature and moisture content dependencies on mycotoxin degradation are most commonly studied. Therefore, temperature and moisture content were varied in this study in the range typically found during industrial oat processing and were extended to the possible maximum range of the used extruder to intensify these effects. The experiments were performed using samples individually fortified with either T-2 or HT-2 to detect any conversion of T-2 to HT-2 during processing.23 By installing different compression screws and dies, high-pressure and lowpressure instrument configurations were applied. These data were compared with results obtained from extrusion cooking experiments using naturally contaminated flour, in the laboratory as well as in industry. All extrudates were analyzed for their T-2 and HT-2 contents and for the presence of the T-2 congeners neosolaniol, 3; T-2 triol, 4; and T-2 tetraol, 5 (Figure 1), as well as for the previously described degradation products. Results from all experiments were compared and the impact of different processing parameters evaluated.
■
MATERIALS AND METHODS
Chemicals. All chemicals were purchased from either VWR (Darmstadt, Germany), Fisher Scientific (Loughborough, UK), or Grüssing (Filsum, Germany). All solvents were of gradient grade. ASTM type 1 water was produced with a Purelab Flex 2 system from Veolia Water Technologies (Celle, Germany). Whole oat flour (particle size < 0.5 mm, 6.6% fat) and maize grits were purchased from Herrnmühle (Reichelsheim, Germany), and white rice flour was bought from Ziegler & Co. (Wunsiedel, Germany). Oat dust was kindly provided by H. & J. Brüggen (Lübeck, Germany). T-2 was isolated out of cultures of F. sporotrichioides DSM 62423 (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany) according to a previously described procedure.24 Some of the resulting 4.3 g of T-2 (2.39 g/kg medium) was hydrolyzed, yielding HT-2, T-2 triol, neosolaniol, and T-2 tetraol according to the same publication.24 Stable isotopic standards of T-2 and HT-2, namely, d9-T-2 (unpublished data) and d3-HT-2,24 and the 4178
DOI: 10.1021/acs.jafc.7b01484 J. Agric. Food Chem. 2017, 65, 4177−4183
Article
Journal of Agricultural and Food Chemistry Table 1. Monitored SRM Transitions and Conditions analyte
a
chemical formula
retention time (min)
ion species +
DPa (V)
parent ion mass (m/z)
fragment ion mass (m/z)
CEb (V)
110
321
202 102
22 24
46
400
245 215
15 22
T-2 tetraol
C15H22O6
3.58
[M + Na]
neosolaniol
C19H26O8
4.57
[M + NH4]+
T-2 triol
C20H30O7
5.72
[M + Na]+
116
405
345 303
22 20
degradation product 1
C24H36O10
5.82
[M + Na]+ [M + NH4]+
220 60
507 502
405 407
43 20
d3-HT-2
C221H292H3O8
6.14
[M + Na]+
110
450
348 285
27 30
HT-2
C22H32O8
6.16
[M + Na]+
110
447
345 285
27 30
degradation product 2
C24H34O9
6.31
[M + Na]+
188
489
387 327
38 41
d9-T-2
C241H252H9O9
6.54
[M + Na]+
145
498
327 245
32 36
T-2
C24H34O9
6.59
[M + Na]+
145
489
327 245
32 36
degradation product 3
C24H34O9
7.11
[M + Na]+
100
489
387 267
28 32
Declustering potential. bCollision energy.
rotation speeds were the same as for the other experiments for comparative reasons. The extruder was not stopped between the experiments to achieve cleaning but was purged with maize grits. Extrusion cooking on an industrial scale was performed using a 400 kW twin-screw extruder operating in batch mode. The dimensions of the modulated screws were 2400 × 120 mm, and screw rotation speeds remained constant during the operation. Extrusion was performed at 150 °C with a corresponding die pressure of 60 bar. The total extruder throughput was 930 kg/h, and the residence time was 50−60 s. Sample Preparation. Sample preparation was carried out according to an existing protocol with slight modifications.25 The ground sample was weighed (5.0 ± 0.1 g) in a 50 mL polypropylene tube and extracted with an Ultra-Turrax T25 disperser (Ika Werke, Staufen, Germany) in 20 mL of acetonitrile/water/formic acid (79:20:1, v/v/v) for 90 s at 12,500 rpm. The resulting suspensions were allowed to settle for 3 h, and 125 μL of extract was subsequently diluted with 875 μL of water in 1.5 mL Eppendorf tubes, followed by a centrifugation step at 2885g for 15 min. Upon centrifugation, the supernatant was transferred to 1.5 mL glass vials for high-performance liquid chromatography coupled with tandem mass spectrometry analysis (HPLC-MS/MS). For all samples containing oat dust to simulate natural contamination and for industrially processed samples, 25 μL of water was replaced by the stable isotope labeled standard solution (20 ng/mL) in the dilution step to give final concentrations of d9-T-2 and d3-HT-2 of 0.5 ng/mL each. Matrix-matched calibrations for T-2 and HT-2 were carried out in a range of 0.5−20 ng/mL at nine different concentration levels. Calibrations with stable isotopic standards were done without matrix in the range of 0.05−50 ng/mL at 11 different concentration levels. Blank extrudate samples of both screw and die configurations were independently spiked with a standard solution of T-2 and HT-2,
imitating a toxin degradation of 0, 50, and 90% according to a contamination level of 500 μg/kg dwb to determine the recovery rates for both analytes. These samples were left open overnight to achieve complete solvent evaporation. Sample preparation took place as described above. As we observed no change in the instrument sensitivity by the extrudate matrix in terms of die temperature and water content, four different matrix-matched calibration curves were used, namely, two each for T-2 and HT-2 for each screw and die setup with the experimental conditions of 160 °C and 30% water content, which represents the average of the parameters observed here. HPLC-MS/MS Parameters. A 1260 Infinity series HPLC (Agilent Technologies, Waldbronn, Germany) coupled to a QTrap 6500 mass spectrometer with an ESI source (Sciex, Darmstadt, Germany) was used. Data acquisition and quantitation were carried out with Analyst 1.6.2 software (Sciex). For chromatographic separation, a ReprosilGold column (Dr. Maisch, Ammerbuch-Entringen, Germany) was used at a column oven temperature of 40 °C. Column dimensions were as follows: column size, 150 × 2 mm; guard column, 5 × 2 mm; particle size, 3 μm; pore size, 120 Å. A binary gradient was applied consisting of methanol (A) and water (B) at a flow rate of 300 μL/ min. The initial conditions of 10% A were held for 0.5 min, increased to 70% A within the following 1.3 min, and further increased to 100% A until 6 min. One hundred percent A was applied for 1 min, followed by an equilibration step at 10% A from 7.5 to 10 min. The mass spectrometer was operated in positive selected reaction monitoring (SRM) mode. The column effluent of the first 2 min was diverted to the waste via a synchronized six-port valve. The ESI source temperature was set to 450 °C and the ion spray voltage was 5500 V. The curtain gas, the collision-activated dissociation gas, and ion source gases 1 and 2 were set to 30 psi, “medium”, 35 psi, and 45 psi, respectively. The entrance potential was 10 V, and unit resolution was applied. With neosolaniol and degradation product 1 as ammonium 4179
DOI: 10.1021/acs.jafc.7b01484 J. Agric. Food Chem. 2017, 65, 4177−4183
Article
Journal of Agricultural and Food Chemistry
Figure 2. Degradation rates of (A, B) T-2 and (C, D) HT-2 during extrusion cooking of a mixture of fortified oat and rice flours at various water contents and temperatures on a single-screw laboratory extruder. Either a 3:1 compression screw and a 3 mm round-section die (low-pressure configuration, A, C) or a 4:1 compression screw and a 2 mm round-section die (high-pressure configuration, B, D) were used to simulate various degrees of mechanical stress.
120 and 200 °C and samples with four different water contents were used. The results of these experiments on the T-2 and HT-2 levels are summarized in Figure 2. Toxin Degradation in Fortified Samples. Using the lowpressure configuration, degradation rates for T-2 slightly rose with temperature from 20.2 ± 0.84% at 120 °C to 29.4 ± 3.59% at 200 °C in the case of 20% moisture content (Figure 2A). Adjusting the water content to 25% yielded no significant change, whereas further increases to 30 and 35% moisture resulted in significant increases of the degradation rates to 44.7 ± 1.02% at 35% of water and 120 °C with a maximum of 50.1 ± 3.22% occurring at 35% moisture and 140 °C. Pressure decreased with increasing moisture and temperature levels. At 20% moisture, it was 57.9 ± 2.99 bar at 120 °C and 26.8 ± 12.4 bar at 200 °C. Increasing moisture to 35% resulted in 5.63 ± 0.23 bar at 120 °C and 3.78 ± 1.17 bar at 200 °C. High relative standard deviations for the pressure were observed when the extruder was operated at temperatures of ≥180 °C. This can be explained by sticking of the sample fluid to the surface of the die and its intermittent release, causing a high amplitude in the pressure curve. Using the high-pressure configuration, degradation of T-2 ranged between 42.5 ± 2.00% at 20% moisture and 160 °C and 59.6 ± 1.51% at 35% moisture and 180 °C (Figure 2B). No significant change was observed in conjunction with the temperature in the range of 120 and 140 °C with the tendency of elevated degradation at higher water contents in the range of 160 and 200 °C. With this configuration, the pressure range generally increased compared to the low-pressure configuration with values of 93.7 ± 10.3 bar for 20% moisture and 120 °C
adducts as an exception, the sodium adducts of all analytes were monitored. Two SRM transitions were chosen for each analyte, according to the highest signal-to-noise ratios (S/N) in a matrixmatched standard solution. Table 1 displays the SRM transitions that were monitored at dwell times of 6 ms each. Performance Characteristics. All data acquisition was carried out via HPLC-MS/MS using either standards of T-2 and HT-2 in matrixmatched calibration or deuterated standards of both toxins for quantitative purposes whenever no blank matrix was at hand, as is the case for the naturally contaminated samples. The whole oat flour used was found to contain T-2 and HT-2 at concentration levels that were neglected for this study in view of the 100-fold higher spiking level. Matrix-induced signal suppression amounted to 32−35% for T-2 and to 40−44% for HT-2, calculated via the slope of the calibration curves with and without matrix. The recovery rates were 96.5 ± 11.4, 92.4 ± 8.51, and 88.1 ± 6.21% for T-2 and 83.5 ± 7.53, 80.0 ± 6.44, and 80.3 ± 4.93% for HT-2, at spiking levels according to toxin degradation of 90, 50, and 0%, respectively. The limit of detection and the limit of quantitation were 0.16 and 0.80 μg/kg for T-2 as well as 0.8 and 1.6 μg/kg for HT-2. All results are reported on dwb and were not corrected for recovery.
■
RESULTS AND DISCUSSION
Extrusion Parameters. The impact of extrusion cooking on T-2 and HT-2 spiked samples was performed with two different extruder configurations. A 3:1 compression screw with a 3 mm round section die was chosen as a low-pressure, lowshear system, whereas a 4:1 compression screw with a 2 mm die was used for high-pressure, high-shear extrusion. Residence times of the sample material in the extruder were 30−50 and 35−70 s for the low- and high-pressure configurations, respectively. On both systems, temperature was varied between 4180
DOI: 10.1021/acs.jafc.7b01484 J. Agric. Food Chem. 2017, 65, 4177−4183
Article
Journal of Agricultural and Food Chemistry
Table 2. Degradation Rates of T-2 and HT-2 during Extrusion Cooking of Oat and Rice Flours Spiked with Naturally Contaminated Oat Dust HT-2 degradation ratea (%)
T-2 degradation rate (%)
a
temperature (°C)
water content = 20%
water content = 30%
water content = 20%
water content = 30%
120 160 200
26.4 ± 4.09 29.0 ± 3.25 17.2 ± 0.93
34.7 ± 3.40 34.9 ± 3.08 35.1 ± 1.55
−12.1 ± 4.38 5.89 ± 4.68 −29.1 ± 0.55
18.4 ± 4.38 22.0 ± 4.68 15.1 ± 0.55
Negative values indicate that increasing amounts of mycotoxins were detected upon extrusion cooking under these conditions.
130 °C is observed. Degradation of up to 38.4 ± 3.48% occurred at 30% moisture content and a temperature of 120 °C, which decreased water-independently until 160 °C with a minimum degradation rate of 2.52 ± 9.66% at a water content of 20%. The maximum degradation of HT-2 was 47.2 ± 0.53% at 35% of water and 200 °C. For both configurations the pressure curves recorded for the extrusion of the HT-2-spiked samples matched those recorded for the T-2-spiked samples. Comparison of the degradation curves for HT-2 revealed a different picture compared to T-2. Data obtained with the lowpressure configuration show that moisture content has no impact in the range studied. Instead, degradation reactions of HT-2 seem to be strongly affected by the temperature, as a minimum of degradation could be observed at 160 °C. This drop in toxin degradation can be explained by the impact of mechanical shear that goes along with the viscosity, which is higher at low temperatures, combined with a progressing degradation at high temperatures. This implies that for temperatures 60%. Comparable results were found by Wu et al. 26 for deoxynivalenol on the same single-screw extruder, with deoxynivalenol being degraded by 27.0 ± 0.01% at 17% moisture and by 37.7 ± 0.62% at 34% moisture at a temperature of 170 °C. However, the closely related type B trichothecene, nivalenol, did not show this moisture dependency on degradation rates at this high temperature.27 When the high-pressure configuration was applied, reactions of T-2 that are induced by the high shear forces seem to be favored at low moisture levels. At 20 and 25% water contents, pressure is increased by 73% on average compared to the lowpressure configuration when temperatures between 120 and 160 °C are applied, leading also to an increase of the degradation rate from 20.2 to 27.1 to 42.5−52.5%. At 30 and 35% moisture contents, the pressure was almost unaffected by the extruder configuration, which is in good accordance with the comparable extent of degradation observed for both configurations. These results support the assumption that T-2 is hydrolyzed during the process of extrusion cooking. Therefore, the samples were analyzed for the presence of HT-2, T-2 triol, neosolaniol, T-2 tetraol, and degradation products 1−3.20 T-2 tetraol and degradation products 2 and 3 were not detected in any sample. The concentrations of T-2 triol and neosolaniol corresponded to
